Far-infrared observations of an unbiased sample of gamma-ray burst host galaxies

3.6 MB·English

by  Saul A. Kohn

Mon.Not.R.Astron.Soc.000,1–11(2014) Printed30January2015 (MNLATEXstylefilev2.2)
Far-infrared observations of an unbiased sample of
gamma-ray burst host galaxies
S. A. Kohn,1,2(cid:63) M. J. Micha(cid:32)lowski,1 N. Bourne,1 M. Baes,3 J. Fritz,3,4 A. Cooray,5
5
I. de Looze,3,6 G. De Zotti,7,8 H. Dannerbauer,9 L. Dunne,1,10 S. Dye,11 S. Eales,12
1
0 C. Furlanetto,11 J. Gonzalez-Nuevo,13,7 E. Ibar,14 R. J. Ivison15,1 S. J. Maddox,1,10
2
D. Scott,16 D. J. B. Smith,17 M. W. L. Smith,12 M. Symeonidis,18,19 E. Valiante12
n
a 1Institute for Astronomy, University of Edinburgh, Royal Observatory, Blackford Hill, Edinburgh EH9 3HJ, UK
J 2Department of Physics and Astronomy, University of Pennsylvania, Philadelphia, PA, 19104, USA
9 3Sterrenkundig Observatorium, Universiteit Gent, Krijgslaan 281 S9, B-9000 Gent, Belgium
2 4Centro de Radioastronom´ıa y Astrof´ısica, CRyA, UNAM, Campus Morelia, A.P. 3-72, C.P. 58089, Michoaca´n, Mexico
5Center for Cosmology, Department of Physics and Astronomy, University of California, Irvine, CA 92697, USA
] 6Institute of Astronomy, University of Cambridge, Madingley Road, Cambridge, CB3 0HA, UK
E 7SISSA, Via Bonomea 265, I-34136, Trieste, Italy
H 8INAF-Osservatorio Astronomico di Padova, Vicolo dell’Osservatorio 5, I-35122 Padova, Italy
9Institut fu¨r Astrophysik, Universit¨at Wien, Tu¨rkenschanzstraße 17, 1180 Wien, Austria
h. 10University of Canterbury, Department of Physics and Astronomy, Private Bag 4800, Christchurch, 8041, New Zealand
p 11School of Physics & Astronomy, University of Nottingham, University Park, Nottingham NG7 2RD, UK
- 12Cardiff School of Physics and Astronomy, Cardiff University, Queen’s Buildings, The Parade, Cardiff, CF24 3AA, UK
o 13Inst. de Fisica de Cantabria (CSIC-UC), Avda. los Castros s/n, 39005 Santander, Spain
r 14Instituto de F´ısica y Astronom´ıa, Universidad de Valpara´ıso, Avda. Gran Bretan˜a 1111, Valpara´ıso, 5030, Chile
t
s 15European Southern Observatory, Karl Schwarzschild Strasse 2, D-85748 Garching, Germany
a 16Department of Physics & Astronomy, University of British Columbia, Vancouver, BC, V6T 1Z1 Canada
[ 17Centre for Astrophysics, Science & Technology Research Institute, University of Hertfordshire, Hatfield, Herts, AL10 9AB, UK

18Astronomy Centre, Department of Physics & Astronomy, University of Sussex, Brighton BN1 9QH, UK
2
19Mullard Space Science Laboratory, University College London, Holmbury St. Mary, Dorking, Surrey RH5 6NT, UK
v
4
0
0
1
0
. ABSTRACT
1
Gamma-ray bursts (GRBs) are the most energetic phenomena in the Universe; be-
0
5 lieved to result from the collapse and subsequent explosion of massive stars. Even
1 though it has profound consequences for our understanding of their nature and selec-
: tion biases, little is known about the dust properties of the galaxies hosting GRBs.
v
We present analysis of the far-infrared properties of an unbiased sample of 20 Bep-
i
X poSAX and Swift GRB host galaxies (at an average redshift of z = 3.1) located in
the Herschel Astrophysical Terahertz Large Area Survey, the Herschel Virgo Cluster
r
a Survey, the Herschel Fornax Cluster Survey, the Herschel Stripe 82 Survey and the
Herschel Multi-tiered Extragalactic Survey, totalling 880 deg2, or ∼3% of the sky in
total. Our sample selection is serendipitous, based only on whether the X-ray posi-
tion of a GRB lies within a large-scale Herschel survey – therefore our sample can
be considered completely unbiased. Using deep data at wavelengths of 100–500µm,
we tentatively detected 1 out of 20 GRB hosts located in these fields. We constrain
their dust masses and star formation rates (SFRs), and discuss these in the context
ofrecentmeasurementsofsubmillimetregalaxiesandultraluminousinfraredgalaxies.
The average far-infrared flux of our sample gives an upper limit on SFR of < 114
M(cid:12)yr−1.ThedetectionrateofGRBhostsisconsistentwiththatpredictedassuming
that GRBs trace the cosmic SFR density in an unbiased way, i.e. that the fraction
of GRB hosts with SFR > 500M(cid:12) yr−1 is consistent with the contribution of such
luminous galaxies to the cosmic star formation density.
Key words: gamma-ray burst: general – dust, extinction – galaxies: high redshift –
galaxies: star formation – infrared: galaxies
(cid:13)c 2014RAS
(cid:63) E-mail:[email protected]
2 S. A. Kohn et al.
1 INTRODUCTION
Table 1.DetailsofHerschelsurveysused
Long-duration gamma-ray bursts (GRBs) have, in many
Survey Area Totalnoise(mJybeam−1)a Ref.
cases, been proven to be connected with supernovae (e.g.
