MUSE integral-field spectroscopy towards the Frontier Fields cluster Abell S1063

45 Pages·2017·17.97 MB·English

A&A599,A28(2017) Astronomy
DOI:10.1051/0004-6361/201629055
&
(cid:13)c ESO2017 Astrophysics
MUSE integral-field spectroscopy towards the Frontier Fields
cluster Abell S1063
II. Properties of low luminosity Lyman α emitters at z >3
W.Karman1,K.I.Caputi1,G.B.Caminha2,M.Gronke3,C.Grillo4,5,I.Balestra6,7,P.Rosati2,E.Vanzella8,D.Coe9,
M.Dijkstra3,A.M.Koekemoer9,D.McLeod10,A.Mercurio11,andM.Nonino7
1 KapteynAstronomicalInstitute,UniversityofGroningen,Postbus800,9700AVGroningen,TheNetherlands
e-mail:[email protected]
2 DipartimentodiFisicaeScienzedellaTerra,UniversitàdegliStudidiFerrara,viaSaragat1,44122Ferrara,Italy
3 InstituteofTheoreticalAstrophysics,UniversityofOslo,Postboks1029Blindern,0315Oslo,Norway
4 DarkCosmologyCentre,NielsBohrInstitute,UniversityofCopenhagen,JulianeMariesVej30,2100Copenhagen,Denmark
5 DipartimentodiFisica,UniversitàdegliStudidiMilano,viaCeloria16,20133Milano,Italy
6 UniversityObservatoryMunich,Scheinerstrasse1,81679Munich,Germany
7 INAF–OsservatorioAstronomicodiTrieste,viaG.B.Tiepolo11,34143Trieste,Italy
8 INAF–BolognaAstronomicalObservatory,viaRanzani1,40127Bologna,Italy
9 SpaceTelescopeScienceInstitute,3700SanMartinDrive,Baltimore,MD21208,USA
10 SUPA,InstituteforAstronomy,UniversityofEdinburgh,RoyalObservatory,EdinburghEH93HJ,UK
11 INAF–OsservatorioAstronomicodiCapodimonte,viaMoiariello16,80131Napoli,Italy
Received3June2016/Accepted23August2016
ABSTRACT
InspiteoftheirconjecturedimportancefortheEpochofReionization,thepropertiesoflow-massgalaxiesarecurrentlystillvery
muchunderdebate.Inthisarticle,westudythestellarandgaseouspropertiesoffaint,low-massgalaxiesatz>3.Weobservedthe
FrontierFieldsclusterAbellS1063withMUSEovera2arcmin2 field,andcombinedintegral-fieldspectroscopywithgravitational
lensingtoperformablindsearchforintrinsicallyfaintLyαemitters(LAEs).Wedeterminedintotaltheredshiftof172galaxiesof
which14arelensedLAEsatz=3–6.1.Weincreasedthenumberofspectroscopically-confirmedmultiple-imagefamiliesfrom6to17
andupdatedourgravitational-lensingmodelaccordingly.Thelensing-correctedLyαluminositiesarewithL (cid:46)1041.5erg/samong
Lyα
thelowestforspectroscopicallyconfirmedLAEsatanyredshift.WeusedexpandinggaseousshellmodelstofittheLyαlineprofile,
andfindlowcolumndensitiesandexpansionvelocities.Thisis,toourknowledge,thefirsttimethatgaseouspropertiesofsuchfaint
galaxies at z (cid:38) 3 are reported. We performed SED modelling to broadband photometry from the U band through the infrared to
determinethestellarpropertiesoftheseLAEs.Thestellarmassesareverylow(106−8 M ),andareaccompaniedbyveryyoungages
(cid:12)
of1–100Myr.Theveryhighspecificstar-formationrates(∼100Gyr−1)arecharacteristicofstarburstgalaxies,andwefindthatmost
galaxieswilldoubletheirstellarmassin(cid:46)20Myr.TheUV-continuumslopesβarelowinoursample,withβ < −2forallgalaxies
with M < 108 M . We conclude that our low-mass galaxies at 3 < z < 6 are forming stars at higher rates when correcting for
(cid:63) (cid:12)
stellarmasseffectsthanseenlocallyorinmoremassivegalaxies.Theyoungstellarpopulationswithhighstar-formationratesand
i
lowH columndensitiesleadtocontinuumslopesandLyC-escapefractionsexpectedforascenariowherelowmassgalaxiesreionise
theUniverse.
Keywords. galaxies:high-redshift–galaxies:distancesandredshifts–galaxies:clusters:individual:AbellS1063–
gravitationallensing:strong–galaxies:evolution–techniques:imagingspectroscopy
1. Introduction Steideletal. 1996, 2003; Bouwensetal. 2011) and spectral-
energy-distribution (SED) fitting codes (e.g. Caputietal. 2011;
TheevolutionofthebrightestgalaxiesintheUniversehasnow Ilbertetal.2013),whicharewellprovenforintermediate-mass
beenstudiedinsignificantdetailouttoz∼8(e.gBouwensetal. and massive galaxies, are not applicable to these faint sources
2014; Salmonetal. 2015; Caputietal. 2015), and is in accor- in all but the deepest multiwavelength studies (e.g. Ouchietal.
dance with the now well-established ΛCDM model. The study 2010; Schenkeretal. 2013) or until The James Webb Space
of low-mass, faint galaxies at high-z is, instead, almost a com- Telescope is operating (e.g. Gardneretal. 2009; Bisigelloetal.
pletelyunknownterritory.Gainingagreaterknowledgeofthese 2016). Therefore, other approaches are needed to understand
faintgalaxiesisimportantastheyarethebuildingblocksofthe the faint end of the galaxy population. One possible approach
observedmoremassivegalaxiesatlowerredshifts,andtheyare is looking for counterparts of absorbers in quasar lines-of-
currently seen as the main candidates for reionizing the Uni- sights (e.g. ArrigoniBattaiaetal. 2016), but this is only fea-
verse at z = 6−10 (Wiseetal. 2014; Kimm&Cen 2014, but sible for bright quasars with long spectroscopic observations
seeSharmaetal.2016). (e.g Rauchetal. 2008). Fortunately, the Lyα line is redshifted
Observationally, high-redshift, low-mass galaxies have in the optical domain for galaxies at z (cid:38) 3. Although stars in
been elusive to date. The Lyman break technique (e.g. massive galaxies are often surrounded by a dusty inter-stellar
ArticlepublishedbyEDPSciences A28,page1of45
A&A599,A28(2017)
and circum-galactic medium which absorbs all Lyα photons LAEs (e.g. Amorínetal. 2010, 2015). Another indication for
(e.g.Laursenetal.2009),less-massivestar-forminggalaxiesare a close resemblance between these galaxies is the finding that
often found with significant Lyα emission (e.g. Oyarzúnetal. lowstellarmass,highSFR,andlowdustcontentcorrelatewith
2016). Therefore, searching for galaxies with strong emission Lyα emission both at low (e.g. Cowieetal. 2011; Henryetal.
linesintheopticalcanbeusedtoidentifylow-masshigh-redshift 2015) and high redshift (e.g. Jiangetal. 2016). In addition,
galaxies. Vanzellaetal. (2016b) and Izotovetal. (2016) found Lyman
Another possibility are optical narrowband studies, which continuum leakage for two of these galaxies, making them im-
searchforgalaxieswithstrongemissionlines(e.g.Nilssonetal. portantcandidatesforreionization.
