published ApJ article - Harvard-Smithsonian Center for Astrophysics

17 Pages·2009·3.36 MB·English

TheAstrophysicalJournal,704:324–340,2009October10 doi:10.1088/0004-637X/704/1/324
(cid:2)C2009.TheAmericanAstronomicalSociety.Allrightsreserved.PrintedintheU.S.A.
LESSTHAN10PERCENTOFSTARFORMATIONINz ∼ 0.6MASSIVEGALAXIESISTRIGGEREDBY
MAJORINTERACTIONS
AdayR.Robaina1,EricF.Bell1,RosalindE.Skelton1,DanielH.McIntosh2,3,RachelS.Somerville4,
XianzhongZheng5,Hans-WalterRix1,DavidBacon6,MichaelBalogh7,FabioD.Barazza8,MarcoBarden9,
AsmusBo¨hm10,JohnA.R.Caldwell11,AnnaGallazzi1,MeghanE.Gray12,BorisHa¨ussler12,CatherineHeymans13,
KnudJahnke1,ShardhaJogee14,EelcovanKampen9,15,KyleLane12,KlausMeisenheimer1,CaseyPapovich16,Chien
Y.Peng17,Sebastia´nF.Sa´nchez18,RaminSkibba1,AndyTaylor19,LutzWisotzki10,andChristianWolf20
1Max-Planck-Institutfu¨rAstronomie,Ko¨nigstuhl17,D-69117Heidelberg,Germany;[email protected]
2DepartmentofAstronomy,UniversityofMassachusetts,710NorthPleasantStreet,Amherst,MA01003,USA
3DepartmentofPhysics,UniversityofMissouri-KansasCity,KansasCity,MO64110,USA
4SpaceTelescopeScienceInstitute,3700SanMartinDr.,Baltimore,MD21218,USA
5PurpleMountainObservatory,ChineseAcademyofSciences,Nanjing210008,China
6InstituteofCosmologyandGravitation,UniversityofPortsmouth,HampshireTerrace,PortsmouthPO12EG,UK
7DepartmentofPhysicsandAstronomy,UniversityofWaterloo,Waterloo,OntarioN2L3G1,Canada
8Laboratoired’Astrophysique,E´colePolytechniqueFe´de´raledeLausanne(EPFL),Observatoire,CH-1290Sauverny,Switzerland
9InstituteforAstro-andParticlePhysics,UniversityofInnsbruck,Technikerstr.25/8,A-6020Innsbruck,Austria
10AstrophysikalischesInstitutPotsdam,AnderSternwarte16,D-14482Potsdam,Germany
11UniversityofTexas,McDonaldObservatory,FortDavis,TX79734,USA
12SchoolofPhysicsandAstronomy,UniversityofNottingham,NottinghamNG72RD,UK
13DepartmentofPhysicsandAstronomy,UniversityofBritishColumbia,6224AgriculturalRoad,VancouverV6T1Z1,Canada
14DepartmentofAstronomy,UniversityofTexasatAustin,1UniversityStation,C1400Austin,TX78712-0259,USA
15EuropeanSouthernObservatory,Karl-Schwarzschild-Str.2,D-85748Garching,Germany
16StewardObservatory,TheUniversityofArizona,933NorthCherryAvenue,Tucson,AZ85721,USA
17NRCHerzbergInstituteofAstrophysics,5071WestSaanichRoad,VictoriaV9E2E7,Canada
18CentroHispanoAlemandeCalarAlto,C/JesusDurbanRemon2-2,E-04004Almeria,Spain
19TheScottishUniversitiesPhysicsAlliance(SUPA),InstituteforAstronomy,UniversityofEdinburgh,BlackfordHill,EdinburghEH93HJ,UK
20DepartmentofPhysics,DenysWilkinsonBldg.,UniversityofOxford,KebleRoad,OxfordOX13RH,UK
Received2009April20;accepted2009August26;published2009September22
ABSTRACT
Both observations and simulations show that major tidal interactions or mergers between gas-rich galaxies can
