NASA Technical Reports Server (NTRS) 20130013631: The Radial Distribution of Star Formation in Galaxies at Z approximately 1 from the 3D-HST Survey

0.54 MB·English

by  NASA Technical Reports Server (NTRS)

TheAstrophysicalJournalLetters,763:L16(6pp),2013January20 doi:10.1088/2041-8205/763/1/L16
(cid:2)C2013.TheAmericanAstronomicalSociety.Allrightsreserved.PrintedintheU.S.A.
THERADIALDISTRIBUTIONOFSTARFORMATIONINGALAXIESATz ∼ 1FROMTHE3D-HSTSURVEY
EricaJuneNelson1,PieterG.vanDokkum1,IvelinaMomcheva1,GabrielBrammer2,BrittLundgren3,
RosalindE.Skelton1,KatherineE.Whitaker4,ElisabeteDaCunha5,NataschaFo¨rsterSchreiber6,
MarijnFranx7,MattiaFumagalli7,MariskaKriek8,IvoLabbe7,JoelLeja1,ShannonPatel7,Hans-WalterRix5,
KasperB.Schmidt9,ArjenvanderWel5,andStijnWuyts6
1AstronomyDepartment,YaleUniversity,NewHaven,CT06511,USA
2EuropeanSouthernObservatory,AlonsondeCo´rdova3107,Casilla19001,Vitacura,Santiago,Chile
3DepartmentofAstronomy,UniversityofWisconsin-Madison,Madison,WI53706,USA
4AstrophysicsScienceDivision,GoddardSpaceFlightCenter,Greenbelt,MD20771,USA
5MaxPlanckInstituteforAstronomy(MPIA),Ko¨nigstuhl17,D-69117,Heidelberg,Germany
6Max-Planck-Institutfu¨rextraterrestrischePhysik,Giessenbachstrasse,D-85748Garching,Germany
7LeidenObservatory,LeidenUniversity,Leiden,TheNetherlands
8DepartmentofAstronomy,UniversityofCalifornia,Berkeley,CA94720,USA
9DepartmentofPhysics,UniversityofCalifornia,SantaBarbara,CA93106,USA
Received2012October15;accepted2012December18;published2013January4
ABSTRACT
Theassemblyofgalaxiescanbedescribedbythedistributionoftheirstarformationasafunctionofcosmictime.
Thanks to the WFC3 grism on the Hubble Space Telescope (HST) it is now possible to measure this beyond the
local Universe. Here we present the spatial distribution of Hα emission for a sample of 54 strongly star-forming
galaxiesatz∼1inthe3D-HSTTreasurysurvey.BystackingtheHαemission,wefindthatstarformationoccurred
inapproximatelyexponentialdistributionsatz ∼ 1,withamedianSe´rsicindexofn = 1.0±0.2.Thestacksare
elongated with median axis ratios of b/a = 0.58±0.09 in Hα consistent with (possibly thick) disks at random
orientationangles.Keckspectraobtainedforasubsetofeightofthegalaxiesshowclearevidenceforrotation,with
inclinationcorrectedvelocitiesof90–330kms−1.Themoststraightforwardinterpretationofourresultsisthatstar
formationinstronglystar-forminggalaxiesatz∼1generallyoccurredindisks.Thedisksappeartobe“scaled-up”
versionsofnearbyspiralgalaxies:theyhaveEW(Hα)∼100Åouttothesolarorbitandtheyhavestarformation
surfacedensitiesabovethethresholdfordrivinggalacticscalewinds.
Key words: galaxies: evolution – galaxies: formation – galaxies: high-redshift – galaxies: kinematics
anddynamics–galaxies:starformation–galaxies:structure
Online-onlymaterial:colorfigures
of cosmic time. In particular, Epinat et al. (2009) and Kassin
1.INTRODUCTION
et al. (2012) suggest that galaxies become cooler and more
Galaxyformationisacomplexprocessinvolvingstarbursts, rotation-dominatedwithtime,z∼1–0beingtheepochof“disk
mergers,andstronggasflows(e.g.,Hopkinsetal.2006;Brooks settling.” This is interesting because most of the stars in the
et al. 2009; Dekel et al. 2009). Furthermore, stellar migration disksofgalaxiesliketheMilkyWaywereformedinthisepoch.