2009;Nakajimaetal.2012;Mattheeetal.2016)bylookingfor TheFrontierFieldsprogramme(hereafterFF,PI:J.Lotz;see
sourceswithstrongcoloursbetweenthenarrowbandsandover- Lotzetal.2016;andKoekemoeretal.2016)providesanexcel-
lapping broadband observations. By applying additional colour lentopportunitytostudyintrinsicallyfaintgalaxiesathighred-
cutsrepresentativeofhigh-redshiftgalaxies,reliablecandidates shifts. Massive galaxy clusters provide a boost in depth thanks
for Lyα emitters (LAEs) can be found. However, it has been to the effect of gravitational lensing. The deep HST coverage
shown that low-redshift extreme line-emitters can contaminate over7differentbandsprovidesphotometryforintrinsicallyfaint
thissample(e.g.Ateketal.2011;Péninetal.2015),andgalax- sourceswhichallowsustostudytheirproperties.Combiningthis
ies with intermediate Lyα line strengths will not survive the deepgravitionally-lensedphotometricsurveywithspectroscopy
colourcuts.Anotherdisadvantageofusingnarrow-bandstudies allows us to determine accurate stellar and gaseous properties
is that these selections are only useful for very narrow redshift down to an intrinsic faintness which is otherwise currently un-
ranges. achievable within a reasonable observing time. Abell S1063
Although Lyα has become the most important line to iden- (AS1063),theclusterstudiedhere,isamongthebeststudiedFF
tify galaxies with redshifts between 2.5 < z < 7 (e.g. clustersforwhichwehaveoneofthebestconstrainedandmost
Shimasakuetal. 2006; Dawsonetal. 2007; Díazetal. 2015; precise strong lensing model available so far (e.g. Monnaetal.
Trainoretal. 2015), it is still unclear what governs whether 2014; Johnsonetal. 2014; Richardetal. 2014; Caminhaetal.
a galaxy is a LAE or not. It has been found that LAEs 2016b;Diegoetal.2016).
are in general less dusty than LBGs (e.g. Ateketal. 2014), In Karmanetal. (2015, hereafter Paper I), we showed that
but they have very similar stellar properties at fixed luminos- usinggravitationallensingincombinationwiththeintegralfield
ity (Shapleyetal. 2001; Korneietal. 2010; Yumaetal. 2010; spectrograph Multi Unit Spectroscopic Explorer (MUSE) we
Malleryetal. 2012; Jiangetal. 2016). There is evidence how- have been able to identify previously undetected, intrinsically
ever, that the prevalence of Lyα emission is much higher in faint LAE. In this work we expand on our previous results by
less luminous systems (Starketal. 2010; Forero-Romeroetal. adding observations of a second MUSE pointing covering the
2012) and less massive systems (Oyarzúnetal. 2016). A sim- secondhalfofthecluster,andusingLyαlineprofilemodelling
ilar trend is found for the equivalent width (EW) of Lyα, incombinationwithbroadbandphotometrytostudytheproper-
which anticorrelates with UV luminosity (e.g. Shapleyetal. ties of LAEs at 3 < z < 6. In addition, we present an updated
2003; Gronwalletal. 2007; Korneietal. 2010). Further, the redshiftcatalogueusingthefullMUSEdataset.
fraction of LBGs with Lyα emission increases with redshift The layout of this paper is as follows. In Sect. 2 we give
out to z ∼ 6 (e.g. Ouchietal. 2008; Cassataetal. 2011, 2015; a brief overview of the MUSE performance and the obtained
Pentericcietal. 2011; Curtis-Lakeetal. 2012; Schenkeretal. data,followedbythedatareductionprocess.InSect.3,wede-
2012; Henryetal. 2012), but experiences a rapid decrease scribe our spectroscopic results, including the determined red-
afterwards (e.g. Kashikawaetal. 2011; Caruanaetal. 2012, shiftsandemissionlineproperties.Weusedspectralenergydis-
2014; Onoetal. 2012; Schenkeretal. 2012; Starketal. 2010; tribution (SED) fitting to the broadband photometry to study
Pentericcietal.2014).Thisdrophastheoreticallyonlybeenex- the stellar properties of these objects in Sect. 4. We summarise
plained succesfully as arising from reionization (Dijkstraetal. anddiscussourfindingsinSect.5,andpresentourconclusions
2011;Jensenetal.2013;Mesingeretal.2015;Choudhuryetal. in Sect. 6. Throughout this paper, we adopt a cosmology with
2015), although additional processes might be involved (e.g. H0 = 70 kms−1 Mpc−1, ΩM = 0.3, and ΩΛ = 0.7. Un-
Dijkstra2014;Choudhuryetal.2015). lesswespecifyotherwise,allgivenstarformationrates(SFRs)
While the broadband photometry can reveal much about are derived from spectral energy distribution (SED) modelling.
the stellar and dust properties of galaxies, the Lyα line pro- All magnitudes refer to the AB system, and we use a Chabrier
fileprovidesimportantinformationonthepropertiesofthegas initial mass function (IMF) over stellar masses in the range
(e.g. Verhammeetal. 2006, 2008; Sawickietal. 2008). Since 0.1–100M(cid:12).
only Lyα photons shifted out of resonance can effectively es-
cape the galaxy, moving gas clouds such as outflows allow
Lyα photons to escape (e.g. Schaereretal. 2011; Laursenetal. 2. Observations
2013; Dijkstra 2014). Dust absorbs the Lyα photons and emits
2.1. Photometry
them at longer wavelengths, while a patchy distribution of the
surrounding medium allows the photons to escape. Therefore, The Hubble Frontier Fields programme1 (FF, PI: J. Lotz;
by careful modelling of the Lyα line, one can learn about the see Lotzetal. 2016; and Koekemoeretal. 2016) targets six
properties of the gaseous medium in and surrounding galax- galaxy clusters with large magnification factors, among which
ies. Recently, it has been demonstrated that galaxies with ex- is AS1063. The programme targets each cluster for a total of
treme optical and near-UV emission lines are often exhibit- 140orbits,dividedover7bandsintheopticalandnearinfrared
ingnarrowLyαemission(Cowieetal.2011;Henryetal.2015; (NIR),reachinga5σdepthof∼29magineachofthesebands.
