Probing Very Bright End of Galaxy Luminosity Function at z >~ 7 Using Hubble Space Telescope Pure Parallel Observations

0.3 MB·English

by  Haojing Yan

Draftversion January11,2011
PreprinttypesetusingLATEXstyleemulateapjv.03/07/07
PROBING VERY BRIGHT END OF GALAXY LUMINOSITY FUNCTION AT Z &7 USING HUBBLE SPACE
TELESCOPE PURE PARALLEL OBSERVATIONS*
Haojing Yan 1, Lin Yan 2, Michel A. Zamojski 2, Rogier A. Windhorst 3, Patrick J. McCarthy 4, Xiaohui Fan 5,
Huub J. A. Ro¨ttgering 6, Anton M. Koekemoer 7, Brant E. Robertson 8,10, Romeel Dav´e 5, Zheng Cai 9
Draft version January 11, 2011
ABSTRACT
1 We report the first results from the Hubble Infrared Pure Parallel Imaging Extragalactic Survey,
1 which utilizes the pure parallelorbits of the Hubble Space Telescope to do deep imaging along a large
0 number of random sightlines. To date, our analysis includes 26 widely separated fields observed by
2 the Wide Field Camera 3, which amounts to 122.8arcmin2 in totalarea. We havefound three bright
n Y098-dropouts, which are candidate galaxies at z & 7.4. One of these objects shows an indication
a of peculiar variability and its nature is uncertain. The other two objects are among the brightest
J candidate galaxies at these redshifts known to date (L > 2L∗). Such very luminous objects could

9 be the progenitors of the high-mass Lyman break galaxis (LBGs) observed at lower redshifts (up to
z ∼ 5). While our sample is still limited in size, it is much less subject to the uncertainty caused by

] “cosmicvariance”thanothersamplesbecauseitis derivedusing fieldsalongmanyrandomsightlines.
O We find that the existence of the brightest candidate at z ≈ 7.4 is not well explained by the current
C luminosity function (LF) estimates at z ≈8. However,its inferred surface density could be explained
. by the predictionfromthe LFs atz ≈7 if itbelongsto the high-redshifttailofthe galaxypopulation
h at z ≈7.
p
Subject headings: cosmology: observations — galaxies: luminosity function, mass function
-
o
r
t 1. INTRODUCTION augmented by the shallower but 8× wider observations
s
(Windhorstet al. 2010)of the WFC3 EarlyRelease Sci-
a The Wide Field Camera 3 (WFC3) on board the
[ Hubble Space Telescope (HST) has begun breaking new ence Program 2 (ERS2; PI: O’Connell), which provide
grounds in our exploration of the early universe. Us- us a unique opportunity to study the bright end of the
2 luminosity function (LF) at z & 7 because of the much
v ing the ultra-deepobservations taken by the “HUDF09”
increasedsamplesizeofbrightcandidates(Wilkinsetal.
1 program (PI: Illingworth), samples of candidate galax-
6 ies at z ≈ 7 have been greatly enlarged (Oesch et al. 2010a, 2010b; Bouwens et al. 2010d; Yan et al. 2010b).
Even with the ERS2 data, however, we are only barely
2 2010; Bunker et al. 2010; McLure et al. 2010; Yan et
2 al. 2010a), and candidates have been found at z ≈ 8 able to probe the bright end at L ∼L∗. It is important
to extend our investigation to the brighter regime, as it
. (Bouwens et al. 2010b; McLure et al. 2010; Yan et al.
0 2010a) and possibly out to z ≈ 10 (Yan et al. 2010a; could be where the progenitors of the high stellar-mass
01 Bouwens et al. 2010c). The studies of the stellar pop- (&1010M⊙)galaxiesatz ≈6–5wouldlocate(Yanetal.
1 ulations of galaxies at z & 7 are now underway (e.g., 2006). Suchstudies willneed to surveylargerareas,and
preferably along different sightlines in order to reduce
: Bouwensetal. 2010a;Labb´eetal. 2010a,2010b;Finkel-
v the bias caused by the underlying large scale structures
stein et al. 2010). The pencil-beam HUDF09 data are
i (a.k.a. “cosmic variance”; see, e.g., Robertson 2010).
