The Lyman alpha reference sample: IV. Morphology at low and high redshift

6.9 MB·English

by  L. Guaita

Astronomy&Astrophysicsmanuscriptno.LGuaita˙ArXiv (cid:13)c ESO2015
January13,2015
The Lyman alpha reference sample: IV. Morphology at low and high
redshift (cid:63)
L.Guaita1,J.Melinder1,M.Hayes1,G.O¨stlin1,J.Gonzalez1,G.Micheva1,2,A.Adamo1,J.M.Mas-Hesse4,A.
Sandberg1,H.Ot´ı-Floranes5,14,D.Schaerer3,6,A.Verhamme6,E.Freeland1,I.Orlitova´7,P.Laursen8,J.M.Cannon9,
F.Duval1,T.Rivera-Thorsen1,E.C.Herenz10,D.Kunth11,H.Atek12,J.Puschnig1,P.Gruyters13,andS.A.Pardy14
1 DepartmentofAstronomy,OskarKleinCentreforcosmoparticlephysics,StockholmUniversity,AlbaNova,StockholmSE-10691,
Sweden
2 SubaruObservatory,Hilo,Hawai,USA
3 CNRS,IRAP,14,avenueEdouardBelin,F-31400Toulouse,France
5 4 CentrodeAstrobiolog´ıa(CSIC–INTA),DepartamentodeAstrof´ısica,POB78,VillanuevadelaCan˜ada,Spain
1 5 InstitutodeAstronom´ıa,UniversidadNacionalAuto´nomadeMe´xico,Apdo.Postal106,EnsenadaB.C.22800,Mexico
0 6 GenevaObservatory,UniversityofGeneva,51ChemindesMaillettes,CH-1290Versoix,Switzerland
2 7 AstronomicalInstituteoftheAcademyofSciencesoftheCzechRepublic,Bocˇn´ıII1401/1a,CZ-14100Praha4,CzechRepublic
n 8 DarkCosmologyCentre,NielsBohrInstitute,UniversityofCopenhagen,JulianeMariesVej30,DK-2100Copenhagen,Denmark
a 9 DepartmentofPhysicsandAstronomy,MacalesterCollege,1600GrandAvenue,SaintPaul,MN55105,USA
J 10 Leibniz-InstitutfurAstrophysik(AIP),AnderSternwarte16,D-14482Potsdam,Germany
11 Institutd’AstrophysiquedeParis,UMR7095CNRS&UPMC,98bisBdArago,F-75014Paris,France
0
1 12 Laboratoired’Astrophysique,E´colePolytechniqueFe´de´raledeLausanne(EPFL),Observatoire,CH-1290Sauverny,Switzerland
13 DepartmentofPhysicsandAstronomy,UppsalaUniversity,Box515,75120,Uppsala,Sweden
] 14 CentrodeRadioastronom´ıayAstrof´ısica,UNAM,CampusMorelia,Mexico
A 15 AstronomyDepartment,UniversityofWisconsin-Madison,475N.CharterStreet,Madison,Wisconsin,53706,USA
G
Accepteddate:Dec19th,2014
.
h
ABSTRACT
p
-
o Aims.Withtheaimofidentifyingrest-frameUVandopticalproperties,typicalofLyαemitters(LAEs,galaxieswithEW(Lyα)>20
r Å)atbothlowandhighredshift,weinvestigatedthemorphologicalpropertiesoftheLARSgalaxies,inparticulartheonesthatexhibit
t intenseLyαradiation.
s
a Methods.Wemeasuredsizesandmorphologicalparametersinthecontinuum,Lyα,andHαimages.Westudiedmorphologybyusing
[ theGinicoefficientvsM20andasymmetryvsconcentrationdiagrams.WethensimulatedLARSgalaxiesatz∼2and5.7,performing

thesamemorphologicalmeasurements.WealsoinvestigatedthedetectabilityofLARSgalaxiesincurrentdeepfieldobservations.
1
The subsample of LAEs within LARS (LARS-LAEs) was stacked to provide a comparison to stacking studies performed at high
v
redshift.
7
Results.LARSgalaxieshavecontinuumsize,stellarmass,andrest-frameabsolutemagnitudetypicalofLymanbreakanaloguesin
8 thelocalUniverseandalsosimilarto2 < z < 3star-forminggalaxiesandmassiveLAEs.LARSopticalmorphologyisconsistent
3 withtheoneofmergingsystems,andirregularorstarburstgalaxies.ForthefirsttimewequantifythemorphologyinLyαimages:
2 evenifavarietyofintrinsicconditionsoftheinterstellarmediumcanfavourtheescapeofLyαphotons,LARS-LAEsappearsmall
0 inthecontinuum,andtheirLyαiscompact.LARSgalaxiestendtobemoreextendedinLyαthanintherest-frameUV.Itmeansthat
1. LyαphotonsescapebyforminghaloesaroundHIIregionsofLARSgalaxies.
0 Conclusions.ThestackofLARS-LAELyαimagesispeakedinthecentre,indicatingthattheconditions,whichmakeagalaxyan
5 LAE,tendtoproduceaconcentratedsurfacebrightnessprofile.Ontheotherhand,thestackofallLARSgalaxiesisshallowerand
1 moreextended.ThiscanbecausedbythevarietyofdustandHIamountanddistribution,whichproducesamorecomplex,patchy,
: andextendedprofile,liketheoneobservedforLymanbreakgalaxiesthatcancontributetothestack.Wecannotidentifyasingle
v morphological property that controls whether a galaxy emits a net positive Lyα flux. However, the LARS-LAEs have continuum
i propertiesconsistentwithmergingsystems.
X
r Keywords.techniques:imaging–galaxies:starformation–galaxies:starburst
a
1. Introduction Hayes et al. (2005, 2009) and O¨stlin et al. (2009) devel-
opedamethodtoseparatetherest-frameUVandLyαemission
Originating mainly in recombining gas being ionized by hot O
in Hubble Space Telescope (HST) data. In these papers it was
andBstars,Lymanalpha(Lyα)radiationhasprovedanexcellent
demonstratedobservationallythat,innearbygalaxies(z < 0.1),
probe of star-forming galaxies at both low (e.g., Cowie et al.
Lyαemissionextendsawayfromthestar-formingregionswhere
2011)andhigh(e.g.,Ouchietal.2010)redshift.
the Lyα photons were originally generated, forming the so-
calledLyαhaloes.