Izotovetal. 2016; deBarrosetal. 2016; Vanzellaetal. 2016a). We used the available public HST data from this programme,
The fact that these galaxies are found to have Lyα emission retrieved from the Frontier Fields page at the STScI MAST
both at low and high redshift, indicates that these so-called
”Green Peas” might be good analogues of the high-redshift 1 https://archive.stsci.edu/prepds/frontier/
A28,page2of45
W.Karmanetal.:PropertiesoflowluminosityLAEsat z>3
Archive,todetectsourcesandmeasurelocationsandmagnitudes is calculated locally using the weight maps provided by the FF
ofsources,adoptingthecurrentzeropoints,providedbytheACS team. We checked each individual detection if it was contami-
and WFC3 teams at STScI, which are tabulated on the same nated by other closeby galaxies, and removed detections when
MAST Archive page for these specific FF filters. At the time dubious, however we note that some galaxies might still suf-
of writing, the optical bands were fully observed for AS1063, ferfromcontaminationduetoinaccuratebackgroundestimates.
while the NIR observations have had only a single orbit expo- Wenotedthatvisuallydetectedsourcesremainedundetectedby
sure.Weusedthev0.5dataproducts,whichdonotcontainself SExtractorintheF814WandHawk-IKsobservations.Weused
calibrationforthiscluster.Weusedtheimageswithaspatialres- more aggressive detection settings for these bands, and added
olutionof0.060(cid:48)(cid:48),inordertohaveauniformpixelscale,without therelevantdetectionstoourcatalogue.FortheHSTimages,we
xtractor
oversamplingtheNIRimages. comparedtheerrorsprovidedbySE withthosemea-
InadditiontobeingaFFcluster,AS1063isalsopartofthe suredfromtheRMSimagesprovidedbytheFFteam.Wefound
xtractor
Cluster Lensing and Supernova Survey with Hubble (CLASH, thatmultiplyingtheSE errorsbyafactorof1.4rec-
Postmanetal. 2012) survey, which targets 25 gravitationally onciledthedifferentmethods.
lensing clusters with HST in 16 bands. We supplement our FF We also measured photometry in the available Spitzer In-
datawiththeCLASHdatain5additionalbands.Thesedataare fraredArrayCamera(IRAC)imaginginchannel1(λ=3.6µm)
significantlylessdeep,butprovideadditionalinformationforthe andchannel2(λ=4.5µm)7,whichwemosaiced.Thisimaging
brightest objects. For all these filters, which are in addition to coversadepthoftypically∼24.9magnitudesat5σ,althoughthis
thoseusedintheFFprogramme,weadoptthecurrentzeropoints isinhomogeneousacrosstheimagingasaresultoftheincreased
providedbytheACSandWFC3teamsatSTScI2,3. crowding and intracluster light near the centre of the field of
AstheLAEsdiscussedherealllieatz>2.8,theNIRimages view. These depths are also subject to being able to extract re-
fromHSTdonotcoverthewavelengthrangeabove4000Årest- liablephotometryviadeconfusiontechniques.
frame.Informationatlongerwavelengthsisthereforecrucialto Thephotometryinthisimagingwasmeasuredusingthede-
better constrain older stellar populations. We collected Hawk-I confusioncodetphot8.Briefly,theuserprovidesthecodewith
data in order to complement our data at longer wavelengths. spatialandsurfacebrightnessinformationforacatalogueofob-
TheHawk-Iimages4wereretrievedfromtheESOArchive5.The jects as detected in the high-resolution imaging (in this case,
wholedatasetincludes997imagesobtainedinSeptember2015. theHSTF160Wimaging).Thecodeconvolvesgalaxytemplates
After dark and flat correction, a first sky subtraction was per- taken from the high-resolution image with a transfer kernel in
formed without source masking. Sources extracted from these ordertocreatethecorrespondingtemplateinthelow-resolution
background subtracted images were used to solve the astrom- image.Thefluxesoftheselow-resolutiontemplatesareallfitted
etry, where we used Scamp (Bertin 2006) in combination with together, in order to produce a best fitting model of the low-
a catalogue from an ESO-WFI-Rc stacked image as reference. resolution image. For further details, the reader is referred to
warp
UsingS (Bertinetal.2002)wecreatedacoaddedimage, Merlinetal.(2015).
which was used to create a segmentation map. We masked all
With this approach, we found two clear IRAC detections
sourcepixelsintheoriginalframeusingthesingle-frameastro- amongourLAEsample,andanadditionalsixwith∼2−4σde-
metricsolutionandthesegmentationmap,andestimatedanew
tections.Fortheremainingobjects,weusedthelocallyestimated
background from the masked image. Finally, we subtracted the
depthof theimage toset anupper limitat 3times thedepth of
estimated background and created a new final coadded image,
theobservationtobetterconstraintherestframeopticalproper-
witha3σdepthof25.9mag 6.
AB ties.Acaveattoourapproachistheissueofexcessivecrowding
We extracted magnitudes from the optical and NIR images
intheclustercentre,wheresomeofthecandidatesaresituated.
xtractor
using SE . As most of these images have irregular
Fortheseobjects,extractingreliablephotometrywasparticularly
morphologies due to lensing, see Fig. 2, we adopt Kron-like
challenging, even when additionally attempting to fit the back-
apertures rather than spherical apertures. We constructed a de-
groundtoaccountfortheclusterlight,butwefoundthatgiven
tection image for the FF photometry by combining the F435, their relatively large uncertainties, they had little effect on our
F606, and F814W images, and required that each source is de-
results.
tectedatmorethan1σinmorethaneightconnectedpixels.For
the CLASH images, we used the detection image provided by
the CLASH collaboration as a detection image, due to a differ- 2.2. Integralfieldspectroscopy
ent spacing and resolution. We note that this might introduce
an offset in the colours of the galaxies between CLASH and TheMUSEinstrumentmountedontheVLT(Baconetal.2012)
FF detections, but this effect will be small compared to the er- is a powerful tool to blindly look for LAEs behind clus-
ters. Its relatively large field of view (1 arcmin2), spectral
ror bars obtained from the shallower CLASH observations. We
tested the validity of using Kron-radii, different detection im- range(4750–9350Å),relativelyhighspatial(0.2(cid:48)(cid:48))andspectral
agesresultinginpossiblecolourdifferencesduetoourapproach (∼3000)resolution,andstabilityallowedustofindLAEsdown
in Appendix C. We used 32 deblending sub-threshold levels, toanobservedfluxof10−18 erg/s/cm2 ina1×1(cid:48) fieldwithonly
witharelativeminimumcontributionof0.1%.Thebackground 4hofexposure.
