A Tentative Detection of an Emission Line at 1.6 μm for the z~12 Candidate UDFj-39546284

0.97 MB·English

by  Gabriel B. Brammer

APJL,ACCEPTED
PreprinttypesetusingLATEXstyleemulateapjv.2/16/10
ATENTATIVEDETECTIONOFANEMISSIONLINEAT1.6µmFORTHEZ 12CANDIDATEUDFJ-395462840
∼
GABRIELB.BRAMMER1,PIETERG.VANDOKKUM2,GARTHD.ILLINGWORTH4,RYCHARDJ.BOUWENS3,IVOLABBÉ3,MARIJN
FRANX3,IVELINAMOMCHEVA2,PASCALA.OESCH4
ApJL,accepted
ABSTRACT
WepresentdeepWFC3grismobservationsofthecandidatez 12galaxyUDFj-39546284intheHST Ultra
∼
Deep Field (UDF), by combining spectroscopic data from the 3D-HST and CANDELS surveys. The total
exposure time is 40.5 ks and the spectrum covers 1.10<λ<1.65µm. We search for faint emission lines
3
bycross-correlatingthe2DG141spectrumwiththeobservedH morphology,atechniquethatisuniqueto
1 160
slitlessspectroscopyatHSTresolution.Wefinda2.7σdetectionofanemissionlineat1.599µm—justredward
0
oftheJH filter—withflux3.5 1.3 10−18ergs−1cm−2. Assumingthelineisreal,itcontributes110 40%
2 140
± × ±
n oftheobservedH160 fluxandhasanobservedequivalentwidth>7300Å.Ifthelineisconfirmed,itcouldbe
Ly-α at z=12.12. However, a more plausible interpretation, given current results, could be a lower redshift
a
J featuresuchas[OIII]λ4959,5007atz=2.19. Wefindtwoother3D-HST[OIII]emitterswithin1000kms−1
ofthatredshiftintheGOODS-Southfield. Additionalsupportforthisinterpretationcomesfromthediscovery
3
of a bright “[OIII] blob” with a secure G141 grism redshift of z=1.605. This object has a strikingly large
2
observedequivalentwidthofnearly9000Åthatresultsinsimilar“dropout”colorsasUDFj-39546284.

] Keywords:galaxies: formation—galaxies: high-redshift
O
C
1. INTRODUCTION first robust z(cid:38)10 candidate by Bouwens et al. (2011a): it
.
h was only detected in H in the HUDF09 dataset (see also
Over the past few years deep imaging programs with the 160
p Oesch et al. 2012). Recent observations from the HUDF12
WFC3 camera on the Hubble Space Telescope (HST) have
- program (Ellis et al. 2013) confirm the H detection and
o opened up the Universe at z>8 for systematic study (e.g., 160
showthatitisundetectedinJH (Ellisetal.2013;Bouwens
r Bouwens et al. 2011a, 2012a; Oesch et al. 2012; Ellis et al. 140
st 2013; Coeet al.2013). Photometric data canprovide strong et al. 2013). The >1.4 mag dropout (2σ) in the JH140 filter
pushesitslikelyredshifttoz 12,althoughalternativeexpla-
a constraints on the redshifts and properties of very high red- ∼
[ shift galaxies (e.g., Bouwens et al. 2011b, 2012b). How- nationscannotberuledout(Ellisetal.2013). Herewereport
a2.7σ detectionofanemissionlineforUDFj-39546284that
ever, the interpretation is hampered by the lack of spectro-
2 would account for most or all of its H flux. Though this
scopicinformationthatprovidesunambiguousredshiftdeter- 160
v line could be Ly-α at z=12, we argue that it is more likely
minationandthecontributionsofemissionlinestothebroad
7
to be [OIII]λ4959,5007 at z=2.19 based in part on the dis-
1 bandfluxes. Theequivalentwidthsofrest-frameopticallines
coveryofabrightlow-redshiftanalogwithoneofthehighest
3 increase with redshift (e.g., Fumagalli et al. 2012), and the
0 emissionlinecontributioncanbesignificanteveninthebroad [OIII] equivalent widths ever observed (8700Å) and broad-
. IRACfilters(e.g.,Labbéetal.2010;Onoetal.2010). band colors similar to UDFj-39546284. Magnitudes are in
