Extragalactic Astronomy with a deep NIR Large Area Survey - some thoughts about...
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
some thoughts about... Extragalactic Astronomy with a deep NIR Large Area Survey Vladimir Avila-Reese Instituto de Astronomía, UNAM
The context (only galaxies/ see other talks for QSO’s, clusters, lensing) There are several competitive current and future ‘static’ NIR surveys with science drivers focused on EG astronomy (UKIDSS www.ukidss.org/index.html ,VISTA www.vista.ac.uk/index.html VIRMOS VDS www.oamp.fr/virmos/virmos_vvds.htm, etc.) • NIR traces Ms, the galaxy stellar structure, hot dust, and in combination with opt. bands, provide key information about the stellar pop’s. • High-z pop’s: e.g., already-in-place at z~1-3 L✴ E’s are revealed in NIR surveys; SF-ing giant galaxies at z>4-5; growth of structure and bias from z=3 to z=0 // but z is crucial!
What might be Sensitivity the SASSIR z~0 z>0 strengths? -Low L, low SB pop’s -Sample completeness (dIrr, BCD, dSph, LSB). down to low L’s & SB’s (SMF up to z~4) -Outer galaxy regions -Galaxy evolution (structure, dust) (K~22 for a L✴ E at z=3) Sky area (large statistics) z~0 z>0 -Morphology (T, bars, -Clustering of different rings, shells,...) & global pop’s at different z’s structure of gal’s (SB profiles, radii, B/D ratio, -Statistics of rare objects CAS, lopsidness) -LSS studies: correlations -Stellar scaling relations functions, BAO’s, voids & -Environment corrtn’s superclusters (ISW)
Spectral range z~0 -YJHK color-color traces z>0 - N I R c o l o r- c o l o r SF activity & hot dust diagram may select high- -By combining with z galaxy pop’s (gal. evol.). opt. bands, full SP -Opt. bands important for description, age-Z photometric z’s. degeneracy broken Sasir The dashed line shows the Euclidian number counts relation, which goes as 10-0.6K, i.e. the position of a survey relative to the line is a relative measure of the volume of space surveyed. This Ultra Deep Survey? illustrates, for example, that for surveys for brown dwarfs 2MASS will detect many more than the KPNO survey, and the LAS is an order of magnitude bigger than 2MASS.
The local universe NIR M/L ~ insensitive to o Bell et al. 03 ApJS, 149, 249 galaxy or stellar type NIR light ~ insensitive to dust extinction K-correction in K-band ~ insensitive to galaxy type NIR traces well galaxy Ms & stellar structure BUT, NIR sky SB (K:14-15) >> but even in K-band, to estimate Ms, a color is needed! outer SB of gal’s or even central SB’s of low-L and LSB gal’s bar ‘ba ryo yon nic ic g ’C NIR bands are ideal for inferring ala xie DM ha s los stellar (and eventually baryonic) stell ar ga laxie galaxy properties and s correlations, (i) to be compared with models, and (ii) to Baldry et al. 08, MNRAS determine the ‘efficiency’ of galaxy and star formation
The local universe NIR M/L ~ insensitive to o Bell et al. 03 ApJS, 149, 249 galaxy or stellar type NIR light ~ insensitive to dust extinction K-correction in K-band ~ insensitive to galaxy type NIR traces well galaxy Ms & stellar structure BUT, NIR sky SB (K:14-15) >> outer SB of gal’s or even central baryon and stellar fractions, a key constraint SB’s of low-L and LSB gal’s to the galaxy formation models bar NIR bands are ideal for inferring ? yon ic g stellar (and eventually baryonic) ala xie s galaxy properties and correlations, (i) to be compared with models, and (ii) to Baldry et al. 08, MNRAS determine the ‘efficiency’ of galaxy and star formation
overed affect- from Fig. 2, it is assumed that any bias induced by this cut Galaxy L and Ms functions hat, for Pet- will be negligible. der, d when in- By estimating the area in this way, the assumption isK-band galaxies from UKIDSS LAS 5 2 A. J. Smith, J. Loveday & N. J. G. Cross ntly brighter that all SDSS target galaxies would be detected in the LAS, ture fluxes, the ellipticity and the seeing, all made available icro-stepped, 1) Table 1. Sample sizesinofpart if that the K-bandWSA ofgalaxy the sky (theluminosity pipeline has doesbeen functions. surveyed. not measure 2 DATA If this is not the the half-light this problem case, it will have two effects: (1) particular radius). types used All of the galaxies of galaxies in this sample are drawn from the ) Given that will be underrepresented in the sample main galaxy (those samplewithin of Data Release the 5 of the Sloan Digital effect of peculiar velocities: et Sky Survey (SDSS; Adelman-McCarthy et al. 2007), from ,rs we have re- Paper SDSS2.4 Loveday (2000) completenessNumber of galaxies in sample Surface brightness limits but outside 345 thethe which limitsopticalfor the LAS) photometry and spectroscopic redshifts K r -Petrosian K mag and r limits--> SB fainter than 20.4m/ in the UKIDSS Large Area Survey, Table 2 shows investigation 23 mag the various of surface P brightnesslimits while leaving detailed arcsec−2 , so we assume the limit in SDSS surface on theand completeness sample. the very spectroscopic redshifts ☐ Petrosian magnitudes brighter (zConf > 0.8) and with de-reddened effect than of peculiar r = 17.6velocities, which are not taken into account. is 391 052. brightness faint-end of the luminosity function to future adds no further incompleteness work. (in principle) ofOfbuilding to the these,such sample, 384a617 Note that this are spectroscopically limitaswould classified need to be relaxed in order to galaxies. 1 INTRODUCTION once the magnitudeThe advantages limits in r and K and the surface census bright-