lead to intense bursts of star formation. Yet, the average enhancement in star formation rate (SFR) in major
mergers and the contribution of such events to the cosmic SFR are not well estimated. Here we use photometric
redshifts,stellarmasses,andUVSFRsfromCOMBO-17,24μmSFRsfromSpitzer,andmorphologiesfromtwo
deep Hubble Space Telescope (HST) cosmological survey fields (ECDFS/GEMS and A901/STAGES) to study
the enhancement in SFR as a function of projected galaxy separation. We apply two-point projected correlation
function techniques, which we augment with morphologically selected very close pairs (separation <2(cid:4)(cid:4)) and
merger remnants from the HST imaging. Our analysis confirms that the most intensely star-forming systems are
indeed interacting or merging. Yet, for massive (M∗ (cid:2) 1010M(cid:6)) star-forming galaxies at 0.4 < z < 0.8, we
find that the SFRs of galaxies undergoing a major interaction (mass ratios (cid:3)1:4 and separations (cid:3)40 kpc) are
only 1.80 ± 0.30 times higher than the SFRs of non-interacting galaxies when averaged over all interactions
and all stages of the interaction, in good agreement with other observational works. Our results also agree with
hydrodynamicalsimulationsofgalaxyinteractions,whichproducesomemergerswithlargeburstsofstarformation
on ∼100 Myr timescales, but only a modest SFR enhancement when averaged over the entire merger timescale.
Wedemonstratethattheseresultsimplythatonly(cid:4)10%ofstarformationat0.4(cid:3)z(cid:3)0.8istriggereddirectlyby
majormergersandinteractions;theseeventsarenotimportantfactorsinthebuild-upofstellarmasssincez = 1.
Key words: galaxies: evolution – galaxies: general – galaxies: interactions – galaxies: starburst – galaxies:
statistics–infrared:galaxies
Online-onlymaterial:colorfigure
intensity star formation events at low redshifts, are almost in-
1.INTRODUCTION
variablyhostedbymerginggalaxies(Sandersetal.1988).Fora
Observationalevidencefromavarietyofanglesindicatesthat numberofapplications,thequantityofinterestistheaverageen-
galaxyinteractionsandmergersofgalaxiescanleadtodramati- hancementinstarformation(SF)triggeredbymerging(ensem-
callyenhancedstarformation(Sandersetal.1988;Bartonetal. bleaverageoverthepopulationofmajormergers/interactions,
2000, 2007; Lambas et al. 2003). This appears to hold true at orequivalently,temporalaverageovermajormergereventsdur-
all redshifts where one can recognize mergers through galaxy ing a merger lifetime), not the high-intensity tail (e.g., Barton
morphologies (z (cid:4) 1 with rest-frame optical morphologies; etal.2000;Lambasetal.2003;Herna´ndez-Toledoetal.2005;
Melbourne et al. 2005; Hammer et al. 2005; 1 (cid:4) z (cid:4) 3 using Linetal.2007;Lietal.2008;Jogeeetal.2009).Bartonetal.
