ebook img

Evolution of Balmer jump selected galaxies in the ALHAMBRA survey PDF

14 Pages·2016·2.36 MB·English
by  
Save to my drive
Quick download
Download
Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.

Preview Evolution of Balmer jump selected galaxies in the ALHAMBRA survey

A&A588,A132(2016) Astronomy DOI:10.1051/0004-6361/201527552 & (cid:2)c ESO2016 Astrophysics Evolution of Balmer jump selected galaxies (cid:2) in the ALHAMBRA survey P.TroncosoIribarren1,2,L.Infante1,2,N.Padilla1,2,I.Lacerna1,2,5,S.Garcia1,2,A.Orsi3,A.MuñozArancibia1,14, J.Moustakas4,D.Cristóbal-Hornillos3,M.Moles3,6,A.Fernández-Soto9,10,V.J.Martínez9,11,12,M.Cerviño6,7,8, E.J.Alfaro6,B.Ascaso13,15,P.Arnalte-Mur11,12,L.Nieves-Seoane10,11,andN.Benítez6 1 InstitutodeAstrofísica,PontificiaUniversidadCatólicadeChile,Avda.VicuñaMackenna4860,782-0436Macul,Santiago,Chile e-mail:[email protected] 2 CentrodeAstro-Ingeniería,PontificiaUniversidadCatólicadeChile,Avda.VicuñaMackenna4860,782-0436Macul,Santiago, Chile 3 CentrodeEstudiosdeFísicadelCosmosdeAragón,PlazadeSanJuan1,44001Teruel,Spain 4 DepartmentofPhysicsandAstronomy,SienaCollege,515LoudonRoad,Loudonville,NY12211,USA 5 MaxPlanckInstituteforAstronomy,Königstuhl17,69117Heidelberg,Germany 6 InstitutodeAstrofísicadeAndalucía(IAA-CSIC),GlorietadelaAstronomía,18008Granada,Spain 7 InstitutodeAstrofísicadeCanarias,VíaLácteas/n,38200LaLaguna,Tenerife,Spain 8 DepartamentodeAstrofísica,FacultaddeFísica,UniversidaddeLaLaguna,38206LaLaguna,Spain 9 UnidadAsociadaObservatorioAstronómico(IFCA-UV),46980Paterna,Spain 10 InstitutodeFísicadeCantabria(CSIC-UC),39005Santander,Spain 11 ObservatoriAstronòmic,UniversitatdeValència,C/CatedràticJoséBeltrán2,46980Paterna,Spain 12 Departamentd’AstronomiaiAstrofísica,UniversitatdeValència,46100Burjassot,Spain 13 GEPI,ObservatoiredeParis,CNRS,UniversitéParisDiderot,61,Avenuedel’Observatoire75014Paris,France 14 InstitutodeFísicayAstronomía,UniversidaddeValparaíso,Avda.GranBretaña1111,Valparaíso,Chile 15 APC,AstroParticuleetCosmologie,UniversitéParisDiderot,CNRS/IN2P3,CEA/lrfu,ObservatoiredeParis, SorbonneParisCité,10rueAliceDomonetLéonieDuquet,75205ParisCedex13,France Received13October2015/Accepted30January2016 ABSTRACT Context. Samples of star-forming galaxies at different redshifts have been traditionally selected via color techniques. The ALHAMBRA survey was designed to perform a uniform cosmic tomography of the Universe, and we here exploit it to trace the evolutionofthesegalaxies. Aims.OurobjectiveistousethehomogeneousopticalcoverageoftheALHAMBRAfiltersystemtoselectsamplesofstar-forming galaxiesatdifferentepochsoftheUniverseandstudytheirproperties. Methods.Wepresentanewcolor-selectiontechnique,basedonthemodelsofspectralevolutionconvolvedwiththeALHAMBRA bandsandtheredshiftedpositionoftheBalmerjumptoselectstar-forminggalaxiesintheredshiftrange0.5<z<1.5.Thesegalaxies aredubbedBalmer-jumpGalaxies(BJGs).WeappliedtheiSEDfitBayesianapproachtofiteachdetailedspectralenergydistribution anddeterminedstar-formationrate(SFR),stellarmass,age,andabsolutemagnitudes.Themassofthehalosinwhichthesesamples residewerefoundthroughaclusteringanalysis. Results. Five volume-limited BJG subsamples with different mean redshifts are found to reside in halos of median masses ∼1012.5±0.2M(cid:4) slightly increasing toward z = 0.5. This increment is similar to numerical simulations results, which sug- geststhatwetracetheevolutionofanevolvingpopulationofhalosastheygrowtoreachamassof∼1012.7±0.1atz=0.5.Thelikely progenitorsofoursamplesatz∼3areLyman-breakgalaxies,whichatz∼2wouldevolveintostar-formingBzKgalaxies,andtheir descendantsinthelocalUniversearegalaxieswithluminositiesof1–3L∗.Hence,thisallowsustofollowtheputativeevolutionof theSFR,stellarmass,andageofthesegalaxies. Conclusions.Fromz ∼ 1.0toz ∼ 0.5,thestellarmassofthevolume-limitedBJGsampleschangesalmostnotatallwithredshift, suggestingthatmajormergersplayaminorroleintheevolutionofthesegalaxies.TheSFRevolutionaccountsforthesmallvariations ofstellarmass,suggestingthatstarformationandpossibleminormergersarethemainchannelsofmassassembly. Keywords.Galaxy:evolution–Galaxy:halo–galaxies:high-redshift–galaxies:evolution–galaxies:general–galaxies:photometry 1. Introduction observational constraints have usually been summarized as galaxyscalingrelationsthatmightormightnotchangewithred- Most of the recent observational efforts to understand galaxy shift (Mannucci et al. 2010; Elbaz et al. 2011; Bouwens et al. evolutionhavebeenfocusedondeterminingthehistoryofcos- 2014; Troncoso et al. 2014), in high- or low-density environ- mic star formation, gas density evolution, metallicity evolu- ments, in extreme physical conditions (starburst, AGN galax- tion, and mass growth of the Universe (Daddi et al. 2004; ies), and in spatially resolveddata due to internalvariationsof Mannucci et al. 2010; Madau & Dickinson 2014; Tomczak the galaxy properties (Sanchez et al. 2013). In parallel, theo- et al. 2014; Bouwens et al. 2015). These multiwavelength retical works and simulations have tried to explain the physi- (cid:2) BasedondataobtainedattheCalarAltoObservatory. cal mechanisms that reproduce the measured global properties ArticlepublishedbyEDPSciences A132,page1of14 A&A588,A132(2016) (Daddi et al. 2010; Davé et al. 2011; Lilly et al. 2013; Lagos Theyincludedourlackofknowledgeoftheprecisegalaxyred- etal. 2014;Padilla et al. 2014).Despite these efforts, the com- shiftand selected the sample accordingto a certain probability pletenessandcleannessofthesamplearestillchallengingprob- thresholddefinedbytheauthors.Therefore,thismethodbydef- lems that depend on the sample selection-method, instrument initionselectsacleanbutnotacompletesample. limits, and telescope time. These problems make the compari- In this work, we aim to develop a technique based purely sonbetweenobservationalandtheoreticalworksevenmoredif- on photometric data to select star-forming galaxies. This type ficult.Forexample,Campbelletal.