1 Obtaining high quality spectra of these very distant ob- theABsystemthroughout.
0
jectsisaformidablechallengebecauseoftheirfaintnessand
3 2. OBSERVATIONS&ANALYSIS
because key emission and absorption features are redshifted
1
into the observer’s near-IR. Apart from a gamma ray burst TheUDFwasobservedwiththeWFC3/G141grismaspart
:
v atz=8.1(Salvaterraetal.2009;Tanviretal.2009)theonly of the 3D-HST survey (GO-12177; PI: van Dokkum; Bram-
i claimed(anddisputed;Bunkeretal.2013)detectionisaLy-α mer et al. 2012a). The observations comprise four sepa-
X
lineatz=8.6inaz-dropoutgalaxyintheUDF(Lehnertetal. rate visits at the same orientation with G141 integrations of
r 2010). 4.7ks (two orbits) each. The UDF was also observed by
a
Here we describe a deep WFC3 G141 grism spectrum of theCANDELSsupernovafollow-upprogram(GO-12099,PI:
one of the most intriguing high redshift candidates yet dis- Riess)whenasupernovaatz=1.55,nicknamed“Primo”,was
covered: UDFj-39546284. This object was identified as the discovered in the first epoch of CANDELS imaging in the
GOODS-South field(see Rodney et al. 2012). The field was
[email protected] observedintwovisitson2010Oct26and2010Nov01with
0BasedonobservationsmadewiththeNASA/ESAHubbleSpaceTele- G141 integrations of 6.6 and 15 ks, respectively. The posi-
scope,programsGO-12099,12177and12547,obtainedattheSpaceTele- tion angle of the grism dispersion axis is rotated by 6.4 de-
scopeScienceInstitute,whichisoperatedbytheAssociationofUniversi-
grees between the two visits, which together are rotated 50
tiesforResearchinAstronomy,Inc.,underNASAcontractNAS5-26555.
1European Southern Observatory, Alonso de Córdova 3107, Casilla degreeswithrespecttothe3D-HSTvisits. Therelationofthe
19001,Vitacura,Santiago,Chile G141pointingstothedeepHUDF09+HUDF12imagingarea
2DepartmentofAstronomy,YaleUniversity,NewHaven,CT06520, isshowninFig.1.
USA The Primo and 3D-HST visits were pre-processed sep-
3LeidenObservatory,LeidenUniversity,NL-2300RALeiden,Nether- arately and identically using the threedhst pipeline6
lands
4UCO/Lick Observatory, University of California, Santa Cruz, CA
95064 6http://code.google.com/p/threedhst/
2 Brammeretal.
(Brammer et al. 2012a). The individual exposures of each
0
visit were then combined into a mosaic by interlacing the 5:3
4
pixels to a subgrid of pixels that are exactly half their orig-
inal size. Interlacing pixels from input images which have
0
0
0.5 pixel offsets (by design in 3D-HST and rounded for 6:
4
CANDELS-Primo) results in a one-to-one correspondence
between input and output pixels (e.g., van Dokkum et al. 0 UDFj-39546284
3
2000) and offers the key benefit of preserving the individual 46: SN“Primo”
pixelerrors.
sitWiesitahntdhespHaUtiaDlFm1o2rpHh1o6l0oigmieasg,ewdeecfionminpguttheea1f.u6llµqmuaflnutxitadteivne- Dec.27:47:00 UDF-40106456
-
contamination model of all objects in the field to H <28.