(deg2) 100 160 250 350 500
Galamaetal.1998;Hjorthetal.2003),suggestingthattheir
progenitorstarsareshort-lived(Hegeretal.2003).Asare- H-ATLAS 600 25 30 7.2 8.1 8.8 1,2
sult,GRBsareexpectedtobelocatedinregionsundergoing HeFoCS 16 9.9 9.2 8.9 9.4 10.2 3
active star formation, making GRBs potentially ideal trac- HerMESb 100 – – 6.4 6.8 7.6 4
ers of star formation across a large range of redshifts and HerSc 79 – – 10.7 10.3 12.3 5
HeViCS 84 23 13 6.6 7.3 8.0 6
helpingustochartthestar-forminghistoryofourUniverse
(e.g. Wijers et al. 1998; Yonetoku et al. 2004). a Thewavelengthinmicronsofeachbandisgivenintheheader.
Many properties of GRBs are useful for accomplishing b HerMES data have different noise levels depending upon
theabovetask.Theirextremelyhighluminosityatgamma- the field surveyed. The average noise levels for each band are
presentedhere,giveninreference(4).
ray wavelengths allows for their detection out to very high
c Coverage of the HerS maps is not uniform. The average noise
redshift, with minimal extinction from gas or dust. The av-
levelsarepresentedhereforthedeeperpartofthesurvey(where
erage GRB redshift is z ∼ 2.2 (Fynbo et al. 2009; Jakobs-
GRB060908islocated,seeTable2andreference5).
son et al. 2006, 2012; Hjorth et al. 2012), but in principle
Hyphens indicate that data were not taken or not available in
GRBs should be visible out to z (cid:62) 15–20 (Lamb & Re- thatband.
ichart2000),andinpracticetheyhavebeenobservedoutto References: (1) Ibar et al. (2010); (2) Pascale et al. (2011);
z ∼9 (Tanvir et al. 2009; Salvaterra et al. 2009; Cucchiara (3) Fuller et al. (2014); (4) Smith et al. (2012); (5) Viero et al.
etal.2011).Moreover,detectionsofGRBsareindependent (2014),and(6)Auldetal.(2013).
of the luminosity of their host galaxies, which allows us to
probe fainter galaxies than in typical flux-limited samples cated inside wide-area Herschel1 (Pilbratt et al. 2010) sur-
(e.g. Tanvir et al. 2012). Previous studies have found that veys. It is important to study this aspect at both ultravio-
GRBhostsaremostlyfaintandblue(LeFloc’hetal.2003), let/optical and infrared wavelength, since these give access
with the GRB occurring in (rest-frame) UV-bright regions to unobscured and dust-obscured star formation activity,
(Bloom et al. 2002; Fruchter et al. 2006; Leloudas et al. respectively. The latter is still poorly understood for GRB
2010, 2011), which is consistent with active star formation. hosts, where deep observations and detections have been
More recent studies (e.g. Greiner et al. 2011; Kru¨hler et al. largely limited to low redshifts and biased samples (Frail
2011; Rossi et al. 2012; Perley et al. 2013) show that GRB etal.2002;Bergeretal.2003;Tanviretal.2004;deUgarte
hostsspanawiderrangeofproperties,suggestingthatprevi- Postigoetal.2012;Wangetal.2012;Hatsukadeetal.2014;
ousoptically-selectedhostsurveyswerebiasedtowardsthese Huntetal.2014;Michal(cid:32)owskietal.2014;Schadyetal.2014;
faint blue hosts. Symeonidis et al. 2014). The objectives of this paper are
Before we can use GRBs to quantitatively trace star to:(1)measurethedust-obscuredSFRsanddustmassesof
formation across cosmic history, such biases of GRB and GRB host galaxies using an unbiased sample; and (2) test
GRB host galaxy samples must be well understood. Us- whether or not GRBs are unbiased tracers of cosmic star
ing The Optically Unbiased GRB Host survey (TOUGH; formation at z>2.
Hjorth et al. 2012), Michal(cid:32)owski et al. (2012) and Perley Thelayoutofthispaperisasfollows.Section2contains
etal.(2014)analysedtheradio-derivedstarformationrates an overview of the Herschel surveys whose data we used to
(SFRs) of GRB hosts at z <2. They found that the distri- acquire our sample. We then discuss the measured proper-
butions of SFRs of GRB hosts were consistent with that of ties of the GRB hosts in Section 3, before moving on to a
other star-forming galaxies at z <2, suggesting that GRBs more extensive discussion of these properties as compared
may be able to trace cosmic star formation; however they to recent analyses of other galaxy types in Section 4. We
madeclearthatfurtherstudyofpotentialbiasesinthemor- provide out conclusions in Section 5.
phologyandmetallicityofGRBhostsisrequired.Similarly, We use a cosmological model of H0=70kms−1Mpc−1,
Hunt et al. (2014) and Schady et al. (2014) concluded that ΩΛ = 0.7 and Ωm = 0.3, and a Salpeter (Salpeter 1955)
SFRs in GRB hosts are consistent with other star-forming initial mass function.
galaxies, and found no strong evidence for GRB hosts be-
ing biased tracers of the global star formation rate den-
sity(SFRD).However,theirsamplefavouredinfrared-bright
hosts. In their study of the hosts of dust-obscured GRBs, 2 SAMPLE AND DATA
Perley et al. (2013, 2014) found that they are not massive
The Herschel Astrophysical Terahertz Large Area Survey
enough to bring the overall GRB host population into the
(H-ATLAS;Ealesetal.2010)isthelargestopen-timesurvey
expected territory of an unbiased SFR-tracing population
conductedbytheHerschel Space Observatory.Itcoversap-
for z ∼1. They found that GRB hosts appear to be biased
proximately 600 deg2 in the far-infrared (FIR) and submil-
towardslow-massgalaxiesandsuggestedthattheGRBrate
limetrewavelengths,usingPACS(100µm,160µm;Poglitsch
relativetoSFRishighlydependentonhost-galaxyenviron-
etal.2010)andSPIRE(250µm,350µmand500µm;Griffin
mentatthisredshift,andhighlightedthatfurtherstudiesof
metallicity are needed. These findings are corroborated by
Boissier et al. (2013).
1 HerschelisanESAspaceobservatory withscienceinstruments
Here we attempt to study the distribution of SFRs of provided by European-led Principal Investigator consortia and
GRB hosts using the unbiased sample of GRBs that are lo- withimportantparticipationfromNASA.