Sendoffprintrequeststo:LuciaGuaita,e-mail:[email protected]
(cid:63) Based on observations made with the NASA/ESA Hubble Space Young starburst galaxies are expected to be very bright in
Telescope.Theseobservationsareassociatedwithprogramme12310. Lyα (Partridge & Peebles 1967). For the past 15 years, star-
1
L.Guaita:LARSatlowandhighz
forming galaxies have been successfully detected at z > 2 by ACS/WFC F502N/F505N/F551N, F656N/F716N/F782N)
identifyingtheirstrongLyαemissionline.Theprincipalmethod filters. Lyα maps were generated by estimating the continuum
used is the narrow-band technique (e.g., Cowie & Hu 1998; at rest-frame λ(Lyα) = 1216 Å, through modelling the galaxy
Rhoads et al. 2000; Ouchi et al. 2008; Gronwall et al. 2007; spectrum as a composite population of young stars, old stars,
Nilssonetal.2009):Lyαemitters(LAEs)presentanexcessina and nebular gas. LARS images were published in Hayes et al.
narrowband(coveringtheredshiftedLyαwavelength)withre- (2013) (hereafter Paper 0) and further analysed in Hayes et al.
specttoabroad-bandfilter(coveringtherest-frameUVcontin- (2014)(hereafterPaperII).WefoundthattheLyαemissionpro-
uum). Because Lyα photons are easily absorbed by dust grains file appeared different from the rest-frame UV and it flattened
andarescatteredbyneutralhydrogen(HI),LAEswerethought onscaleslargerthantherest-frameUV.Themajorityofthe14
tobea specialpopulationofgalaxieswith specialdustandHI galaxies showed a negative Lyα equivalent width at small radii
amountsanddistribution.Althoughextensivestudieshavebeen and then an increase farther out. We concluded that this was
carried out to characterize LAE physical properties and their due to scattering on neutral hydrogen, which is able to shape
special conditions (Nilsson et al. 2011; Acquaviva et al. 2012; the Lyα emission into the form of haloes. Also, by comparing
McLinden et al. 2014; Vargas et al. 2014; Hagen et al. 2014, LARSLyαwithglobalphysicalproperties,itappearedthatthe
amongthemostrecentones),theresultshavebeeninconclusive. Lyαphotonescapewasfavouredinthesystemwithweakerdust
The mechanisms (e.g., interstellar medium geometry and kine- reddeningandsmallstellarmass.
matics)controllingtheescapeofLyαphotonarestilldebated. The neutral hydrogen content of LARS galaxies was pre-
The morphology ofthe rest-frame UV and optical continua sented in Pardy et al. (2014) (hereafter Paper III). The spec-
providesinformationaboutgalaxyformationandevolution(e.g., tral lines of HI were detected in 11 of the 14 observed LARS
star-formingregiondistribution,mergerevents,Conselice2003; galaxiesanditwasalsofoundthattheLyαescapewasfavoured
Lotz et al. 2004). From the ground LAEs were observed to be inlowHI-masssystems.LARSinterstellarmediumkinematics
compact in the rest-frame UV, but multiple components were will be presented in Rivera-Thorsen & Hayes (2015) (submit-
identified in deep HST-resolution images (Bond et al. 2009, ted),inDuvaletal.(inprep),andinOrlitova´ etal.(inprep).
2012).Therehavealsobeenafewattemptstoquantifythemor- Inthispaper,numberIVoftheseries,weaddresstheques-
phologyoftheLyαemissionitself.Bondetal.(2010)explored tionwhetherspecificgalaxymorphologicalpropertiescouldbe
a sample of seven LAEs placed at z (cid:39) 3.1 (observed-frame relatedtotheescapeofLyαphotonsandescapeinhaloes.Note
λ(Lyα)∼5000Å)byusingHSTWFPC2(WideFieldCamera2) thatinPaperIthepresentpaperwastermedpaper7,duetoapre-
F502N narrow-band imaging. They found that, for one source, viousnumbering.InSec.2,webrieflyexplainhowwemeasured
Lyα emission extended till 1.5 kpc (≤ 1 kpc for the other six), morphologicalparameters(detailsaregiveinAppendixA)and
a just-slightly-larger scale than the UV continuum. Also, this theprocessadoptedtosimulatehowsomelocal(z<0.2)galax-
sourcewascomposedoftwomainclumpsbothintherest-frame ieswouldappearathighredshift(z > 2).InSec.3,wedescribe
UVandinLyα.Finkelsteinetal.(2011a)spatiallyresolvedthree themorphologicalpropertiesofthesampleoflocalgalaxiesand
spectroscopicallyconfirmedLAEsplacedatz (cid:39) 4.5(observed- compare them with local-Universe and high-z galaxy popula-
frame λ(Lyα) ∼ 6570 Å) by using the HST ACS (Advanced tions. In Sec. 4, we study the morphological properties of the
CameraforSurveys)F658Nnarrowband.Twooutofthethree high-z-simulatedgalaxies.InSec.5,wepresentthestackingof
systemsshowedLyαemissionsignificantlymoreextendedthan the high-z-simulated sampleand comparewith high-z stacks in
theUVcontinuum. the literature. In Sec. 6 and 7, we discuss and summarize the
mainresultsofthepaper.
Recently, evidence of extended Lyα emission was found in
ThroughoutweadoptABmagnitudesandassumeaΛCDM
the stack of a large sample of Lyman break galaxies, (Steidel
et al. 2011), generally more massive and dustier than LAEs, cosmologyof(H0,Ωm,ΩΛ)=(70kms−1 Mpc−1,0.3,0.7)asin
Hayesetal.2013,2014.
andofgalaxieslocatedinoverdenseandnot-overdenseregions
(Matsudaetal.2012).Bystackinghundredsofz(cid:39)2.2,z(cid:39)3.1,
z (cid:39) 3.7, and z (cid:39) 5.7 LAEs from deep ground-based imaging,
2. Method
Momose et al. (2014) discovered extended Lyα emission, with
scalelengthsof5−10kpc.However,bystackingtheirsampleof Wepresentthemorphologyofthelocal(0.03 < z < 0.2)LARS
LAEsFeldmeieretal.(2013)justfoundamarginaldetectionat galaxiesandinvestigatehowitwouldchangeifthesamegalax-
z ∼ 3.1andanon-detection(z (cid:39) 2.07).Itisclearthatdepthand ieswereobservedathighredshift.Asexplainedabove,inPaper
imageresolutionwerethemainfactorsaffectingtheirresults. II we isolated the contributions of the rest-frame UV (∼ 1220
Instead, giant Lyα nebulae, powered by active galactic nu- Å),optical(∼6570Å),Lyα(1216Å),andHα(6563Å).Inthis
clei, have been studied by a few authors to assess the role of work, we measure the morphological parameters of these con-
HIscatteringandLyαradiativetransfereffects(e.g.,Humphrey tributions. The LARS galaxies are hereafter referred to as Ln,
etal.2013a;Prescottetal.2014) wherenrangesfrom01to14(seePaper0).