160
0
The modeling technique, first presented in Brammer et al. 7:3
4
(2012b), includesthefirst4spectralorders, amongthemthe
compact zeroth order that can resemble emission lines. The
0
modelsofbrightobjectswithH160<23arerefinedbasedon 48:0
their observed spectra, while fainter objects are assumed to
HUDF09/HUDF12
havecontinuaflatinunitsof fλ. Therefore,emissionlinesin 30 3D-HST,GOODS-S34,36,37,38 5 10 20 30 40ks
faintergalaxiesarenotincludedinthecontaminationmodel. 48: CCAANNDDEELLSS,,PPRRIIMMOO--11012061
Togenerateadeep2DspectrumfromallUDFgrisminte-
grations, we first extract two-dimensional spectra from each 46 43 03:32:40 37 34 31
R.A.
visit.Thesespectraarealignedintwodimensionstothenear-
estpixelinthesubsampledgridandthealignedpixelsarethen Figure1. Exposuremapofthe3D-HSTandCANDELSG141grismpoint-
combined with inverse variance weights including a term to ingsintheHUDF.Thevectorsindicatethedirectionandextentofthefirst-
orderG141spectra. ThelocationsoftheCANDELSsupernovaandthetwo
down-weight contaminated pixels. Stacked spectra from the
faintUDFgalaxiesdescribedbelowareindicated. Theseobjectslieinthe
central3arcmin2oftheUDFhaveacombinedexposuretime areaofdeepestG141grismcoverage,withatotalexposuretimeof40.5ks
of40.5ks. (17orbits).
tothatofthebroad-bandimage.
3. ATENTATIVEEMISSIONLINEINUDFj-39546284
While there are no obvious emission lines in the raw 2D
3.1. TheDeepNear-IRSpectrum spectrum, there is a clump of pixels at 1.6µm which is
∼
The two-dimensional 40.5 ks G141 spectrum of UDFj- marginally enhanced compared to its surroundings. To im-
39546284 is shown in Fig.2. While some contamination prove the signal-to-noise of this feature, we cross-correlate
remains in the stacked spectrum, the orientations of the the2DspectrumwithakerneldefinedbythecentralR=0.(cid:48)(cid:48)3
UDF/G141visitsturnouttobenearlyperfectforavoidingsig- of the deep H thumbnail of UDFj-39546284. This kernel
160
nificantcontaminationoftheUDFj-39546284spectrum,par- isconstructedbyextractingkernelsforeachoftheindividual
ticularly at λ>1.4µm. The contamination-subtracted spec- UDF/G141 visits with different orientations and combining
trum is remarkably clean with none of the systematics that themweightingbythemedianerrorintheircorresponding2D
typically plague ground-based NIR spectroscopy (see, e.g., grismspectra.Thus,theobjectprofileisslightlysmoothedbut
Lehnert et al. 2010). The cluster of bright pixels at the top will reflect the 2D morphology of lines in the stacked spec-
ofthespectrumislikelyafaintcontaminantemissionlinenot trum. The cross-correlation spectrum is shown in the right-
includedinthemodel;itlieswellawayfromanycontribution handpanelsofFig.2,andthestrengthofthefeatureat1.6µm
fromUDFj-39546284. is enhanced compared to the raw 2D spectrum. We verified
The “salt-and-pepper” appearance of the 2D spectrum is thattheweakfeatureisnotvisibleinasimilaranalysisofany
the result of the uncorrelated noise of the individual pix- single UDF/G141 visit, confirming that the feature does not
els. Thedistributionofcleanedpixelvaluesisgaussianwith arisefromafluxexcessinoneofthemindividually,suchasa
σ =0.0033electronss−1. Randomgaussiandeviatesscaled groupofhotpixelsoranun-flaggedcosmicray.
obs
bythepropagatedpixelerrorshaveσ =0.0036electronss−1. We show a one-dimensional spectrum extracted along the
err
Thisslightoverestimationoftheerrorsby10%issignificantly traceinFig.3. Wecomputeanassociateduncertaintyateach
lowerthanthefactorsof60–70%thatresultfromtypicalap- pixelalongthetrace,highlightedinblue,bycross-correlating
plicationsoftheDrizzlealgorithm(Fruchteretal.2009). the squared error array with the same H160 kernel. Further-
more, we compute a 2D model spectrum as in Fig.2 with
3.2. A2σEmissionLineDetectionatλ=1.599µm knownpositionandintegratedline flux, and extractitscross
correlation spectrum in the same way as the observed spec-
SlitlessHST grismspectraofferadistinctadvantagecom-
trum. Thefeatureatλ=1.599 0.004µmisdetectedat2.7σ
pared to ground-based data for detecting weak emission line withanintegratedlinefluxof±3.5 1.3 10−18ergs−1cm−2.