overed affect- from Fig. 2, it is assumed that any bias induced by this cut Galaxy L and Ms functions hat, for Pet- will be negligible. der, d when in- By estimating the area in this way, the assumption isK-band galaxies from UKIDSS LAS 5 2 A. J. Smith, J. Loveday & N. J. G. Cross ntly brighter that all SDSS target galaxies would be detected in the LAS, ture fluxes, the ellipticity and the seeing, all made available icro-stepped, 1) Table 1. Sample sizesinofpart if that the K-bandWSA ofgalaxy the sky (theluminosity pipeline has doesbeen functions. surveyed. not measure 2 DATA If this is not the the half-light this problem case, it will have two effects: (1) particular radius). types used All of the galaxies of galaxies in this sample are drawn from the ) Given that will be underrepresented in the sample main galaxy (those samplewithin of Data Release the 5 of the Sloan Digital effect of peculiar velocities: et Sky Survey (SDSS; Adelman-McCarthy et al. 2007), from ,rs we have re- Paper SDSS2.4 Loveday (2000) completenessNumber of galaxies in sample Surface brightness limits but outside 345 thethe which limitsopticalfor the LAS) photometry and spectroscopic redshifts K µe,K - 21 mag arcsec−2 ing good agreement The with previous results. Possible improvements a are discussed e ! targeted for spectroscopy in the SDSS µthat ce -fainter than K=16, deviates from the Euclidean slope discussion about possible low-surface SDSS brightness main incomplete- galaxy sample hasnumber limit of of galaxies could be implemented when extending this analysis LAS. galaxy sample. But firstµthe to the full main e,r size of this sample must - 24.5 mag arcsec−2 effectness in 2MASS of peculiar (Andreon24.5 2002), velocities, magwhich which arcsec are−2 would not affect taken (Strauss etthe intoal.low- account. 2002). This limit is taken e, -sky brightness in K + LAS depth--> r -Petrosian K mag and r limits--> SB fainter than 20.4m/ in the UKIDSS Large Area Survey, Table 2 shows investigation 23 mag the various of surface P brightnesslimits while leaving detailed arcsec−2 , so we assume the limit in SDSS surface on theand completeness sample. the very spectroscopic redshifts ☐ Petrosian magnitudes brighter (zConf > 0.8) and with de-reddened effect than of peculiar r = 17.6velocities, which are not taken into account. is 391 052. brightness faint-end of the luminosity function to future adds no further incompleteness work. (in principle) ofOfbuilding to the these,such sample, 384a617 Note that this are spectroscopically limitaswould classified need to be relaxed in order to galaxies. 1 INTRODUCTION once the magnitudeThe advantages limits in r and K and the surface census bright-
P = 0.18, and (3) making the LF 0.22 mag brighter in mag- K-band galaxies from UKIDSS LAS 9 nitude, to convert from 0.1 r to r. Schechter function parame- ? ters are M ∗ − 5 log h = −20.32 ± 0.04, α = −0.87 ± 0.05 and φ∗ = (0.0216 ± 0.0010)h3 Mpc−3 . red K-band galaxies from UKIDSS LAS 13 Intrinsic or incompleteness? Detection of red- Figure 19. K-band (left) and r-band LFs for red and blue galax- f Figure Figure 12. K-band luminosity function for the whole sam- 12. K-band luminosity function for the whole sam- core low-Lthe ies, showing gal’s is LF total affected as well. by the mag, r, and SB limits e ple, with a compendium of published results from observations a e ple, with a compendium or semi-analytic or semi-analytic Schechter function models. models. of Only fit, Only published i.e., Mthe results the filled pointsfrom observations are used in the filled points are used in the K − 5 log h < −20; the unfilled faint Stellar mass function - Schechter points function Figure are likelyBBD 17. fit, to i.e.,from suffer for MK red some− 5incompleteness galaxies, h 2.35. d pointsbrightness likelyorto red, arebest-fitting low-luminosity suffer from some galaxies. Schechter incompleteness function pa- − ing off of the luminosity–surface brightness relation at high The ∗ Cho" loniewski function, estimatedof low-surface using M K t rameters brightness or are red, − 5 log h = −23.17 Mlow-luminosity ± 0.04,Schechter galaxies. α = −0.81function ± 0.04 andpa- luminosities, suggesting that this division reflects a property ? 5 log h < −20 and µe,abs ∗ = (0.0176 ± 0.0009)h < 19, is shown by the red dotted 3 Mpc−3 . f φ ∗ − 5 log hof=the −23.17 ± 0.04, of the underlying population. Moreover, the Cho"loniewski rameters are M contours. Parameters fit are M ∗ −α5 = log−0.81 ± 0.04mag, h = −22.88 and φ∗ =α(0.0176 ∗ = 0.0121h = 0.17,±φ0.0009)h 3 Mpc −3 .−3 3 Mpc , µ∗e,abs = 17.09 mag arcsec−2 , function appears to fit the blue BBD much better than it fits 0.570 mag arcsec−2 the BBD for the whole sample. However, we caution that the σµe,abs = Equation (1) can be writtenand as β = −0.012. lack of red-core galaxies with MK − 5 log h > −20 could be “ r ”2 Equation (1) can be written a symptom of the incompleteness identified in Fig. 10, while .asfunction. Only the filled points P µe " K + 2.5 Figure log 20. 