(cid:13)c 2014RAS,MNRAS000,1–11
Unbiased FIR observations of GRB hosts 3
Table 2.GRBsample
GRB α δ 90%error Herschel z Reference
(J2000) (J2000) (arcsec) Field
990308 185.7976667 6.7347500 0.3 HeViCS Schaeferetal.1999
050412 181.1053333 −1.2001667 3.7 GAMA12 Hjorthetal.2012
050522 200.1440417 24.7890833 3.8 NGP Butler2007
051001 350.9530417 −31.5231389 1.5 SGP 2.4296 Butler2007;Hjorthetal.2012
060206 202.9312917 35.0507778 0.8 NGP 4.048 Butler2007;Fynboetal.2009
060908 31.8267500 0.34227778 1.4 HerS 1.884 Hjorthetal.2012;Jakobssonetal.2012
070611 1.99241667 −29.7554722 1.8 SGP 2.0394 Butler2007;Hjorthetal.2012
070810A 189.9635000 10.7508889 1.4 HeViCS 2.17 Hjorthetal.2012;Schadyetal.2012
070911 25.8095417 −33.4841389 3.1 SGP Butler2007
071028B 354.1617917 −31.6203611 0.3 SGP Clemensetal.2011
080310 220.0575000 −0.1748889 0.6 GAMA15 2.42 Littlejohnsetal.2012;Foxetal.2008
091130B 203.1481250 34.0885278 0.6 NGP Butler2007
110128A 193.8962917 28.0654444 0.4 NGP 2.339 Butler2007;Sparreetal.2011
110407A 186.0313917 15.7118417 1.0 HeViCS Chuangetal.2011
111204A 336.6283750 −31.3748056 1.9 SGP Sonbasetal.2011
120703A 339.3567083 −29.7232500 0.5 SGP Xuetal.2012
120927A 136.6137500 0.4161944 1.4 GAMA09 Beardmoreetal.2012
121211A 195.5332917 30.1485000 0.5 NGP Chester&Mangano2012
130502A 138.5689583 −0.1233056 1.5 GAMA09 Beardmoreetal.2013
140102A 211.9193750 1.3332778 0.5 GAMA15 Hagen2014
140515A 186.0650000 15.1045556 1.8 HeViCS 6.327 Butler2007;Chornocketal.2014
et al. 2010). The Herschel Virgo Cluster Survey (HeViCS; long as close to 100% of the gamma-ray triggered GRBs
Daviesetal.2010)usedPACSandSPIREtosurveyatotal then have good X-ray localizations obtained for them. In-
area of 84 deg2 (Auld et al. 2013; Baes et al. 2014) of the deed, nearly all GRBs are detected in X-ray when prompt
Virgo galaxy cluster, with diminishing sensitivities beyond observations are available.
the central 55 deg2. Similarly, the Herschel Fornax Cluster We selected all GRBs with X-ray positions (Butler
Survey(HeFoCS;Daviesetal.2013)usedPACSandSPIRE 2007) inside the H-ATLAS, HeViCS and HerS fields, which
to survey a total area of 16 deg2 in the Fornax cluster. The amounted to 21 GRBs. GRB 990308 occurred within the
HerschelStripe82Survey(HerS;Vieroetal.2014)covered HeViCS field (Schaefer et al. 1999). As this is a pre-Swift
79 deg2 along the Sloan Digital Sky Survey (SDSS; York burst, in order to not corrupt the unbiased nature of our
et al. 2000) ‘Stripe 82’ field using SPIRE. The Herschel samplewesimplyprovideourmeasurementsofitsfluxden-
Multi-tiered Extragalactic Survey (HerMES; Oliver et al. sity here, and it does not enter into any other calculations.
2012; Roseboom et al. 2010; Smith et al. 2012; Viero et al. NoGRBpositionsoverlappedwithanyHeFoCSorHerMES
2013; Wang et al. 2014) used SPIRE to observe 29 fields, fields.Thecataloguecontains780GRBsovertheentiresky,
covering ∼100 deg2 in total. soin880deg2weexpecttofind780×880/41253(cid:39)17GRBs,
Inthisstudy,weuseH-ATLASGAMA,NGPandSGP which is close to the number of 20 that we found (i.e. ex-
data(Ibaretal.2010;Pascaleetal.2011;Rigbyetal.2011; cludingGRB990308).Oursampleselectionprocessmaybe
Smith et al. 2011, Bourne et al. in prep, Valiante et al. in considered unbiased, since the only selection criterion ap-
prep.), all observed HeViCS fields (V1 to V4), the HeFoCS plied was positional. Table 2 shows the positions and any
field(Daviesetal.2013),HerSmaps2 andthesecondmajor knownredshiftsoftheGRBsinoursample;Fig.1showsthe
HerMES data release3 (Smith et al. 2012). It is important Herschel data. For seven of our GRBs redshifts are known,
tonotethateachofthesesurveysuseddifferentnumbersof and only for these cases do we provide physical properties.
cross-scanswiththeinstrumentsaboardHerschel,resulting
in different levels of sensitivity. These differences are sum-
marized in Table 1. 3 RESULTS
Butler (2007) presented a catalogue of refined X-ray
positions of Swift/XRT (Gehrels et al. 2004; Burrows et al. We determined the FIR flux densities of GRB hosts by fit-
2005) GRBs4. Half of the 90% confidence error region radii ting Gaussian functions at the GRB positions in the Her-
are <2.2 arcsec. It has been shown by Fynbo et al. (2009) schel maps using the respective beam sizes5. If a GRB host
and Hjorth et al. (2012) that such X-ray selection, as op- appeared to be detected at > 3σ we further investigated
posed to optical selection, does not introduce any bias, as possible blending. We identified any obvious optical or FIR
sources clearly distinct from the GRB host, and repeated
the flux density measurement by fitting simultaneously two
2 www.astro.caltech.edu/hers/ (or more) Gaussian functions at the GRB positions and at
3 hedam.lam.fr/HerMES/
4 Thecatalogue(butler.lab.asu.edu//Swift/xrt_pos.html)is
constantlyupdated.Weaccessedthecatalogueon2014July25. 5 Thesewere9.4,13.4,18.2,24.9and36.3arcsecforthe100,160,
Atthistime,ithad780entries. 250,350and500µmbands,respectively.