features: the2Dspectrumprovidesanimageofthegalaxyat ± ×
Theprobabilityoffindingagaussiannoisefeaturewithequal
thewavelengthoftheemissionline(see,e.g.,Brammeretal.
orgreatersignificanceis 10%for30independentresolution
2012b), and the morphology of the galaxy is known a priori ∼
elements( 3000Å).
from the deep broad-band WFC3 imaging. As long as the
∼
emission line morphology and the broad-band morphology
3.3. TheRealityofthe1.599µmFeature
aresimilar(whichisgenerallythecaseatz 1;Nelsonetal.
∼
2012), or the line contributes significantly to the broad-band There are regions in the 2D cross-correlation spectrum
flux (which is the case here; see below), we can search for withapparentlysimilarsignificancetothefeatureat1.6µm.
features in the 2D spectrum whose morphology corresponds Many of them can be rejected because they do not fall pre-
HST spectroscopyofUDFj-39546284 3
k
c
a
St
1
4
1
G
m.
a
nt
o
C
d
e
n
a
e
Cl
el
mod 100
e
n
Li G141 F140W F160W
1.1 1.2 1.3 1.4 1.5 1.6 1.1 1.2 1.3 1.4 1.5 1.6
λ /µm λ /µm
Figure2. ThedeepG141grismspectrumofUDFj-39546284. Theleftpanelsshowthepixelsasobservedandmodeled,andtherightpanelsshowthespectra
cross-correlatedwiththeH160 objectprofiletoenhanceweakfeatures. Thefirstrowshowsthestackedspectrumasobserved. Thesecondrowshowsthe
contaminationmodel(§2),withsomealiasingduetotheunequalweightsoftheinterlacedPrimomosaics. Thethirdrowshowstheobservedspectrumcleaned
ofthiscontamination. Thereisaweakemissionfeatureinthecross-correlationspectrumatλ=1.599µm. Thebottomrowshowstheemissionlinemodelof
UDFj-39546284 wherethemorphologyisderivedfromtheobservedH160image(seealsoFig.3).
1” 1”
15 15
2m 2m
g/s/c10 g/s/c10
er er
180− 5 180− 5
1 1
/ /
x x
u u
fl fl
ne 0 ne 0
li li
10 10 F160Wkernel Contam.
Uniformkernel(R=0.30”)
8 UDFj-39546284 8
UDF-40106456
6 Lyα, z = 12.12? 6
N N (Hβ)+[OIII]+Hα, z = 1.303
S/ 4 [OIII]λ5007, z = 2.19? S/ 4
2 2
0 0
1.1 1.2 1.3 1.4 1.5 1.6 1.1 1.2 1.3 1.4 1.5 1.6
λ/µm λ/µm
Figure3. OptimalextractionoftheUDFj-39546284spectrum(left)andthatofacomparisonobject,UDF-40106456(right).Thetoppanelsshowthestacked2D
spectraandthespectracross-correlatedwiththeH160imagekernel,whichisindicatedatleft.Theupperspectrashowthe2Dcross-correlationextractedalongthe
traceandscaledbytheG141sensitivitycurveand∆λofthespectrum,whichforauniformkernelwouldgiveanintegratedlinefluxwiththeunitsaslabeled.The
samecross-correlationisappliedtothe2Dpixelvariancearray,yieldingtheuncertaintiesshadedingray.Thebottompanelsshowthesignal-to-noisespectrum.
Whilethereareadditionalapparentlysignificantfeaturesatλ<1.5µm,thesearelikelyduetocontaminationratherthancomingfromUDFj-39546284(§3.3).