Local starbursts (Overzier et al. 2008, 2009, 2010; Hayes
etal.2013,2014;Pettyetal.2014;O¨stlinetal.2014)areunique
2.1. Morphologicalparameterestimation
laboratoriestostudytherest-frameUVindetailandopticallight
distribution, morphology, and to investigate the mechanisms, WiththeaimofquantifyingthemorphologyofLARSgalaxies,
which allow Lyα photons to escape. In O¨stlin et al. (2014) we calculated their sizes and performed non-parametric mea-
(hereafter Paper I) we presented the Lyman alpha reference surements of morphological parameters (see Appendix A and
sample(LARS),whichiscomposedof14star-forminggalaxies Fig.A.1fordetails).
at z < 0.2. These galaxies were observed during HST cycle 18 We calculated sizes, in terms of Petrosian semi-major axis
(P.I. G. O¨stlin) in a set of rest-frame UV (ACS/SBC F125LP, (rP20,e.g.,Lotzetal.2004;Lisker2008),circularPetrosianra-
F140LP, F150LP) and optical (WFC3/UVIS F336W/F390W, dius (Petrosian 1976), and radii containing 20%, 50%, 80% of
F438W/F475W, F775W/F850LP, F502N, F656N, and theflux(r ,r ,r ).Acomparisonbetweentheseradiigivesan
20 50 80
2
L.Guaita:LARSatlowandhighz
idea of the distribution of the light in the galaxy. We also esti- – Asymmetryvsconcentration,asymmetryvsclumpiness,to-
mated asymmetry (A), concentration (C), clumpiness (S), Gini gether with clumpiness vs concentration, as presented by
coefficient (G), and second-order moment of the brightest 20% Conselice(2003)
of the galaxy’s flux (M20, see e.g., Conselice 2003; Lotz et al. – GinicoefficientvsM20bright-pixelmoment,aspresentedin
2004;Scarlataetal.2007;Michevaetal.2013). Lotzetal.(2004).
Theasymmetryquantifiesthesymmetryofagalaxywithre-
specttoa180-degreerotation;theconcentrationdescribeshow Theconcentrationdependsonthegalaxystar-formationhis-
much the light is concentrated in the centre of a galaxy; the toryinthesensethatarapidgravitationalcollapsecanproduce
clumpinessmeasurestheamountofsmall-scalestructureswithin high concentration. The presence of disk and intergalactic gas
a galaxy; the Gini coefficient provides the information on how whichcoolsontothedisktendstoproducealowerconcentration
uniformisthelightdistribution;M20tracesthespatialdistribu- value.Diskgalaxiesarecharacterizedby3 <C< 4,ellipticals
tionofanybrightknots,andalsooff-centreclumps,itsdefinition by C > 4 (Bershady et al. 2000). The asymmetry is sensitive
isverysimilartothatofC,butM20ismoresensitivetomerger toanyfeaturethatproducesasymmetriclightdistributions(e.g.,
structures,suchasoff-centrecomponents. star-formationknots,interactions,andmergers).Itiscommonly
We first ran the Source Extractor (SExtractor) software assumedalsoathighzthatlargeasymmetry(A>0.38)indicates
(Bertin & Arnouts 1996). It provided the galaxy centroid and a major merger (Aguirre et al. 2013; Conselice 2003). Spiral
the elliptical aperture, containing the entire galaxy and charac- galaxies and systems composed of more than one component
terizedbysemi-majoraxis(sma)equaltorP20.Thephotometry arecharacterizedbyA> 0.1.Theclumpinessissensitivetothe
wasperformedwithinthisSExtractordetectionaperture. presence of star-forming clumps as well, but background noise
We adopted configuration parameters like in Bond et al. canmakeitdifficulttodetectlowsurfacebrightnessregionsand
(2009) (DETECT THRESH =1.65, DETECT MINAREA=30, increase the appearance of the galaxy as a mix of clumps. The
DEBLEND MINCONT=1). They were optimized to provide GinicoefficientcanbestronglycorrelatedwithC.Bydefinition,
morphological measurements in deep HST rest-frame UV ob- G=1meansthatthelightisallconcentratedinonepixel,G=0
servationsatz > 2.TopreventSExtractorfrombreakingupthe that the light is equally distributed across the galactic body. In
clumpy, resolved z ∼ 0 LARS galaxies into smaller fragments, thecaseofashallowlightprofile,bothGandCarelow.When
weassumedalargervalueofDETECT MINAREA.Thisparam- more than one clump contains a significant fraction of light, G
etersetsthenumberofcontiguouspixelsrequiredforadetection can be much larger than zero, but C still low. M20 traces the
tobeacceptedbySExtractor.WemeasuredfluxesatSExtractor spatialdistributionofoff-centrebrightregions.
centroid within elliptical apertures, by using the ELLIPSE task In general, starburst and irregular galaxies are expected to
in iraf.stsdas.isophote and within circular apertures, by using havelargeA,largeS,andintermediateC,mergingsystemsand
the PHOT task in iraf.digiphot.apphot. ELLIPSE and PHOT perturbeddisksshowlargeM20andintermediateG.
outputs served to infer sizes, A, and C at minimum asymmetry
(CminA),as explainedin AppendixA andpreviously adoptedin
2.3. High-redshiftsimulation
Bershadyetal.(2000);Conselice(2003);Michevaetal.(2013).