andsecularprocessescanchangethestructureofgalaxiesatlate Itisnowpossibletoobtainhighspatialresolution((cid:3)1kpc)
times (e.g., Rosˇkar et al. 2008; Grand et al. 2012). Therefore, informationonHα emissionatz ∼ 1withahighStrehlratio,
evenifwecouldperfectlylocateandagedateeverystarinthe owingtothenear-IRslitlessspectroscopiccapabilitiesprovided
MilkyWay,westillcouldnotsaywhereandwithwhatstructural by the WFC3 camera on the Hubble Space Telescope (HST).
and kinematic properties those stars formed. The only way to In Nelson et al. (2012), we used data taken as part of the
establishwhereagalaxy’sstarsformed,andhencehowitwas 3D-HST survey to build on previous ground-based studies by
assembled,istomapthestarformationwhilethosestarswere mappingtheHαandstellarcontinuumwithhighresolutionfor
forming. asampleof57galaxiesatz∼1andshowedthatstarformation
ObtainingHαandstellarcontinuummapsofz(cid:2)1galaxies, broadlyfollowstherest-frameopticallight,butisslightlymore
withthe∼1kpcresolutionnecessarytoputconstraintsonthe extended. Here we stack the Hα maps of these galaxies to
spatialdistributionofstarformation,ischallengingandhasso construct high signal-to-noise ratio (S/N) radial profiles and
far only been possible using a combination of adaptive optics measure structural parameters of stellar continuum emission
andintegralfieldunitson8–10mclasstelescopes.Thesestudies and star formation. These unique, high spatial resolution data
paintacomplexpicture:theyfindthatstar-forminggalaxiesat from 3D-HST are combined with kinematics from the Near
z ∼ 2 are a mix of “puffy” and often clumpy rotating disks, InfraredSpectrometer(NIRSPEC)ontheW.M.Kecktelescope
mergers,andmorecompactdispersion-dominatedobjects(e.g., (McLean et al. 1998). We assume H = 70 km s−1 Mpc−1,
0
Genzeletal.2008;Shapiroetal.2008;Crescietal.2009;Fo¨rster ΩM =0.3,andΩΛ =0.7.
Schreiberetal.2009,2011;Lawetal.2009;Jonesetal.2010;
Mancinietal.2011).
2.3D-HSTANDSAMPLESELECTION
Some studies claim that there are trends in the structural
propertiesofHαasafunctionofredshift,implyingthattheway The 3D-HST survey, a 248-orbit Treasury program on the
galaxiesassembletheirstarsvariesfundamentallyasafunction HST,suppliesthetwo-dimensionalemissionlinemapsneeded
1
TheAstrophysicalJournalLetters,763:L16(6pp),2013January20 Nelsonetal.
thehistogramofFigure1;e.g.,Noeskeetal.2007;Daddietal.
2007; Whitaker et al. 2012). These galaxies were selected in
Hα not IR; while this selection may bias the sample, it may
possiblyalsohavetheadvantagethatitaddslessuncertaintyto
interpretingthespatialdistributionofstarformationduetodust
extinction.
3.THESPATIALDISTRIBUTIONOFSTARFORMATION
3.1.StackedHαandStellarContinuumEmission
WestackthehighspatialresolutionHα mapsfrom3D-HST
tocreateaverageHα maps—increasingtheS/Nandproviding
forareliablemeasurementofthestructuralpropertiesofHαto
largeradiiforthissampleatz∼1.Wedividethesampleinthree
binsbasedontheirHαsize10inordertohomogenizethesample
andinvestigatetrendswithsize.11Themedianstellarmassesof
the stacks are log(M∗) = 10.0,10.2,10.2. The galaxies are
normalized by their F140W flux and centered according to
their luminosity-weighted F140W image centers. We masked
the [Sii] λλ6716,6731 Å emission in the fit. At our spectral
resolution, the FWHM of a line is ∼100 Å, meaning that Hα
Figure1.LocationofselectedgalaxiesintheSFR–massplane.Sixsources
λ6563 Å and [Nii] λ6583 Å are unresolved but Hα and [Sii]
withoutIRphotometryarenotshowninthefigure.Thehistogramshowsthe
positionoftheselectedgalaxiesrelativetotherestofthegalaxiesbydividing areseparatedby∼3resolutionelements.Werotatedtheimages
apower-lawfitoutofthedistribution.Thepurplelineshowsthefit,thelight basedontheirGALFIT-derivedF140Wpositionanglestoalign
purpleregiondelineates±1σ.Theselectedgalaxiesappeartolielargelyonthe them along their major axes, summed them, and divided the
upperhalfofthe“mainsequence”inSFR(IR+UV).