(2014)comparedthestellar of selection allows us to directly compare our results with the massofGALFORMgalaxiespredictedbythemodelwiththose previously mentioned works that also used a drop-out tech- obtained through the fit of their predicted broad-band colors. nique to select their BzK, LBG, etc. samples. We use the They foundthat both quantitiesdiffer for an individualgalaxy, uniform separation between two contiguous medium bands of hencetheclusteringofmass-selectedsamplescanbeaffectedby the ALHAMBRA survey to register the redshifted position of systematicbiases.Therefore,mass-selectedsamplesmightpro- the3646ÅBalmerjumpwithintheopticaldomain,allowingus vide erroneousconclusionsregardingtheir progenitorsand de- toselectgalaxysamplesintheredshiftrange0.5<z<2. scendants.Inaddition,theevolutionofscalingrelationsiscon- We based ourtwo-colorselection techniqueon theBruzual strained with observations of galaxy samples that are selected & Charlot (2003) models and applied it to the GOLD withluminosityorstellar-massthresholdsandarelocatedatdif- ALHAMBRA catalogs (Molino et al. 2014). In the following, ferent redshifts, which does not necessarily constitute causally the galaxies selected by this method are dubbed Balmer-jump connectedpopulations(i.e.,theydonotfollowa progenitor-to- Galaxies(BJGs), and theirphysicalpropertiesare investigated. descendantrelation).Clustering-selectedsamplesovercomethis Basedonaclusteringstudy,wefindtheprogenitorsanddescen- problembecauseinahierarchicalclusteringscenario,acorrela- dantsofgalaxiesinthesesamples,allowingustostudytheevo- tionanalysisallowsustoestimatethebiasandhencestatistically lutionoftheproperties,derivedthroughaSEDfit,asafunction determine the progenitors and descendants of galaxy samples. of redshift for halos of a certain mass. This paper is organized The bias parametermeasuresthe clusteringdifference between asfollows:inSect.2wesummarizetheALHAMBRAobserva- the galaxy spatial distribution and underlying dark-matter dis- tions and introduce the nomenclature of the ALHAMBRA fil- tribution. Thus, it relates the typical mass of halos hosting the ter system used throughout the paper. In Sect. 3 we describe galaxies(Shethetal.2001).Hence,measuringitingalaxysam- the selection method and justify it on the Bruzual & Charlot ples at differentredshiftsdetermineswhether we are following (2003) models. The BJG samples are defined here. In Sect. 4 theevolutionofbaryonicprocessesoccurringinhalosofsimilar each galaxySED is modeled using iSEDfit(Moustakaset al. massesornot.Thisfactisofextremeimportancebecauseonceit 2013), and the physical properties of each sample are charac- isdetermined,wecanusethemultiwavelengthdatatostudythe terized as a whole. In Sect. 5 the clustering properties are cal- evolutionofthebaryonicprocessesatcertainhalomass,estab- culated,and in Sect. 6 we discuss the main results. Finally, we lishingadirectlinkbetweenobservationsandgalaxyformation concludeinSect.7.Throughoutthepaper,weuseastandardflat models.Padillaetal.(2011)selectedearly-typegalaxiesaccord- cosmology with H = 100 hkms−1Mpc−1, Ω (z = 0) = 0.3, 0 m ing to their clustering and luminosity function in the MUSYC ΩΛ(z = 0) = 0.7,σ8 = 0.824±0.029,andthe magnitudesare survey.Sofar,nostudythatselectsstar-forminggalaxiesaccord- expressedintheABsystem. ingtheirclusteringandluminosityfunctionhasbeenreported. Star-forminggalaxiesareofparticularinterestbecausethey 2. Data:theALHAMBRA survey allowustostudythemechanismsthatswitchthestarformation onoroffanditsevolutionwithredshift.Consideringthelackof The Advanced Large Homogeneous Area Medium-Band widespectroscopicsurveysinthesenseofwavelengthcoverage RedshiftAstronomical(ALHAMBRA1)surveyprovidesakind andsurveyedarea,themajorityofthestar-forminggalaxysam- of cosmic tomography for the evolution of the contents of the pleshavebeenchosenusingtwo-colorselectiontechniques.The Universeovermostofthe cosmichistory.Benitez etal. (2009) so-called drop-out technique is based on recording the differ- have especially designed a new optical photometricsystem for encebetweentwo distinctpartsofthespectrumthatgeneratea theALHAMBRAsurveythatmaximizesthenumberofobjects breakonit(e.g.,the912ÅLymanbreakandthe3646ÅBalmer withanaccurateclassificationbySEDandphotometricredshift. jump). This difference is strong enough that it has been mea- Itemploys20 contiguous,equal-width∼310Å, mediumbands sured in broad bands, selecting star-forming galaxies at early that cover the wavelength range from 3500 Å to 9700 Å, plus periods of the Universe (z > 1.4), such as the BzK, BX, BM, the standard JHK near-infraredbands. Moleset al. (2008)and DOGs,andLBGs(Steideletal.1996;Daddietal.2004;Steidel Aparicio Villegas et al. (2010) presented an extensive descrip- etal.2004;Deyetal.2008;Infanteetal.2015).Severalauthors tion of the survey and filter transmission curves. The observa- havemeasuredthebias(Gawiseretal.2007;Blancetal.2008; tionsweremadeatCalarAltoObservatory(CAHA,Spain)with Guaitaetal.2010)bydeterminingthemassofthehaloinwhich the3.5mtelescopeusingthetwowide-fieldimagersintheop- each galaxy sample resides and by connecting the progenitors tical(LAICA)andNIR(Omega-2000).Thetotalsurveyedarea anddescendantsofthesegalaxysamples.Otherworksselected is2.8deg2 distributedineightfieldsthatoverlapareasofother the samples by fully relying on their photometric redshift and surveys such as SDSS, COSMOS, DEEP-2, and HDF-N. The the physical properties determined through fitting the spectral typical seeing of the optical images is 1.1arcsec, while for the energy distribution (SED; Tomczak et al. 2014), far-IR lumi- NIRimagesitis1.0arcsec.Fordetailsaboutthesurveyanddata nosity (Rodighiero et al. 2011), or Bayesian approaches such releasewerefertoMolinoetal.