less certain UV morphologies; Chapman et al. 2004). Ultra- (2007) carefully quantified the star formation rate (SFR) en-
luminousinfraredgalaxies(ULIRGs),representingthehighest- hancementinmergersinlow-masshalosatlowredshift,using
324
No.1,2009 GALAXYINTERACTIONSANDSTARFORMATION 325
the Two-Degree Field Galaxy Redshift Survey (Colless et al. Therefore, to constrain galaxy evolution models and to
2001). They found that roughly 1/4 of galaxies in close pairs understand the physical processes responsible for the main
(separated by <50 kpc) in low-mass halos with M < −19 mode of star formation at z < 1, it is of interest to determine
bJ
haveSFRenhancementsofafactorof5ormore.21 observationallythetypicalenhancement22inSFRaveragedover
Ithasalsobeennotedthatthestrongdecreaseofthecosmic thedurationoftheentiremajor(stellar-massratiobetween1:1
SFRdensitybetweenz = 1andz = 0(e.g.,Lillyetal.1996; and1:4)galaxymergerorinteractionandtoconstraintheoverall
Madauetal.1996;Hopkins2004;LeFlochetal.2005)wasnot fractionofSFtriggeredbymergers/interactionsatintermediate
dissimilarfromtherelativelyrapiddropinmergerrateinferred redshift.Inacompanionpaper(Jogeeetal.2009)wefocuson
(at that time) from close pairs and morphologically selected the rate of merging and also present a preliminary exploration
mergers(LeFe`vreetal.2000).Ifmuchofthestarformationat oftheaveragechangeintheSFRcausedbylate-stagemajorand
z > 0.5 were triggered by merging, the apparent similarity in minormerging(seealsoKavirajetal.2009),findinganaverage
evolution between SFR and merger rate would be a natural mildenhancementwithintherestrictionsimposedbythesample
consequence. More recently, studies of the fraction of star size.Inthispaper,wepresentastatisticallyrobustanalysisofthe
formationinmorphologicallyselectedinteractingandmerging propertiesofstar-forminggalaxiesat0.4 < z < 0.8including
galaxiesatintermediateredshiftsz<1havedemonstratedthat, allrelevantmergerphasesandaimedatprovidingasatisfactory
infact,thebulkofstarformationisinquiescentlystar-forming answertotwokeyquestions.Whatistheaverageenhancement
disk-dominatedgalaxies(Hammeretal.2005;Wolfetal.2005; inSFRasafunctionofgalaxypairseparationcomparedtotheir
Belletal.2005;Jogeeetal.2009). SFR before the interaction? What fraction of star formation is
Similarly, it has long been argued that early-type (elliptical directlytriggeredbymajormergersandinteractions?
andlenticular)galaxiesareanaturaloutcomeofgalaxymergers There are a number of conceptual and practical challenges
(e.g., Toomre & Toomre 1972; Schweizer & Seitzer 1992). In in such an experiment. Enhancements in SFR produce both
any hierarchical cosmogony, mergers are expected to play a a boost in luminosity, but also increase dust content and
largerole;awiderangeofwork—observationsoftheincreasing extinction.Ataminimum,onethereforeneedsdust-insensitive
number density of non-star-forming early-type galaxies from SFR indicators. In addition, simulations have indicated that
z=1tothepresent(Belletal.2004;Brownetal.2007;Faber SF can be enhanced at almost all phases of an interaction
etal.2007),thekinematicandstellarpopulationsoflocalearly- from first passage through to after coalescence (e.g., Barnes
typegalaxies(Trageretal.2000;Emsellemetal.2004),orthe &Hernquist1996;diMatteoetal.2007);althoughclosepairs
jointevolutionofthestellarmassfunctionandSFRsofgalaxies will inevitably include some fraction of galaxies before first
(Bell et al. 2007; Walcher et al. 2008; Pe´rez-Gonza´lez et al. pass and galaxies with unbound orbits. Therefore, an analysis
2008)—hasgivensupporttothenotionthatatleastsomeofthe needs to include both close pairs of galaxies (those before
early-type galaxies assembled at z < 1 have done so through coalescence)andmorphologicallyclassifiedmergers(primarily
galaxymerging.Insuchapicture,theaverageSFRenhancement thosenearoraftercoalescence).Morphologicalclassificationis
frommergingisofinterestforinterpretingtheSFandchemical notastraightforwardart(seeJogeeetal.2009,foracomparison
enrichmenthistoryofearly-typegalaxies,inasmuchasitgives between automated classifications and visual morphologies),
an idea of what kind of fraction of stars in present-day early- even in ideal cases (Lisker 2008). Finally, galaxy mergers
typegalaxieswecanexpecttohaveformedintheburstmode, are rare and short-lived, necessitating large surveys to yield
andwhatfractionwecanexpecttohaveformedinaquiescent substantivesamplesofmergers.