(cid:13)c 2014RAS,MNRAS000,1–11
4 S. A. Kohn et al.
990308
050412
050522
051001
060206
060908
070611
070810A
070911
100 µm 160 µm 250 µm 350 µm 500 µm Optical
Figure 1.HerschelimagesatthepositionsoftheGRBsinoursample.Atthecentreofeachframe,ablueboxindicatesthepositionof
theGRB,aslistedinTable2.Additionalsources,whichweresimultaneouslyfittedtogetherwiththeGRBpositions,aremarkedwith
redcircles(seeSection3).PACSimagesareshownfrom−2(white)to2mJypixel−1 (black).Thepixelsizesare3and4arcsecfor100
and 160um, respectively. SPIRE images from −10 (white) to 30 mJy beam−1 (black). Each panel is 2 arcmin on a side. The first five
columnsshow100,160µmPACSbands,and250,350,500µmSPIREbands,respectively,whileblankframesindicatethatPACSmaps
oftheseregions(HerS)arenotavailable.Thelastcolumnshowstheopticaldata(fromHjorthetal.(2012)for050412,051001,060908,
070611and070810A;fromTh¨oneetal.(2008)for060206;fromdeUgartePostigoetal.(2011)andSparreetal.(2011)for110128A;from
SDSS (Ahn et al. 2012) for 990308, 050522 and 110407A and from GAMA (Driver et al. 2009) for 070911, 080310, 091130B, 111204A,
120703A,120927A,121211A,130502A,140102Aand140515A).
(cid:13)c 2014RAS,MNRAS000,1–11
Unbiased FIR observations of GRB hosts 5
071028B
080310
091130
110128A
110407A
111204A
120703A
120927A
121211A
100 µm 160 µm 250 µm 350 µm 500 µm Optical
Figure 1.(continued.)
the positions of these other sources (this was performed for ous optical source was found close to the positions of GRB
GRBs 060206, 070911, 071028B, 091130 and 110128A). For 060206(Thoeneetal.2007),071028Band091130B(GAMA;
GRB 110128A in the deep optical image from de Ugarte Driver et al. 2009) but on the 250µm images we identified
Postigo et al. (2011) and Sparre et al. (2011) we found a sources that are clearly distinct from the GRB hosts.
low-z galaxy contributing to the Herschel flux density close ThefluxdensitiesineachbandarepresentedinTable3.
to the GRB position (red circle on Fig. 1). No such obvi- Noise was estimated by measuring the flux at 100 random
(cid:13)c 2014RAS,MNRAS000,1–11
6 S. A. Kohn et al.
130502A
140102A
140515A
100 µm 160 µm 250 µm 350 µm 500 µm Optical
Figure 1.(continued.)
positions around the GRB position (and in the same way
Table 3.HerschelfluxdensityatGRBpositions
as for the GRB position) and the error is the 3σ clipped
mean of these measurements; it was consistent with being
GRB F100 F160 F250 F350 F500
dominated by the confusion noise (Nguyen et al. 2010). (mJy) (mJy) (mJy) (mJy) (mJy)
Only GRB 091130B was located close to a source from
990308 −13±9 7±12 −1±7 2±8 −2±7
theH-ATLAScatalogue(9arcsecseparation,corresponding
050412 −6±13 −24±18 7±7 2±8 9±7
to half of the beam size at 250µm). This source (HATLAS
050522 12±13 −17±17 −10±8 −7±8 0±9
J133235.2+340526) has measured flux densities of 43±6,
051001 10±7 3±12 −5±8 −4±10 −6±8
21±7 and 7±8mJy in the 250, 350 and 500µm bands, 060206 20±14 2±20 8±8 6±11 7±9
respectively(Rigbyetal.2011).Anassociationbetweenthe 060908a – – −8±8 −3±9 −2±9
GRB and the source is unlikely – their 9 arcsec separation 070611 −14±20 14±20 11±5 12±6 −3±7
is large compared to the 2 arcsec positional accuracy of the 070810A −30±17 −19±24 5±10 15±10 28±12
Herschel source. 070911 8±11 22±20 9±12 3±13 −7±11
071028B 6±14 0±20 −2±9 0±8 3±10
SFRs were calculated according to Kennicutt (1998),
080310 −3±11 19±14 −12±9 −19±10 −24±13
basedontheinfraredluminosityintegratedover8−1000µm
091130 −9±14 −23±18 −11±9 4±8 4±8
usingtheSEDofULIRGArp220(Silvaetal.1998),scaled
110128A 13±14 −7±18 15±8 6±9 5±9
to the flux densities of our GRB hosts. The use of Arp 220
110407A −10±11 11±18 −15±5 −14±9 −13±8
isappropriate,sinceitsdusttemperatureisclosetothatof 111204A 23±14 −34±17 −17±9 −3±13 5±11
typical GRB hosts (Priddey et al. 2006; Micha(cid:32)lowski et al. 120703Ab 33±13 -39±20 4±10 0±9 4±11
2008,2009,2014;Watsonetal.2011;Hatsukadeetal.2014; 120927A −20±13 0±22 16±11 −1±10 −9±12
Hunt et al. 2014; Symeonidis et al. 2014). However, since 121211A −3±13 −16±16 −14±7 −3±8 −2±8
we probe close to the peak of dust emission, the choice of 130502A 22±11 7±14 −10±5 −21±8 −31±9
thetemplatedoesnotinfluencetheresultssubstantially.For 140102A −18±12 −13±17 −1±8 −6±9 −3±8
140515A 0±9 −8±9 −14±8 −13±10 −20±9
example,themediandifferencebetweentheSFRscalculated
NWMc 2±3 −4±4 −3±2 −2±2 −3±2
using the SED of Arp 220 and the SED of a typical SMG
Stacking 1±3 −2±4 3±2 3±4 3±2
(Micha(cid:32)lowski et al. 2010) is ∼10%.