(2014)andCristóbal-Hornillos as iSEDfit (Moustakas et al. 2013). Recently, Viironen et al. etal.(2009).We hereusedthepublicGOLDcatalogs2,which (2015) implemented a method in the ALHAMBRA survey to containsdataofsevenoutoftheeightALHAMBRAfields.The select galaxy samples using the probability distribution of the magnitudelimits ofthese catalogsare (cid:6)m (cid:7) ∼ 25 forthe four AB photometric redshift (zPDF). The quality of the detailed SED distribution,providedbythemediumbandsoftheALHAMBRA 1 http://www.alhambrasurvey.com survey,allowedthemtoperformanaccuratestatisticalanalysis. 2 http://cosmo.iaa.es/content/ALHAMBRA-Gold-catalog A132,page2of14 P.TroncosoIribarrenetal.:TheevolutionofBalmerjumpselectedgalaxiesintheALHAMBRAsurvey Table1.NameandeffectivewavelengthofeachALHAMBRAfilter. Name λeff Name λeff Name λeff [nm] [nm] [nm] U 365.5 U 396.5 B 427.5 1 2 3 B 458.5 B 489.5 B 520.5 4 5 6 B 551.5 R 582.5 R 613.5 7 8 9 R 644.5 R 675.5 R 706.5 10 11 12 I 737.5 I 768.5 I 799.5 13 14 15 I 830.5 z 861.5 z 892.5 16 17 18 z19 923.5 z20 954.5 Fig.1. ALHAMBRA filter transmission curve covering the optical to NIR. From top to bottom, the red curves show the typical redshifted Notes.Columns1,3,and5indicatetheadoptedfilternamethatisused spectrum of a star-forming galaxy at z ∼ 1.5, z ∼ 1.0 and z ∼ 0.5. throughoutthepaper;Cols.2,4,and6showtheeffectivewavelengthof ThegrayshadedareasmarkthepositionoftheALHAMBRAfiltersR , eachALHAMBRAband. 8 I ,andz ,whichhighlightsthepositionoftheBalmerjump.Toselect 14 20 theBJGsamples,weusethefilterdirectlyredwardofthegrayfilterto sampletheendpositionoftheBalmerjump. blue bands and from (cid:6)m (cid:7) ∼ 24.7 mag to 23.4 mag for the AB red bands. The NIR limits at AB magnitudesare J ≈ 24 mag, H ≈ 23 mag, and K ≈ 22 mag. In the following, the fil- ter nomenclature presented in Table 1 is used and the ALH-4 and showed that the data are complete until K = 20 or VEGA andALH-7fields are excludedfromour analysis.Previousau- K = 21.8. We have checked the K-band numbers counts AB thorshaveshownthatoverdensitiesresideinthesefields,which of the 48 individual pointing that comprise the ALHAMBRA might alter the redshift distribution of the selected samples as GOLDcatalog.EverysinglepointingtendstofallatK ∼ 22. AB wellasthe clusteringmeasurements(see Sect. 5).Arnalte-Mur WeusedtheK =22limittodefineourcompletesamples. AB etal.(2014)obtainedtheclusteringpropertiesofALHAMBRA galaxiesand studied the sample variance using the seven inde- 3.1.Selectionofstar-forminggalaxies pendentALHAMBRAfields.Theyquantifiedtheimpactofin- dividualfields on the final clustering measurements. Using the WeusedthemodelsofBruzual&Charlot(2003)convolvedwith Norberget al. (2011)method,they determinedtwo outliersre- theALHAMBRAfiltersystemtodefineatwo-colorcriterionto gions,whicharetheALH-4andALH-7fields.Inaddition,part select star-forming galaxies analogously to the work of Daddi oftheALH-4fieldspatiallycorrespondstotheCOSMOSfield. etal.(2004).Theycreatedthe BzK color-selectiontechniqueto Previous works have shown that this region is dominated by cull star-forming galaxies at z > 1.4. This technique is based somelarge-scalestructures(LSS),themostprominentarepeak- on the Balmer jump, which is an indicator of recent star for- ing at z ∼ 0.7, and 0.9 (Scoville et al. 2007). These structures mation. They used the models of Bruzual & Charlot (2003) to haveX-raycounterpartsandaprobabilityhigherthan30%ofbe- identifythe redshiftedBalmer jumpin the z-band,creating the ingLSS. Guzzoet al. (2007)studied theclusterslocated atthe color-selectioncriterion BzK ≡ (z−K) −(B−z) = −0.2; center of the LSS (at z ∼ 0.7), while Finoguenovet al. (2007) where BzK ≥ −0.2 selects star-forminAgBgalaxies aAtBz > 1.4. found diffuse X-ray emission in the most compact structures. Ourapproachisanalogoustothe BzK methodinthesensethat There are other LSS foundin the COSMOS field, butthey fall we also used the Bruzual & Charlot (2003) models, but we outoftheredshiftrangeofthesamplesstudiedinthispaper. recordedredshifted Balmer jumpsin various(diverse)medium bands to select galaxies at different redshifts. This situation is illustratedinFig.1,whichshowsthetypicalspectrumofastar- 3. Sampleselection forming galaxy redshifted to z ∼ 0.5, z ∼ 1.0 and z ∼ 1.5. To exclude the stars from the original ALHAMBRA Gold The ALHAMBRA bands are overplotted from the optical to catalog, we used the stellarity index given by SExtractor NIR ranges. We chose the U2-band to sample the region be- (CLASS-STAR parameter C(K)) and the statistical star and fore the Balmer jump instead of the B-band. We selected the galaxyseparation(Molinoetal.2014,Sect.3.6)encodedinthe U2bandbecauseitreachesahighercompletenesslevelthanU1. variablestellarflag(Salh)ofthecatalogs.Throughoutthepresent ObjectsthatwerenotdetectedintheU2 bandwerealsoconsid- paper, we define as galaxies those ALHAMBRA sources with eredinourselection,andtheirmagnitudelimitwasused.Redder C(K) ≤ 0.8 andS ≤ 0.5.TheALHAMBRA Gold catalogis bandsmightsampleweakfeaturesbluewardoftheBalmerjump alh an F814W (i.e., almost an I-band) selected catalog. This band of low-redshift galaxies z < 0.4. The K band is exactly the wascreatedbytheALHAMBRA teamasa linearcombination sameaswasusedbyDaddietal.(2004),whilethez-bandwas ofALHAMBRAbands(seeEq.