resultingstackbythesummedmasks.Finally,inordertocorrect
(Acolorversionofthisfigureisavailableintheonlinejournal.)
fortheeffectsofthepoint-spreadfunction(PSF),thestackswere
deconvolvedbyaddingtheGALFIT-derivedPSF-deconvolved
to measure the spatial distribution of star formation. The modeltotheresidualsofthefit,amethoddevelopedinSzomoru
WFC3 G141 grism provides spatially resolved spectra of all etal.(2010).
sources in the field. The wavelength range of the G141 grism, Figures 2 and 3 show the stacked, PSF-corrected Hα emis-
1.15μm < λ < 1.65μm, covers the Hα emission line for sion.TheS/Ninthestacksishigh,andtheemissionisclearly
galaxies in the redshift range 0.7 < z < 1.5. The survey also spatiallyextended.Furthermore,boththerest-frameRbandand
providesbroadbandnear-infraredimagingintheF140Wfilter, Hα maps are clearly elongated. We quantified the structural
which samples the rest-frame R band for z ∼ 1. Using the propertiesofthestackswithGALFIT(Pengetal.2002)using
same camera under the same conditions to map both Hα and empirical PSFs. For the continuum and Hα, we find weighted
continuum emission allows us to compare their distributions mean axis ratios of b/a(R) = 0.55 ± 0.07 and b/a(Hα) =
directly, a crucial advantage of this strategy. The maps have 0.58±0.09,respectively,consistentwithexpectationsforran-
a spatial sampling of ∼0.5 kpc at z ∼ 1 (0(cid:4).(cid:4)06 pixels). The domly oriented disks with thickness c/a = 0.26, 0.33. This
data presented in this Letter are based on the ∼70 pointings, thicknesscouldbeattributedtodisksbeingintrinsicallythickat
roughlyhalfofthefulldataset,thatwereobtainedpriorto2011 z ∼ 1,heterogeneouspropertiesofthegalaxiesbeingstacked,
June (see Nelson et al. 2012). The survey, data reduction, and theeffectofbulges,andaselectionbiasedtowardface-ongalax-
determinationofderivedquantities(e.g.,redshiftsandmasses) ies,ornoise.
aredescribedinBrammeretal.(2012)andvanDokkumetal.
(2011). 3.2.RadialProfilesareNearlyExponential
The 3D-HST grismspectra have high spatialresolution and
Figure 3 shows the radial profiles derived from the stacks.
lowspectralresolution(R ∼ 130),meaningthatemissionline
Radial profiles of the PSF-corrected stacks were created by
structurereflectsalmostexclusivelyspatialstructure(morphol-
summing the flux in elliptical radii. A 30% correction was
ogy) in contrast to data with high spectral resolution where
applied to account for [Nii] (see Section 4). All uncertainties
structurereflectsvelocity(rotationordispersion).Emissionline
werederivedbybootstrapresamplingthestacks.
mapsofgalaxiesaremadebysubtractingthecontinuumemis- The Hα emission ranges from exponential to much less
sion from the two-dimensional spectrum. The details of the
centrallyconcentratedthanexponential.Thestellarcontinuum
procedurearedescribedinNelsonetal.(2012),Lundgrenetal.
emissioninallthestacksisnearlyexponential.
(2012),andK.B.Schmidtetal.(2013,submitted).
Theradialequivalentwidthprofiles(rightpanelsofFigure3)
We selected galaxies with redshifts 0.8 < z < 1.3, total are flat to within a factor of two and EW(Hα) > 100 Å
F140W AB magnitude <21.9, and rest-frame EW(Hα) >
for all stacks within 1.5r . This suggests that the high global
100Å.Wealsolimitedthesampletogalaxieswitheffectively EW(Hα)s seen in these gealaxies are not driven primarily by
nocontaminatingfluxfromthespectraofothernearbyobjects.
centralstarburstsbutbystrongstarformationatallradii.