ThegraycurvesinthebottompanelsshowtheS/NinR=0.(cid:48)(cid:48)3aperturesalongthetrace.Theredcurvesindicatethepotentialcontaminationfromnearbysources.
Thefilledorangeregionsaremodelemissionlinespectra,whoseshapesmatchthoseoftheobservedlines.
4 Brammeretal.
cisely along the trace of the grism spectrum. There are en- 4.2. Lowredshiftsolutions—[OIII]atz=2.2
hancementsfoundnearthetraceatλ=1.14µmand1.22µm,
Although we cannot exclude other low redshift identifica-
though these wavelengths suffer from higher contamination
tionssuchas[OII]atz=3.28orHαatz=1.44,alikelypossi-
perhapsconsistingoffaintemissionlinesnotincludedinthe
bilityisthatthelineis[OIII]λ4959,5007atz=2.19 0.01.A
contaminationmodel. TheHST/WFC3broad-bandphotome- ±
populationofextreme[OIII]emissionlinegalaxiesatz 1.7
tryplacesstrongconstraintsonthepossibilitythattheselines ∼
was recently identified by van der Wel et al. (2011) from
come from UDFj-39546284: either of the bluer lines alone
their significant line contribution to the J photometry in
wouldresultinY orJ twomagnitudesbrighterthanthe 125
105 125 theCANDELSsurvey. Additionalgalaxieswith[OIII]rest-
2σHUDF12limits.
frame equivalent widths reaching 2000Å have been identi-
In contrast, the 1.599µm line is consistent with the
HUDF12photometry. Withinthesame0.(cid:48)(cid:48)3apertureusedto fied in WFC3 grism spectroscopy by Atek et al. (2011) and
Brammer et al. (2012b). The strongest emitters in van der
measure the line flux, we measure H =29.04 for the H
kernel. Thisis27%brighterthanthe1E60llisetal.(2013)ma1g60- Wel et al. (2011) have J125−H160 ∼−1, with [OIII] in the
nitude measured within an 0.(cid:48)(cid:48)25 aperture, but is only 73% J125band.Whilesuchcolorscanmimicthoseofhigh-redshift
dropoutgalaxies(Ateketal.2011),evenmoreextremeequiv-
of the total magnitude measured by Bouwens et al. (2013).
alent widths at (cid:38)3000Å (rest-frame) are required to reach
These differences are consistent with aperture corrections of
J −H >1.4observedforUDFj-39546284.
theextendedsource,whichshowsafaintextendedtailtothe 125/140 160
NEintheH image(Bouwensetal.2013). Apureemission
160
line with the observed flux would result in H =28.92; the 4.3. DiscoveryofaPossibleAnalog: AnExtreme[OIII]
160
lineaccountsfor110 40%(1σ)oftheH flux. Thiscor- Emitteratz=1.605
160
±
responds to a (1,2)σ limit on the observed-frame equivalent We have discovered a very bright example of such an ex-
widthofthelineofEW>(7300,1300)Å. treme emission line galaxy in G141 grism observations of
We test our technique by analyzing faint lines of another the EGS field (GO-12547, PI: Cooper). This remarkable
UDFgalaxywithasecureredshift,UDF-40106456(rightpan- object, EGS-XEW-1, located at α=14:17:58.2, δ=+52:31:35
els of Fig.3). This galaxy has H =27.4 and unambigu- and shown in Fig.4, is an “[O III] blob” at z = 1.605 with
160
ousemissionlinesofHαand[OIII]λ4959,5007atz=1.303 JH140 =21.5 and a complex morphology comprised of dif-
visible in the stacked 2D spectrum before cross-correlation. fuse (A) and compact (B) components separated by 1(cid:48)(cid:48). The
Using the cross-correlation technique we find a significant exposure times and dithered pixel sampling for the Cooper
detection of Hβ even though the line is barely visible in etal.grismprogramarenearlyidenticaltothoseof3D-HST,
the original spectrum. Two spurious features are detected at and the spectra were processed with the 3D-HST interlac-
1.24µm and 1.29µm with comparable flux as the tentative ing software as described above. The diffuse component,
line in UDFj-39546284. These are both clearly associated which is extended over more than 1(cid:48)(cid:48) (8.5 kpc), has a com-
with residual contamination from another source at the top bined observed-frame equivalent width of nearly 9000Å for
of the 2D spectrum. The Hα line flux of UDF-40106456 is the blended Hβ and [OIII]λ4959,5007 emission lines. The
9.5 1.0 10−18ergs−1cm−2, just three times brighter than morecompactcomponenthasEW=3000 160Å.