modeintheprogenitorgalaxies. In this work, we address these challenges as far as possible
DirectobservationalconstraintsontheenhancementinSFR (seealsoLinetal.2007andLietal.2008).Weuseestimatesof
causedbymergingprovideanimportantcalibrationformodel- redshift and stellar mass from the COMBO-17 survey (Wolf
ingtriggeredstarformationincosmologicallymotivatedgalaxy et al. 2003; Borch et al. 2006) to define and characterize
formation models. Hydrodynamic simulations of interacting the sample. Stellar mass selection should limit the effect
galaxiesinwhichgasandstarformationareexplicitlymodeled of enhanced star formation and dust content on the sample
have demonstrated that torques resulting from the merger can definition. We use SFR indicators that are constructed to be
efficientlystripgasofitsangularmomentum,drivingittohigh dustextinctioninsensitive,bycombiningultraviolet(UV;direct,
densitiesandleadingtosignificantenhancementinstarforma- unobscured light from young stars) and infrared (IR; thermal
tion(e.g.,Barnes&Hernquist1996;Mihos&Hernquist1996; emissionfromheateddust,poweredprimarilybyabsorptionof
Cox2004;Coxetal.2006a,2008;diMatteoetal.2007).How- UVlightfromyoungstars)radiation(Belletal.2005).Finally,
ever,state-of-the-artcosmologicalsimulationslackthedynamic we study a very well-characterized sample of galaxy pairs at
rangetoaccuratelysimulatetheinternalstructureofgalaxiesin 0.4 < z < 0.8usingweightedprojectedtwo-pointcorrelation
significant volumes, so estimates of the global implications of functions (Skibba et al. 2006; Li et al. 2008), supplementing
merger-drivenstarformationenhancementhavehadtorelyon them at very small separations (cid:4)15 kpc with very close pairs
semianalytic calculations (e.g., Somerville et al. 2001; Baugh or merger remnants morphologically selected from two wide
etal.2005;Somervilleetal.2008).Furthermore,astheprogeni- HST mosaics, GEMS (Rix et al. 2004) and STAGES (Gray
torpropertiesplayakeyroleinthesimulatedSFRenhancements et al. 2009), in an attempt to account for all stages of galaxy
(e.g., di Matteo et al. 2007; Cox et al. 2008), inaccurate pro- interactions.
genitor property values (e.g., incorrect gas fraction or internal Theplanofthispaperisasfollows.InSection2,wediscuss
structure) will lead to incorrect estimates for the average frac- the data and the methods used to estimate the stellar masses
tionofSFinmergerseveniftheSFineachindividualmerger and the SFRs. In Section 3, we describe the sample selection
weremodeledperfectly. andthemethodusedfortheanalysis.InSection4,wepresent
21 Thiscorrespondsroughlytoamasscutof5×109M(cid:6),assumingastellar 22 WhenwerefertoSFRenhancement,wedefinethisastheratioofSFRin
M/LbJ ∼1,appropriateforastar-formingbluegalaxywithaChabrier(2003) somesubsample(e.g.,closepairs)totheaverageSFRofallsystemsinthat
stellarIMF. massbin.