The non-parametric measurements and signal-to-noise es- WesimulatedtheobservationsofLARSgalaxies(allatz<0.2)
timations were performed counting the flux of pixels belong- at higher redshift by transforming their original science- and
ing to a segmentation map. We defined the segmentation map weight-map images (Paper I) according to the following steps
in two ways, one is an ellipse with semi-major axis equal to (seealsoOverzieretal.2008;Adamoetal.2013).
rP20(Scarlataetal.2007)andorientationgivenbySExtractor;
one contains the pixels with surface brightness larger than the 1. Theimageswereresampledpreservingtheflux(IDLfrebin
valueatthePetrosianradius(Lotzetal.2004)measuredinthe function).Thesizeoftheoutputimagewasdefinedbyfixing
smoothed image (smoothed by a kernel of width rP20/5). We the physical size of the galaxies. We chose mainly a z ∼ 2
calculated M20, S and Gini coefficient by considering the pix- sampling to be able to compare with the interesting results
elswithinthesesegmentationmaps.TheGinicoefficientsmea- obtained by surveys of Lyman alpha emitters in the last re-
suredinthesetwosegmentationmapsaredenotedbyGrP20 and cent5years(Nilssonetal.2009;Guaitaetal.2010;Hayes
GSB−rp20S respectively. As described in Scarlata et al. (2007), etal.2010;Nakajimaetal.2012,Sandberginprep.).Also,
GrP20wasdefinedtobeconsistentforredshiftcomparisons,thus thesizechangesalittlewithredshift.
wepreferitoverGSB−rP20S throughoutthepaperwhenwecom- 2. Continuum subtraction (Hayes et al. 2009) was applied to
parewithhighredshift. theresampledimagestogeneraterest-frameUVcontinuum
To test our code, we applied it to template galaxies with andLyαline,rest-frameopticalcontinuumandHαlineim-
known profiles and compared the output to the results by ages. The line images are in units of flux (erg sec−1cm−2),
Bershadyetal.(2000)andLotzetal.(2006).Werecoveredthe whilethecontinuumimagesareinunitsoffluxdensities(erg
expectedvaluesasdescribedinAppendixA. sec−1cm−2Å−1).
3. The image pixel values were scaled based on luminosity
distance and surface brightness dimming (i.e., Hubble &
2.2. Combinationofmorphologicalparameters
Tolman1935;Bouwensetal.2004).
AsshowninConselice(2003)andLotzetal.(2004),combina- 4. Gaussiannoise,correspondingtoacertainsimulatedsurvey
tions of morphological parameters can give information about depth, was added to the resampled and rescaled images by
galaxyhistory(e.g.,star-formationandmergingepisodes).First runningtheMKNOISEtaskiniraf.artdata.Tocalculateun-
of all the rest-frame UV morphology is sensitive to the current certainty on galaxy sizes and morphological parameters we
star formation; the rest-frame optical traces the structure of the performed Monte Carlo simulations by repeating 100 real-
entire galaxy stellar population (Lee et al. 2013; Bond et al. izationsofanoisyimage.Thenoiseappliedwasdefinedas
2014). The combinations of parameters (see previous section) the10σdetectionwithina∼50pixel(equivalenttoasquare
weadoptedare, apertureof∼0.2”onasideforHSTACSopticalfilters)area,
3
L.Guaita:LARSatlowandhighz
similartothelimitsgivenfortheHUDF(HubbleUltraDeep – Star-forminggalaxiesselectedbasedontheirB−zandz−K
Field,Beckwithetal.2006).Wedonotshowsimulationsin colour(sBzK,Yumaetal.2012;Leeetal.2013);passiveand
whichweonlyresampledthepixelscaletothatofaground- star-forminggalaxiesselectedbasedontheirB−zandz−K
based telescope and instrument, because the main effect on colour(pBzKandsBzKLeeetal.2013)
continuum and line images was produced by survey depth – Star-forming galaxies at z ∼ 2 − 3 by Law et al. (2012).
andground-basedpointspreadfunction(PSF,seeSec.5). These authors found a typical value of the Gini coefficient
(GSB−rP20S = 0.4) for the sources with the strongest Lyα
Tochoosereasonablerangesofdetectionlimits(Table1)to emission,characterizedbyM ∼1.5×1010M(cid:12)
∗
apply, we referred to the MUSYC (MUlti-wavelength Survey
– GOODS (Great Observatories Origins Deep Survey) and
byYale-Chile)NB3727narrowband(Guaitaetal.2010;Bond UDF(UltraDeepSurvey)z∼4andGOODSz∼1.5sources
etal.2012),tothetriplenarrowbandbyNakajimaetal.(2012),
fromthestudyofLotzetal.(2006)
CANDELS/HUDF(McLureetal.2013)broadband,andtothe
– Sub-millimetergalaxies(SMGs,Aguirreetal.2013)
dualnarrow-bandsurveybyLeeetal.(2012). – Narrow-bandselectedLymanalphaemittersatz (cid:39) 2.07and
z (cid:39) 3.1 (Bond et al. 2009, 2012) belonging to the MUSYC
3. LARSgalaxiesatz ∼ 0 survey.Weconsideredthestackofthez∼2.07entiresample
andofsubsamplesseparatedbyphotometricproperties,UV-
To be able to compare the Lyα, Hα, and continuum properties faint(UV-bright)withR > 25.5(< 25.5),IRAC-faint(IRAC-
ofLARSgalaxieswiththoseofhigh-zLyαemitters(LAEs),we bright) with f < 0.57(> 0.57) µJ, low-(high-)EW with
3.6µm
focused on the twelve LARS galaxies with EW(Lyα) > 1 Å EW(Lyα)<66(>66)Å,red-(blue-)LAEwithB−R>0.5(<
as measured in Paper II. Thus, we excluded from this analysis 0.5)(Guaitaetal.2011).
the two galaxies of the sample (L04 and L06) characterized by – Narrow-bandselectedLymanalphaemittersatz ∼ 5.7,6.5,
null Lyα maps. LARS galaxies with integrated EW(Lyα) > 20 and7.0(Jiangetal.2013),thefirstvery-high-redshiftsample
Å composed the subsample of LARS-LAEs (consisting of six where non-parametric morphological measurements were
galaxies).VariousphysicalcharacteristicsoftheLARSgalaxies performed.
(includingtheircoordinates)arediscussedinPaperII.