Thepositionofthesampleinthe0.8<z <1.3starformation
rate(SFR)–stellarmassplaneisshowninFigure1.SFRswere
10 ThesizesaremeasuredusinggrowthcurvesasdescribedinNelsonetal.
derivedfromtheIR+UVluminositiesasdescribedinWhitaker
(2012).
et al. (2012). The selected galaxies lie predominantly on the 11 TheresultsaresimilarwhenthebinsaredefinedaccordingtotheF140W
upper half of the “star-forming main sequence” (as shown in sizes.
2
TheAstrophysicalJournalLetters,763:L16(6pp),2013January20 Nelsonetal.
Hα Line Emission N =10 N =35 N = 9
s stack stack stack
ck r(Hα)=1.2±0.1 r(Hα)=3.7±0.2 r(Hα)=6.3±1.8
a e e e
t
S
d
e
v
ol
s
e
R
-
y
all
ti
a
Sp b/a=0.50±0.05 n(Hα)=1.2±0.6 b/a=0.60±0.06 n(Hα)=0.9±0.1 b/a=0.61±0.13 n(Hα) < 1.0
]
2
c
p
k 1.00
/
r
y
/
M
[ 0.10
)
R
F
S
(
Σ 0.01
−10 −5 0 5 10 −10 −5 0 5 10 −10 −5 0 5 10
r [kpc] r [kpc] r [kpc]
Figure2.Stacksofthehighspatialresolution,PSF-correctedmapsofHαemission(top).StackingwasdonebasedonHαsize;thenumberofgalaxiesincluded
ineachsizebinislisted(Nstack)asisthemajoraxiseffectiveradius(inkpc)ofeachstackmeasuredbygalfit(re).ThestackshaveaweightedmeanSe´rsicindex
n(Hα)∼1andaxisratioofb/a(Hα)=0.58±0.09,consistentwithdisksatrandomorientationangles.Bottompanelsshowcorrespondingradialprofileswithdark
purple—small,mediumpurple—mid-sized,andlightpurple—large(Figure3,Section3.2).ThestackedHαemissionalwayshasSe´rsicindexn(cid:2)1.
(Acolorversionofthisfigureisavailableintheonlinejournal.)
Figure3.StacksofHα (left)andrest-frameR-band(middle)emissionhavenearlyexponential(orshallower)radialprofiles.TheEW(Hα)profile(discussedin
Section3.2)isshownintherightpanel.Thebottompanelsshowtheprofilesnormalizedbytheireffectiveradius.ThehorizontallineistheSFRsurfacedensity
criterionfordrivinglarge-scaleoutflows(Heckman2002).AsinFigure2,thedark,medium,andlightcolorscorrespondtothesmall,medium,andlargestacks,
respectively,asshownatthetop.ThespatialdistributionofHαemissionisexponentialorshallower.
(Acolorversionofthisfigureisavailableintheonlinejournal.)
3
TheAstrophysicalJournalLetters,763:L16(6pp),2013January20 Nelsonetal.
2.5
2] 10.00
c
p
2.0 /k
yr
or /
Col 1.5 [M 1.00
V y
U- sit
n
d e
e 1.0 D
v
Deri ace 0.10
urf
0.5 S
R
F
S
0.01
0.0
0 2 4 6 8 10 12 0 2 4 6 8 10 12
r[kpc] r[kpc]
Figure4.Leftpanelshowstheradialcolorprofilesforeachofthestackswithblackforthesmallstack,darkgrayforthemediumstack,andlightgrayforthelarge
stack.TherightpanelshowstheimplieddustcorrectedradialprofilesofstarformationsurfacedensitywiththecolorsasinFigures2and3.Errorbarontheright
paneldenotesatypicaluncertainty.Dust-correctedradialprofilesofstarformationremainnearlyexponential.
(Acolorversionofthisfigureisavailableintheonlinejournal.)