2π Stellar Figs. 17–19 show 2 mass the K-band BBD and K- and (15)are r-band Figure 21. ofStellar the lack suchmass function, galaxies using with various µe,abs 19.5 mag arcsec−2 > mass-to-light rare “ ” ? LFs used in the Schechter split by the SDSS rP rest-frame 2 function fit, i.e., u−r stellar mass greater than −2 PSF colour. The space ratios and the default kcorrect mass, which is derived from the could be due to the low-surface brightness limit for de Vau- µe " K +102.5 hlog M For K=16 9.5 and −2 µ = 19.5 . points are likely 2π" ;e the unfilled mag arcsec , this to corresponds (15) to suffer from some template fit to the input (ugriz) absolute magnitudes. Masses density a limit is Petrosianof2by inestimated incompleteness summing radius of 4.0 low-surface the weights arcsec, brightness of galaxies considerably galaxies. Masses lesswith calcu- couleursfrom calculated profile galaxies. the K-band kcorrect mass-to-light ratios have − rthe For uK=16 than >and lated 62.35µe or from arcsecthe u − intrinsic K-band =limit 19.5 rmag < 2.35 kcorrectto for arcsec red −2data; and mass-to-light the , −2this only blue ratios galaxies have for surface corresponds been to The LFs been increased show by 0.1 dex. a sharp division, with red-core galax- increased respectively, by 0.1 with than dex. jackknifeSchechter function parameters are found brightness a limit fainter intoPetrosian be log(M∗ hradius 2 /M ) = oferrors 20.38 mag 10.44 ± estimated 4.0 arcsec arcsec, doessubsequently. the 6 arcsec 0.02, α considerably = −1.02 ± 0.04 less ies more abundant than blue-core galaxies by an order of largeThe radiusBBD for dominate. limit " red galaxies, excluding outliers, showsand no magnitude masses derivedat high from theluminosity K-band M/L (and vice ratios versa at compared low lumi- with than the 6φ arcsec ∗ = (0.0112limit intrinsic ± 0.0007)h 3 Mpcto the −3 data; . Stellar onlybased masses for surface on other evidence Fig. IMFs 9ofalso a correlation have shows between the best-fitting been reduced luminosity Cho" for comparison −2 lwith our and oniewski fit tosurface results, the based nosity), the masses and with derived fromthethebright opticalend of the bands. The LF around precise cause 0.8 mag brightness fainter than 20.38 mag arcsec does the 6 arcsec
• A NIR survey with both a very large area and high sensitivity (depth of K~21)* will allow to reconstruct the local NIR bivariate L/SB function (i) down to the dwarf galaxy pop and including LSB galaxies, (ii) with enough statistics for determining accurately the LF bright end (rare gal’s) and for (iii) allowing to separate the LF function by galaxy types, colors, and environments (1200 nights for 3E4deg2 at J=20, K=18.4 in a 4m telescope; UKIDSS-LAS will be for only 4E3deg2) •Redshifts are necessary; for local galaxies (z
The dwarf galaxy population (building blocks) from Vaduvescu PhD thesis •What is the abundance and the LF faint-end of different types of dwarf gal’s? What is the minimum Ms of galaxies? •How is the distribution of dwarfs in clusters, groups, filaments and voids? •How is the stellar structure of dIrr’s, BCDs and dS’s? NIR: mass, structure. Deep, LA surveys are necessary to explore dwarfs in the local volume and in different environments.
Some results from Vaduvescu thesis; Vaduvescu, Richer & McCall 06, AJ, 131 SPM observations. Careful photometric analysis. NIR SB profiles: sech law. dE and dSph are probably old fossils (or formed by tidal destruction?) What about dIrr, BCD, dS? BCD are dIrr in bursting phase and dE are fading dIrr? Key constraints to the kinematics! LCDM model (see Valenzuela’s talk)
Baryonic quantities, the ultimate goal To fully understand galaxy formation & evolution, we need to constrain stars, gas, and dust Synergy of NIR surveys Baryonic with HI surveys --> the real mass backbone of Stellar galaxies
Disk galaxy scaling relations 4.2. The stellar scaling relations The main changes as passing from the baryo stellar Avila-Reese, Zavala, Firmani, Hernandez-Toledo, AJ, in scaling correlations press; (Zavala, A-R,H-T,are the 03, Firmani decrease A&A) of and intrinsic scatter in the Vm -M correlations. 2.6 B 2.6 Kters of the R-M and R-Vm correlations also 2.4 2.4 but the Very significances reduced of the number differences are very More interesting,of the‘useful’ galaxies (weak) anticorrelation be 2.2 2.2 residuals of the baryonic Vm -Mbar and Rbar -M 2 2 The TF relations relations tends to dissapear in the stellar case, lated to other of our results, namely that the 1.8 sl=0.314 (2.77) 1.8 sl=0.261 disk surface (3.67) •The bar TFR is steeper than the density) is a third parameter in σintr=0.063 σintr=0.049 onic TFR st one. but not anymore in the stellar one. 1.6 1.6 10 differences •Intrinsic are scatters: change from bar to st 9 10 11 9 11 and to K-bandto related the gas (or stellar) m cases tion of galaxies, •Crucialshowing that the for constraining modelsmass infa histories vary systematically 16 among galaxies. Avila-Reese et al. variations consistent with galaxy models bas 0 2.6 St 2.6 Bar 2.4 2.4 ΛCDM framework? Let us first discuss in more detail the chan ? -0.2 2.2 2.2 slope of the Vm -M relation. If the stellar m 2 2 tion, fs = 1 − fg , depends on Ms as fs ∝ M Mbar = Ms /fs ∝ -0.4 Ms 1−β . Therefore, f M from Vm s 1.8 sl=0.274 (3.40) 1.8 sl=0.306 (3.00) = s σintr=0.045 one passes σintr=0.051 to V m ∝ M s . In the left pan α(1−β)M + M s g 1.6 1.6 7, fs vs Ms is plotted -0.6 for our galaxy sample. 9 10 11 9 10 11 with a high scatter, fs9 correlates 10 11 with 1.5 M 2 2.5rou s filled: HSB empty: LSB € ❍: red ∆: blue fs = 0.65(Ms /1010M" )0.13 , Fig. 7.— Stellar fraction fs vs Ms , Σs,0 , and (B − K). Solid and empty sy galaxies are represented with circles, while blue galaxies with triangles. The solid are from models by FA00.