In Fig. 11 we present the RGB images of the twelve LARS The local-Universe studies, we adopted for comparison, in-
galaxies. Most of the galaxies show localized knots of star for- clude,
mationsuperposedonextendedrest-frameopticalemission;the
Lyα emission is extended on larger angular scales (the Lyα – SloanDigitalSkySurvey(SDSS)early-andlate-typegalaxy
haloes). relationsobtainedfromtheanalysisofimagesinthezband
WeinvestigatedthepropertiesofLARSgalaxiesinthecon- (Shenetal.2003)
text of other galaxy populations, to assess the fairness of our – Lyman break analogues (LBAs) at z (cid:39) 0.2. These are
comparison. In Figs. 2 and 3 we show the location of the local starbursts that share typical characteristics of high-z
LARS galaxies in the half-light radius vs stellar mass (r vs LBGs,suchasstellarmass,metallicity,dustextinction,star-
50
log(M /M )) and the half-light radius vs UV absolute magni- formation rate, and physical size. We considered a sample
∗ (cid:12)
tude (r vs M ) diagrams, to understand if the LARS galax- of 30 LBAs from Overzier et al. (2009, 2010). They were
50 UV
ies harbour stellar masses and UV magnitudes comparable to characterizedbyamedianabsoluteUVmagnitudeof-20.3,
valuesintheliterature.Thesediagramshavebeendesignedfor almostonemagnitudefainterthantypicalLBGs.
local galaxies, for which sizes could be easily measured in the
rest-frameopticalbands(Shenetal.2003).However,measure- WefoundthattheLARSgalaxiesoccupyaquitewiderange
mentsintherest-frameUVcouldalsobeperformedathighred- of r . Their rest-frame UV and optical sizes (Table 2) are
50
shift. Following the method described in Sec. 2, we estimated broadly consistent with LBAs, LBGs, and SMGs. Their stellar
the half-light radius as r in the rest-frame optical and also in mass tend to be larger than LAEs, consistent with LBAs and
50
therest-frameUVimages. LBGs. However, there is an overlap in stellar mass between
Thehigh-zstudiesweadoptedforcomparisonallperformed LARS-LAEs (M∗ < 1010 M(cid:12)) and the most massive LAEs in
size and morphological measurements by using HST images. thesampleofBondetal.(2012).Also,LARSgalaxiesareless
Theseinclude, massivethancSFGs.Thelargesthalf-lightradiicharacterizethe
LARS galaxies with the most distorted morphology (see also
– Continuum-selectedLymanbreakgalaxies(LBGs)atz ∼ 3, Fig. B.2). LARS M magnitudes (and so star-formation rate,
UV
withandwithoutLyαinemission(Pentericcietal.2010);at SFR )arecomparablewiththoseofz (cid:39) 2.07LAEsandz ≥ 7
UV
z ∼ 1,2,and3(Moslehetal.2011),at1.5 < z < 3.6(Law LBGs.Thereisanoverlapwithz>5LAEs.However,themea-
etal.2012);atz∼1.8(Lotzetal.2004);z-dropoutsatz∼7 surements of LARS sizes in the rest-frame UV are larger than
(Grazianetal.2012);highsignal-to-noisez-andY-dropouts thoseofz≥7LBGs.
detectedintheHubbleUltraDeepField,UDF12(Onoetal. Therefore, LARS galaxies could be considered as LBAs,
2013) with size, stellar mass, and star-formation rate similar to 2 <
– Compactstar-forminggalaxies(cSFGs)at2 < z < 3(Barro z<3star-forminggalaxies.
et al. 2013). These authors have pointed out that, based on
their number densities, masses, sizes, and star formation
rates, z ∼ 2−3 compact, star-forming galaxies were likely 3.1. ContinuummorphologyofLARSgalaxiesatz∼0
progenitorsofcompact,quiescent,massivegalaxiesatz<2
FollowingthemethoddescribedinSec.2,weestimatedthenon-
1 We took advantage of this codification of the Lupton parametricmeasurementsfortheLARSgalaxies(Table3).
et al. (2004) prescription to produce RGB images: Combinations of morphological parameters (see Sec. 2.2)
http://dept.astro.lsa.umich.edu/∼msshin/science/code/ can give information about a galaxy’s star-formation history.
Python fits image/ Lotz et al. (2004) proposed a criterion for separating perturbed
4
L.Guaita:LARSatlowandhighz
Table1:10σdetectionlimitsappliedtohigh-zsimulatedLARSimages
F(Lyα) m(rest-frameUV) F(Hα) m(rest-frameoptical)
ergsec−1cm−2 AB ergsec−1cm−2 AB
5E-19 30 2E-19 29
3E-18 29 6E-19 28
8E-18 28 1E-18 27
2E-17 27 3E-18 26
5E-17 26 1E-17 25
Notes. Noise corresponding to the detection limits given in the table was added to LARS continuum and line images. As LARS Lyα and
Hα(rest-UVandopticalcontinua)imagesareinunitsofflux(fluxdensity),theimagedepthsaregiveninunitsofergsec−1cm−2(ABmagnitudes).
The MUSYC (Guaita et al. 2010; Bond et al. 2012) NB3727 survey implied a 10σ detection limit of F(Lyα)=5E-17 erg sec−1cm−2. MUSYC
U,B(HUDFV606)10σdetectionlimitswereabout26(29.5).AssumingaNB3727filterwidthandtransmissionprofile,asourcewithm (rest-
AB
frame UV)=30 and EW(Lyα) = 20 Å is characterized by F(Lyα)=5E-19 erg sec−1cm−2. m (rest-frame UV) = 30,29,28 are consistent with
AB
HUDF09,GOODS,andGEMSsurveydepths(Bondetal.2009).Leeetal.(2012)surveywascharacterizedbya10σdetectionlimitof22.9
in NB210 and 23.7 in K. CANDELS wide(deep) F160W 10σ detection limit was 25.8(26.5). Assuming a NB210 filter width, a source with
m (rest-frameoptical)=27andEW(Hα)=20ÅischaracterizedbyF(Hα)=2E-19ergsec−1cm−2.
AB
Fig.1:False-colourimagesoftheLARSgalaxiesanalysedinthispaper.Redencodesrest-frameopticalcontinuum,greenrest-frameUVcontin-
uum,andblueshowscontinuum-subtractedLyα.Scalesinkpcaregivenontheside.Intensitycutlevelsaresettoshowdetails.
disks or merging systems from normal galaxies, by studying identify merging systems, GSB−rP20S > −0.14 × M20 + 0.33.
local ultra luminous infrared galaxies (ULIRGs). The crite- Also,Conselice(2003)distinguishedtheregionwhereirregular
rion identifies a region in the G vs M20 diagram, which is orstarburstgalaxieswerelocatedintheA-CminAandA-Splanes.