The derived Se´rsic indices (Figure 2) of the Hα and stellar corrections:
continuumprofilesaren(Hα)=1.2±0.6,0.9±0.1, n(Hα)<
1andn(F140W)=1.8±0.4,1.4±0.1,1.1±0.2withweighted log(SFR(IR+Hα))−log(SFR(Hα))∼0.3×(U −V). (1)
means of n(Hα) = 1.0±0.4 and n(F140W) = 1.4±0.2. As
Thedust-correctedradialprofilesareshownintherightpanel
showninvanDokkumetal.(2010),structuralparametersmea-
ofFigure4.Thesmallestgalaxiesappeartohavestarformation
suredfromstacksareclosetothemeanvaluesfortheindividual
galaxies going into the stacks. The upper limit of n(Hα) < 1 that is marginally steeper than exponential. The large galaxies
forthelargestHα stackreflectsthefactthatthederivedn(Hα) havethelargestradialgradientsbuttheirimpliedstarformation
isstilllesssteepthanexponential.Althoughthisanalysisassigns
depends somewhat on the details of the fit (treatment of sky,
fittingregion)butalwaysn(Hα)<1. the entire observed color gradient to extinction, the centers of
WefindthatthestackedspatialdistributionofHαforgalaxies the radial color profiles could be red because of dust or age.
inthissampleisexponentialorshallower.Allstackshaven<2, Importantly, the Se´rsic indices of the star formation in these
stacks based on the implied dust correction remain close to
implyingabulgefractionoflessthan20%(vanDokkumetal.
1998). The radial Hα profiles (Figure 3, left), seem to show a one:n=1.4,1.1,0.6foreachofthestacks,weightedmean=
trend toward lower n(Hα) with increasing size. However, this 1.0 ± 0.4. So, even when accounting for dust, the averaged
starformationinthesegalaxieshasanearlyexponentialspatial
trendisnotstatisticallysignificantgiventheerrorsinthederived
distribution.
Se´rsic indices. Additionally, note that the median galaxy with
EW(Hα) > 100 Å has a high enough star formation surface
densitytodrivewindsoutto∼1re (Heckman2002). 4.KINEMATICS
The flattening of the stacks and the exponential Hα profiles
3.3.EffectsofDust suggestthatthestarformationoccursinrotatingdisks.Totest
this,wemeasuredkinematicsforasubsetofthissampleusing
A major uncertainty in the interpretation of the radial Hα NIRSPEC on the Keck II telescope on 2012 April 9–10. The
profileistheeffectofdifferentialdustextinction:ifsomeparts sample comprises eight galaxies chosen from the small (1),
ofthegalaxiesaremoreobscuredthanothers(seee.g.,Wuyts middle(5),andlarge(2)stack,whichhavesizes,masses,Se´rsic
et al. 2012), the derived radial profile of Hα would not reflect indices, SFRs, and axis ratios representative of the sample as
theradialprofileofstarformation.Toassesstheimportanceof awhole.
this effect, we estimate the extinction as a function of radius. We used the low dispersion mode of NIRSPEC with 0(cid:4).(cid:4)5
Wedeterminetheextinctionfromtheradialcolorprofile. seeing and a slit width of 0(cid:4).(cid:4)7, giving a spectral resolution of
Thecolorprofileisdeterminedfromthecombinationofthe σ ∼80kms−1intheJband(comparedto∼500kms−1inthe
F140WstackswithACSF814Wstacks,whichcanbeconverted grismspectra).Theslitwasalignedalongthemajoraxisofeach
intorest-frameU−Vcolorprofiles.Asshownintheleftpanel galaxyandobservationswereconductedinaseriesoffour900s
ofFigure4,formostoftheradialextentofthestacksthecolor exposures,ditheringalongtheslit.Thedatareductionfollowed
isfairlyconstant,butitbecomesredderinsidearadiusof2kpc standard procedures for long-slit spectroscopy (see, e.g., van
(seealsoSzomoruetal.2012).Toestimateroughlyhowcolor Dokkumetal.2004).Weextractedkinematicinformationfrom
translates into missed star formation, we assume that the star thetwo-dimensionalspectrabyfittingaGaussiantotheHαand
formationthatisnotcapturedbyHαwillbecapturedbytheIR [Nii] emission simultaneously at each spatial position along
(Kennicutt et al. 2009). Using photometry from R. E. Skelton the slit. The median [Nii]/Hα is 0.3. Assuming the velocity
et al. (2013, in preparation), in addition to rest-frame U−V shearinthetwo-dimensionalspectraisduetorotation,wetake
colorsandIR-basedSFRsfromtheNEWFIRMMediumBand therotationalvelocitytobethevelocitydifferencebetweenthe
Survey(Whitakeretal.2012),empiricalrelationswerederived geometrical center of the galaxy and the maximum velocity.
forthetranslationofrest-frameU−Vcolorsintostarformation We correct the rotation velocities for inclination angle using
4
TheAstrophysicalJournalLetters,763:L16(6pp),2013January20 Nelsonetal.