X
*Based on observations made with the NASA/ESA Hubble Our Hubble Infrared Pure Parallel Imaging Extra-
ar SpaceTelescope,obtainedattheSpaceTelescopeScienceInstitute, galactic Survey, or “HIPPIES”, was designed for this
which is operated by the Association of Universities for Research purpose. This program utilizes the unique “pure paral-
in Astronomy, Inc., under NASA contract NAS 5-26555. These lel”observingmode ofHST,whichenablessimultaneous
observations areassociatedwithprograms#11700and11702.
1Center for Cosmology and AstroParticle Physics, The Ohio observations using instruments that are not doing the
State University, 191 West Woodruff Avenue, Columbus, OH primary observations. As the pointings of the primary
43210,USA observations are drawn from a wide variety of programs
2SpitzerScienceCenter,CaliforniaInstituteofTechnology, MS
of different objectives, the associated parallel observa-
220-6,Pasadena,CA91125,USA
3SchoolofEarthandSpaceExploration,ArizonaStateUniver- tions are in effect observing completely random fields,
sity,Tempe,AZ85287,USA and hence are ideally suited for minimizing the impact
4Observatories of the Carnegie Institution of Washington, 813 of cosmic variance. In HST Cycle 17, there were two
SantaBarbaraStreet, Pasadena,CA91101,USA
5Astronomy Department, The University of Arizona, Tucson, pure parallel programs using WFC3 as the imaging in-
AZ85721, USA strument, one led by our team (PID 11702, PI: H. Yan)
6LeidenObservatory,UniversityofLeiden,P.O.Box9513, Lei- and the other led by M. Trenti et al. (PID 11700). Our
den2300RA,TheNetherlands programhasbeenextendedtothecomingCycle18(PID
7SpaceTelescopeScienceInstitute,3700SanMartinDrive,Bal-
12286). HIPPIESwillincorporatethedatafromallthese
timore,MD21218, USA
8AstronomyDepartment,CaliforniaInstituteofTechnology,MS programs for its science objectives, one of which being
249-17,Pasadena, CA91125,USA addressing the very bright-end of galaxy LF at z ≈ 7
9Physics Department, The University of Arizona, Tucson, AZ and beyond. Here we report our initial results based
85721,USA
10HubbleFellow on the data taken by the end of Cycle 17. Throughout
2
this Letter, we use the following cosmological parame- selection, and hence such image defects in UVIS do not
ters: ΩM =0.27,ΩΛ =0.73andH0 =71kms−1Mpc−1. likely cause any contamination.
The quoted magnitudes are all in the AB system. The MultiDrizzle software (Koekemoer et al. 2002)
was then used to correct for the geometric distortions
2. DATA DESCRIPTION
and to stack the images in each band. For each field,
InCycle 17,pure parallelswereonlyallowedwhenthe the World Coordinate System (WCS) of the first im-
primeinstrumentwaseithertheCosmicOriginsSpectro- age in Y was always chosen as the reference to align
098
graph (COS) or the Space Telescope Imaging Spectro- all images. The final pixel scale that we adopted is
graph(STIS).Programs11700and11702onlyrequested 0′.′09 pixel−1. The pixel noise of the drizzle-combined
the highgalactic latitude opportunities at|b|>20o, and images is correlated at small scale because of the sub-
only requested long-duration opportunities (& 3 orbits) pixel sampling. We followed the procedure utilized in
so that the observations could be sufficiently deep. The the GOODS program (Dickinson et al. 2004) to calcu-
observations and data from these two programs are de- late the correlationamplitudes in our mosaics, and then
scribed below. toderivetheso-called“RMSmaps”thatregistertheab-
solute root-mean-squarenoise in each pixel.
2.1. Observations
These two programs used both the infrared (IR) and 2.3. Photometry
the UV-optical (UVIS) channels of WFC3. The IR im-
Matched-aperturephotometrywascarriedoutbyrun-
ages were taken in the same three bands in both pro-
ningtheSExtractorprogramofBertin&Arnouts(1996)
grams, namely, F098M, F125W, and F160W (hereafter in dual-image mode. The J -band mosaics were used
Y ,J ,andH ,respectively). IntheUVISchannel, 125
098 105 125 as the detection images. We adopt the SExtractor
program11700usedF606W(V ),whileprogram11702
606 MAG AUTO magnitudes, which were measured using
used F600LP (LP ). The left panel of Figure 1 shows
600 the default (Kron factor, minimum radius) of (2.5, 3.5).
the system response curves in these passbands.