The radius-M (-L) relations 1 1 •Scattered and segregated by SB, and less by color 0.5 0.5 •Slopes around 0.3, but for the B band case, 0.44. ☛ 0 0 sl=0.441 sl=0.285 σintr=0.19 σintr=0.20 9 10 11 9 10 11 A wide range in SB is crucial for this relation. Most of previous similar 1 1 samples did not include LSB gal’s 0.5 0.5 Scaling relations are fossils of: cosmological initial conditions 0 sl=0.310 0 sl=0.322 + gal. evolution + SP evolution. σintr=0.19 σintr=0.20 Observational constraints 9 10 11 9 10 11 (including early-types) are yet filled: HSB empty: LSB very limited. Even scatters are ❍: red ∆: blue highly informative
Trends among the residuals ✤In the bar case there is a (weak) anti-correlation: sl=-0.15. The level -0.5 baryonic of dependence of Vm on the disk surface density decreases as SB decreases (the halo becomes dominant). Could MOND explain this? ✤The anti-corr’n disappears in the st and L cases. Why? (A SF effect) stellar ✤The B-band TFR residuals increase as the gal’s are deviated to the redder side (Kannappan+ 02). For a given LK, the bluer the color, K band the larger LB. ✤The bar TFR residuals are larger for gal’s that end with higher st. fractions (formed earlier and/or had an efficient SF) B band The scaling rel’s change for luminous, stellar, and baryonic quantities: clues to models of gal. formation and cosmology
The stellar structure of galaxies in the NIR •NIR morphology is different to the optical ones (de Blok, Puerari,...): morphological reclassification. B/D ratio and dust corrections •NIR bar and bulge statistics (z~0 and z>0 and as a function of environment): secular evolution of galaxies & dynamical history (Hernandez-Toledo+) NIR •NIR CAS parameters & •not floculent arms lopsideness statistics: •a clear bar galaxy assembly history, dark matter halo structure and sub-structure, etc. (Hernandez-Toledo+, Valenzuela+) •Outer stellar disks structure -cutoff?, warps •Color gradients: inside out formation, secular mass redistribution, dust gradient? NGC 253 •LSB galaxy stellar structure
CB07 models MNRAS, 384,930 Colours of SF-ing galaxies are poorly understood. 5800 SDSS/UKIDSS late-type local galaxies -More strongly SF-ing galaxies have redder H-K color. -The more dust attenuation, the redder H-K -TP-AGB stars dominate the H & K bands following a SF burst. -TP-AGB stars are the main source of dust in nuclear region of the galaxy SFing non-SFing -Models: g-r vs Y-K could break divided by age divided by Z the age-Z degeneracy. -Data: marginally. Y-K is metal sensitive (see also Galay+02)
NIR photometry is very useful for understanding the physics of interacting galaxies (several of their underlaying properties are distorted in optical bands): a deep and extensive NIR catalogue of these galaxies is desirable NIR colors as probes of interacting galaxies (Geller et al. 06, AJ, 132, 2243) interacting isolated control sample Hot dust could be a new tracer of SF in compact dust-enshrouded bursts. The z=0 SFRD could be underestimated
The z>0 universe A large area survey of well-defined mass-tracing objects (e.g., galaxies in the NIR) in a given z-range is useful for several LSS and structure formation problems. --Baryonic Acoustic Oscillations (e.g. SDSS-III): dA at z=0.3, 0.6 & 2.5 ~150 Mpc, z=0.3 Used to measure dark energy properties
Imprint of superstructures in the CMBR due to the ‘late-time’ Integrated Sach-Wolfe (ISW) effect Accelerated expansion due to dark energy (z depends on DE gravitational z--> secondary anysotropies in the CMBR: cross-correlate with LSS at the z of the effect. Rudnick+07ApJ, 671 The ISW signal peaks at l~20 and z=0.5: superstructures of ~4deg or 100 Mpc/h Granett+ 0805.3695 (LRG in SDSS) 1.4Ghz radio sources from NVSS 140 Mpc hole @ the CMBR cold spot Measurement of DE parameters Test of cosmological models
Evolution of the K-band LF (UKIDSS-UDS) For high-z studies, z determination is demandatory Cirasuolo + 08, 08043471 Downsizing: bright/massive gal’s are assembled at high z (1-3) and then passively evolve, while the formation of less L gal’s is progressively shifted to lower z’s
Models vs observations: a great challenge
et al. (2001) in Figure 1, we estimate that the effective sky 2001), there is evidence of clustering, particularly a dip around coverage of the parent sample is 650 deg2, with an error of z p 0.04 and a sharp peak around z p 0.06. Therefore, the Evolution of the galaxy pair fraction (merging rate) ∼5%. We ignore this error because it is much smaller than other fluctuations of the LF of paired galaxies around the smooth errors. Given our pair selection criteria, both components of a Schechter function (e.g., the excess at MK p !22.75) are pos- pair have the same Vmax determined by the Ks magnitude of the sibly1094due to this effect. RAWAT ET AL. Vol. 681 The stellar masses, corresponding to the absolute eterized magnitude by their redshift flag). So in cases in which the neighbor bins of the LF, are also listed in Table 1. Following has Kochanek Local pair fraction as a function a VVDS redshift flag 2 (75% confidence) or worse, there is a et al. (2001) and Cole et al. (2001), we first translate themuch greater risk of the pair being designated as having discor- isophotal dant redshifts. In addition to this, one also has to allow for the Ks magnitude to the “total” Ks magnitude (dK s p 0.2 fact of L or M in K-band (2MASS) assume the conversion factor of Mstar /L K p 1.32 M, shift mag), thenof the pairs using redshifts from the other two red- that some /L catalogs whichcan also be erroneously counted as discordant (for the ,same reason as above). Such pairs account for "15/45 discor- is Thefor a Salpeter Astrophysical Journal,initial mass2008 681:1089Y1098, function July 10 (Cole et al. 2001). The dant redshift