Figure5.Shownherearetherotationcurves(middlerow)derivedfromtheNIRSPECspectra(insetsincorners),oneexamplefromeachstackwiththeimplied
rotationvelocities(vcorr—correctedandvuncorr—uncorrectedforinclination)andvelocitydispersion(σ)listedinkms−1.Thetoprowshowsthecorresponding
rest-frameR-bandimagesandthebottomrowshowsafalsecolorimagewiththeHαemissioninredandstellarcontinuuminblue.Intheserows,ellipsesmarkthe
R-bandandHαeffectiveradii(re(R),re(Hα)),respectively,andgrayarrowshaveascaledlengthcorrespondingtotheprojectedspecificangularmomentum.
(Acolorversionofthisfigureisavailableintheonlinejournal.)
the3D-HSTaxisratios.Wecalculateone-dimensionalvelocity appeartoberotating.(e.g.,Wisnioskietal.2011;Epinatetal.
dispersions by fitting a Gaussian to the Hα emission in the 2012;Swinbanketal.2012).
central row of each spectrum. We calculate the instrumental
σ to be ∼80 km s−1 by fitting a Gaussian to the sky lines. 5.DISCUSSION
The intrinsic velocity dispersion is calculated by subtracting
the instrumental dispersion from the measured dispersion in ThecentralresultofthisLetteristhattheradialdistributionof
quadrature. stackedHαemissioninz∼1galaxiesisclosetoexponentialout
We find that all galaxies show velocity shear in their to∼10kpc.Combinedwiththeaxisratiosofthestacksandthe
two-dimensional spectra, with derived rotation velocities of kinematicsofasubsetofthesample,themoststraightforward
90–235kms−1 uncorrectedand110–330kms−1 correctedfor interpretationisthatstarformationseemstypicallytooccurin
inclination.12 Figure5showstherotationcurvesforoneexam- disks at z ∼ 1 at least for galaxies with high EW(Hα). The
plegalaxyforeachstack.Thesespectraareofthebestquality factor of ∼10 increase in the SFRs of galaxies from z = 0
butappeartoberepresentativeoftheeight.Thevelocityprofile to z = 1 (e.g., Damen et al. 2009; Fumagalli et al. 2012) is
of the smallest galaxy in Figure 3 does not show evidence for apparently driven by increased star formation activity in disks
a turnover, which means the measured maximum velocity is a ratherthanamuchgreaterprevalenceofmerger-drivencentral
lowerlimitontherotationvelocityatlargeradii.Weconclude starbursts—consistentwithotherstudies(e.g.,Rodighieroetal.
thatthekinematicsareconsistentwiththediskinterpretationof 2011;Wuytsetal.2011,2012).
thestructuralpropertiesoftheHαemission.Althoughdifferent Thisresultraisesanumberofquestions.First,itisnotclear
classification methods make it difficult to perform a quantita- how these galaxies are related to z ∼ 0 galaxies. An average
tivecomparison,ourresultsarequalitativelyconsistentwiththe galaxy in this sample has a distribution of stars with a Se´rsic
finding that a large fraction of star-forming galaxies at z ∼ 1 indexofn=1.4.Ifthisaveragegalaxyformsstarswithn=1.0,
it will have a lower Se´rsic index at later times. Either we are
12 Measuredvelocitydispersionsarealsorelativelyhigh,butaredifficultto witnessing the build-up of only the latest of late-type galaxies
interpretgiventhelowspatialresolutionoftheKeckspectra. (thesegalaxiessubstantiallychangewheretheirstarformationis
5
TheAstrophysicalJournalLetters,763:L16(6pp),2013January20 Nelsonetal.