To derive dust masses of GRB hosts, we assumed Errors are from the clipped average of the flux densities of ran-
T = 50K (a similar temperature to Arp 220; Klaas dompositionsaroundtheGRBpositions.
dust a GRB060908isinsideHerS,aSPIRE-onlysurvey.
et al. 1997; Dunne et al. 2000; Lisenfeld et al. 2000; Rang-
b Tentativedetection.
wala et al. 2011), emissivity β = 1.5 (which yields con-
c Noise-weightedmean.
servative upper limits) and a dust absorption coefficient
κ = 0.67cm2g−1 (Silva et al. 1998). Assuming lower
1.2mm
Tdust = 30K results in masses larger by a factor of ∼ 10. spondingtotheJ2000coordinatesofeachGRBhostinthe
Derived properties of the GRB hosts (at Tdust = 30 and SPIREandPACSimages.UsingPSF-filteredimagesineach
50K) with known redshift are presented in Table 4. Fig. 2 band, a region of 15×15 pixels was extracted around each
showsthespectralenergydistributionsofthoseGRBswith position, and these images were combined using a weighted
known redshifts, and are overlaid on the SED of Arp 220, mean.Eachpixelinthestackedimageistheweightedmean
which was used to calculate the SFR of each host. of the corresponding pixels from all the cut-outs, weighted
StackedimagesinthefiveHerschelbandswereobtained by the inverse variance of the flux density measurement at
by averaging cut-out images around the central pixel corre- the GRB host position. The size of the stacked images cor-
(cid:13)c 2014RAS,MNRAS000,1–11
Unbiased FIR observations of GRB hosts 7
respondsto45,60,90,120and180arcsecateachofthefive moresignificantsamplesof GRB hostsobservedintheFIR
bands,respectively.Wedidnotdetectanysignificantsignal are needed to investigate this issue.
atanyofthestackedimages(fluxdensitiesarepresentedin The average 2σ upper limit of the dust mass for our
the last row of Table 3). sampleislog(Mdust/M(cid:12))=7.4±0.3–8.1±0.2(50 −30K).
This suggests that GRB hosts are less dusty than SMGs,
as Micha(cid:32)lowski et al. (2010) found the typical SMG to con-
tain log(Mdust/M(cid:12))=9.02±0.36 (also see Chapman et al.
2005; Kova´cs et al. 2006; Laurent et al. 2006; Coppin et al.
4 DISCUSSION
2008; Magnelli et al. 2012; Swinbank et al. 2014). The dust
GRBs are closely linked with star-formation, but they can massesofGRBhostsinoursampleareonlyconsistentwith
be used as a probe of star formation only if the number thoseofSMGsifweassumeaverylowtemperatureof30K
of GRBs at a given redshift is proportional to the cosmic (e.g.Micha(cid:32)lowskietal.2010foundandaveragetemperature
SFRDatthisepoch.InsuchacasethedistributionofSFRs of 35K for SMGs). Calzetti et al. (2000) and Smith et al.
of GRB hosts should reflect the contribution of galaxies to (2013) found that dust emission at much lower tempera-
the SFRD. In particular, the fraction of hosts above some tures (20−30K) could account for a large fraction of total
SFRthresholdshouldbeequaltothefractionalcontribution flux in starburst galaxies, but other studies (Priddey et al.
of such luminous galaxies to the SFRD. If the fraction of 2006;Micha(cid:32)lowskietal.2008,2009,2014;Symeonidisetal.
luminousGRBhostsislower(higher)thanthiscontribution, 2014) suggest this is an unrealistically low temperature for
then we would conclude that GRBs are biased towards less an infrared-bright GRB host (however none of these stud-
(more) actively star-forming galaxies. ies is based on an unbiased GRB sample). Moreover, high
Integrating the 1.8<z <2.3 infrared luminosity func- dust temperature was found to be typical among Herschel-
tion of Magnelli et al. (2013) above LIR > 5 × 1012L(cid:12) selectedULIRGsathighredshifts(Symeonidisetal.2013).
(SFR>500M yr−1)weinferthatsuchgalaxiescontribute
(cid:12) The noise-weighted average of the flux densities at all
8+6% (a 2σ range of 3-24%) to the cosmic SFRD at z ∼2.
−3 20GRBpositionsandfluxdensitiesmeasuredinthestacked
IfGRBstracestarformationdensityinanunbiasedway,we
maps are shown in Table 3. At an average GRB redshift of
would expect this percentage of GRB hosts (1.7+1.3 hosts)
−0.6 z=2.14(Hjorthetal.2014),thiscorrespondstoa2σ limit
in our sample to be detected, because Herschel sensitivity of SFR < 114M(cid:12)yr−1 (LIR < 2×1012L(cid:12)) using the Arp
allowsustodetectgalaxieswithLIR >5×1012L(cid:12) (assum- 220 template, and 2σ limits on dust masses of M <2×
dust
ingtheArp220templateatz=2–2.5scaledtothe3σlimit 107M(cid:12)forTdust =50K,andMdust <2×108M(cid:12)forTdust =
of 15–20mJy).
30K. Measurements of stacked images provide consistent
Out of our 21 GRB hosts, we found just one tenta- limits on the dust mass of typical GRB hosts: M <6×
dust
tive detection (5±5%), namely GRB120703A with a sig- 107M(cid:12) for Tdust = 50K, and Mdust < 2.4×108M(cid:12) for
nificance of ∼ 2.7σ, suggesting that GRBs may trace star- T = 30K. The deepest limit on SFR comes from the
dust
formation density in an unbiased way, at least for redshift stacked 250µm-band, giving a typical GRB host SFR of
2<z<4(andtheexpectationof1−3detectedhostsisan <114 M(cid:12)yr−1.
upper limit, since we are assuming that GRBs in our sam-
If GRBs trace the cosmic SFRD in an unbiased way,
ple with no redshift information are at similar redshifts as
then the average luminosity of GRB hosts should be equal
the other ones). This is supporting evidence for the much
totheaverageluminosityofothergalaxiesweightedbytheir
sought-after relationship between GRB rate and SFR (e.g.