pairs in the z band. This can significantly swing the Xu+04, ApJ 603 # 2008. The American Astronomical Society. All rights reserved. Printed in U.S.A. differential pair fraction function (Table 1 and Fig. 2) isofcal- number pairs with discordant redshifts. If we discard these 15 discordant cases and recalculate the contamination rate, we culated using the Schechter functions of the paired get galaxies 60% # 14% as the contamination rate, which is in much bet- and of 2MASS galaxies (Kochanek et al. 2001) interthe lumi-with the contamination rate of 42% # 6:3% that agreement we have obtained using our statistically estimated foreground/ nosity/mass rangeTOWARD coveredAbyROBUST the paired ESTIMATEgalaxyOF sample, THE with MERGER background RATE EVOLUTION contamination. the error estimated from the quadratic sum of the error USING NEAR-IR PHOTOMETRY In a of the our contamination calculation is robust, and the nutshell, LF of paired galaxiesA. and Rawat, 1, 2 its deviation Francois Hammer, 2 from Ajit contamination the K. Kembhavi,1rate Schechter andcompares Hectorwell with2 the contamination rate Flores reported in literature, Received 2007 June 19; accepted 2008 March 31 as well as with the subsample of our major function. The last bin is not included because therepairs is where onlyweahave spectroscopic redshifts for both members of single galaxy in the bin, and therefore the value is toothe pair. Also, as a result of our eagerness to throw out prospec- uncertain. tive pairs as having discordant redshifts due to spurious redshift ABSTRACT Although error bars are substantial, some general trends canin some cases, we are going to end up with an es- be identified We use solid: in Figure K-band a combination of deep, high angular resolution imaging 2. Unlike the LF ofthe2MASS determination data galaxies, timate of fromfraction merger the CDFS that is(HST/ACS GOODS a bit too low, and thussurvey) it should the LFFig. 0:2 ! dashed: and ground-based of6.—Histogram paired z ! 1:2. galaxies of We z-band near-IR $ z has (z select host Ks images )/(1 þ za negative galaxies neigh solely ) for host to derive the 65on slope the major basis pairs evolution in be of the faint of their where the treated J-bandend, galaxy as a major lower merger limit. rate in the redshift range rest-frame absolute magnitude, which is a good % redshift, 5.3. Source Blendin g and /(1 þ z) 3:43#0:49 Photometric Completeness suggesting both tracer that galaxies with membersof the have stellar mass. a spectroscopic We redshift. find Note !z/(1 þ z) % 0:006 as seen in the figure. Most low stellar mass are less likely steep that only 20 pairs have $0:006 evolution with other pairs have !z/(1 þ z) much 2:18#0:18 with the merger rate for optically largerselected than 0.006pairs so that and /(1 they are þthe out of z) frame. for pairs selected in the near-IR. A major Ourbias result is unlikely when studying close to pairs be affected by may of objects lumi-be nosity evolution that is relatively modest when using rest-frame causedJ-band selection. by the fact that sourcesTheare apparently blended when more rapid their evo- separation lution that we find in the visible TABLE 2 is likely caused by biases is comparable relating to to the incompletenessPSF. It particularly and spatial affects ground-based resolution affectingKs 00 Inthe Figure 6 many objects ground-based near-IR have !z/(1 þ photometry, z) much larger than underestimating images because pair counts at higher the K s PSF ("0.5 ) is much larger redshifts in the near-IR. The major than the HST 0.006 (so Paremeters much of Schechter out ofatFunction Ks LF of 6). 00 Fig. 2.—Ks-band LFs and stellar mass functions and differential pair fraction merger ratesowas that most $5.6oftimes them are higher the z$frame 1:2inthan Fig. at the opticalepoch. current PSF ("0.1 ). This Overall, 41% is illustrated ; (0:5 Gyr/!) in Figure of all7,galaxies where we The members of such pairsof are Paired treated as Galaxies having ‘‘discordant red- have plotted the angular separation between the main galaxy (right coordinates). The lines are Schechter functions of the LF of paired with MJ ! %19:5 have undergone a major merger in the last $8 Gyr, where ! is the merger timescale. Interestingly, shifts’’ and are essentially foreground/background superpositions. and neighbors identified in the z filter against the z magnitude we findplot, no we effect canon the 5derived h major merger rate (fdue/h3to ) theError presence of thesegregated large-scale structure at z ¼bins. 0:735 in the galaxies (solid ), of the LF of 2MASS galaxies by Kochanek et al. (2001; From a this Error ∗ !derive Mthen logthe Error fraction log of pairs that 0are of the neighbor, into three redshift A neighbor CDFS. actually foreground/background superpositions (i.e., the contami- here is identified as any galaxy within 5 h100 kpc % r % 20 h$1 $1 dotted ), and of the LF of 2MASS galaxies by Cole et al. (2001; dashed ). The 100 0.30 nation rate).0.56 This Subject headingg !22.55 contamination rate is 0.25 found to be !3.46 "69% s: galaxies: evolution — galaxies: formation # 0.13— kpc of the primaryinteractions galaxies: galaxy, regardless of its magnitude. — galaxies: statisticsThe re- shaded area presents the differential pair fractions and the errors. 