theU±DFj-×39546284emissionline.Thedetectionsignificance The broad-band SED of EGS-XEW-1±is compared to that
is S/N 10 and the cross-correlation technique enhances the of UDFj-39546284 in the right panel of Fig.4. The spec-
∼
detectionsignificancebyafactoroftwocomparedtosimple tral break that results from a line such as the one observed
photometry along the trace within an equivalent 0.(cid:48)(cid:48)30 aper- inEGS-XEW-1isjustlargeenoughtosatisfythe2σlimiton
ture. This example demonstrates that the cross-correlation theJH −H dropoutcolorofUDFj-39546284. However,
140 160
techniqueworksandisabletorecoverextremelyfaintemis- if UDFj-39546284 had exactly the same spectrum as EGS-
sionlines. XEW-1, the [OII]λ3727 line would probably be detected in
the bluer WFC3 HUDF12 photometry. Interestingly, EGS-
XEW-1 has much redder UV colors than f λ−2 observed
λ
4. LINEIDENTIFICATION for the high equivalent width starbursts at s∼imilar redshifts
4.1. IsUDFj-39546284atz=12? (vanderWeletal.2011;Brammeretal.2012b),which,ifalso
trueforUDFj-39546284,couldhelpexplainthelackofz=2
Ifthegalaxyisatz 12,asissuggestedbasedonitsstrong
∼ Ly-α observed for that source (Ellis et al. 2013). While not
photometric break between the JH and H filters (Ellis
140 160 a perfect match, EGS-XEW-1 provides a directly-observed
et al. 2013; Bouwens et al. 2013), the line—if real—can be
routetoaplausiblelow-redshiftinterpretationfortheUDFj-
identified as Ly-α redshifted to z = 12.12. The rest-frame
39546284photometry.
equivalent width of Ly-α would not be unreasonably high at
>170Å(accountingforthez=12LymanbreakinH160);such 5. DISCUSSIONANDCONCLUSIONS
values can be reached in young, low metallicity starbursts at
WehavedescribedadeepHST grismspectrumofthecan-
high redshift (Schaerer 2003). However, such strong Ly-α
didatez=12galaxyUDFj-39546284. Usingtheknownemis-
emissionmightbeunexpectedearlyinthereionizationepoch
sion line morphology to increase the line sensitivity of the
when the neutral fraction is high (Santos 2004; but see also
slitless grism spectrum, we detect a tentative emission line
Dijkstraetal.2011).Moretothepoint,ifthelineisrealitcan
at λ=1.599µm with flux 3.5 1.3 10−18ergs−1cm−2 and
accountformostoralloftheH160flux,meaningthatthepho- ± ×
tometricbreakdoesnotnecessarilyreflectastrongcontinuum observed-frameequivalentwidth>7300Å.
breakandthephotometricredshiftisunreliable.Thereforethe While the current observations do not conclusively forbid
linecouldalsobealonger-wavelengthfeatureatmuchlower thez=12interpretationofUDFj-39546284,anumberofinde-
redshift,apossibilitydiscussedbyEllisetal.(2013). pendentfactorsconspiretosuggestthatthe2.7σlineisinfact
real and that the true redshift of UDFj-39546284 is z 2.2.