AS1063wastargetedwithMUSEtosearchforhigh-redshift
2 ACS zeropoints: http://www.stsci.edu/hst/acs/analysis/ galaxies (Paper I) and simultaneously aid in constraining the
zeropoints
lens properties, see Caminhaetal. (2016b, hereafter C16) for
3 WFC3 zeropoints: http://www.stsci.edu/hst/wfc3/phot_
a detailed description of the used lensing models. The data on
zp_lbn
thesouth-westernhalfofAS1063wasdescribedinPaperIand
4 ESOProgramme095.A-0533,PIBrammer.
5 http://archive.eso.org
6 Aftersubmissionofthispaper,Brammeretal.(2016)releasedapub- 7 PISoifer,programmeID10170.
licversionoftheHawk-Idata.Weperformedacomparisonofthedata, 8 tphotispubliclyavailablefordownloadingfromwww.astrodeep.
andfoundasimilarquality. eu/t-phot/,seealsoMerlinetal.(2015).
A28,page3of45
A&A599,A28(2017)
Fig.1.Distributionmapoftheidentifiedgalaxies,intheforeandbackground(left)andinthecluster(right).Thegalaxiesareoverplottedon
an HST-RGB image, consisting of the F435W (blue), F606W (green), and F814W (red) filters from the FF programme. In the left panel, the
background galaxies are shown with red circles, while foreground objects are shown with blue circles. In the right panel, squares correspond
topassiveclustergalaxies,whilestarsindicateactiveclustergalaxies,wheretheirclassificationisbasedonthepresenceorabsenceofoptical
emissionlines.Thegalaxieshavebeencolouredaccordingtotheirvelocityrelativetothecluster(z=0.3475),withbluercoloursmeaninghigher
velocitiestowardsus,andreddercolourscorrespondingtohighervelocitiesawayfromus,seealsothecolourbarontheright.
consists of 8 exposures of 1400 s each9, or a total integration athighredshift.Finally,weusedthepredictionsfromourlens-
time of 3.1 h. In this paper we add the north-eastern half of ing models to search for additional images of lensed LAEs in
AS1063totheavailabledata,whichhas12exposuresof1440s, MUSE observations. At each of these positions, we then ex-
oratotalintegrationtimeof4.8h10.Wenotethat4ofthelater tracted a spectrum with an aperture of 1(cid:48)(cid:48) radius to determine
exposureswereratedwithagradeC,meaningtheobservational theredshiftofthegalaxy.
requirements were not met. However, we did include these ex-
posurestoourdatacube,astheydidnotdecreasethespatialres-
olution,anddidimprovethedepthofthefinaldatacube. 3. Spectralanalysis
Eachpointingemployedthesameobservingstrategy,where
Wepresentedtheredshiftsobtainedfromthefirstsouth-western
we used observation blocks of 1440s which followed a dither
pattern with offsets of a fraction of an arcsecond and rotations (SW)pointinginPaperI.Herewecomplementthosemeasure-
ments with the new redshifts determined for the north-eastern
of90degreestobetterremovecosmicraysandtoobtainabet-
(NE) half of AS1063. In Appendix A we provide a complete
ter noise map. We followed the data reduction as described in
compilation of all the redshifts obtained from our two MUSE
Paper I for both pointings, and refer to that paper for details.
pointings. We determined redshifts for three additional high-
Here we provide only a brief description of the data reduction.
redshift galaxies with multiple images, of which two were de-
We used the standard pipeline of MUSE Data Reduction Soft-
scribedinC16.Thethirdsystemisaz=3.606LBG,withweak
ware version 1.0 on all of the raw data. This pipeline includes
Lyα emission and two images within the south western MUSE
the standard reduction steps like bias subtraction, flatfielding,
pointing.Thissystem,labelledasSW-70,hasaclearcontinuum
wavelengthcalibration,illuminationcorrection,andcosmicray
and several UV absorption features clearly visible in both im-
removal. We checked all wavelength calibrations for accuracy
ages,seeFig.4.
and verified the wavelength solutions. The pipeline then com-
WeselectedalltheLAEsthatwefoundintheobservations,
bines the raw datainto a datacube that includes the variance of
seeTable1,resultingin6and8LAEsbehindthesouth-western
every pixel at every wavenlength. Consequently, we subtracted
half and the north-eastern half of AS1063 respectively. Two
the remainder of the sky at every wavelength by measuring the
median offset in 11 blank areas at every wavelength, and sub- of these LAEs were discussed in more detail in Vanzellaetal.
(2016a) and Caminhaetal. (2016a). The first is an optically-
tractingthisfromtheentirefield.WemeasuredaspatialFWHM
of1.1(cid:48)(cid:48) inthesouth-westerndatacubeandaFWHMof<1.0(cid:48)(cid:48) in thin, young, and low-mass galaxy that is a good candidate for
a Lyman continuum emitter, which we studied using the ex-
the north-eastern datacube on a point like source selected from
pandedwavelengthrangeandhigherresolutionspectroscopyof
HSTimagesinbothpointings.
X-SHOOTER.ThesecondLAEisaccompaniedbyanextended
For each pointing, we used a spectrally collapsed image of
Lyα nebula, for which we found the most likely origin is scat-
the datacube to find sources. In addition to this, we visually
teredLyαphotonsemittedbyembeddedstarformation.
inspected the datacube to find sources with emission lines that
We determined the redshifts for 116 objects in the NE of
werenotvisibleinthestackedimage.Further,weusedtheHST
AS1063, belonging to 102 individual galaxies. We found 6
images to look for bright galaxies not included in our list, or
foreground objects, 74 galaxies that belong to the cluster, and
galaxiesthatwereonlyvisibleineitherorbothoftheF606Wand
22 galaxies behind the cluster. We identified 10 galaxies that
F814Wbands,asthisisoftenagoodindicationthatthesourceis
showmultipleimages,foratotalof25images,includingthetwo
9 ESOProgramme060.A-9345,PICaputi&Grillo. imagesofthequintiplylensedz = 6.11LAEwhichstilllacked
10 ESOProgramme095.A-0653,PICaputi. spectroscopicconfirmation.Combiningtheseredshiftswiththe
A28,page4of45
W.Karmanetal.:PropertiesoflowluminosityLAEsat z>3
Fig.2.HSTF814WstampsofallLAEsintheMUSEfootprint.TheIDofeachLAEisplottedinthetoprightcornerofeachpanel.Eachstamp
is4(cid:48)(cid:48)oneachside,andascalebarwithasizeof1(cid:48)(cid:48)isshowninthetopleftimage.