A 2-pixel, 5×5 Gaussian filter was used to convolve the
Ourcurrentworkincludes26WFC3pureparallelfields J -band images for detection; the detection threshold
125
acquired by the end of Cycle 17 (2010 August). The to- was set to 1.5σ, and a minimum of four connected pix-
tal exposure time in these fields ranges from 6 to 34 ks,
els above the threshold was required. The RMS maps
with the median of 10.8 ks (∼ 5 orbits). The distribu-
were used to measure the noise in the pixels within
tion of exposure time in the four bands roughly follows
the MAG AUTO apertures. We only include sources at
the ratio of UVIS:F098M:F125W:F160W= 1 : 2 : 1 : 1, S/N≥5inthe J -band(withinthe MAG AUTO aper-
125
but varies significantly among the fields. The CCD ob-
ture) for further analysis.
servations in UVIS always have at least two CR-SPLIT
The depth of our mosaics varies significantly because
exposuresthatenable the rejectionofcosmic rays. Most
theexposuretimeandthebackgroundconditionvarysig-
of the IR observations also have at least two exposures
nificantly among the HIPPIES fields. For reference, the
each band in each field. As the IR array was always median5σdepthsmeasuredwithina0′.′2-radiusaperture
non-destructivelyread-outmultiple times during the ex- are 27.36, 27.08, 27.16, 27.05, and 26.68 mag in V ,
606
posure(atleastfivetimesfortheseobservations),cosmic LP , Y , J , and H , respectively. The source
600 098 125 160
rayrejectiongenerallyisnotaproblem,evenwithsingle count at J ≤25.2 mag is complete in every field, and
125
exposures. iscompleteatJ ≤26.0magfor20outofthe26fields.
125
2.2. Data Reduction
3. CANDIDATE GALAXIES AT Z &7.4
Our data reduction starts from the outputs from the
In this section, we describe the selection of candidate
HST On-the-Fly Reprocessing (OTFR) pipeline, which
galaxiesatz &7.4asY -dropoutsusingHIPPIESdata.
are photometrically calibrated images in units of count 098
rate (i.e., the “* flt.fits” images fetched from the HST
3.1. Y -dropout selection
data archive). 098
Thevastmajorityofourparallelobservationsarenon- Because Y does not overlapwith J , it is possible
098 125
dithered because the primaryCOSor STIS observations to adopt a large color decrement such that the selection
are not dithered. This makes it difficult to reject im- islesspronetotheconfusionwiththe4000˚A breakofred
age defects using the usual outlier clipping method in galaxiesatlowredshiftsandlessaffectedbyphotometric
the stacking process. Most of the defects are flagged in errors. Following Yan et al. (2010b), our main color cri-
the “data quality” (DQ) extension of the processed im- teria are Y −J >1.3 mag and non-detections (i.e.,
098 125
age, which records the pixels that are diagnosed by the S/N<2)inthe “veto”V and/orLP images. Based
606 600
processing software to be problematic for various rea- on our simulation using a large number of galaxy tem-
sons (e.g., bad pixels, problems in telemetry, unstable plates fromthe models of Bruzual& Charlot(2003)and
response, etc.). In order to reduce their chance of con- the line-of-sight H I absorption recipe of Madau (1995),
taminatingourdropoutsample,suchimagedefectsinthe the adopted Y −J color decrement is appropriate
098 125
IRwerefilledbyinterpolationofnearbypixels. Usingthe for selecting galaxies at z & 7.4. As our UVIS images
archivalERS2 data, we also identified severalthousands are not very deep, and we do not have data at longer
of additional outlier pixels that deviate from linear re- wavelength(such as Spitzer IRAC data; see, e.g. Yan et
sponse at 5σ level, and these pixels were also fixed in al. 2010a, 2010b), we further require that a legitimate
the same way. The defected pixels were not treated for candidate should have J −H < 0.3 mag in order
125 160
the UVIS images in the current work, because these im- to further reduce the chance of contamination from red
agesareonly usedfor the “veto”purposeinthe dropout galaxies at low redshifts. The right panel of Figure 1
3
Fig. 1.—(left)SystemresponsecurvesoftheWFC3passbands
usedinthecurrentpureparallelobservations. Bothprogramsused
the samefilters intheIR. Inthe UVIS,program 11700 usedV606
twhheilleopcarotigornamof1L1y70α2luinseedreLdPsh60if0t.edThtoevze=rtic7a.4l.do(rtitgehdtl)inCeoilnodri-ccaotloesr ImFaigge.s2a.—re I5m′.′4ag×e5st′.′a4mipnssoizfet.heTthweorYe0d98c-idrcrloepsoaurtesi0n′.′5ouirnsraamdpiules..