326 ROBAINAETAL. Vol.704
our estimates of the enhancement in SFR as a function of 2.2.GEMSandSTAGESHSTImagingData
projected separation. In Section 5, we compare with previous
F606W (V-band) imaging from the GEMS and STAGES
observations, constrain the fraction of SF triggered by major
surveys provides 0(cid:4).(cid:4)1 resolution images for our sample of
mergers and interactions at 0.4 < z < 0.8, and compare
COMBO-17galaxies.UsingtheAdvancedCameraforSurveys
with simulations of galaxy merging. Finally, in Section 6, we
(ACS; Ford et al. 2003) on board the Hubble Space Telescope
summarize the main findings of this paper. All the projected
(HST), areas of ∼30(cid:4) × 30(cid:4) in each of the ECDFS and the
distancesbetweenthepairsusedhereareproperdistances.We
assumeH0 =70kms−1Mpc−1,ΩΛ0 =0.7,andΩm0 =0.3. dAe9te0c1t/io9n02tofiealdlhimaviteinbgeemnasgunrviteuydeedotof amdAeBp(tFh6a0ll6oWwi)ng=ga2la8x.y5
lim
(Rixetal.2004;Grayetal.2009;Caldwelletal.2008).These
2.THEDATA imagingdataarelaterusedtovisuallyclassifygalaxies,allowing
very close pairs (separations <2(cid:4)(cid:4)) and merger remnants to be
2.1.COMBO-17:RedshiftsandStellarMasses
includedinthisanalysis.WechoosenottouseF850LPHSTdata
COMBO-17 has to date fully surveyed and analyzed three availablefortheGEMSsurveyinordertobeconsistentinour
fields to deep limits in 5 broad and 12 medium passbands classification between the two fields (only F606W is available
(Extended Chandra Deep Field South (ECDFS), A901/2 and fromSTAGES).
S11;seeWolfetal.2003andBorchetal.2006).Usinggalaxy,
star, and quasar template spectra, objects are classified and 2.3.MIPS24μm,TotalInfraredEmissionandStar
redshiftsassignedfor∼99%oftheobjectstoalimitofm ∼ FormationRates
R
23.5(Wolfetal.2004).Thephotometricredshifterrorscanbe
The IR observatory Spitzer has surveyed two of the
describedas COMBO-17fields:a1◦×0◦.5scanoftheECDFS(MIPSGTO),
σ and a similarly sized field around the Abell 901/902 galaxy
z ∼0.007×[1+100.8(mR−21.6)]1/2, (1) cluster(MIPSGO-3294:PIBell).Thefinalimageshaveapixel
1+z
scaleof1(cid:4).(cid:4)25pixel−1andanimagepointspreadfunction(PSF)
FWHMof(cid:7)6(cid:4)(cid:4).Sourcedetectionandphotometryaredescribed
and rest-frame colors and absolute magnitudes are accurate to
∼0.1mag(accountingfordistanceandk-correctionuncertain- indepthinPapovichetal.(2004)andcatalogmatchinginBell
ties).Theastrometryisaccurateto∼0(cid:4).(cid:4)1andtheaverageseeing et al. (2007).23 Based on those works, we estimate that our
is0(cid:4).(cid:4)7.ItisworthnotingthatEquation(1)leadstotypicalred- sourcedetectionis80%completeatthe5σ limitof83 μJyin
shifterrorsofσ (cid:7)0.01forbright(m <21)andσ (cid:7)0.04for the 24 μm data in the ECDFS for a total exposure of ∼1400
z R z s pixel−1. The A901/902 field has similar exposure time, but
faint(21<m <23.5)galaxiesinthe0.4<z<0.8interval.
R
owing to higher (primarily zodiacal) background the 5σ limit
ThestellarmasseswereestimatedinCOMBO-17byBorch
(80%completeness)is97μJy,withlowercompletenessof50%
et al. (2006) using the 17-passband photometry in conjunc-
at83μJy.Weusebothcatalogstoalimitof83μJy.
tion with a non-evolving template library derived using the
Pe´gasestellarpopulationmodel(seeFioc&Rocca-Volmerange Toincludebothobscuredandunobscuredstarformationinto
theestimateoftheSFRofgalaxiesinoursample,wecombine
1997, 1999) and a Kroupa et al. (1993) initial mass function
UV emission with an estimate of the total IR luminosity in
(IMF). Note that the results assuming a Kroupa (2001) or a
concert. As the total thermal IR flux in the 8–1000 μm range
Chabrier (2003) IMF yield similar stellar masses to within
∼10%.Thereddesttemplateshavesmoothlyvaryingexponen- is observationally inaccessible for almost all galaxies in our
sample,wehaveinsteadestimatedtotalIRluminosityfromthe
tiallydecliningstarformationepisodes,intermediatetemplates
observed 24 μm flux, corresponding to rest-frame 13–17 μm
haveacontributionfromalow-levelconstantlevelofstarfor-
emission at the redshifts of interest z = 0.4–0.8. For this
mation, while the bluer templates have a recent burst of star
exercise, we adopt a Sbc template from the Devriendt et al.
formationsuperimposed.