occurringbetweenz∼1andz∼0)ortheirstars,onceformed, REFERENCES
must migrate into a different configuration. In other words, if
these are the ancestors of typical z ∼ 0 spiral galaxies, their Bournaud,F.,Elmegreen,B.G.,&Martig,M.2009,ApJL,707,1
bulgesneedtobebuiltinsomewaybesidesthestarformation Bovy,J.,Hogg,D.W.,&Rix,H.-W.2009,ApJ,704,1704
Brammer,G.B.,vanDokkum,P.G.,Franx,M.,etal.2012,ApJS,200,13
weareseeing.Thiscouldbeaccomplishedbymergersorsecular
Brooks,A.M.,Governato,F.,Quinn,T.,Brook,C.B.,&Wadsley,J.2009,ApJ,
evolution(e.g.,Bournaudetal.2009;Ellisonetal.2011).These 694,396
observationscouldbeunderstoodinthecontextofthefollowing Cresci,G.,Hicks,E.K.S.,Genzel,R.,etal.2009,ApJ,697,115
picture: at z > 1, disks are gas-rich and turbulent. As redshift Daddi,E.,Dickinson,M.,Morrison,G.,etal.2007,ApJ,670,156
Damen,M.,Labbe´,I.,Franx,M.,etal.2009,ApJ,690,937
decreases,thegasfractiondecreases,andabarinstabilitygrows,
DeBuhr,J.,Ma,C.-P.,&White,S.D.M.2012,MNRAS,426,938
increasingthecentralconcentrationandcentralverticalvelocity
Dekel,A.,Sari,R.,&Ceverino,D.2009,ApJ,703,785
dispersion. (e.g., van den Bosch 2001; DeBuhr et al. 2012; Drimmel,R.,&Spergel,D.N.2001,ApJ,556,181
Forbesetal.2012). Ellison,S.L.,Patton,D.R.,Nair,P.,etal.2011,inIAUSymp.277,Tracing
Wealsonotethesomewhatsurprisingpresenceofapopula- theAncestryofGalaxies,ed.C.Carignan,F.Combes,&K.C.Freeman
(Cambridge:CambridgeUniv.Press),178
tion of large, rapidly rotating disks with relatively low stellar
Epinat,B.,Contini,T.,LeFe`vre,O.,etal.2009,A&A,504,789
massesatz ∼ 1.Thesegalaxieshavemeanre(Hα)=7.3kpc, Epinat,B.,Tasca,L.,Amram,P.,etal.2012,A&A,539,A92
v = 240 km s−1, σ = 89 km s−1, and M∗ = 2.2×1010 M(cid:6), Forbes,J.,Krumholz,M.,&Burkert,A.2012,ApJ,754,48
meaningtheyaredisks8Gyrago,largerinsizethantheMilky Fo¨rsterSchreiber,N.M.,Genzel,R.,Bouche´,N.,etal.2009,ApJ,706,1364
Fo¨rsterSchreiber,N.M.,Shapley,A.E.,Erb,D.K.,etal.2011,ApJ,731,65
Way (Drimmel & Spergel 2001), and thicker (Freeman 1987),
with∼1/3ofthestellarmass(McMillan2011)andsimilaror Freeman,K.C.1987,ARA&A,25,603
Fumagalli,M.,Patel,S.G.,Franx,M.,etal.2012,ApJL,757,22
higher maximum rotation velocities (Bovy et al. 2009). Both Genzel,R.,Burkert,A.,Bouche´,N.,etal.2008,ApJ,687,59
whatthesegalaxiesbecomeinthelocaluniverseandhowsuch Genzel,R.,Newman,S.,Jones,T.,etal.2011,ApJ,733,101
extended star-forming disks were made in an epoch of high Grand,R.J.J.,Kawata,D.,&Cropper,M.2012,MNRAS,421,1529
Heckman,T.M.2002,inASPConf.Ser.254,ExtragalacticGasatLowRedshift,
diskturbulenceareopenquestions.Thesmalldisksontheother
ed.J.S.Mulchaey&J.Stocke(SanFrancisco,CA:ASP),292
end of the size distribution are also of interest. With median Hopkins,P.F.,Hernquist,L.,Cox,T.J.,etal.2006,ApJS,163,1
re(Hα) = 1.9 kpc, re(R) = 2.1 kpc, M∗ = 8.9×109M(cid:6), and Jones, T. A., Swinbank, A. M., Ellis, R. S., Richard, J., & Stark, D. P.
stackedn(Hα)=1.2,theymorecloselyresemblecompactver- 2010,MNRAS,404,1247
Kassin,S.A.,Weiner,B.,Faber,S.M.,etal.2012,ApJ,758,106
sionsofdiskystarformersthanmerger-drivenstarformationin
Kennicutt,R.C.,Hao,C.-N.,Calzetti,D.,etal.2009,ApJ,703,1672
spheroids(seealsovanderWeletal.2011).