13%. On first sight this might appear larger than the contamina- sultant list of 1085 neighbors includes major as well as minor tion rate of 42% # 6:3% that we have obtained using our statisti- neighbors and constitutes the pool from which the major pairs cally estimated foreground/background contamination (although are identified. In Figure 7 the open circles denote neighbors that note the large error 1. bars since we are dealing with small number INTRODUCTION have Ks magnitudes, technique whereas the of pair counting. green 2 Section crosses denote lists the dataneigh- sets we have statistics here). However, we checked the spectra of the pairs bors that do not have Ks magnitude. As is shown in Figure 7, used in this work and briefly explains the methodology that we Galaxy havingmergers are believed discordant to be the It redshifts ourselves. chief turnsmechanism driv- out that in many there are a large number ("43%) of such neighbors that do not cases,evolution the redshift determination can be quite poor indeed, espe- have employed in this paper for deriving the merger have Ks -band magnitudes. This has the consequence of reduc- rate. Section 3 ing galaxy within the hierarchical framework. Although cially details the sample selection criterion employed by us and the mergers areinrare the at low-exposure the current and low-resolution epoch, VVDSframework the hierarchical (as param- ing the pool of neighbors from which the major pairs are identified possible biases in our sample. Section 4 explains the details of predicts that the merger rate must have been higher at earlier the photometry performed by us in the Ks filter. Section 5 explains epochs. Despite its importance in understanding galaxy evolu- in detail how we identify major pairs of galaxies, along with cor- The merging rate history is an tion, the quantification of the galaxy major merger rate and its recting for possible contamination and incompleteness issues. evolution with redshift is still an ill-constrained and hotly de- bated issue. Patton et al. (1997) have derived a clear increase in Section 6 compares the differences in pairs identified in the vis- ible and the near-IR. Section 7 deals with the calculation of the the merger fraction with redshift until z $ 0:33 with a power-law important constrain to models. index of 2:8 # 0:9. However, extension of this work to higher redshifts has been riddled with controversy. Le Fevre et al. (2000) reported steep evolution of the major merger rate until z $ 1:0, major merger rate from our identified pairs and its evolution with redshift. We conclude by discussing the implications of our re- sults obtained in this paper in Section 8. We adopt a cosmology High ang. resolution needed with H0 ¼ 70 km s%1 Mpc%1, !M ¼ 0:3, and !" ¼ 0:7. with a power-law index of 3:2 # 0:6 using pair-counting in the optical band for identifying the merger candidates. Bundy et al. 2. THE DATA (2004) reported a much more modest evolution of merger rate using K 0 -band images for identifying major merger candidates. We have used version v1.0 of the reduced, calibrated images of Lin et al. (2004) also reported very weak evolution in the merger the Chandra Deep FieldYSouth (CDFS) acquired with HST/ACS rates using the DEEP2 redshift survey. Bell et al. (2006) reported as part of the GOODS survey (Giavalisco et al. 2004). The a fairly rapid evolution in merger fraction of massive galaxies SExtractor ( Bertin & Arnouts 1996) based version r1.1 of the
On the evolution of clustering of 24-µm-selected galaxies M. Magliocchetti,1,2,3! M. Cirasuolo,4 R. J. McLure,4 J. S. Dunlop,4 O. Almaini,5 S. Foucaud,5 G. De Zotti,3,6 C. Simpson7 and K. Sekiguchi8 1 INAF, Osservatorio Astronomico di Trieste, Via Tiepolo 11, 34100 Trieste, Italy 2 ESO, Karl-Schwarzschild-Str.2, D-85748 Garching, Germany 1134 M. Magliocchetti et al. 3 SISSA, Via Beirut 4, 34014 Trieste, Italy 1138 M. Magliocchetti et al. intermedium z high z intermedium z 4 SUPA, Scottish Universities Physics Alliance, Institute for Astronomy, University of Edinburgh, Royal Observatory, Edinburgh EH9 3HJ 5 School of Physics and Astronomy, University of Nottingham, University Park, Nottingham NG7 2RD high z 6 INAF, Osservatorio Astronomico di Padova, Vicolo dell’Osservatorio 5, 35122 Padova, Italy 7 Astrophysics Research Institute, Liverpool John Moores University, Twelve Quays House, Egerton Wharf, Birkenhead CH41 1LD 8 National Astronomical Observatory of Japan, Mitaka, Tokyo 181-8588, Japan Accepted 2007 October 16. Received 2007 October 9; in original form 2007 June 29 Figure 5. Angular correlation function w(θ) for the low-z (left-hand panel) and high-z (right-hand panel) samples. The solid (blue) curves represent the best HON fits to the data obtained under the A Bassumption S T R A C T that the distribution of galaxies within their dark matter haloes mirrors that of the dark matter (chosen to be described by a NFW profile, cf. Section 4), while This paper the dashed investigates the (green) curves clustering are theofbest properties fits obtained a complete forof sample a steeper galaxy profile, ρ ∝ r−3 . The long-short 1041 24-µm-selected dashed (cyan) curves correspond to the best one-parameter 2 sources brighter than fits F 24obtained by setting in equation (5) α = 0 and N 0 =the µm= 400 µJy in the overlapping region between 1. Spitzer The negative Wide-w(θ ) data points in the χ analysis of the low-z sample have been treated Area as in Extragalactic Infrared Section 3.1 (see text forand (SWIRE) details). UKIRT Infrared Deep Sky Survey (UKIDSS) Ultra Deep Survey (UDS) surveys. With the help of photometric redshift determinations we have Figure 3. Sky distribution of UDS-SWIRE, F 24 µm ! 400 µJy sources (small dots). The filled (red) circles in the left-hand concentrated panel illustrate the 350 the two interval ranges z = [0.6–1.2] (low-z sample) and z ! 