SW MUSE observations, results in a total of 9 foreground ob- The total number of spectroscopically confirmed multiple
jects, 121 cluster galaxies, and 42 background galaxies with image systems in AS1063 has been increased from 10 to 18,
MUSEredshiftsinthecentralregionofAS1063.Wedonotfind with17systemshavingatleast2redshiftsspectroscopicallyde-
anyhigh-ionizationUVemissionlinesforanyofthenewLAEs. terminedwithMUSE,seeTableB.1.Wefindoneadditionalfaint
A28,page5of45
A&A599,A28(2017)
Table 1. LAEs behind AS1063, see Table A.1 for quality flags and a clearfromthisfigurethattheobservedfluxesvarywidely,from
crosscorrelationwithmultipleimages.
verybright(SW-49)toveryfaint(NE-97)Lyαlines.Weseethat
all profiles have an asymmetry typical for Lyα lines, and most
ID RA(J2000) Dec(J2000) z showaclearsmallerbluepeakorsuggestthepresenceofasmall
NE-91 342.19238 –44.52505 2.9760 blue peak. The width of the lines varies amongst our sample,
SW-49a 342.17505 –44.54102 3.1169c,e but we find that most LAEs have narrow emission lines. Such
SW-49b 342.17315 –44.53999 3.1169a,b,c,e,f narrowlinessuggestthepresenceoflowcolumndensitygas,and
SW-50 342.16225 –44.53829 3.1160 are suggested to be candidate Lyman continuum emitters (e.g.
SW-68a 342.18745 –44.53869 3.1166a,h Jonesetal. 2013; Verhammeetal. 2015; Vanzellaetal. 2016a;
SW-68b 342.17886 –44.53587 3.1166a,h Dijkstraetal.2016).
NE-93a 342.18283 –44.52028 3.1690
NE-93b 342.19196 –44.52409 3.1690
3.1. PhysicalpropertiesdeducedfromLyα
SW-51 342.17402 –44.54124 3.2275
NE-94a 342.18935 –44.51871 3.2857 The LAEs studied here belong to the intrinsically-faintest
NE-94b 342.19615 –44.52291 3.2857 galaxies (f = 36−2500 × 10−20 ergs−1cm−2) spectroscopi-
λ
NE-96 342.19709 –44.52483 3.4514 cally confirmed at these redshifts, see Fig. 5. We measured
NE-97 342.19100 –44.52679 3.7131 the Lyα luminosity by summing the flux over the spectro-
SW-52a 342.18150 –44.53936 4.1130c scopic width of the Lyα line, and subtracting the average of
SW-52b 342.17918 –44.53870 4.1130c uncontaminated spectral regions redwards and bluewards of
NE-98a 342.19015 –44.53093 5.0510 Lyα. Subsequently, we used our lensing models to correct the
NE-98b 342.19085 –44.53566 5.0510 luminositiesforthemagnificationsduetothegalaxycluster,see
NE-99a 342.18378 –44.52122 5.2373 Table C.1 for the adopted magnification factors. The errors on
NE-99b 342.18874 –44.52276 5.2373 themagnificationsaretypicallyoftheorderof5–10%,whichare
NE-100 342.19701 –44.52212 5.8940 generallylargerthanthephotometricerrors.Forclarity,wehave
SW-53a 342.18106 –44.53462 6.1074b,c,g notpropagatedtheerrorsonourmagnificationintoourerrores-
SW-53b 342.19088 –44.53747 6.1074b,c,g timates of other properties in any table or figure. An additional
NE-118c† 342.18402 –44.53159 6.1074 errorbasedonthemagnificationisthereforeapplicableforflux
NE-118d† 342.18904 –44.53004 6.1074 derivedproperties,suchasluminositiesandstellarmasses.
Aswehavemultipleimagesforsomegalaxies,wecanper-
SW-70a 342.18586 –44.53883 3.6065
form a test on our luminosities and magnifications. We com-
SW-70b 342.17892 –44.53668 3.6065
pared the obtained luminosities for these objects, and used the
Notes. The last galaxy, SW-70, is no LAE, but a Lyman Break meanluminositywhentheyagreedwithin2σ.Forthosesources
galaxy with minimal Lyα emission, and is therefore not considered where a larger difference was found, we reinvestigated the dat-
in the remainder of this paper. Previous redshift determinations by: acube,andfoundthatthelower-luminosityobjectswereunder-
(a) C16;(b) Balestraetal.(2013);(c) PaperI; (d) Richardetal.(2014); estimated.Fortwoofthese,theunderestimationwasduetothe
(e) Vanzellaetal. (2016a); (f) Johnsonetal. (2014); (g) Booneetal. proximityoftheedge,whichresultedinonlyapartialcoverage.
(2013);and (h) Caminhaetal.(2016a). (†) NE-118isthesameimage
For3otherobjects,contaminationbynearbyclustermembersre-
family as SW-53 but located in the NE rather than the SW. To avoid
sultedinanoversubtractionofthecontinuum,whilefor1object
confusion with object NE-53, we listed the objects with the NE-118
alowerS/Nincombinationwithproximitytoaskylineresulted
identifier.
inlowerfluxes.
lineemitterwhichweassociatewithLyαemissionatz=5.894,
3.1.1. Lineprofilemodelling
which has no clear counterpart in the FF images. The lensing
modelpredictsadditionalimagesoutsideoftheobservedMUSE In addition to L , we used our spectra to obtain physical
Lyα
field, but their magnifications are too low to be detected in the properties of the gas surrounding these faint galaxies. We used
HSTimages.Theadditionof2andpossibly3spectroscopically the Lyα line fitting pipeline described in detail in Gronkeetal.
confirmedsystemsatz > 5shouldhelptofurtherconstrainthe (2015) which consists of a pre-computed grid of Lyα radiative
cosmological parameters, see C16, while the increased number transfer models on an expanding shell and a Bayesian fitting
ofz=3−4imageswilldecreasethedegeneraciesanduncertain- framework.
tiesinthemodels.Wecorrectedallpropertiesinthemainbody The expanding shell model (first used by Ahnetal. 2003)
of this paper for gravitational lensing magnification, using the consists of a central Lyα (and continuum) emitting source sur-
modeldescribedinAppendixB. rounded by an outflowing shell of hydrogen and dust. Such a
Due to lensing distortions, most of the galaxies discussed modelhassixfreeparameters:twodescribingthephotonemit-
here have irregular morphologies in the image plane. To opti- ting source (the intrinsic line width σ and equivalent width
misetheS/NoftheLyαline,wecreatedaspatialmaskforeach EW), three for the shell content (the neiutral hydrogen column
i
source,andextractedthespectrumwithinthismask.Eachmask densityNHi,thedustopticaldepthτd andtheeffectivetempera-
was constructed by collapsing the cube over the spectral width tureT whichincludestheapproximateeffectofturbulence)and
oftheLyαline,smoothingthisstackedimagebya3pixelwide theoutflowvelocityv .
exp
boxcar function, and masking out every pixel with values <5σ The pre-computed grid mentioned above consists of
offthebackground.Weverifiedthatthiseffectivelymaskedout 10800 models11 covering the three parameters T, NHi and vexp
nearbycontaminatingsources,whilealsoselectingtheentirere- as they shape the spectrum in a complex, non-linear fashion
gionofLyαemission.