SFRs(becauseagalaxywithahigherSFRwillhaveahigher
Blain & Natarajan 2000; Lloyd-Ronning et al. 2002; Hern-
probability to host a GRB). From the luminosity function
quist & Springel 2003; Christensen et al. 2004; Yonetoku
(φ)ofMagnellietal.(2013)wecalculatedtheweightedmean
et al. 2004; Yu¨ksel et al. 2008; Kistler et al. 2009; Elliott (cid:104)L(cid:105) =(cid:82)Lmaxφ·SFR×LdL/(cid:82)Lmaxφ×SFRdL.Assum-
etal.2012;Micha(cid:32)lowskietal.2012;Robertson&Ellis2012; SFR Lmin Lmin
Hunt et al. 2014). ingSFR∝LIR thisgives(cid:104)log(L/L(cid:12))(cid:105)SFR =12.23±0.15,or
However, our analysis suffers from low number statis- (cid:104)SFR(cid:105)SFR = 290+−19200M(cid:12)yr−1 using the Kennicutt (1998)
tics as none of our Herschel flux measurements is of high conversion and assuming the Saltpeter IMF (propagating
significance,weconsiderherehowourconclusionschangeif the errors on luminosity function parameters using the
theyarenotreal(0+4%detectionrate).Thatwouldproduce MonteCarlomethod).Thisvalueisonlyweaklydependent
−0
a < 2σ tension between the data and the expectation that on the adopted cut-off luminosities, log(Lmin/L(cid:12)) = 7 and
GRBstracethestarformationinanunbiasedwayatz∼2 log(Lmax/L(cid:12))=13,asitismostlyconstrainedbytheshape
(i.e. 8+6% detection rate). This demonstrates that an FIR oftheluminosityfunctionclosetoitsknee.Again,thisvalue
−3
survey of a larger unbiased sample of GRB hosts is crucial isonly<2σawayfromthelimitwemeasuredforGRBhosts
to reach a statistically significant conclusion. (<114M(cid:12)yr−1),sowecannotruleoutthatGRBhostsare
These conclusions are not uncontested. Perley et al. simply drawn based on the SFRs from the general popu-
(2013)foundabiasedcorrespondencebetweentheGRBhost lation of galaxies, and we highlight that larger samples of
galaxy population and SFRD (at least at z <1.5). In their unbiased GRB hosts observed at the FIR are needed.
extensive study of the hosts of Swift-observed GRBs, they The average flux density suggests that a typical GRB
found that GRB rate is a strong function of host-galaxy host has a flux below the confusion limit of Herschel, so
properties. In particular, GRB hosts were found to be less surveyswithbetterresolutionareneededtodetectthema-
massiveandbluerthanwhatwouldbeexpectediftheGRB jorityofthepopulation.Suchsurveyswillbeundertakenin
rate is proportional to SFR in galaxies at z < 1.5 (Perley the future with the Large Millimeter Telescope and Cerro
et al. 2013). We probe GRB hosts at higher redshifts, but Chajnantor Atacama Telescope.
(cid:13)c 2014RAS,MNRAS000,1–11
8 S. A. Kohn et al.
GRB051001 GRB060206
100.0 100.0
z=2.4296 z=4.048
mJy) 10.0 mJy) 10.0
Flux density ( 1.0 Flux density ( 1.0
0.1 0.1
10 100 1000 10000 10 100 1000 10000
Wavelength (microns) Wavelength (microns)
GRB060908 GRB070611
100.0 100.0
z=1.884 z=2.0394
Flux density (mJy) 101..00 Flux density (mJy) 110..00
0.1 0.1
10 100 1000 10000 10 100 1000 10000
Wavelength (microns) Wavelength (microns)
GRB070810A GRB080310
100.0 100.0
z=2.17 z=2.42
Flux density (mJy) 110..00 Flux density (mJy) 110..00
0.1 0.1
10 100 1000 10000 10 100 1000 10000
Wavelength (microns) Wavelength (microns)
GRB110128A GRB140515A
100.0 100.0
z=2.339 z=6.327
mJy) 10.0 mJy) 10.0
Flux density ( 1.0 Flux density ( 1.0
0.1 0.1
10 100 1000 10000 10 100 1000 10000
Wavelength (microns) Wavelength (microns)
Figure 2. 2σ upper limits on the SEDs of GRBs with known redshifts overlaid with the SED of Ar(cid:13)pc222001,4uRseAdSt,oMcaNlRcuAlaSte00th0e,1S–F1R1
of each GRB host galaxy. The SED of Arp220 has been redshifted to match each GRB, shown on the upper right of each plot. Upper
limitsarefromthebaseofthearrow.
Unbiased FIR observations of GRB hosts 9
Table 4.PropertiesofGRBswithknownredshifts
GRB z Dustmass(T =30K) Dustmass(T =50K) SFR
(108 M(cid:12)) (108 M(cid:12)) (M(cid:12)yr−1)
051001 2.4296 <11 <0.9 <500
060206 4.048 <24 <2.7 <1500
060908 1.884 <6.2 <0.7 <320
070611 2.0394 5.1±2.6 0.5±0.2 280±120
070810A 2.17 <9.8 <0.9 <520
080310 2.42 <11 <1.0 <560
110128A 2.339 <9.4 <0.9 <550
140515A 6.327 <83 <4.9 <2800
NWMa 2.14 <1.8 <0.3 <114
Stackinga 2.14 <2.4 <0.6 <114
Dustmassescalculatedassumingemissivityβ=1.5.Errorsreflectonlystatisticaluncertaintypropagatedfromthefluxdensityerrors,
andlimitsareto2σ.
a Thesearethepropertiescalculatedfromthenoise-weightedmeanfluxdensityofall21GRBpositions,atanaverageGRBredshiftof
z=2.14(Hjorthetal.2014;Section4).