(5)inMolinoetal.2014),and replaced with the variable Xn-band, which covers the optical itwasusedfortheirsourcedetection.Objectswithfaintfeatures range from 613 Å to 954 Å. This means that X can be any n in this band are notdetected,which mightaffectthe complete- ALHAMBRA filter from R to z (see Table 1). The X -band 9 20 n ness of the selected sample. We discuss this in more detail in samplestheregiondirectlyredwardoftheBalmerjump,hence Sects.4.3and6.1.Thiscatalogiscompleteupto F814W = 23 theU −X colorindicatesthegalaxyredshiftrangedepending 2 n (Molinoetal.2014),henceweusedthislimittobuildourcom- ontheselectedn.Ontheotherhand,theX −K colorregisters n plete sample. Cristóbal-Hornilloset al. (2009) studied the NIR the duration of the star formation age. The different panels in completeness using the early release of the first ALHAMBRA Figs.2and3showthetheoreticalevolution(Bruzual&Charlot field. They determined a change in the slope of the K-band 2003) of the U X K ≡ (X − K)−(U − X ) color as a func- 2 n n 2 n number counts in the magnitude range 18.0 < K < 20.0 tionofredshiftwith9 < n < 20(X = R ,....,z ).Red,green, Vega n 9 20 A132,page3of14 A&A588,A132(2016) Fig.2. Criteria for selecting star-forming galaxies. The solid and dashed lines show the color evolution of star-forming and passive galaxies, respectively. Red,green,andbluedashed linesshow thepassivelyevolvingmodelswithaformationredshiftofz = 2,3,6,respectively.Red, f green,andbluesolidlinesshowmodelsthatevolvewithaconstantstar-formationrateforagesof0.2,1,and2Gyr.Thehorizontaldashedline indicatesthecolorcutU X K ≡(U −X )−(X −K)=0.Ineachpanel,thestar-forminggalaxiesarelocatedabovethisthreshold.Theyellow 2 n 2 n n shaded regions mark the redshift range of the selected BJG samples in each color combination. Dashed vertical linesindicate the redshift 0.5 and1.5. andbluesolidlinesshowtheU X K-colorevolutionofconstant selectgalaxiesatredshiftz >0.5,0.6,0.7,0.8,0.85,0.95,1.05, 2 n star-formationratemodelsforages0.2,1,and2Gyrandredden- 1.1,1.2,1.25,and 1.3.Ineach ALHAMBRA filter setU X K, 2 n ingE(B−V)=0.3.Red,green,andbluedashedlinesshowthe this color condition selects star-forming and passive galaxies U X K-colorevolutionofpassivelyevolvingmodelsforforma- in the wide redshiftrange.Theoretically,these selected passive 2 n tionredshiftofz = 2,3and6.In eachpanel,the star-forming galaxiesalways lie at higher redshift(i.e., Δz > 0.25)than the f models always lie in the region U X K > 0. Consequently, to selected star-forming galaxies. To select galaxies in a narrow 2 n select a sample at redshift higher than z ∼ 0.5, a combination redshiftrange,weusedmorethanonecolorcondition;firstthe thatinvolvesaX filterredderthanR mustbeused,asweshow (U −X )−(X −K)>0conditionwithn= j,toselectallgalaxies n 9 2 n n inthebottompanelsofFigs.2and3. abovecertain redshiftz , and second we subtractedthe higher j redshiftsamples,selectedwithn≥ j+1(passiveandstarform- 3.2.BJGsamples ing),fromthefirstgalaxyselection.Forexample,ourlowestred- shiftsamplewasselectedbyimposingtheconditionU R K =0 2 9 WeusedtheconditionU X K ≡(U −X )−(X −K)>0andthe (i.e.,selectgalaxiesatz > 0.5)andsubtractingthegalaxysam- 2 n 2 n n homogeneouscoverageintheopticalrangeoftheALHAMBRA ples selected by U X K > 0 with n ≥ 10, which selects all 2 n filter system, where n ranges from 9 to 20 (X = R ,..,z ), to galaxiesat z > 0.6. In this way, only the star-forminggalaxies n 9 20 A132,page4of14 P.TroncosoIribarrenetal.:TheevolutionofBalmerjumpselectedgalaxiesintheALHAMBRAsurvey Fig.3. Criteria for selecting star-forming galaxies. The solid and dashed lines show the color evolution of star-forming and passive galaxies, respectively. Red,green,andbluedashed linesshow thepassivelyevolvingmodelswithaformationredshiftofz = 2,3,6,respectively.Red, f green,andbluesolidlinesshowmodelsthatevolvewithaconstantstar-formationrateforagesof0.2,1,and2Gyr.Thehorizontaldashedline indicatesthecolorcutU X K ≡(U −X )−(X −K)=0.Ineachpanel,thestar-forminggalaxiesarelocatedabovethisthreshold.Theyellow 2 n 2 n n shaded regions mark the redshift range of the selected BJG samples in each color combination. Dashed vertical linesindicate the redshift 0.5 and1.5. in the redshift range 0.5 < z < 0.6 were selected (see yellow 4. PhysicalpropertiesoftheBJGsamples shadedregionin Figs.2 and3).Followingthismethodandus- In this section, we characterize each BJG sample by providing ing the homogeneous separation between each ALHAMBRA a mean characteristic value of its redshift, absolute magnitude, band, we selected eleven star-forming galaxy samples peak- ingatz ∼ 0.55,0.65,0.75,0.8,0.9,1.0,1.05,1.15,1,2,1.25,and stellarmass,age,starformationrate,etc. z∼1.4bycullingthegalaxiesthatsatisfiedthestar-formingcri- terionU X K ≡(U −X )−(X −K)>0,whereX =R ,...,z 2 n 2 n n n 9 19 4.1.Photometricredshifts andcleanedof higherredshiftgalaxiesby subtractingthe sam- plesthatsatisfiedU X K > 0,withn ≥ 10,...,20.Table2sum- WeusedtheBayesianphotometricredshift(zBPZ)publishedin 2 n marizesthepropertiesoftheselectedsamplesusingthemethod the Gold ALHAMBRA catalogs (Molinoet al. 2014)to verify describedabove.Weensuredthatallsamplesareroughlyinde- our selection method. The BPZ code was optimized to deter- pendentofeachotherbyremovingthehigh-redshiftsamplesthat minephotometricredshifts,fordetailsonthezBPZcalculations wereselected usingthe X filters withn ≥ j+1 (see Col. 3of see Molino et al. (2014) and Benitez et al. (2009). Figure 4 n Table2)fromthesampleselectedwiththefilterXn=j. shows the zBPZ distribution of the eleven BJG samples. To A132,page5of14 A&A588,A132(2016) Table2.PropertiesoftheBJGselectedsamples. Table3.MedianphysicalpropertiesoftheBJGsamples. Sample Initialset Cleansets N (cid:6)z(cid:7) Name N zBPZ (cid:6)M∗(cid:7) SFR Age name U2XnK U2XnK Two-color M(cid:4) M(cid:4)yr−1 [Gyr] BJG R n≥10 5489 0.5–0.6 BJG 5489 0.52+0.16 10.15+0.11 0.08+0.22 5.52+1.62 1 9 1 −0.14 −0.13 −0.23 −1.87 BJG2 R10 n≥11 5174 0.6–0.7 BJG2 5174 0.64−+00..1181 10.25−+00..1132 0.27−+00..2223 5.13−+11..7560 BJG3 R11 n≥12 4497 0.7–0.8 BJG3 4497 0.74−+00..1170 10.30−+00..1143 0.38−+00..2243 4.83−+11..6480 BJG4 I12 n≥13 4012 0.8–0.85 BJG4 4012 0.81−+00..1142 10.37−+00..1153 0.42−+00..2265 4.67−+11..6229 BJG5 I13 n≥14 3550 0.85–0.95 BJG5 3550 0.90−+00..1182 10.36−+00..1164 0.49−+00..2254 4.37−+11..5244 BJG6 I14 n≥15 2878 0.95–1.05 BJG6 2878 0.98−+00..1192 10.41−+00..1175 0.52−+00..2275 4.12−+11..4165 BJG7 I15 n≥16 2231 1.05–1.1 BJG7 2231 1.07−+00..2102 10.46−+00..1175 0.63−+00..2265 3.83−+11..4009 BJG8 I16 n≥17 2325 1.1–1.2 BJG 2325 1.15+0.12 10.45+0.16 0.69+0.26 3.63+1.05 BJG9 z17 n≥18 2058 1.2–1.25 BJG8 2058 1.24−+00..2114 10.45−+00..1198 0.72−+00..2275 3.41−+11..3061 BJG10 z18 n≥19 2140 1.25–1.3 BJG9 2140 1.29−+00..2251 10.38−+00..2109 0.72−+00..2254 3.16−+11..3025 BJG11 z19 n=20 3391 1.3–1.5 BJG10 3391 1.33−+00..3238 10.34−+00..2230 0.72−+00..2223 2.99−+11..2095 11 −0.42 −0.24 −0.21 −1.27 Notes.Column1,nameoftheselectedsample;Col.2,initialfilterset Notes.Column1,nameoftheselectedsample;Col.2,numberofgalax- usedforselection;Col.3,secondfiltersetsusedtosubtracthigherred- iesselected;Col.3,BPZphotometricredshift;Col.4,logarithmofthe shift galaxies from the initial sample; Col. 4, number of galaxies se- stellarmass(ChabrierIMF);Col.5,logarithmofthestarformationrate; lected;Col.5,expectedredshiftrangeofthesample. Col.6,galaxyage. BJG ,BJG ,BJG ,BJG , and BJG samples. The median of 2 4 6 8 10 zBPZ distribution, reported in Table 3, clearly agree with the medianoftheexpectedredshiftrangeestimatedthroughthetwo- colorselectioncriteriareportedinTable2. 4.2.SEDfitting ForeachgalaxyoftheBJGsamples,wefitthe20opticalbands plusthethreeNIRbandsusingtheBayesianSEDmodelingcode iSEDfit (Moustakas et al. 2013). After we fixed the redshift to the best-fit value given by BPZ in the ALHAMBRA cata- log (Molino et al. 2014), iSEDfitcalculated the marginalized posterior probability distributions for the physical parameters in a certain model space that was previously defined. Using a MonteCarlotechnique,wegenerated20,000modelSEDswith delayed star formation histories SFH ∼ te−t/τ, where τ is the star formationtimescale. The SEDs were computedemploying the flexible stellar population synthesis models (FSPS, v 2.4; Conroyetal.2009;Conroy&Gunn2010)thatarebasedonthe miles (Sanchez-Blázquezetal.2006)andBasel(Lejeuneetal. 1997,1998;Westeraetal.2002)stellarlibraries.Weassumeda Chabrier(2003)initialmassfunctionfrom0.1−100 M(cid:4) andthe time-dependentattenuationcurveofCharlot&Fall(2000).We adopteduniformpriorsonstellarmetallicityZ/Z(cid:4) ∈ [0.04,1.0], galaxy age t ∈ [0.01,age(z )] Gyr, rest-frame V-band at- BPZ tenuation A ∈ [0−3] mag, and star formation timescale τ ∈ V [0.01,age(z )]Gyr,whereage(z )istheageoftheUniverse Fig.4.Photometricredshiftdistributionoftheselectedsamples.Upper BPZ BPZ ateachgalaxy’sphotometricredshift. panel:red,pink,yellow,lightgreen,cyan,andbluehistogramsshowthe zBPZ distribution of the BJG ,BJG ,BJG ,BJG ,BJG , and BJG Figures9and10showtheSEDsofagalaxyrandomlypicked 1 3 5 7 9 11 samples, respectively. Lower panel: magenta, orange, dark green, fromeachBJGsample.Thegreenfilledsymbolsshowthedata green, and light blue histograms show the zBPZ distribution for the and their error bar. The model that minimize the χ2 is drawn BJG2,BJG4,BJG6,BJG8, and BJG10 samples, respectively. The ver- with the red line. The universe of models, generated using the tical colored lines show the median of the zBPZ distribution for the aforementionedsetup,and scaled by their reducedχ2 is shown selectedsamples.Thelabellocateintheupper-rightcornerreportsthe withtheblueshading. medianredshift. ForeachBJGsampleandcertainphysicalproperties,theme- dian value of the sum of the posterior probability distributions better visualize the distribution tails, the upper panel shows were calculated and are reported in Table 3. The uncertainties the BPZ distribution of the BJG ,BJG ,BJG ,BJG ,BJG , indicatethe1σconfidencelevel;toaccountforasymmetricdis- 1 3 5 7 9 and BJG samples, while the lower panel presents the tributions,wedeterminedthepercentiles16and84. 11 A132,page6of14 P.TroncosoIribarrenetal.:TheevolutionofBalmerjumpselectedgalaxiesintheALHAMBRAsurvey Table4.MedianphysicalpropertiesoftheBJGKsamplesconsidering theabsolutemagnitudelimitK=−21.2. Name N zBPZ (cid:6)M∗(cid:7) SFR Age M(cid:4) M(cid:4)yr−1 [Gyr] BJGK 3322 0.56+0.18 10.50+0.10 0.24+0.29 5.80+1.43 1 −0.14 −0.11 −0.30 −1.87 BJGK 3778 0.67+0.11 10.46+0.11 0.40+0.27 5.31+1.37 2 −0.16 −0.12 −0.26 −1.74 BJGK 3576 0.75+0.10 10.45+0.12 0.46+0.26 4.92+1.31 3 −0.12 −0.13 −0.26 −1.67 BJGK 3480 0.83+0.11 10.45+0.13 0.49+0.27 4.69+1.24 4 −0.09 −0.14 −0.28 −1.60 BJGK 3150 0.92+0.10 10.44+0.14 0.55+0.26 4.37+1.19 5 −0.12 −0.15 −0.26 −1.52 Notes.Column1,nameoftheselectedsample;Col.2,numberofgalax- Fig.5.Absolutemagnitudeasafunctionofredshift.Thecoloreddots iesselectedfromtheBJGsamplesaftertheK-bandabsolutecut;Col.3, representagalaxyofeachBJGsample,usingthesamecolorcodethan BPZphotometricredshift;Col.4,stellarmass(ChabrierIMF);Col.5, in Fig. 4. The horizontal solid line shows the absolute magnitude cut starformationrate;Col.6,galaxyage. K =−21.2. abs redefinethefirstfiveBJGsamplesinthefollowingbyselecting 4.3.ComparisonbetweenBJGsamples allgalaxiesbrighterthan K = −21.2.Furthermore,the BJG abs n samples with n ≥ 6 are not consideredin the following analy- Inthissubsection,wecomparethepropertiesderivedfromSED sis.Table4presentsthepropertiesofthefinalsampledefinition fitting of the BJG samples for galaxies of similar K-band ab- (BJGK ,withn < 6),whichtakesintoaccounttheabsolutelu- solute magnitude, i.e. similar stellar