∼
HST spectroscopyofUDFj-39546284 5
26
UDFj-39546284,H160 BlobA,z 2.2(NMBS)
AB 100=8.5kpc UDFj-39546284,2σlimits →
27
EGS-XEW-1(A)
1.2 EWHβ+[OIII]=8700±700A˚ 28 Hβ+[OIII]Hα
˚A] 1.0 [OIII]/Hβ=11.4±1.3 g
2m/ z=1.605 ma 29
erg/s/c 0.8 AB [OII] 600
180−0.6 [OIII]λ4959,5007 30
1
[
λ0.4
f Hβ N
31
E
0.2
V606/I814/JH140
0.0 32
1.1 1.2 1.3 1.4 1.5 1.6 1 2 3 4
λ/µm λ/µm
Figure4. WFC3/G141spectrumandbroad-bandSEDofEGS-XEW-1,an“[OIII]blob”atz=1.605withextreme[OIII]emissionlinesandcolorssimilar
toUDFj-39546284. TheasymmetriclineprofileofcomponentAisconsistentwithlinesofHβandthe[OIII]λ4959,5007doubletasindicated;thelinesare
unambiguouslyresolvedforcomponentB,atthesameredshift. TherightpanelshowstheSEDoftheblobnormalizedtothatoftheUDFj-39546284H160
magnitude.Non-detectionsforUDFj-39546284areshownwith2σupperlimitstakenfromBouwensetal.(2013).AlongwiththeG141spectrum,ground-based
photometryoftheblobfromtheNEWIRMMediumBandSurvey(NMBS)(Whitakeretal.2011)isshowninblue,redshiftedsuchthatthe[OIII]linewould
havebeenobservedat1.6µmandscaledtoaccountforthedifferentNMBS-J3,JH140,andH160passbands.Theinsetshowsacolorthumbnailoftheblobcreated
fromHSTACS/V606,ACS/I814,andWFC3/JH140images.
First,thelineisobservedatjusttherightwavelengthandflux and 5007 Å lines of the doublet separately, either in (much)
to satisfy both the HUDF H detection and the >1.4 mag deepergrismdataorindeepground-basedspectroscopy. Hα
160
dropoutatλ<1.595µminthebluerHSTbands.Ifthefeature may also be visible in the K-band at a similar flux as [OIII]
were observed just 4 pixels (93Å) bluer it would violate the (Fig.4). Regardless of the reality of the emission line, we
constraintsonJH byamagnitude. Second,Bouwensetal. haveshownthatthebroad-bandSEDofUDFj-39546284pro-
140
(2013)demonstratethatwithoutasignificantcontributionof vides a reasonable match to a low redshift SED, specifically
Ly-α to the H flux, the implied rest-frame UV luminosity the observed spectrum of EGS-XEW-1, an “[OIII] blob” at
160
forz 12issome20 higherthanwouldbeexpectedfrom z=1.605,redshiftedtoz=2.2. Thisstudydemonstratesthat
theev∼olutionofthelum×inosityfunctionatz=8–10.Third,re- theG141grismonWFC3canprovidecleanspectrawithwell-
centdeepspectroscopicsurveysofphotometricdropoutsfind understoodnoisepropertiesindeepintegrationsuptoatleast
a decreasing fraction Ly-α emitters with increasing redshift 40ks(17orbits).
when the universe was increasingly neutral (Pentericci et al.