5 CONCLUSION on data taken from the SDSS and the UKIRT Infrared
DeepSkySurvey.ComplementaryimagingoftheGAMAre-
We have measured the FIR flux densities of 20 GRB host
gions is being obtained by a number of independent survey
galaxies.Thesamplewasselectedinanovel,unbiasedfash-
programmes including GALEX MIS, VST KIDS, VISTA,
ion,inwhichweusedadatabaseofGRBpositionstocross-
VIKING,WISE,Herschel-ATLAS,GMRTandASKAPpro-
referencethepositionssurveyedbytheH-ATLAS,HeViCS,
vidingUVtoradiocoverage.GAMAisfundedbytheSTFC
HeFoCS, HerMES and HerS data releases. For 8 out of the
(UK), the ARC (Australia), the AAO, and the partici-
20GRBs,redshiftsareavailable,andforthesewewereable
pating institutions. The GAMA website is: http://www.
to calculate or place an upper limit on the SFR and dust
gama-survey.org/.
massofthehostgalaxy.Onehostwas(tentatively)detected,
This research has made use of data from HerMES
consistent with the contribution of such bright galaxies to
project(http://hermes.sussex.ac.uk/).HerMESisaHer-
the SFRD in the Universe at z ∼ 2 (Magnelli et al. 2011,
schel Key Programme utilizing Guaranteed Time from the
2013). This may support the thesis that the GRB rate and
SPIREinstrumentteam,ESACscientistsandamissionsci-
SFR are fundamentally related, and that GRBs may trace
entist. The HerMES data were accessed through the Her-
SFRDinanunbiasedway.AnalysingalargersampleofGRB
schel Database in Marseille (HeDaM - http://hedam.lam.
hosts selected in such an unbiased way is necessary to give
fr) operated by CeSAM and hosted by the Laboratoire
a definite answer on whether this is indeed the case.
d’Astrophysique de Marseille.
Funding for SDSS-III has been provided by the Al-
fredP.SloanFoundation,theParticipatingInstitutions,the
ACKNOWLEDGEMENTS National Science Foundation, and the US Department of
Energy Office of Science. The SDSS-III web site is http:
We thank Daniele Malesani and Christina Tho¨ne for their //www.sdss3.org/.
contributions of additional optical data, and the anony-
mous referee for their helpful comments, which improved
the quality of the paper. MJM acknowledges the support
of the UK Science and Technology Facilities Council. NB is
REFERENCES
supported by the EC FP7 SPACE project ASTRODEEP
(Ref. no. 312725). EI acknowledges funding from CONI- Ahn C. P., et al., 2012, ApJS, 203, 21
CYT/FONDECYT postdoctoral project N◦:3130504. LD, Auld R. et al., 2013, MNRAS, 428, 1880
RJI and SJM acknowledge support from ERC Advanced Baes M. et al., 2014, A&A, 562, A106
Grant COSMICISM. JGN acknowledges financial support BeardmoreA.P.,EvansP.A.,GoadM.R.,OsborneJ.P.,
from the Spanish CSIC for a JAE-DOC fellowship, co- 2012, GRB Coordinates Network, 13824, 1
fundedbytheEuropeanSocialFund,bytheSpanishMinis- BeardmoreA.P.,EvansP.A.,GoadM.R.,OsborneJ.P.,
terio de Ciencia e Innovacion, AYA2012-39475-C02-01, and 2013, GRB Coordinates Network, 14536, 1
Consolider-Ingenio 2010, CSD2010-00064, projects. BergerE.,CowieL.L.,KulkarniS.R.,FrailD.A.,Aussel
TheHerschel-ATLASisaprojectwithHerschel,which H., Barger A. J., 2003, ApJ, 588, 99
is an ESA space observatory with science instruments pro- Blain A. W., Natarajan P., 2000, MNRAS, 312, L35
vided by European-led Principal Investigator consortia and BloomJ.,KulkarniS.,DjorgovskiS.,2002,ApJ,123,1111
with important participation from NASA. The H-ATLAS Boissier S., Salvaterra R., Le Floc’h E., Basa S., Buat V.,
website is http://www.h-atlas.org/. Prantzos N., Vergani S. D., Savaglio S., 2013, A&A, 557,
GAMA is a joint European-Australasian project A34
based around a spectroscopic campaign using the Anglo- Burrows D. N. et al., 2005, SSRv, 120, 165
Australian Telescope. The GAMA input catalogue is based Butler N., 2007, AJ, 133, 1027
(cid:13)c 2014RAS,MNRAS000,1–11
10 S. A. Kohn et al.
CalzettiD.,ArmusL.,BohlinR.C.,KinneyA.L.,Koorn- Laurent G. T. et al., 2006, ApJ, 643, 38
neef J., Storchi-Bergmann T., 2000, ApJ, 533, 682 Le Floc’h E. et al., 2003, A&A, 400, 499
Chapman S. C., Blain A. W., Smail I., Ivison R. J., 2005, Leloudas G. et al., 2011, A&A, 530, A95
ApJ, 622, 772 Leloudas G., Sollerman J., Levan A. J., Fynbo J. P. U.,
ChesterM.M.,ManganoV.,2012,GRBCoordinatesNet- Malesani D., Maund J. R., 2010, A&A, 518, A29
work, 14063, 1 LisenfeldU.,IsaakK.G.,HillsR.,2000,MNRAS,312,433
Chornock R., Berger E., Fox D. B., Fong W., Laskar T., Littlejohns O. M. et al., 2012, MNRAS, 421, 2692
Roth K. C., 2014, ApJ, submitted Lloyd-RonningN.M.,FryerC.L.,Ramirez-RuizE.,2002,