In Fig. 3, we show the observed line profiles of all LAEs 11 The spectra can be accessed online at http://bit.ly/
discoveredinthedatacubes,uncorrectedformagnification.Itis man-alpha/
A28,page6of45
W.Karmanetal.:PropertiesoflowluminosityLAEsat z>3
10
NE-91 SW-49 SW-50 SW-68
8
6
4
2
0
10
NE-93 SW-51 NE-94 NE-96
8
6
4
)
1
−
A◦ 2
2
−
m
0
c
10
1− NE-97 SW-52 NE-98 NE-99
s
8
g
r
e
8 6
1
−
0
1 4
(

λ
f 2
0
10
NE-100 SW-53
8
6
4
2
0
1200 1210 1220 12301200 1210 1220 12301200 1210 1220 12301200 1210 1220 1230
λ (A◦)
Em
Fig.3.LyαlinesfortheLAEsextractedfromtheMUSEdatacube,shownbythebluelines.Thespectraareshiftedtorestframewavelengths,and
thefluxesarenotcorrectedforthegravitationalmagnification.Thegreybandsshowwavelengthswithsignificantskyinterference,whiletheblack
dashedlineshowstherestframewavelengthofLyα.ThesystemicredshiftsofLAEsSW-49andSW-68havebeendeterminedfromthenarrow
UV-emissionlines,whileweadoptedtheredshiftsbasedontheLyαlinefortheotherobjects.
Emitted wavelength (A◦)
) 1200 1400 1600 1800
1
A◦− 120
2 z: 3.6063
m−100
c
80
1
−
s 60
g
er 40
0
2
− 20
0
ux (1 0 CIII Lyα NVSiII OI CII SiIV SiII CIV FeII HeIIOIII]OIII]AlII NiIINIV NiIINiII SiII SiIAlIIIAlIII CIII [NII] [NII] [NII]
Fl 5000 6000 7000 8000 9000
Observed wavelength (A◦)
Fig.4.SpectrumofobjectSW-70a,anewlyidentifiedLBGintheSWofAS1063.Thespectrumhasbeensmoothedforillustrativepurposes,the
greybandscorrespondtowavelengthswithsignificantskyinterference.
A28,page7of45
A&A599,A28(2017)
1044 Shimasaku+2006 21
Dawson+2007
Ouchi+2008
1043 Rauch+2008
Sawicki+2008 20
Verhamme+2008
Blanc+2011
]1042
s Cowie+2011 )
rg/ Kashikawa+2011 2m−19
[eα1041 CHuerntriys-+L2ak0e1+22012 /cH
Ly Mallery+2012 N
L Wofford+2013 g(
1040 Erb+2014 Lo18
Hayes+2014
Hashimoto+2015
1039 Henry+2015 17 Verhamme+2008
Trainor+2015
Hashimoto+2015
This work
0 1 2 3 z 4 5 6 7 This work
1040 1041 1042 1043 1044
Fig.5.DelensedluminosityoftheLyαlineagainsttheredshiftofour
targets, marked by the red stars. We compare this to previously large LLyα [erg s−1]
samples of spectroscopically-confirmed LAEs in the literature, which
Fig. 6. Delensed luminosity in Lyα against the column density deter-
areshownbydotsofdifferentcolours.WeoverplotthevaluesofL(cid:63)at
minedfrommodellingtheLyαprofile.Legendandsymbolsareasin
variousredshiftsfromOuchietal.(2008)withablackdashedline.
Fig.5.
(Verhammeetal. 2015). The grid was created using the radia-
timethatthegaseouspropertiesofsuchfaintsourcesarestudied
tive Monte Carlo code tlac (Gronke&Dijkstra 2014) which
throughLyαmodelling.Therefore,thispresentsoneofthefirst
traces individual photon packages in real- and frequency space studiestodeterminetheeffectstarformationhasongasinthese
(foracomprehensivereviewonLyαradiativetransfersee,e.g.,
faintgalaxies.
Dijkstra2014).
Theeffectoftheremainingthreeparametersismodelledin
post-processingbyassigningaweighttoeachindividualphoton 3.1.2. Shellproperties
packet,whichmeansthattheprocedureaffectstheshapeofthe
WemodelledtheLyαprofilesof12LAEsusingtheapproachde-
lineandnotonlythenormalization.Thisstrategydoesnotonly
scribedaboveyieldingexcellentfitstotheobservedLyαspectra
save computational time but allows to model these parameters
(see Appendix D). We present the properties based on the
continuously,andthus,leadstoamoreprecisesamplingofthe
Lyα line in Table 2, where we already combined the multiple
likelihoodwhencomparingthemodelleddatatoobservations.
imagesintoasingleresult.WemodelledtheLyαlineprofileof
The actual fitting procedure is done by sampling the Gaus-
sian likelihood using the affine invariant Monte-Carlo sampler each image, and combined the results of each modelling into a
single result per LAE. We used the average of all images, af-
emcee (Foreman-Mackeyetal. 2013) using 400 walkers and
600steps12.Inadditiontotheminimalsetofthesixshell-model terdiscardingmultipleimageswhichwerepossiblyaffectedby
closegalaxiesorartefactsinthedatacube.Wedidnotinclude2
parameters, we also fit simultaneously for the redshift z and
LAEs in the modelling, as their Lyα lines were too faint and
thefull-widthathalfmaximumoftheGaussiansmoothingker-
spectrallyunresolved.
nel FWHM. Note that the former adds immense complexity to
In Fig. 6, we show that the column density of the expand-
the fitting procedure as shifting z by a small fraction can alter
ing shell is low in all of the galaxies. In Vanzellaetal. (2016a)
the quality of the fit tremendously. We used the redshift esti-
we reported the results of Lyα line modelling for SW-49 us-
mate from UV emission lines if available or otherwise the red-
shift of Lyα with an intrinsic uncertainty of ∼200 kms−1 (see ing higher resolution X-shooter spectra and an updated shell-
model fitting pipeline. We find that the MUSE and X-shooter
Sect. 3.1) as a prior. Alternatively, the latter, i.e. smoothing the
resultsarebroadlyconsistentwithintheerrorbars,althoughwe
spectrum, makes the likelihood function better behaved. How-
used a larger database reaching lower column densities for the
ever, the width of the smoothing kernel is a function of the ac-
X-shooterspectrum.
tual size of the Lyα halo as well as the measurement aperture.