diagram demonstrating the selection of Y098-dropouts. The black NorthisupandEastistoleft.
dots arethe fieldobjects. Theshaded areaindicates theselection
area,wherethetwoobjects inourfinalsampleareshowninsolid
redsquaresandthepeculiarY098-dropoutthathasanindicationof
variabilityisshowninopenredsquare. Theconnectedcirclestoleft
showthecolortrackofatypicalyounggalaxyathighredshiftsus-
ingthemodelfromBruzual&Charlot(2003(BC03);τ =0.3Gyr
and age of 100 Myr), and attenuated by the intervening H I ab-
sorptionalongthesightlineusingtheformalismofMadau(1995).
Different colors are used to code different redshift ranges along
thesetracks: blueisforz<6.4,magenta isfor6.4≤z≤7.3,and
blackis forz≥7.4. Theselection areaiswell separated fromthe
regions occupied by the potential sources of contaminations. The
yellowstarsshowthecolorsofGalacticbrowndwarfs(e.g. Leggett
et al. 2002). The green symbols show the colors of a typical red
galaxy at z ≈ 1–3 simulated using a BC03 model (τ = 50 Myr
andageof2.0Gyr). Thecyansymbolsrepresentthecolorsofthe
sameredgalaxybutwithastrong[OIII]emissionlineofrestframe
equivalentwidthof100˚A.Theblackandthegreenarrowsrepresent
the reddening vectors in these passbands appropriate for z = 7.4
andz=2,respectively,usingthereddeninglawofCalzetti(2001)
andE(B−V)=0.5mag.
shows the selection of Y -dropouts on the J −H
098 125 160
versus Y −J color-color diagram.
098 125
The selected candidates were visually inspected to ex-
cludeobjectsthatwereformallyreportedtohaveS/N<2
in the UVIS image but still visible to eyes. In this step,
we also rejected objects that were caused by artifacts
and image defects that are not identified in the data re-
duction process (see Section 2.2). A last concern about
the contaminationis thatthe interpolationofthe identi-
fieddefectedpixelsmightleavesomeresidualsthatcould
mimic the colors of Y -dropouts by chance. To deal
098
with this problem, we created masks of the interpolated
pixels (see Section 2.2) for the individual images, and Fig. 3.— (top) Same as Fig. 2, but for the variable object
ran MultiDrizzle to project them to the same reference (YsDrop03). Thefieldwas observedbyboth programs11700 and
frame of the science images. The locations of the candi- 11702, and hence has UVIS data in both V606 and LP600. (bot-
dates were examined on these projected masks, and we tom) Light curves of HIPPIES YsDrop03. The J125 and H160
data points are matched-aperture MAGAUTO magnitudes ob-
onlykeptthe objectsthatdonothaveanydefects inthe tained using the master J125 stack as the detection image, while
central 6×6 pixel area. the arrows are the 2σ upper limits measured within an r = 0′.′2
aperture. The errors are 1σ error within the MAGAUTO aper-
3.2. Sample ture.
Our selection above resulted in a total of three Y -
098
dropouts over 26 HIPPIES fields that amount to 122.8 taineditslightcurvesasshowninFigure3. Thevariabil-
arcmin2 in area. The colorsofthese three Y -dropouts ity is the most obvious in H , where it is the brightest
098 160
are superposed in the right panel of Figure 1 (solid red in the first image, declines by ∼ 1.0 mag within 2 days,
squares). Their imageareshowninFigures2and3,and andremainsnearly constantin the following ∼1.2 days.
Table 1 lists their photometric information. SimilartrendisalsoseeninJ ,albeitlesssignificantly.