mass. Therefore, in addi- minositny cut (K = −21.2). In Fig. 6 the comparison of the tion to restricting the BJG samples within the magnitude sur- abs physicalpropertiesof galaxieswith similar K-band luminosity veylimit (see Sect. 3), we appliedan absolutemagnitudelimit is shown. Red squares show the median value of the probabil- for all samples. In the following, we determine the most ap- itydistributionforeachBJGK sample(withn < 6),whilethe propriateabsolutemagnitudelimitthatallowsustobuildthese n coloredareasindicatetheirdispersion.Theerrorbarsshowthe samples. Figure 5 shows the K-band absolute magnitude, ob- medianerrorofthepropertiesderivedthroughtheiSEDfit.In tainedthroughtheiSEDfit,asafunctionofBPZredshift.The Sect.6.3theseresultsarediscussed. red, orange, green, cyan, light blue, and blue dots indicate the galaxiesintheBJG ,BJG ,BJG ,BJG ,BJG ,andBJG sam- 1 4 6 7 9 11 pleswithapparentmagnitudebrighterthanthemagnitudesurvey 5. Halomassesthroughclusteringanalysis limitK =22,respectively.Wecannotethatpartlyasaresultof Inhierarchicalclustering,structuresbuildupintimefromsmall the Malquist bias, the farthest sample BJG is roughly com- 11 pleteonlyuntilK =−22.5(bluedots).However,bychoosing densityfluctuations.Smallstructuresagglomeratetobuildlarge abs structures. Dark matter halosare biased tracers of the underly- this bright absolute limit to compare all BJG samples, we re- ing matter density field. Massive halos lie in higher and rarer strictourstudytothemostmassiveandbrightobjectsandalso densitypeaksandare more clusteredthanlower masshalos. If enormously reduce the statistics for each sample, especially at agalaxypopulationishostedbyhalosofagivenmass,thenthe low redshift. Hence, we decided to study all objects brighter than K = −21.2, which corresponds to the absolute limit clusteringamplitudeof the galaxypopulation,comparedto the abs where the BJG sample at z ∼ 1 is complete and comparable, expecteddarkmatterclusteringatthesameredshift,canbeused 5 toderivethetypicalhalomasscorrespondingtothatgalaxypop- in termsof K luminosity,to theotherlowerredshiftBJG abs n<6 ulation.Thisisencapsulatedinthebiasparameter,b,definedas samples.InFig.5theblacksolidlineshowstheabsolutemag- nitudelimitK = −21.2.Wechosethez ∼ 1limitbecausethe b2 =ξgal(r)/ξDM(r),whereξgalandξDMarethetwo-pointspatial abs correlationfunctionsforgalaxiesanddarkmatter,respectively. ALHAMBRAGolddataarean F814W (i.e.,almostan I-band) selected catalog andthusare less sensitive to galaxiesat z ≥ 1 Thecorrelationfunctionatacertainredshift,ξ(r,z),canbechar- withapronouncedBalmerjump.Forgalaxiesatz≥1,thespec- acterizedwithapowerlawandacorrelationlength,r0,through ξ(r,z) = (r/r (z))γ,whereγisthepowerindex.Forsparsesam- tralregiondirectlybluewardoftheBalmerjumpisbarelyornot 0 plesandatsmall scalesthis representationis typicallyused.In at all detected in the F814W because the Balmer jump falls in largersurveysandsimulationsthecorrelationfunctionisusually the I band and the region directly blueward of it falls in the 13 modeledbycombiningtermsforgalaxiesinthesame haloand R band.Since the F814W image detectionis a linear combi- 12 indifferentones(Zehavietal.2011;Contrerasetal.2013). nationinvolvingtheR band,thesetypesofgalaxiesarebarely 12 Inthefollowing,wemeasurethebiasandmassesofthehalos or notat all detected as sourcesin the final catalog.We expect thathostthe BJGK samples.Specifically,foreach BJGK a deeper selection of galaxieswith a pronouncedBalmer jump n<6 n<6 atz < 1.Hence,theBJGsamplesatz (cid:2) 1 selectedfrom Xn≤13 sample, we calculate the two-point angular correlation func- tion, and then, using the Limbers deprojection, given a red- bandsareoptimizedtobecompleteaccordingtotheF814Wse- lection,whereasthesamplesatz ≥ 1areincomplete.Thelevel shift distribution and a cosmological model, we calculate the otiffyinbceocmaupsleettehneersesmofaythbeeBmJGanny≥6effseacmtsplwesoriksindgiffitocgueltthtoerq(uea.gn.-, cbo2r(cid:13)relσa28ti,goanl(rl0e,nzg)t/hσ.28T,DhMe(bz)i,aswphaerraemσe8t,gearlf(oσl8lo,DwMs)fisrotmherr0oothtrmoueganh theF814WbandisalinearcombinationofALHAMBRAbands square fluctuation amplitude in 8 h−1 Mpc spheres of galaxies andbluewardregionsofthe Balmerjumpthatareundetectable (dark matter). The halo mass is obtained from the bias by us- asaresultofintrinsicgalaxypropertiessuchasredshiftanddust ing the modelsof Sheth et al. (2001).We chose this procedure excess). To minimize the incompletenessof our samplesin the and the 8 h−1 Mpc scale to directly compare our results with comparisonof the physicalpropertiesat differentredshifts, we previous works that connected progenitors and descendants of A132,page7of14 A&A588,A132(2016) Fig.6.EvolutionofthephysicalpropertiesoftheBJGK samples.Thepanelsshowthemedianoftheprobabilitydistributionof(toptobottom n<6 andlefttoright)thespecificstarformationrate,starformationrate,age, B-bandluminosity, K-bandluminosity,andstellarmass.Redsquares show the evolution of BJGK galaxies brighter than K = −21.2. The colored areas represent the dispersion of each sample. The error bars n<6 correspondtothemedianerrorofthephysicalpropertiesdeterminedbytheiSEDfit.DashedlinesshowthefittotheCANDELSanalysistaken fromvanDokkumetal.(2013)forgalaxiesofstellarmassessimilartotheMilkyWay. samples of star-forming galaxies (see Fig. 7, Adelberger et al. to calculate r we used the power-law approximation ξ(r,z) = 0 2005;Ouchiet al. 2005;Kashikawa et al. 2006;Gawiser et al. (r(z)/r (z))−γthatdependsonthespatialseparationandredshift. 0 2007;Hildebrandtetal.2007;Hayashietal. 