Therefore,weusedanallowedrangeforFWHMcorresponding We compare the best fit column densities in our sample
to the wavelength-dependent spectral resolution of the MUSE to those found by Verhammeetal. (2008) and Hashimotoetal.
instrument. (2015), who also fit Lyα profiles with expanding shell mod-
els. With all except one LAE best fit with Log(N /cm2) < 20
Gronkeetal. (2015) discussed the uncertainties of using H
Lyαlineprofilefittingforvariouseffects,forexamplemorphol- and only five LAEs with Log(NH/cm2) > 19, we find lower
column densities than the other studies. However, the galaxies
ogy and signal-to-noise ratio. They showed that the expansion
studied here are also ∼1 dex fainter than the ones presented in
velocity and column density can be recovered reasonably well
Verhammeetal.(2008)andHashimotoetal.(2015).
in most cases, while degeneracies and uncertainties are more
prominent among the other parameters. Therefore, we focus in These low column densities support the idea that for these
thispaperonthesetwoquantities,althoughwegivethefullfit- faintgalaxies,theLyαisnotsignificantlybroadenedbyscatter-
ting results in Appendix D. This is, to our knowledge, the first ing.ThisisconsistentwiththepicturethatLyαescapebecomes
easierforfaintergalaxieswithsignificantLyαemission.
12 Forparticularlydifficult,multi-modalcasesweusedaparalleltem- We find no trends in the velocity of the expanding shell
peredensembleMCMCsampler(forareviewsee,Earl&Deem2005) withtheLyαluminosity,butwenotethatwefindrelativelylow
with20temperatures,50walkersand3000steps. outflow velocities in all galaxies. The absence of a correlation
A28,page8of45
W.Karmanetal.:PropertiesoflowluminosityLAEsat z>3
Table2.PropertiesderivedfromLyαspectroscopy,afteraveragingresultsformultipleimages.
ID z f a Log(L /(erg/s))b Log(N /cm−2)c v d
Lyα Lyα H exp
10−20erg/s/cm−2
NE-91 2.9760 108±14 40.91±0.06 – –
SW-49 3.1166 1080±129 41.91±0.06 17.07+0.26 66.78+7.07
−0.03 −6.11
SW-50 3.1169 984±18 41.92±0.01 17.53+0.24 123.63+6.57
−0.23 −5.85
SW-68 3.1166 1942±16 42.22±0.01 20.01+0.20 463.18+20.40
−0.25 −48.58
NE-93 3.1690 431±26 41.58±0.02 18.36+1.16 44.65+27.91
−0.65 −1.79
SW-51 3.2271 36±2 41.24±0.01 18.53+0.50 97.99+14.63
−0.71 −17.76
NE-94 3.2857 524±27 41.70±0.02 19.22+0.37 183.33+23.37
−0.20 −8.92
NE-96 3.4514 228±15 41.39±0.03 18.59+0.14 56.68+7.35
−0.25 −8.85
NE-97 3.7131 155±19 41.30±0.06 20.39+0.09 180.58+13.00
−0.14 −12.00
SW-52 4.1130 106±4 41.24±0.02 17.12+0.32 98.10+12.93
−0.17 −10.29
NE-98 5.0510 186±8 41.70±0.02 19.42+0.22 149.27+11.83
−0.21 −11.83
NE-99 5.2373 113±11 41.47±0.05 17.99+0.97 107.87+223.75
−0.56 −127.68
NE-100 5.8940 60±32 41.36±0.33 – –
SW-53 6.1074 2694±67 43.04±0.01 19.81+0.07 150.56+3.69
−0.08 −4.1
Notes.Thecolumnsare(a)lens-correctedLyαflux;(b)luminosity;(c)thehydrogencolumndensity,and(d)expansionvelocity.
Table3.Parameterspaceusedforconstructingstellartemplatesusedin theageoftheUniverse.WeusedaCalzettietal.(2000)extinc-
ourSEDfitting. tioncurvetoattenuateallstellartemplates,with0< E(B−V)<
1.5.We enabledaddingnebularemissionlinestothe templates
Range Nr.steps basedontheirUVluminosityasdescribedinIlbertetal.(2009),
ii
where the line fluxes of [O ] λ3727 are derived using the
Age 0.01Myr–2.3Gyr 50 iii
current SFR, and [O ] λλ4959,5007, Hα, and Hβ are then
E(B−V) 0–1.5 20
scaledfromlocallyderivedlineratios.Theadditionofemission
Z 0.0004–0.02(Z(cid:12)) 4 lines to stellar-population templates is shown to be important
τ 0.001Gyr–5Gyr 5 in the high-redshift Universe (e.g. Schaerer&deBarros 2009;
deBarrosetal.2014).
Notes. The stepsizes are logarithmic distributed for the ages, while
We fitted the SED of each image and combined the re-
we use an irregular spacing for the stepsizes of E(B−V) where we
sults of the multiple images in a similar method as for the
finelysamplethelowvalues,anduselargerstepsforthehighervalues.
Themetallicitiescorrespondtothem32,m42,m52andm62modelsof Lyα luminosities and Lyα line fitting. We used the average re-
BC03. sults of multiple images when the quality of the photometric
data of each image is similar, but adopt the results of only the
bestconstrainedimageifthedifferenceissignificant,forexam-
between the Lyα luminosity and the expansion velocity of pleforNE-94weonlyusedimagea.Forgalaxieswherewesus-
the shell is unexpected, as both are considered to correlate pectcontaminationfromnearbygalaxies,forexampleSW-68b,
with the SFR (e.g. Weineretal. 2009; Bradshawetal. 2013; weusedonlytheimageswithoutcontamination.Weperformed
Chisholmetal.2015).Apossibleexplanationcouldbethatthe tests on the reliability of our results in Appendix C, and found
outflowspeeddoesnotfollowtheSFRatlowmasses,seebelow, that although there are few constraints in the restframe optical,
orthattheLyαluminositydependsmorestronglyontheescape ourresultsdonotchangesignificantly.
fractionofLyαphotonsthanontheSFR.
We present the results from the SED fitting in Table 4. We
remind the reader that most of the photometry is in the rest-
frame UV, but that the Hawk-I and IRAC filters trace the rest-
4. Stellarproperties
frame optical. We detected 6 LAEs in the K-band and 1 LAE
We used the constructed photometric catalogue in combination in the IRAC filters, but the non-detections provide important
withthespectroscopicredshiftstoperformaspectralenergydis- upperlimitswhendeterminingstellarmassesandshowthatthere
tribution (SED) fitting on our selected sample. We used LeP- isnohiddendominantoldpopulationofstars.
hare(Arnoutsetal.1999;Ilbertetal.2006)incombinationwith The masses we derived from our SED fitting are very low,
Bruzual&Charlot (2003, hereafter BC03) templates to fit the varyingfrom∼106 M to∼108 M ,significantlylowerthanthe
(cid:12) (cid:12)
photometry with stellar population models (see Table 3 for our stellarmassesexploredinpreviousstudiesofLAEs.Thisisnot
model parameters). The set of stellar populations consists of surprising,asthesegalaxiesareamongtheintrinsicallyfaintest
an exponentially declining star formation histories, SFR(t) = discovered so far with absolute UV-magnitudes ranging form –
SFR × e−t/τ, with different values for τ. In addition, we cre- 19 to –14, and illustrates once more the advantages of gravita-
0
ated templates with three different metallicities, and ages up to tionallensing.Wenotethatfor2LAEsdiscoveredhere(NE-99
A28,page9of45
A&A599,A28(2017)
e hare
Table4.StellarpropertiesderivedfromSEDmodellingusingL P incombinationwithBC03templates.