125
One of these three objects, YsDrop03 (see Figure 3), However, such a variability does not seem to present in
shows an indication of peculiar variability, which makes Y or the two UVIS bands, where this object always
098
its candidacy as a high-redshift galaxy less certain. We remains undetected although the observations in these
performedphotometryineachindividualimage,andob- bands were executed both before and after the first im-
4
The two Y -dropouts in our final sample are the
098
brightest candidate galaxies at z > 7 discovered by
WFC3 to date. Even assuming that their redshifts are
at the lower end of the selection window (z ≈ 7.4), we
obtainedtheir absolutemagnitudesofM =−21.6±0.16
and −21.2±0.15 mag for YsDrop01 and YsDrop02, re-
spectively, which are at least 2× brighter than any cur-
rent estimates of the M∗ value at z ≈ 7 or 8. Earlier
surveys overcontiguous fields have resulted in a number
ofz-banddropouts(candidategalaxiesatz ≈7)ofcom-
parable brightness (e.g., Ouchi et al. 2009; Hickey et al.
2010;Castellano et al. 2010)or even brighter (Capak et
al. 2009),and our current work indicates that such high
luminosity galaxies could exist at z &7.4.
Using the formalism of Madau et al. (1998), the
rest-frame UV luminosities of our candidates corre-
spond to star formation rates (SFRs) of 23.7+3.5 and
−3.1
16.4+−22..41 M⊙ yr−1. As an order-of-magnitude estimate,
if they keep forming stars at the same rates to z ≈ 5,
theycouldaccumulatestellarmassesof0.7–1.1×1010M⊙
in the next ∼ 473 million years. Therefore, our Y -
098
dropoutscouldbetheprogenitorsofthehigh-massLBGs
observedatz ≈5thathavestellarmassestotheorderof
Fig. 4.— ComparisonofthesurfacedensityofY098-dropoutsat ∼1010M⊙ (see e.g., Yan et al. 2005;Stark et al. 2007).
tahnedvtehreypbrreidgihcttieonndsifnrofemrrvedarfirooumsLoFusrasatmzp≈le7(raendds8ol(idcuprevnetsa).goTnhse) The existing LF estimates of LBGs at z ≈ 7 and be-
inset zooms in the region at the very bright end. The z ≈7 LFs yond are largely based on observations at L . L∗. For
are taken from Bouwens et al. (2008; Bo08), Ouchi et al. (2009; this reason, it is interesting to see whether the num-
Ou09), McLure et al. (2010; M10), Oesch et al. (2010; Oe10), ber density of such very bright candidate galaxies at
Bwohuilweetnhseezt≈al.8(L2F01s0adr;eBtaok1e0nd)f,roamndBYoa1n0de,tMal1.0(a2n0d10Yba;nYe1t0ba)l,. z & 7.4 is consistent with the expectations from such
(2010a; Y10a). The z ≈7 LFs are integrated over 6.4≤z ≤7.7, LFs. This comparison is shown in Figure 4, where the
whilethez≈8onesareintegratedover7.4≤z≤9.4. Thesurface result from this study is shown as the red pentagons.
densityofgalaxiesatz≈7inferredfromthez850-dropoutsinthe As the formal redshift selection window of Y -dropout
HUDF(blackopensquares; Yanetal. 2010a) andtheERS2field spans 7.4.z .9.4,we compare to the predi0c9t8ions from
(blacksolidsquares;Yanetal. 2010b)arealsoplotted.
the LFs at both z ≈ 8 and 7. Our data point is appar-
entlyhigherthanthepredictionsfromthez ≈8LFs,in-
ageswere takenin H andJ . Therefore,this object
160 125 cludingtheoneofBo10dthatpredictsthehighestsurface
stillremainsalegitimateY -dropoutafterthisscrutiny.