Christensen L., Hjorth J., Gorosabel J., 2004, A&A, 425, ApJ, 574, 554
913 Magnelli B., Elbaz D., Chary R. R., Dickinson M., Le
Chuang C. J., Lin C. S., Urata Y., Huang K. Y., 2011, Borgne D., Frayer D. T., Willmer C. N. A., 2011, A&
GRB Coordinates Network, 11901, 1 A, 528, A35
Clemens C. et al., 2011, A&A, 529, A110 Magnelli B. et al., 2012, A&A, 539, A155
Coppin K. et al., 2008, MNRAS, 384, 1597 Magnelli B. et al., 2013, A&A, 553, A132
Cucchiara A. et al., 2011, ApJ, 736, 7 Micha(cid:32)lowski M., Hjorth J., Watson D., 2010, A&A, 514,
Davies J. I. et al., 2010, A&A, 518, L48 A67
Davies J. I. et al., 2013, MNRAS, 428, 834 Micha(cid:32)lowskiM.J.,HjorthJ.,CastroCer´onJ.M.,Watson
de Ugarte Postigo A. et al., 2012, A&A, 538, A44 D., 2008, ApJ, 672, 817
de Ugarte Postigo A., Malesani D., McCormac J., Jakobs- Micha(cid:32)lowski M. J. et al., 2009, ApJ, 693, 347
son P., 2011, GRB Coordinates Network, 11605, 1 Micha(cid:32)lowski M. J. et al., 2014, A&A, 562, A70
Driver S. P., et al., 2009, Astronomy and Geophysics, 50, Micha(cid:32)lowski M. J. et al., 2012, ApJ, 755, 85
12 Nguyen H. T. et al., 2010, A&A, 518, L5
DunneL.,EalesS.,EdmundsM.,IvisonR.,AlexanderP., Oliver S. J. et al., 2012, MNRAS, 424, 1614
Clements D. L., 2000, MNRAS, 315, 115 Pascale E. et al., 2011, MNRAS, 415, 911
Eales S. et al., 2010, PASP, 122, 499 Perley D. A. et al., 2013, ApJ, 778, 128
ElliottJ.,GreinerJ.,KhochfarS.,SchadyP.,JohnsonJ.L., Perley D. A. et al., 2014, ApJ, submitted
Rau A., 2012, A&A, 539, A113 Pilbratt G. L. et al., 2010, A&A, 518, L1
FoxA.J.,LedouxC.,VreeswijkP.M.,SmetteA.,Jaunsen Poglitsch A. et al., 2010, A&A, 518, L2
A. O., 2008, A&A, 491, 189 Priddey R. S., Tanvir N. R., Levan A. J., Fruchter A. S.,
Frail D. A. et al., 2002, ApJ, 565, 829 Kouveliotou C., Smith I. A., Wijers R. A. M. J., 2006,
Fruchter A. S. et al., 2006, Nature, 441, 463 MNRAS, 369, 1189
Fuller C. et al., 2014, MNRAS, 440, 1571 Rangwala N. et al., 2011, ApJ, 743, 94
Fynbo J. P. U. et al., 2009, ApJS, 185, 526 Rigby E. E. et al., 2011, MNRAS, 415, 2336
Galama T. J. et al., 1998, Nature, 395, 670 Robertson B. E., Ellis R. S., 2012, ApJ, 744, 95
Gehrels N. et al., 2004, ApJ, 611, 1005 Roseboom I. G. et al., 2010, MNRAS, 409, 48
Greiner J. et al., 2011, A&A, 526, A30 Rossi A. et al., 2012, A&A, 545, A77
Griffin M. J. et al., 2010, A&A, 518, L3 Salpeter E. E., 1955, ApJ, 121, 161
Hagen L. M. Z., 2014, GRB Coordinates Network, 15690, Salvaterra R. et al., 2009, Nature, 461, 1258
1 Schady P. et al., 2012, A&A, 537, A15
Hatsukade B., Ohta K., Endo A., Nakanishi K., Tamura Schady P. et al., 2014, A&A
Y., Hashimoto T., Kohno K., 2014, Nature, 510, 247 Schaefer B. E. et al., 1999, ApJL, 524, L103
HegerA.,FryerC.L.,WoosleyS.E.,LangerN.,Hartmann SilvaL.,GranatoG.L.,BressanA.,DaneseL.,1998,ApJ,
D. H., 2003, ApJ, 591, 288 509, 103
Hernquist L., Springel V., 2003, MNRAS, 341, 1253 Smith A. J. et al., 2012, MNRAS, 419, 377
Hjorth J., Gall C., Michal(cid:32)owski M. J., 2014, ApJL, 782, Smith D. J. B. et al., 2011, MNRAS, 416, 857
L23 Smith D. J. B. et al., 2013, MNRAS, 436, 2435
Hjorth J. et al., 2012, ApJ, 756, 187 Sonbas E. et al., 2011, GRB Coordinates Network, 12612,
Hjorth J. et al., 2003, Nature, 423, 847 1
Hunt L. K. et al., 2014, A&A, 565, A112 Sparre M. et al., 2011, GRB Coordinates Network, 11607,
Ibar Edo; Ivison R. J. et al., 2010, MNRAS, 409, 38 1
Jakobsson P. et al., 2012, ApJ, 752, 62 Swinbank A. M. et al., 2014, MNRAS, 438, 1267
Jakobsson P. et al., 2006, A&A, 447, 897 Symeonidis M. et al., 2014, MNRAS, 443, L124
Kennicutt R. C., 1998, Ann. Rev. Astronomy Astrophys., Symeonidis M. et al., 2013, MNRAS, 431, 2317
36, 189 Tanvir N. R., Barnard V. E., Blain A. W., Fruchter A. S.,
Kistler M. D., Yu¨ksel H., Beacom J. F., Hopkins A. M., Kouveliotou C., Natarajan P., Ramirez-Ruiz E., et al.,
Wyithe J. S. B., 2009, ApJL, 705, L104 2004, MNRAS, 352, 1073
Klaas U., Haas M., Heinrichsen I., Schulz B., 1997, A&A, Tanvir N. R. et al., 2009, Nature, 461, 1254
325, L21 Tanvir N. R. et al., 2012, ApJ, 754, 46
Kov´acs A., Chapman S. C., Dowell C. D., Blain A. W., Thoene C. C., Perley D. A., Cooke J., Bloom J. S., Chen
Ivison R. J., Smail I., Phillips T. G., 2006, ApJ, 650, 592 H.-W.,BartonE.,2007,GRBCoordinatesNetwork,6741,
Kru¨hler T. et al., 2011, A&A, 534, A108 1
Lamb D. Q., Reichart D. E., 2000, ApJ, 536, 1 Tho¨ne C. C. et al., 2008, A&A, 489, 37
(cid:13)c 2014RAS,MNRAS000,1–11