2007;Blancetal. Likewise,thetwo-pointangularfunctionisw(θ,z)= A (z)θ(1−γ), w 2008;Hartleyetal.2008;Yoshidaetal.2008;Guaitaetal.2010; where A (z) istheangularamplitude.By fittingthe correlation w McCracken et al. 2010; Lin et al. 2012). More detailed treat- function to this power law, between 0.005 and 0.2 degree, the ments,suchasconsideringbothhalotermsanddifferentscales, A (z)isinferred.TheangularamplitudesarereportedinTable5. w arebeyondthescopeofthispaper. To calculate the spatial from the angularfunction,we used the Limber (1953)inversion,which requiresa redshift distribution N(z),andassumedacosmologicalmodel.Limber’sinversionin- 5.1.Angularcorrelationandcorrelationlength volvessolvingthefollowingintegral(Kovacetal.2007): To determine the angular correlation function, we used the (cid:2) 2 LandyandSzalay(1993)prescription rγ = (cid:2)Aw(c/H0)[ N(z)dz] , (2) 0 C F(z)D1−γ(z)N2(z)g (z)dz w(θ)=(N −2N +N )/N , (1) γ θ d gg gr rr rr where N and N are the number of pairs at a separation θ wherethe cosmol(cid:3)ogyplaysa role in the Hubbleparameter H0, of galaxigegs in thercratalog and points in a random catalog with gd(z) = (1 + z)2 1+Ωmz+ΩΛ[(1+z)−2−1], and in the an- gular diameter distance D . The parameter C depends on the the same layout as the galaxy sample, respectively. N is the θ γ cross number of points between the galaxy and randomgr distri- powerindexsuchthatCγ =Γ(0.5)Γ[Γ0.[50(.γ5−γ]1)].Tobeconsistentin- butions. For our calculationswe consideredfive ALHAMBRA ternally and compare our results with other works (Adelberger fields(see Sect. 2) thatencompass36 pointingcatalogs.Errors et al. 2005;Ouchi et al. 2005;Kashikawa et al. 2006;Hayashi were estimated by a jackknife method, which has been shown etal.2007;Hildebrandtetal.2007;Gawiseretal.2007;Blanc to be a robust error estimator (Zehavi et al. 2011; Cabré et al. etal.2008;Hartleyetal.2008;Yoshidaetal.2008;Guaitaetal. 2007),althoughitcanoverestimatethevarianceonsmallscales 2010;McCrackenetal.2010;Linetal.2012),thevalueofγwas (Norberg et al. 2009). We calculated the angular correlation fixedtothecanonicalγ=1.8.Thisvalueisfullyjustifiedbyex- function 36 times, each time eliminating one catalog out of perimentswhereγwasleftfreeandisconsistentwiththecluster- the36available.Theuncertaintywasestimatedasthevariance ingofluminosity-selectedALHAMBRAsamples(Arnalte-Mur ofw(θ).WefollowedthemethodofInfante(1994),Quadrietal. etal.2014).F(z)accountsfortheredshiftevolutionofthecorre- (2007)tocorrectfortheintegralconstraint.Asmentionedabove, lationfunction,whereF(z)=(1+z)−(3+(cid:10)),andweused(cid:10) =−1.2, A132,page8of14 P.TroncosoIribarrenetal.:TheevolutionofBalmerjumpselectedgalaxiesintheALHAMBRAsurvey Fig.7.Evolutionofthebiasfactor,calculatedwiththevarianceat8h−1 Mpc,forsamplesofstar-forminggalaxies.Reddiamondsindicatethe BJGK samples. At z ∼ 2, the gray, blue, red, green, and light green squares show the sBzK samples of Hayashi et al. (2007), Blanc et al. n<6 (2008),Hartleyetal.(2008),Linetal.(2012),andMcCrackenetal.(2010),respectively.YellowsquaresshowtheBMandBXselectedgalaxies ofAdelbergeretal.(2005).Atz∼3,theblue,green,andredsquaresindicatetheLBGsselectedbyYoshidaetal.(2008),Kashikawaetal.(2006), Hildebrandtetal.(2007),Ouchietal.(2005),respectively.GraysquaresshowtheLy-αemittersselectedbyGuaitaetal.(2010),Gawiseretal. (2007).Thelinesshowthebiasevolutionfordifferenthalomasses(Shethetal.2001),whichareindicatedbeloweachlineattherightsideofthe panel. Table5.ClusteringpropertiesoftheBJGKsamples. Name N (cid:6)z(cid:7) (cid:6)K (cid:7) A r Bias logM gal abs w 0 h BPZ mag 10−3 [h−1Mpc] (r=8h−1Mpc) [h−1M(cid:4)] BJGK 3332 0.56+0.2 –22.3 4.2±0.2 4.67±0.33 1.36±0.09 12.71+0.13 1 −0.1 −0.15 BJGK 3778 0.67+0.1 –22.3 4.3±0.5 4.58±0.39 1.41±0.11 12.66+0.15 2 −0.2 −0.17 BJGK 3576 0.75+0.1 –22.4 4.0±0.2 3.79±0.25 1.24±0.07 12.27+0.14 3 −0.1 −0.16 BJGK 3480 0.83+0.1 –22.4 4.2±0.1 4.15±0.23 1.40±0.07 12.45+0.10 4 −0.1 −0.12 BJGK 3150 0.92+0.1 –22.5 3.7±0.2 4.01±0.23 1.40±0.07 12.37+0.11 5 −0.1 −0.12 Notes.Column1,nameoftheselectedsample;Col.2,numberofgalaxiesselectedfromtheBJGsamplesaftertheK-bandabsolutecutandimage mask;Col.3,BPZphotometricredshift;Col.4,mediumK-bandabsolutemagnitude;Col.5,amplitudeofcorrelation;Col.6,correlationlength; Col.7,biasfactorcalculatedwiththevarianceat8h−1Mpc.Column8,logarithmofthehalomassinunitsof[h−1M(cid:4)]. whichcorrespondstothevalueadoptedforaconstantclustering was taken from Peebles (1980) (Eq. (59.3)), where J is a 2 in comoving coordinates(Quadri et al. 2007). To calculate the parameter defined in terms of γ as J = 72/[(3 − γ)(4 − 2 correlationlengtherror,wetooktheerrorcontributionfromthe γ)(6−γ)2γ]. We fixed the spatial scale at 8 h−1 Mpc such that amplitu√deoftheangularcorrelationfunction,thePoissonianer- ξgal(8,z) = (r0(z)/8h−1Mpc)γ.Theevolutionofthedarkmatter rors(∼ N(z)), and the effects of the photometricredshifterror density variance in a comoving sphere of radius 8 h−1 Mpc is inshiftingandbroadeningN(z)intoaccount. σ2 (8,z) = σ D(z), where D(z) is the linear growth factor at DM 8 redshiftz.The biasmeasuredforeachBJGK sampleata scale of8h−1MpcisreportedinTable5. 5.2.Biasandmassmeasurements Figure 7 shows the bias factor as a function of redshift for In turn we estimated the bias parameter, b, from the correla- theBJGKsamples(reddiamonds).Thelinesshowthebiasevo- tion length. As pointed out above, b is related to r0 through lutionfordifferenthalomasses(Shethetal.2001),whicharein- the spatial correlation function by b2 ≈ ξgal(r,z)/ξDM(r,z) or dicatedbeloweachlineattherightsideofthepanel.TheBJGK b2 ≈ σ2 (r,z)/σ2 (r,z). The numeratorξ (r,z) ≈ J σ2 (r,z) samples approximately follows the bias evolution of halos of gal DM gal 2 gal A132,page9of14 A&A588,A132(2016) oftheU andU ,whichreaches97%and99.7%,respectively. 