098 density. If we single out the brightest object YsDrop01,
If it is not at z &7.4, the only viable explanation for its
the discrepancy is more obvious. The Bo10d z ≈ 8 LF
coloranditsvariabilityisthatitisaflaringbrowndwarf
predictsacumulativesurfacedensityof∼2.7×10−3per
starwithinourGalaxy(e.g.,Schmidtetal. 2007). How-
arcmin2 to ≤ 25.6 mag over 7.4 . z . 9.4, which is a
ever, this interpretation is less favored, because this ob-
factor of three lower than the density inferred from our
ject has full width at half maximum (FWHM) of ∼ 0′.′3
YsDrop01(8.1×10−3perarcmin−2). Ontheotherhand,
in both J and H stacks and hence seems to be re-
125 160 the LF of Yan et al. (2010b; M∗ = −20.33, α = −1.80
solved. Anotherpossibilityisthatthisobjectisanactive
andΦ=5.52×10−4Mpc−3)predictsthehighestbright-
galactic nucleus (AGN; see, e.g., Cohen et al. 2006) at
end surface density at z ≈ 7. This LF, like others that
z & 7.4, or a transient at a similar redshift. Without
are mostly based on the WFC3 z -dropout results, is
further data, it is difficult to determine its nature. In 850
applicable over the redshift range of 6.4 . z . 7.7. It
any case, it does not fit in the conventional picture of predictsacumulativesurfacedensityof∼1.5×10−2per
an LBG, and therefore we do not include it in our final
arcmin2to≤25.6mag,whichseemstobeabletoexplain
sample.
the existence of YsDrop01 if it is at z .7.7.
The remaining two are good candidate galaxies at
z & 7.4. None of them show time variability. Both are 5. SUMMARY
resolved (FWHM & 0′.′3–0′.′4) in J and H bands
125 160 In this Letter, we present the initial results from the
and hence are not likely brown dwarfs. Finally, they are
HIPPIES program. To date, our analysis includes 26
not likely red galaxies at z ≈ 1–3 whose 4000˚A break widelyseparatedHSTWFC3pureparallelfieldsat|b|>
could mimic Lyman break, as their J125 −H160 colors 20oobtainedinCycle-17,whichamountto122.8arcmin2
are much bluer than those of such interloppers. This
in total area. Using these data, we search for candidate
is particularly true for YsDrop02, which has a very blue galaxies at z & 7.4 as Y -dropouts. One candidate
colorofJ −H ≈−0.2mag. Afewsimilarcaseshave 098
125 160 thatwe found showsanindication ofpeculiar variability
alreadybeenobservedincandidategalaxiesatz ≈7and in J and H but remains undetected in the bluer
125 160
beyond (e.g., Oesch et al. 2010; Bouwens et al. 2010a;
bands, and its nature is unclear. Excluding this object,
Yan et al. 2010a,2010b).
ourcurrentsampleconsistsoftwoverybrightcandidates
at L>2L∗, which, based on their SFR estimates, could
4. DISCUSSION
belinkedtotheprogenitorsofhigh-massLBGsobserved
5
TABLE 1
Properties of Y098-Dropoutsa
ID R.A.andDecl. (J2000)b V606 LP600 Y098 J125 H160 Y−J J−H
HIPPIES-YsDrop01 16:31:35.24+37:36:14.03 >28.1 ... 27.33±0.71 25.52±0.16 25.33±0.21 1.81 0.19
HIPPIES-YsDrop02 14:36:50.56+50:43:33.69 >28.4 ... >28.4 25.86±0.15 26.07±0.27 >2.5 −0.21
HIPPIES-YsDrop03c 07:50:51.41+29:16:17.54 >28.4 >28.3 >29.0 26.62±0.17 26.46±0.20 >2.4 0.16
a. Allmagnitudelimitsare2σlimitsmeasuredwithinanr=0′.′2aperture.
b. ThequotedcoordinatesarebasedontheastrometricsolutionsprovidedbytheOTFRpipelineforthefirstY098 imageineachfield.
c.
Thisobjectshowsindicationofvariabilityandthusitsnatureisunclear.
atlowerredshifts(uptoz ∼5). Whileitssizeisstillvery Wethanktherefereeforthehelpfulcomments. Weac-
limited,oursampleisconstructedfromalargenumberof knowledgethesupportofNASAgrantHST-GO-11702.*.