1.6 (high-z sam- on objects included in the redshift interval z = [0.6–1.2], while those in the right-hand panel represent the 210 sources with either z > 1.6 or no Figure optical5.orAngular near-IR correlation function w(θ) for the low-z (left-hand panel) and high-z (right-hand panel) samples. The solid (blue) curves represent the best the data, as small-scale measurements ple) asof it isthefits intwo-point these regions correlation were weassumption expect the mid-infrared (IR) population tomatter be dominated identification. function prefer smoother profiles by such intense HON as NFW be described to the data dust-enshrouded by a NFW obtained or ρ ∝ r activity −2.5 profile, under the cf. Section 4), while that the the distribution of galaxies . How-such as star formation and black hole accretion. dashed (green) curves are using the best fits halo within their dark obtained ocupation for a steeper model haloes mirrors that of the dark matter (chosen to Inves- galaxy profile, ρ ∝ r−3 . The long-short dashed (cyan) curves correspond to the best one-parameter fits obtained by setting in equation (5) α = 0 and N 0 = 1. The negative w(θ ) data points in the χ 2 of galaxies undergoing intense star formation which we expect ever, (cf. atgeometry very low redshifts of UDS, and corresponds to tigations it is reasonable about The ofthethe to expect ofclustering halfanalysis angular maximum of the low-z theof correlation majority BzK sample been of have galaxies function produce an amplitude A ∼ 0.010 for the high-z treated as in Section 3.1 (see text for details). earlier in this section and also in Fig. 2) to dominate our sample. dark matterscale probed haloes by thetosurvey. be virialized sample and (Hartley and Aet therefore 0.0055 ∼ al., be in for the low-z to prep.). reasonably Galaxies one. The corresponding correlation lengths are r0 = The low-z sample instead contains 350 sources. Their sky dis- Fig. 4 presents our results for the two15.9 low-z+2.9 (left-hand and panel)+1.5 8.5−1.8 and Mpc,e.g. showing that the high-z population is more strongly clustered. Com- tribution is represented by the filled (red) circles on the left-hand described high-z(at least beyond (right-hand thewhile panel) samples, veryTable insmall −3.4the 1the scales, UDS reports data, the have w(θ see ) as small-scale been Bukert selected measurements of the totwo-point correlation side of Fig. 3. The average redshift of the sample is found to1995) be values by as a function profiles. NFW-like of angular scale Two inparisons both cases. options with The are physical error bars instead models possible forsuch the asformation at < zNFW and evolution of large-scale structure reveal lie in the function redshift prefer smoother rangeprofiles1.4 < or ρ ∝ r−2.5 . How- of massive (M " 10 M# ) haloes, 13 !z" = 0.79, its median zmed = 0.84 and the comoving number den- show Poisson estimates for the points. Since thatthe thedistribution high-z ever, at verysources is clus- low redshifts are exclusively it is reasonableassociated to expect thewithmajority very sity of these sources is n̄ G = (9.9 ± 1.0) × 10−5 Mpc−3 . Evensuch in high tered,redshifts. these estimatesThe first one only provide a lower islimit that 2.5 to andgalaxies the can uncertainties. still be do trace characterised the as this case, the number density is in excellent agreement with underlying comparable dark mattertoofthosehaloeswhich locally and to be virialized hosttherefore groups-to-clusters be to reasonablyof galaxies and are very common the However,distribution it can be shown ofthatthe overdark matter the considered withineither within range described such (rare)(at their angu- star-forming haloes, (sBzK) and structures. Conversely, lower zsee least beyond the very or small scales, e.g. Bukert galaxies are found to reside in smaller haloes findings of Caputi et al. (2007) for their z ∼ 1 sources with ν Lν ! lar scales this estimate is close to those obtained from bootstrap 11.3 it is the dark matter itself which resampling (e.g. Willumsen, Freudling (M doespassive not 1995) 12follow by NFW-like (pBzK) 10highMredshifts. NFW-like profiles. by the Twopro- Daddi optionset are al.instead possible at 10 L& , where ν Lν was calculated as indicated before. The most & Da Costa min ∼ issue such 1997). # ) In and The to be very first one rareis thatingalaxies such systems. still do trace Onthethe other hand, mid-IR photometry relevant properties for the high-z and low-z samples are summarized files. TheParticular second option attention was is instead devoted to the that shows (2004) ofgalaxies that close pairs.are underlying criteria. the low-z simply In and this more redshift of the dark matter within their haloes, andobjects and probe a similar mix- high-z samples include similar in Table 2. fact, the points on the top left-hand corners of both w(θ)distribution esti- radiallymatesconcentrated than scales correspond to angular the dominant turerange close toof isdark theitactive the 5.4-arcsec matter it isdark galactic the matter reso- counterpart. itself which nucleus passive (AGN) does not galaxies andfollow NFW-like pro- star-forming that galaxies. While recent studies have files. The second option is instead that galaxies are simply more Whatever the favourite choice, onedetermined lution of the 24-µm Spitzer channel and thing therefore is could areradially foundclear; be toonbe affected a concentrated strong subhalo evolution thethanmore scales, of the the dominant 24-µm highly luminosity function between z ∼ 2 and 0, they dark matter counterpart. 3 C L U S T E R I N G P R O P E RT I E S by source confusion. However, since optical and near-IR images for 24-µm-selected these sources havegalaxies at redshifts a much better cannot resolution clustered. (∼0.8 ! zWhatever provide 0.6the arcsec), weare Thecan more information favouriteclustering strongly on choice, the one physical of thingaisset nature of such ofon subhalo clear; an evolution. Our clustering results scales, radiallyuseconcentrated this information and than their consider instead low-redshift as ‘good’ indicate thatofbe all thosecounterparts. pairs24-µm-selected made thisdirectly galaxies is due at Such tolinked redshifts the ! 