We gratefully acknowledge funding support from the Marie
2011; Caruana et al. 2012). Fourth, we show that spatially-
CurieActionsoftheEuropeanCommission(FP7-COFUND)
extended objects with EW>7300Å exist at z 2. Finally,
wefindatleasttwo[OIII]emitterswithin1000∼kms−1 ofthe anFdaScTilSitcieIsg:rHanutbGbOle-S1p2a1c7e7.Telescope(WFC3)
z=2.19 solution for UDFj-39546284 in the 3D-HST cover-
ageof thefull GOODS-Southfield, the nearestseparated by
REFERENCES
69(cid:48)(cid:48) (570 kpc); the fact that other galaxies exist at this exact
redshift increases the probability that the line is real and is
Atek,H.,Siana,B.,Scarlata,C.,etal.2011,ApJ,743,121
[OIII]λ4959,5007. Bouwens,R.,Bradley,L.,Zitrin,A.,etal.201a,arXiv/1211.2230,ApJ,
If the physical properties of UDFj-39546284 are similar submitted
to either the [OIII] blob EGS-XEW-1 or to the high equiva- Bouwens,R.J.,Illingworth,G.D.,Labbe,I.,etal.2011a,Nature,469,504
Bouwens,R.J.,Illingworth,G.D.,Oesch,P.A.,etal.2012b,ApJ,754,83
lentwidthstarburstingdwarfgalaxiesstudiedbyvanderWel
—.2011b,ApJ,737,90
et al. (2011), UDFj-39546284 at z=2.2 would represent a
Bouwens,R.J.,Oesch,P.A.,Illingworth,G.D.,etal.2013,
newclassofobject750timesfainterthantheformeror 30 arXiv/1211.3105,ApJL,submitted
∼
times fainter than the latter. Scaling from the typical stel- Brammer,G.B.,vanDokkum,P.G.,Franx,M.,etal.2012a,ApJS,200,13
lar masses of the van der Wel et al. (2011) galaxies, UDFj- Brammer,G.B.,Sánchez-Janssen,R.,Labbé,I.,etal.2012b,ApJ,758,L17
Bunker,A.J.,Caruana,J.,Wilkins,S.M.,etal.2013,arXiv/1301.4477,
39546284wouldhaveastellarmassoforder106M ,similar
(cid:12) MNRAS,inpress
tothemassofasinglemassivestarcluster(e.g.,Reinesetal. Caruana,J.,Bunker,A.J.,Wilkins,S.M.,etal.2012,MNRAS,427,3055
2008). The extended H morphology would then indicate Coe,D.,Zitrin,A.,Carrasco,M.,etal.2013,ApJ,762,32
160
thedistributionofionizedgassurroundingthecluster. Dijkstra,M.,Mesinger,A.,&Wyithe,J.S.B.2011,MNRAS,414,2139
Ellis,R.S.,McLure,R.J.,Dunlop,J.S.,etal.2013,ApJ,763,L7
The most urgent question is whether the line is real and
Fruchter,A.,Sosey,M.,Hack,W.,etal.2009,TheMultiDrizzleHandbook,
associated with UDFj-39546284 as opposed to coming from
v3.0,Tech.rep.
a nearby contaminating object spectrum or simply being a Fumagalli,M.,Patel,S.G.,Franx,M.,etal.2012,ApJ,757,L22
clump of positive noise fluctuations. If the line is real and Labbé,I.,González,V.,Bouwens,R.J.,etal.2010,ApJ,716,L103
is [OIII] at z=2.2, it should be possible to detect the 4959 Lehnert,M.D.,Nesvadba,N.P.H.,Cuby,J.-G.,etal.2010,Nature,467,
940
6 Brammeretal.
Nelson,E.J.,vanDokkum,P.G.,Brammer,G.,etal.2012,ApJ,747,L28 Schaerer,D.2003,A&A,397,527
Oesch,P.A.,Bouwens,R.J.,Illingworth,G.D.,etal.2012,ApJ,745,110 Tanvir,N.R.,Fox,D.B.,Levan,A.J.,etal.2009,Nature,461,1254
Ono,Y.,Ouchi,M.,Shimasaku,K.,etal.2010,ApJ,724,1524 vanderWel,A.,Straughn,A.N.,Rix,H.-W.,etal.2011,ApJ,742,111
Pentericci,L.,Fontana,A.,Vanzella,E.,etal.2011,ApJ,743,132 vanDokkum,P.G.,Franx,M.,Fabricant,D.,etal.2000,ApJ,541,95
Reines,A.E.,Johnson,K.E.,&Hunt,L.K.2008,AJ,136,1415 Whitaker,K.E.,Labbé,I.,vanDokkum,P.G.,etal.2011,ApJ,735,86
Rodney,S.A.,Riess,A.G.,Dahlen,T.,etal.2012,ApJ,746,5
Salvaterra,R.,DellaValle,M.,Campana,S.,etal.2009,Nature,461,1258
Santos,M.R.2004,MNRAS,349,1137