ID Log(M /M ) Log(Age /yr) E(B−V) Log(SFR /(M yr−1)) Log(sSFR/yr) β M
(cid:63) (cid:12) SSP SED (cid:12) UV
NE-91 8.02+0.20 8.26+0.45 0.03+0.03 –0.38+0.16 –8.56+0.39 –2.42±0.09 –17.77
−0.19 −0.67 −0.02 −1.32 −1.10
SW-49 6.91+0.18 6.70+0.89 0.045+0.03 –0.16+0.90 –7.00+0.80 –2.79±0.08 –17.09
−0.08 −0.06 −0.015 −0.30 −0.30
SW-50 8.23+0.05 6.28+1.28 0.30+0.00 1.88+0.69 –6.34+0.69 –1.95±0.05 –18.05
−0.06 −0.63 −0.20 −1.99 −2.82
SW-68 7.78+0.26 7.12+0.69 0.10+0.10 0.36+0.67 –7.39+0.66 –2.09±0.08 –18.18
−0.16 −0.48 −0.05 −0.51 −0.90
NE-93 6.24+0.27 6.49+0.84 0.035+0.036 -0.47+0.91 –6.61+0.80 –2.85±0.46 –15.63
−0.03 −0.03 −0.005 −0.07 −0.03
SW-51 6.81+0.49 7.59+0.82 0.05+0.05 –0.96+0.30 –7.87+0.63 –2.31±0.05 –15.83
−0.34 −0.65 −0.03 −0.27 −0.69
NE-94 8.50+0.16 6.70+0.66 0.30+0.05 1.61+0.93 –6.74+0.64 –1.77±0.08 –18.30
−0.16 −0.66 −0.05 −0.59 −0.75
NE-96 7.33+0.41 7.24+0.87 0.10+0.15 –0.36+1.34 –7.46+1.12 –2.14±0.22 –16.60
−0.44 −0.96 −0.07 −0.64 −1.26
NE-97 6.35+0.18 6.40+0.72 0.03+0.03 –0.18+0.81 –6.47+0.71 –2.90±0.15 –15.98
−0.17 −0.75 −0.02 −0.69 −0.85
SW-52 6.36+0.38 7.12+0.53 0.03+0.03 –0.90+0.61 –7.26+0.74 –2.50±0.28 −15.33
−0.25 −0.66 −0.02 −0.33 −0.88
NE-98 6.46+0.10 6.16+0.36 0.01+0.02 0.28+0.66 –6.19+0.68 –3.17±0.02a –16.19
−0.09 −0.66 −0.01 −0.54 −0.45
NE-99 – – – – – – –14.42
NE-100 – – – – – – >–14.70
SW-53 7.47+0.00 6.04+0.30 0.00+0.00 1.37+0.60 –6.10+0.60 −3.17±0.02a –18.97
−0.00 −0.54 −0.00 −0.30 −0.30
Notes.Thepropertieswerederivedafteraveragingtheresultsofmultipleimagesofthesamesource,ifapplicable.SeeAppendixCforresultsof
allindividualLAEs.(a)ThisisthemaximalUVslopeinourusedtemplates,thephotometricUV-slopeissteeperandthesmallerrorsaretherefore
notrepresentativebutaresultofourmethodofcalculatingβ.
and NE-100), we recovered only a single detection through all 1044
deep FF filters, which is in the filter containing Lyα. The com-
pletion of the NIR-imaging of AS1063 in the summer of 2016
shouldaddmoredetectionstotheseobjects,andwillbettercon-
1043
strainthepropertiesofthesepossiblyevenlessmassivegalaxies.
In Fig. 7, we compare the stellar masses of our galaxies to
theirLyαluminosities.Weseethatthelowermassesarepaired s]
g/ 1042
withlowerluminosities.Thelowluminositiesandmassesfound r
e
here in comparison to previous literature results confirm again [
α
thattheseobjectsprobeanewregionofparameterspace. Ly
L
ThepresenceofsuchnarrowandstrongLyαemissionisal- 1041
ready a clear indication that there is little dust present. We find
thatthemedianmarginalisedE(B−V)valuesareallverysmall,
withonly3galaxieshaving E(B−V) > 0.1,inagreementwith 1040
previousstudies(e.g.Ateketal.2014).
The ages of these very low mass objects are relatively low, 6 7 8 9 10 11
Log(M /M )
thatis1–100Myr,seeFig.8.Wefindthatonlytwogalaxieshave
fl
anage>100Myr,andageseemstodecreasewithredshift.The
Fig.7.Stellarmassversusthelens-correctedLyαluminosity.TheLAEs
youngstellaragesindicatethatthesearesystemsthatarerapidly described hereare comparedto a collectionof previousLAE studies,
buildinguptheirmass.ThisisconfirmedbytheSFRthatweob- with the colours identical to Fig. 5, where we have supplemented the
taingiventhelowstellarmasses,seeFig.9,aswiththecurrent resultsfromBlancetal.(2011)withthoseofHagenetal.(2014).
SFR,mostgalaxieswilldoubletheirmasswithin107 yrs.Asa
consequence of these young ages, the models produced by dif-
ferent values of τ are very similar. Therefore, the current data
are unable to distinguish between the different star formation low mass galaxies, this extrapolation is rather uncertain due to
histories. a degeneracy between the slope of the power-law and its zero-
We compare the SED-derived SFR to the SFR extrapolated point.Ifweusethesteeppower-lawrelationdescribedforz∼4
from the SFR-stellar mass relation determined for more mas- galaxies by Salmonetal. (2015), the number of LAEs above
sive galaxies, both at lower and similar redshift. We find that this relation will decrease by a factor 2. We note however, that
most of the LAEs fall above this relation if an extrapolation manystudiesfavouraslopeofα=1(e.g.Gonzálezetal.2010;
of the power-law relation described by Whitakeretal. (2014) Whitakeretal.2014;Ilbertetal.2015),whichwouldmakemost
is considered. Because very little is known about the SFR in oftheseLAEstarburstinggalaxies.
A28,page10of45