1 2 Since U has a higher detection level than U , we tailored the 2 1 selectiontechniqueusingtheU band. 2 To select samples at z < 0.5, we tried to use filters bluer than the R , whose central wavelength is lower than 613.5 Å. 9 Nevertheless, the theoretical evolution of the color U X K 2 n<9 of passive and star-forming galaxies, based on the Bruzual & Charlot (2003) models, tends to occupy the same locus in the color-redshift(U X K−z)plane.Hence,itdoesnotallowsep- 2 n<9 arating the star-formingfrom the passive galaxiesas clearly as fortheU X K selectionwithn>9(seeFigs.2and3). 2 n ThenumberdensityofeachBJGsampledecreaseswithred- shift, probably because of the nature of these objects or be- causeoftheMalquist-biasselectioneffectintheU -band,which 2 we didnotconsiderhere.Inaddition,we verifiedourselection methodwith the BPZ photometricredshifts,whose uncertainty increaseswithredshift,andthereforethenumberdensitiesmight beunderestimatedaccordingtoδ (BPZ)=0.014z(BPZ). z Based on this color technique (visual inspection of Figs. 2 and 3), the redshift distributions should be narrower than in Fig. 4. The distributions can become wider if the errors of the photometricredshiftsareproperlytakenintoaccount.Themean Fig.8. Halo mass as function of redshift for the BJGK samples. n<6 formal BPZ errors of the BJG and BJG samples are 0.02 Squaresshowthemedianvaluesofthehalomass,whilepolygonsdraw 1 11 and 0.03, which corresponds to 20% and 30% of the total ex- themasslimittakingintoaccounttheerrorofredshiftandbias. pected width (≤0.1), respectively. We performed simulations considering Gaussian redshift distributions filled randomly us- masses∼1012.5h−1M(cid:4).Finally,wecalculatedthehalomassus- ing the same number of galaxies found in each sample, the ingEq.(8)ofShethetal.(2001),whichrelatesthebiaswiththe expected width (0.1), and the formal BPZ error δz(BPZ) = peakheight,ν=δ (z)/σ(M,z),whereδ isthecriticaloverden- 0.014(1+z(BPZ)).Byperturbingthephotometricredshiftwith sc sc sity computed using the sphericalcollapse model. Here we as- itscorrespondingerrorandrandomlychoosingapositiveorneg- sumedδ =1.69.Then,thehalomassM wasobtainedthrough ativevariance,thewidthofthedistributionsincreasesbyafactor sc h σ(M,z)evaluatedattheredshiftsoftheBJGK sampleslistedin of 1.5 for the first BJG sample and up to twice this value for 1 Table4.InFig.8thehalomassoftheBJGKsamplesasafunc- theBJG11 samplewithincreasingredshift.Clearly,higherpho- tion of redshift is shown. This is calculated with the variance tometric redshift errors increase the distribution width. In the at8h−1Mpc.Squaresshowthemedianvaluesofthehalomass, AHAMBRAdata,BPZtendstolowerprecisionatI < 24.5AB whilepolygonsshowa conservativelimit.Thisdrawsthemass (Molino et al. 2014). The global redshift distribution shows a limitsconsideringthe1σdeviationoftheredshiftandbias.The meanof(cid:6)z(cid:7) = 0.86atthismagnitudelimit. Hencethedistribu- resultsarepresentedinTable5. tionsatz≥0.9tendtobroadenevenmorebecausethephotomet- ricredshifterrorsexceed0.014(1+z).Foradistributionatz∼1 withanoriginalwidthof0.1andδ (BPZ) = 0.03(1+z(BPZ)), z 6. Discussion the width increases up to three times. Outliers that are due to either photometricredshiftor the colorselection techniquecan 6.1.Selectiontechnique alsocontributetothetailsandmayalsobroadenthephotometric The new two-color selection technique U X K based on redshiftdistributions. 2 n ALHAMBRA medium-bandsallows us to extract eleven sam- We also investigated the selection method by creating ples of star-forming galaxies at z > 0.5 in narrower redshift an X band composed of two, three, and four consecutive n rangesthanwaspossibleinpreviousstudies(Daddietal.2004; ALHAMBRAbands.Thisincreasedthestatistics,accuracy,and Steidel et al. 2004; Adelberger et al. 2005). This selection width of the redshiftdistributionof eachsample. Nevertheless, method is based purely on the ALHAMBRA photometric data to fully exploitthe capabilitiesof a multi-mediumbandsurvey and does not depend on the models assumptions, method, or inselectinggalaxiesinsmallredshiftranges,wechosetousea templates used to determine the photometric redshifts or prop- uniquebandinourselectionmethod. ertiesderivedfromSEDfit.Wevalidatedthistechniquewiththe Viironenetal.(2015)haveshownthatispossibletoobtaina Bruzual&Charlot(2003)modelsandalsocheckedotherstellar cleangalaxysampleusingtheredshiftprobabilitydistributions, populationsynthesismodels,whichgiveconsistentresults. which is by definition (intrinsically) not complete. The fact of TheALHAMBRAGOLDdataareanF814W-selectedcata- choosingacertainprobabilitythresholdimpliesthatsomegalax- log,thusitislesssensitivetotheregionsbluewardoftheBalmer ieswill notbeconsideredin thefinalsample.To overcomethe jumpofgalaxiesatz>1.TheR bandonlycontributesto10% problemofpurityversuscompleteness,theysuggestedusingthe 12 ofthefinalF814Wdetectionimage,whiletheI ,andI bands information of the whole zPDF to select each sample. In this 13 14 contribute18%each.Hence,thedetectionintheF814W image workwetailoredandusedatwo-colorcriterionthatconsidered becomespoorerforgalaxiesatz≥1withapronouncedBalmer alltheobservationaldataavailable,selectingacompletesample, jump. butatthesame timewas probablymorecontaminatedthanthe Toavoidcontaminationoflow-redshiftgalaxiesinoursam- samplesselectedaccordingtoredshiftprobabilitydistributions. ples,weusedthebluestbandavailabletosampletheregionblue- We used the zBPZ-BJG distributions to calculate this contam- wardoftheBalmerjump.Hence,westudiedthedetectionlevel ination by estimating how many objects have a BPZ redshift A132,page10of14

Description:
12 Departament d'Astronomia i Astrofísica, Universitat de València, 46100 Burjassot, . the galaxies selected by this method are dubbed Balmer-jump.
See more

The list of books you might like

Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.