GSB−rP20S > −0.115 × M20 + 0.384. For z < 1.2 galaxies AsseeninFig.4andTable3,LARSgalaxies,inparticular
observed in a rest-frame optical band (4000 Å) at HST resolu- the LARS-LAEs, tend to avoid the location of normal galax-
tion, Lotz et al. (2008) proposed a slightly different relation to iesandtooccupytheregionofperturbeddisksormergingsys-
temsandofirregularorstarburstgalaxies.ThevaluesofG,M20,
5
L.Guaita:LARSatlowandhighz
Table2:SizeoftheoriginalLARSgalaxiesatz∼0
(1) (2) (3) (4) (5) (6) (7)
LARS rP20ell rP20circ rP20minA r r r
20 50 80
kpc kpc kpc kpc kpc kpc
rest-UV
L01 2.63 1.34 0.81 0.86 1.20 2.11
L02 2.42 2.28 1.49 0.69 1.28 2.17
L03 1.40 0.97 0.77 0.33 0.63 1.17
L05 1.71 1.46 0.84 0.35 0.69 1.45
L07 1.13 0.90 0.81 0.33 0.60 0.95
L08 5.98 3.35 2.08 1.67 3.01 4.96
L09 18.72 1.11 0.59 2.73 8.52 13.81
L10 4.73 3.71 3.09 0.77 2.08 3.93
L11 23.77 18.56 18.56 6.12 12.21 19.41
L12 1.84 1.20 1.05 0.41 0.75 1.76
L13 4.42 3.70 1.23 1.95 2.47 3.75
L14 1.83 1.71 0.98 0.67 0.98 1.52
rest-optical
L01 4.87 2.15 2.44 1.28 2.11 4.01
L02 7.12 2.23 2.51 1.08 2.63 5.62
L03 5.73 1.98 1.83 0.72 2.00 4.59
L05 2.32 1.79 1.84 0.41 0.98 1.79
L07 3.95 2.89 1.96 0.62 1.51 3.18
L08 5.28 3.17 3.29 1.48 2.63 4.22
L09 20.54 3.06 2.07 5.98 10.27 14.74
L10 7.14 3.93 3.93 1.08 2.83 5.57
L11 19.13 14.52 14.52 4.83 9.82 15.43
L12 3.68 3.15 1.20 0.64 1.46 2.97
L13 11.41 5.04 5.35 2.31 5.30 9.15
L14 3.72 3.78 1.22 0.73 1.28 2.80
Notes.SizemeasurementsfromellipticalandcircularaperturephotometryoftheoriginalLARSimages.(1)LARSid,(2)Petrosiansemi-major
axis,(3)circularPetrosianradius,(4)Petrosianradiusatminimumofasymmetry,(5)radiuscontaining20%,(6)50%,and(7)80%ofthetotal
flux.Themeasurementswereperformedinthebandscorrespondingtotherest-frameUV(eitherF140orF150)andtotherest-frameoptical(either
F775,F814,orF850).Thestepinsemi-majoraxisis1pixel(∼0.02kpcatz∼0.03)andincircularradiusis2pixels(∼0.05kpcatz∼0.03).
Table3:MorphologicalparametersoftheoriginalLARSgalaxiesatz∼0
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11)
LARS GrP20 GSB−rP20S M20 ell Ccirc Cell CminA SN A S
pixel
rest-UV
L01 0.70 0.68 -0.87 0.61 1.34 1.94 1.73 251.40 0.42 0.16
L02 0.73 0.72 -0.79 0.53 3.14 2.50 3.49 52.26 0.50 0.26
L03 0.63 0.64 -1.17 0.30 2.20 2.71 2.76 83.18 0.21 0.11
L05 0.72 0.68 -1.26 0.34 3.11 3.07 3.27 408.69 0.33 0.15
L07 0.62 0.62 -1.02 0.40 1.99 2.28 1.24 696.00 0.26 0.09
L08 0.62 0.65 -0.89 0.33 1.95 2.36 1.43 21.98 0.37 0.29
L09 0.71 0.75 -2.33 0.80 1.76 3.52 3.35 23.86 0.21 0.24
L10 0.67 0.68 -1.52 0.52 3.53 3.53 2.76 32.83 0.33 0.24
L11 0.58 0.79 -1.22 0.87 3.40 2.51 3.40 44.05 0.30 0.31
L12 0.67 0.68 -1.50 0.34 2.56 3.15 2.82 583.81 0.28 0.04
L13 0.72 0.67 -0.77 0.21 1.23 1.42 2.61 89.17 0.33 0.23
L14 0.74 0.68 -0.71 0.27 1.51 1.78 2.72 622.68 0.21 0.02
rest-optical
L01 0.59 0.61 -1.09 0.64 1.72 2.48 2.53 21.91 0.29 0.09
L02 0.59 0.64 -1.36 0.69 2.96 3.58 2.89 4.25 0.14 0.31
L03 0.61 0.62 -2.16 0.45 3.32 4.02 3.42 50.72 0.09 0.05
L05 0.64 0.67 -1.57 0.73 3.27 3.22 4.77 64.96 0.33 0.15
L07 0.60 0.61 -1.68 0.60 3.28 3.56 2.97 14.60 0.24 0.10
L08 0.54 0.55 -1.05 0.23 2.10 2.28 1.78 36.82 0.31 0.16
L09 0.62 0.74 -1.13 0.83 2.03 1.96 4.35 11.38 0.45 0.34
L10 0.57 0.58 -1.87 0.54 3.08 3.56 3.31 9.95 0.13 0.15
L11 0.50 0.60 -1.25 0.85 3.46 2.52 3.46 18.70 0.20 0.23
L12 0.64 0.63 -1.90 0.46 3.70 3.34 3.35 54.74 0.33 0.13
L13 0.58 0.62 -1.52 0.55 2.43 2.99 3.67 3.98 0.18 0.30
L14 0.76 0.72 -1.32 0.10 2.92 2.92 3.01 35.24 0.32 0.13
Notes. Morphological parameters estimated for the original LARS galaxies, following the equations in Appendix A. (1) LARS id, (2) Gini
coefficientestimatedwithinthefixed-sizesegmentationmap,(3)Ginicoefficientestimatedwithinthesegmentationmapbuiltfromthepixelswith
surfacebrightnesslargerthanthevaluecorrespondingtothatatthePetrosianradius,(4)M20,(5)SExtractorellipticity1-B/A,(6)concentration
fromcircularapertures,(7)concentrationfromellipticalapertures,(8)concentrationcorrespondingtotheminimumofasymmetry,(9)signal-to-
noiseperpixel,(10)asymmetry,(11)clumpiness.AsdiscussedinAppendixA,wecouldexpectanuncertaintyof∼10%inG,of<5%inM20,
andof∼20%inC.WecouldexpectanevenlargeruncertaintyinS.ThedifferenceinGrP20andGSB−rP20S isgenerallymarginal.