Law,D.R.,Steidel,C.C.,Erb,D.K.,etal.2009,ApJ,697,2057
There are several caveats. First, our sample is not complete Lundgren,B.F.,Siana,B.,Henry,A.L.,etal.2012,ApJL,760,49
in mass or in SFR(UV+IR). In future papers (using the full Mancini,C.,Fo¨rsterSchreiber,N.M.,Renzini,A.,etal.2011,ApJ,743,86
3D-HSTsurvey)wewillstudythedistributionofstarformation McLean, I. S., Becklin, E. E., Bendiksen, O., et al. 1998, Proc. SPIE,
3354,566
as a function of these parameters. Second, the distribution of
McMillan,P.J.2011,MNRAS,414,2446
starformationinthestacksisnotnecessarilyrepresentativeof Nelson,E.J.,vanDokkum,P.G.,Brammer,G.,etal.2012,ApJL,747,28
individual galaxies. In individual galaxies the star formation Noeske,K.G.,Weiner,B.J.,Faber,S.M.,etal.2007,ApJL,660,43
is clumpy (e.g., Genzel et al. 2008, 2011; Fo¨rster Schreiber Peng,C.Y.,Ho,L.C.,Impey,C.D.,&Rix,H.-W.2002,AJ,124,266
Rodighiero,G.,Daddi,E.,Baronchelli,I.,etal.2011,ApJL,739,40
et al. 2009; Nelson et al. 2012; Figure 1) and it is difficult
Rosˇkar, R., Debattista, V. P., Quinn, T. R., Stinson, G. S., & Wadsley, J.
to quantify the structure. Stacking these clumpy objects could
2008,ApJL,684,79
be hazardous, but it may also provide more insight than can Shapiro, K. L., Genzel, R., Fo¨rster Schreiber, N. M., et al. 2008, ApJ,
begleanedfromindividualgalaxies:assumingthattheclumps 682,231
are transient and “light up” a part of the underlying gas disk Swinbank,M.,Smail,I.,Sobral,D.,etal.2012,arXiv:1209.1395
Szomoru,D.,Franx,M.,vanDokkum,P.G.,etal.2010,ApJL,714,244
for a short time (as in, e.g., Wuyts et al. 2012), our stacking
Szomoru,D.,Franx,M.,vanDokkum,P.G.,etal.2012,arXiv:1208.4363
techniqueeffectivelyproducesatime-averagedmapofthestar vandenBosch,F.C.2001,MNRAS,327,1334
formationinthegalaxies.Finally,dustattenuationremainsakey vanderWel,A.,Straughn,A.N.,Rix,H.-W.,etal.2011,ApJ,742,111
uncertainty,bothintheselectionandintheinterpretationofthe vanDokkum,P.G.,Brammer,G.,Fumagalli,M.,etal.2011,ApJL,743,15
van Dokkum, P. G., Franx, M., Fo¨rster Schreiber, N. M., et al. 2004, ApJ,
profiles.ThiscanbeaddressedwithmapsoftheIRemissionat
611,703
thefullALMAresolution,orbymeasuring(spatiallyresolved)
vanDokkum,P.G.,Franx,M.,Kelson,D.D.,etal.1998,ApJ,500,714
Balmerdecrements. vanDokkum,P.G.,Whitaker,K.E.,Brammer,G.,etal.2010,ApJ,709,1018
Whitaker,K.E.,vanDokkum,P.G.,Brammer,G.,&Franx,M.2012,ApJL,
754,29
We thank the referee for a thorough report which improved
Wisnioski,E.,Glazebrook,K.,Blake,C.,etal.2011,MNRAS,417,2601
the Letter. Support from grant HST GO-12177 is gratefully
Wuyts,S.,Fo¨rsterSchreiber,N.M.,Genzel,R.,etal.2012,ApJ,753,114
acknowledged. Wuyts,S.,Fo¨rsterSchreiber,N.M.,vanderWel,A.,etal.2011,ApJ,742,96
6