random fields and thus the impact of cosmic variance is HY is supported by the long-term fellowship program
minimal. Thesurfacedensityinferredfromourbrightest of the Center for Cosmology and AstroParticle Physics
candidate is not well explained by the existing LF esti- (CCAPP) at The Ohio State University. BER is sup-
matesatz ≈8,butcouldbe explainedbytheprediction portedbyHubbleFellowshipProgramnumberHST-HF-
ofLFsatz ≈7ifitbelongstothehigh-redshifttailofthe 51262.01-A. RAW is supported by NASA JWST Inter-
z ≈ 7 galaxy population. Our new HIPPIES data soon disciplinaryScientistgrantNAG5-12460fromGSFC.We
to be taken in the coming HST cycle (∼ 42 pointings) dedicatethisLetter tothememoryofJohnHuchra,who
will enable us to search for different Y -dropouts in a during his life has been a very staunch supporter of the
105
similar way, and the comparison of the two results will Hubble Space Telescope project.
shedlighttotheevolutionofthestarformationactivities
in the early universe.
REFERENCES
Bertin,E.&Arnouts,S.1996, A&AS,117,393 Labb´e,I.,etal.2010b, ApJ,716,L103
Bouwens,R.J.,Illingworth,G.D.,Franx,M.,&Ford,H.2008,ApJ, Leggett, S.K.,etal.2002,ApJ,564,452
686,230 Madau,P.1995,ApJ,441,18
Bouwens,R.J.,etal.2010a, ApJ,708,L69 Madau,P.,Pozzetti, L.&Dickinson,M.,1998,ApJ,498,106
Bouwens,R.J.,etal.2010b, ApJ,709,L133 McLure, R.J., Dunlop, J.S., Cirasuolo, M., Koekemoer, A.M.,
Bouwens,R.J.,etal.2010c,Nature,submitted(arXiv:0912.4263) Sabbi,E.,Stark,D.P.,Targett,T.A.,Ellis,R.S.2010,MNRAS,
Bouwens,R.J.,etal.2010d, ApJ,submitted(arXiv:1006.4360) 403,960
Bruzual,A.G.&Charlot,S.2003,MNRAS,344,1000 Oesch,P.A.,etal.2010,ApJ,709,L21
Bunker,A.,etal.2010, MNRAS,409,855 Ouchi,M.,etal.2009,ApJ,706,1136
Calzetti,D.2001,PASP,113,1449 Robertson,R.E.2010,ApJ,713,1266
Capak,P.,etal.2009,arXiv:0910.0444 Schmidt, S.J.,Cruz, K.L.,Bongiorno, B.J.,Liebert, J., Reid,I.N.
Castellano,M.,etal.2010,A&A,524,28 2007,AJ,133,2258
Cohen,S.H.,etal.2006, ApJ,639,731 Stark, D. P., Bunker, A. J., Ellis, R. S., Eyles, L. P., & Lacy, M.
Dickinson,M.,etal.2004,ApJ,600,L99 2007,ApJ,659,84
Finkelstein, S. L., Papovich, C., Giavalisco, M., Reddy, N.A., Wilkins,S.M.,Bunker,A.J.,Ellis,R.S.,Stark,D.,Stanway,E.R.,
Ferguson, H.C.,Koekemoer, A.M.,& Dickinson, M.2010, ApJ, Chiu,K.,Lorenzoni,S.,Jarvis,M.J.2010a, MNRAS,403,938
719,1250 Wilkins, S.M.,Bunker, A.J., Lorenzoni, S., & Caruana, J. 2010b,
Hickey,S.,Bunker,A.,Jarvis,M.J.,Chiu,K.,&Bonfield,D.2010, MNRAS,accepted ((arXiv:1002.4866)
MNRAS,404,212 Windhorst,R.A.,etal.2010,submittedtoApJ(arXiv:1005.2776)
Koekemoer, A. M., Fruchter, A. S., Hook, R. N., & Hack, W. Yan,H.,etal.2005,ApJ,634,109
2002, in “The 2002 HST Calibration Workshop: Hubble after Yan,H.,Windhorst,R.,Hathi,N.,Cohen,S.,Ryan,R.,O’Connell,
theInstallationoftheACSandtheNICMOSCoolingSystem”, R.,McCarthy,P.2010a, Res.Astron.Astrophys.,10,867
ed.S.Arribas,A.Koekemoer,&B.Whitmore(Baltimore,MD: Yan,H.,etal.2010b, ApJ,submitted
SpaceTelescopeScienceInstitute), 337
Labb´e,I.,etal.2010a,ApJ,708,L26