6
L.Guaita:LARSatlowandhighz
Fig.2:Half-lightradiusmeasuredusingellipticalaperturesintherest-frameUV(upperpanel)andoptical(lowerpanel)asafunctionofstellar
mass. LARS values at z ∼ 0 from this work (red diamonds) are shown together with previous rest-frame UV measurements: Overzier et al.
(2009) individual LBA values estimated in the HST ACS/SBC F150LP and ACS/HRC F330W filter (open blue squares), Yuma et al. (2012)
medianestimationsintheF850WfilterforsBzK(greenstar),Bondetal.(2012)LAEsamplestackandsubsamples(magentasquare,blueand
magentastars,smallblackcircle),Aguirreetal.(2013)individualSMGvaluesmeasuredintheHSTF110Wfilter(bigblackdots),andPentericci
etal.(2010)averagevaluesofLBGswithandwithoutLyαinemission(greendots);andrest-frameopticalmeasurements:Overzieretal.(2009)
individual LBA values estimated in the HST (Wide Field and Planetary Camera2) WFPC2/F606W and ACS/WFC F850LP filter (open blue
squares), Yuma et al. (2012) median estimations in the F160W filter for sBzK (green star), Mosleh et al. (2011) median values of UV-bright
sources(GALEX-LBGsatz ∼ 0.6−1.5,LBGatz ∼ 2.5−3.5,andcontinuum-selectedstar-forminggalaxiesatz ∼ 1.5−2.5,bluediamonds),
Aguirreetal.(2013)individualSMGvaluesmeasuredinF160Wfilter,Lawetal.(2012)meanvalueofallthesampleofstar-forminggalaxiesat
1.5<z<3.6estimatedintheF160Wfilter(yellowtriangles),andBarroetal.(2013)valuesforcompactstar-forminggalaxiesat2<z<3also
calculatedintheF160Wfilter(greentriangles).WealsoshowthecurvederivedbyShenetal.(2003)forlocalSDSSearly-andlate-typegalaxies.
Asthesecurveswereobtainedinz-bandsforlocalgalaxies,itismoremeaningfultocomparethemtotheradiiintherest-frameoptical.However,
forreference,weshowthemintheupperpanelaswell.ThestellarmassesareallcorrectedtoSalpeter-IMFvaluesandthesizemeasurementsare
allscaledtobecomparabletohalf-lightradii.LARSstellarmasseswerecalculatedinPaperII.L09andL11areoutsidethegraph,duetotheir
half-lightradiuslargerthan6kpc.
C,andA,wecalculatedforLARSgalaxies,areconsistentwith ter, characterized by one of the smallest stellar masses and the
the ones measured by Overzier et al. (2010) for LBAs. Even if largestGSB−rP20S amongtheLARS-LAEs.
our sample is just composed of twelve sources, we do not see
any significant dependency between G and EW(Lyα). L08 is
themostmassiveoftheLARS-LAEs,butequallyconcentrated
withinthesegmentationmap.L02isthelargest-EW(Lyα)emit-
7
L.Guaita:LARSatlowandhighz
Fig.3:Half-lightradiusmeasuredusingellipticalaperturesintherest-frameUVimageasafunctionoftheabsoluterest-frameUVmagnitude.
LARSmeasurementsfromthiswork(reddiamonds)areshowntogetherwiththeliteratureestimationsbyOverzieretal.(2010)correspondingto
themedianvalueoftheirz < 0.3LBAs,observedinthenear-infraredbands,byJiangetal.(2013)forasampleofz ∼ 5.7,6.5,7.0LAEs(cyan
dots),byBondetal.(2012)forthestackandsubsamplesofLAEsatz (cid:39) 2−3(blackcircles,magentasquares),byGrazianetal.(2012)which
measuredSExtractorhalf-lightradiiforasampleofz-dropouts(greentriangles),andbyOnoetal.(2013)forasampleofhighsignal-to-noisez-
andY-dropoutsdetectedinHubbleUltraDeepField,UDF12(blackstars).
3.2. LyαmorphologyofLARSgalaxiesatz∼0 survey,thefaintconnectingstructurestendtobelostinthenoise
andagalaxyappearstobecomposedofseparatedclumps.Inthat
One of the goals of our work was to quantify and compare the
caseSExtractoridentifiesmorethanonesourceandphotometry
morphologies of LARS galaxies in Lyα and in the continuum.
is performed by locating the photometric aperture around the
Wepresentmorphologicalparametersmeasuredinthecontinua
brightestclump.InFigs.B.1,B.2,andB.3,weshowhowLARS
and in Lyα of LARS images in Figs. 5 and 6. GrP20, M20,
galaxieswouldappearifdetectedinthedeepestcontinuumand
concentration,andellipticityaresmaller;whileclumpinessand
linesurveyssimulatedhere,whileFigs.B.4,B.5,andB.6show
asymmetry are generally larger in Lyα than in the rest-frame
theresultsforshallowersurveys.InAppendixC,wepresentthe
UV continuum. The LARS-LAEs tend to be characterized by
correspondingsurfacebrightnessprofiles.
thehighestconcentration,lowestasymmetry,andlowestclumpi-
nessinLyα.GrP20 andM20measuredintherest-frameoptical In the following sub-sections, we describe the detection of
LARSgalaxiesinthesimulatedsurveyswith10σdetectionlim-
areconsistentwiththevaluesmeasuredinLyα.
itspresentedinTable1.Inthefirstsub-section,wegivedetails
onthedetectionofL01asanexample.Weproceedtodescribe
4. LARSgalaxiesasseenatz ∼ 2 thecasesoftheLARS-LAEsandofthegalaxieswiththefaintest
Lyα emission. Then, we explain the variations in size and el-
We applied the procedure described in Sec. 2.3 to simulate
lipticity versus clumpiness owing pixel resampling and survey
LARSgalaxiesatz ∼ 2.Wenamedthehigh-zsimulatedgalax-
depth.InSec.4.6,wequantifythemorphologyofz2LARSand
iesasz2LARSandthesubsampleofLyαemittersasz2LARS-
comparewithhigh-zobservationsfromtheliterature.