0.6 are of a zpresence moredifferent stronglypopulations of objects inhabiting 3.1 The angular correlation function galaxies galaxies which are identified in the Subaru/UDSradially can images, concentrated while we than their low-redshift tocounterparts. Such ato be exclusively associated with strong regard concentration is indicative different structures, as active systems halos at z # 1.5 are found The angular two-point correlation function w(θ) gives the excess as potentially spurious those madeof ofthea strong galaxiesmass strong without interaction ofsuch the concentration dark between matter is indicative of a strong interaction between probability, with respect to a random Poisson distribution, of finding galaxiesoptical/near-IR which reside in the counterparts. Forproximitylow-mass the low-z sample, of galaxies, the there galaxies arehalo 11 pairs which whileinvery centres, reside theand massive we of sources proximity appearandtowehave concluded their active phase the halo centres, two sources in the solid angles δ$1 , δ$2 separated by an angle ◦ in which they reside; in the case of with distances between 0.001◦ and 0.0031 expect foundsuchto abe strong before . Of this epoch. these, expect onlysuch a Finally, one was we noteto that strong interaction the small-scale eventually drive a substantial clustering data seem to require steep θ. In practice, w(θ ) is obtained by comparing the actual source closer thaninteraction 5.4 arcsec and bothto theeventually IRAC −3 pBzKs and R-band number drive the pho- dark of mergers. aIt substantial matternothalos is possibly agalaxies random coincidence thathaloes. This is suggestive of close distribution with a catalogue of randomly distributed objects subject (ρ ∝ r ) profiles for the distribution of within their Figure 6. Average number of galaxies $N% per dark matter halo of specified number of tometry mergers. indicate thatIt weisarepossibly dealing with not have a two distinct random bright masssources. in coincidence 10 24-µm-selected excess sources ofat that 10 13 solar intermediate-to-high redshifts are to the same mask constraints as the real data. We chose to use the encounters close (i.e. again with distances between 0.0011 and ◦ and/or mergers 0.0031 associated ◦ ) pairs which could stronglyFigure favour6. both Average AGN andMstar number mass formation of galaxies (expressed in M&$N% activity. perThedark units). solid matter (red) line halo of specified represents the case for estimator (Hamilton 1993) bright 24-µm-selected are instead found in the sources at Of high-z sample. intermediate-to-high masses. these, only one is at a dis- redshifts are in general to very intense star-forming systems; the ex- the high-z sample, while the dashed (green) line is for the low-z sample. cess of proximity between such galaxies wouldmass M (expressed in M units). The solid (red) line represents the case for DD RR in general tance associated below 5.4 arcsec. toUnfortunately, very intense Key the two words: sources of galaxies: star-forming this pair systems; evolution the eventually –ex-galaxies: statistics – cosmology:&observations – cosmol- lead to w(θ ) = 4 − 1, (1) are unidentified both in the optical and in the near-IR gas-rich mergers which could easily trigger enhanced bands, and so the high-z sample, while the dashed (green) line is for the low-z sample. star formation (DR)2 cess of proximity between such galaxies ogy: theory would activity – like large-scale eventually the ones observed. structure leadIn to of Universe passing, we note – that infrared: evidencegeneral. the lowest possible masses where the sources start appearing), and the pair has to be considered as spurious. high-zfor an upturn (although less significative than ours) on scales r " we note that since $N% is a quantity which is averaged all over the where DD, RR and DR are the number of data–data, random– gas-rich mergers The angularwhich could correlation easily function for the trigger h enhanced sample 0.2 −1 Mpc was then has star also formation been recently reported by Gilli et al. (2007) for random and data–random pairs separated by a distance θ. computed by both including and excluding the possibly spurious dark matter haloes more massive than Mmin , in the most extreme activitypair. The two considered UDS-SWIRE samples have an estimated 5σ likeThethe ones (θobserved. small-scale In passing, ∼ 0.0015◦ ) results are their we by thenote shownsample z ∼ that oftwo evidencesources inthe 1, 24-µm-selected lowestfields. the GOODS possiblecase masses where our results imply the sources that more start than one appearing), in two and of all the structures flux completeness F for an upturn " 400 µJy. Furthermore, since the whole horizontal (although less significative dashes in the right-hand than The panel of Fig. 4. The bestours) filled average circle onnumber scales rfrom " the of galaxies $N%canonical hierarchical as a function of halo merging scenario, with masses M ∼ 10that 13 proved to M& host a high-z, F 24 µm # 400 µJy galaxy.
NIR color selection of high-z gal. populations Colour information is also vital for discriminating between objects, and for the determination of photometric redshifts, particularly using the 4000Å break as it moves through the I, J, & H bands over the range 1
Concluding remarks • Many EG problems could be solved with a deep and LA NIR survey • Synergy with SMM/Opt/X-ray deep surveys will be crucial for z determination and selection of high-z objects. • Synergy with HI surveys will allow to get baryonic quantities (the dream of modelers) • Follow-up (spectra, opt. bands, etc.) of discovered high-z galaxies.
You can also read