LAEs.Weestimatedsizesandcalculatedmorphologicalparam-
eters (Tables 5 - 8 and 9-12) in the same way we did for the
originalimagesinSec.2.1. 4.1. DetectionofL01inhigh-redshiftsurveys
Thepurposeofthistestwastounderstandwhetherwecould
expecttodetectLARS-typegalaxiesandLARS-typeLyαhaloes InFig.7,weshowtherest-frameUV,Lyα,optical,andHαim-
incurrenthigh-zsurveys.Inparticular,wewantedtounderstand ages of L01 simulated to be at z ∼ 2 (z2L01) as they would
how galaxy size and morphological parameters changed when beobservedinthedeepestsurveysprobedhere;SExtractorde-
varying the survey depth (Table 1). The results show that, in a tectionaperturesareover-plotted.Thedetectionparameterswe
sufficientlydeepsurvey,faintgalaxystructuresinbetweenbright adoptedaresensitiveenoughthat(atthedeepestsimulatedsur-
knotsremainconnectedtogetherandSExtractorisabletodetect veys)thisgalaxyisdetectedasasinglesource.Asdescribedin
just one source (the entire galaxy) in the image. In a shallower detailinPaperIandII,L01consistsofabrightUVstar-forming
8
L.Guaita:LARSatlowandhighz
Fig.4:Combinationsofrest-frameopticalmorphologicalmeasurements,usedintheliteratureasdiagnosticsofgalaxypastandcurrenthistory.
GSB−rP20S vsM20(upperleft),AvsCminA (upperright),andAvsS(lowerright).Thelowerleftpanelshowsstellarmassvstherest-frame
UVGSB−rP20S.ThetypicalvalueofGSB−rP20S = 0.4forthestrongestLyαemittersofthesamplebyLawetal.(2012)isreportedasaredstar.
ThetwelveLARSgalaxiesanalysedherearepresentedassquares,LARS-LAEsareroundedbyopencircles.Thecolourscalecorrespondsto
EW(Lyα).Forcomparison,greenstarscorrespondtotheFreietal.(1996)sampleofnormalgalaxiesandlightbluetrianglescorrespondtothe
ULIRGsampleofBorneetal.(2000)asprocessedbyLotzetal.(2004).Blacktrianglescorrespondtoasampleofstarburstgalaxiespresentedin
Conselice(2003).DashedandsolidlinescorrespondtotheseparationbetweenULIRGsandnormalgalaxies,proposedbyLotzetal.(2004)and
Lotzetal.(2008)respectively(seetext).
centre with an extended tail, also seen in Hα and in the rest- mlim =28andF(Lyα)lim =8E-18ergsec−1 cm−2.However,
rest−UV
frame optical. The Lyα emission is coincident with the bright on scales larger than 4 kpc, the profiles are indistinguishable
UV knot and extends in a fan-like structure possibly indicating from the background noise. In shallower surveys, the profiles
thepresenceofanexpandingbubble.Themainfeaturesofemis- start to be affected by the simulated-survey noise on smaller
sion (dark red pixels in Fig. 7) and absorption (white pixels), scalesandz2L01couldnotbedetectedbyadoptingaSExtractor
observed in Lyα thanks to the HST resolution and the careful detectionthreshold,DETECT THRESH=1.65.Therefore,size
continuum subtraction presented in Paper II, are clearly visible and morphological parameter measurements could not be per-
inthez∼2simulationaswell.However,theextremelydetailed formed either. We define mlim and F(Lyα)lim as the limits
rest−UV
Lyαstructuresclosetothecentreofthegalaxy(seePaperIFig. fordetectionandmorphologicalparametermeasurement.These
1)arenotvisible.ThelastpanelofFig.7showsL01Lyαimage, limitsaremlim =26andF(Hα)lim =3E-18ergsec−1cm−2
rest−optical
convolved with a ground-based seeing. From the ground L01 forL01rest-frameopticalandHα.
Lyαmorphologywouldappearsmoothed.
Weshowthesurfacebrightnessprofilesofz2L01inFig.8. ThelowerleftpanelofFig.8showsthattherest-frameUV
Therest-frameUVandLyαprofiles(leftcolumnpanels)arepre- continuumprofileissteeperthantheLyαprofile.Thelowerright
served when observed in a survey with sensitivity deeper than panel shows that the rest-frame optical continuum tends to be
9
L.Guaita:LARSatlowandhighz
Fig.5: Non-parametric measurements performed in the Lyα images versus the ones performed in the rest-frame UV. From the upper left to
the lower right: GrP20, M20, CminA, A, S, and SExtractor ellipticity (1-B/A, where A and B are the semi-major and semi-minor axes of the
detectionellipse).Thedashedlineindicatesthe1:1relation.ThenumbersreportedineachpanelcorrespondtotheSpearmantestcoefficient,r,
andprobability,p,ofuncorrelateddatasets.r=0indicatesnocorrelation,r=1(-1)indicatesdirect(indirect)proportionality.
shallowerthantherest-frameUVandtheHα,andmoresimilar outskirts. The Lyα emission is accompanied by regions of ab-
totheLyαprofiles. sorption. As a typical trend, in increasingly shallower surveys
the filaments, seen in the continua, show lower surface bright-
ness,whiletheirLyαemissionbecomeincreasinglylocalizedin
4.2. DetectionoftheLARS-LAEsinhigh-redshiftsurveys thegalaxycentre.OnlyL14,thegalaxybrightestinLyα,could
bedetectedintheshallowestLyαsurveyprobedhere.Themag-
TheLARS-LAEgalaxiesshowmorethanonebrightknot,con- nitudeandfluxlimitsforz2LARS-LAEdetectionareshownin
nectedbyfilaments,inthecontinua.Theyalsoshowanintense Table4.
Lyα emission close to their centres and Lyα structures in their
10