(1999)SEDlibrary(Zhengetal.2007b;Belletal.2007).The
The masses are consistent with those using M/L estimates
resulting IR luminosity is accurate to a factor of (cid:4)2: local
basedonasinglecolor(e.g.,Belletal.2003).Randomstellar
mass errors are <0.3 dex on a galaxy-by-galaxy basis, and galaxies with IR luminosities in excess of > 1010L(cid:6) show
a tight correlation between rest-frame 12–15 μm luminosity
systematicerrorsinthestellarmasseswerearguedtobeatthe
and totalIRluminosity (Spinoglio etal.1995; Chary &Elbaz
0.1 dex level (see Borch et al. 2006, for more details). Bell &
2001; Roussel et al. 2001; Papovich & Bell 2002) with a
deJong(2001)arguedthatgalaxieswithlargeburstsofrecent
scatter of ∼0.15 dex. Furthermore, Zheng et al. (2007b) have
starformationcouldproducestellarM/Lvaluesatagivencolor
thatarelowerbyupto0.5dex;thisuncertaintyismorerelevant stackedluminous(LTIR (cid:5)1011L(cid:6))z∼0.7galaxiesat70μm
and 160 μm, finding that their average spectrum is in good
in this work than is often the case. While this will inevitably
agreementwiththeSbctemplatefromDevriendtetal.(1999),
remainanuncertaintyhere,wenotethattheBorchetal.(2006)
validating at least on average our choice of IR SED used for
templatesdoincludeburstsexplicitly,thuscompensatingforthe
extrapolationofthetotalIRluminosity.
worstoftheuncertaintiesintroducedbyburstingstarformation
We estimate the SFR by using both directly observed UV-
histories. In Section 4.1.2, we will explicitly study the impact
lightfrommassivestarsanddust-obscuredUV-lightmeasured
thatsuchuncertaintieshaveonourresults.
fromthemid-infrared.AsinBelletal.(2005)weestimatethe
In what follows, we use COMBO-17 data for two fields:
SFRψbymeansofacalibrationderivedfromPe´gasesynthetic
the ECDFS and Abell 901/902 fields, because of their com-
plementarydata:deepHST/ACSimagingfromtheGEMSand
STAGES projects, respectively (allowing an investigation of 23 Inthispaper,weareinterestedinSFRenhancementsinclosepairsof
morphologicallyselectedmergerremnantsandveryclosepairs), galaxies,wheretheclosestpairsmayfallwithinasingleSpitzer/MIPSPSF.
Accordingly,inthisworkwechoosetoexplorethetotalSFRinthepair
anddeep24μmimagingfromtheMIPSinstrumentsonboard
(avoidsdeblendinguncertainties)ratherthantheindividualSFRoccurringin
Spitzer,requiredtomeasureobscuredSF. bothgalaxies.
No.1,2009 GALAXYINTERACTIONSANDSTARFORMATION 327
modelsassuminga100Myroldstellarpopulationwithconstant
SFRandaChabrier(2003)IMF
ψ/(M(cid:6) yr−1)=9.8×10−11×(LTIR+2.2LUV). (2)
Here, L is the total IR luminosity and L = 1.5νl
TIR UV ν,2800
is a rough estimate of the total integrated 1216–3000 Å UV
luminosity. This UV luminosity has been derived from the
2800 Å rest-frame luminosity from COMBO-17 l . The
ν,2800
factorof1.5inthe2800Å-to-totalUVconversionaccountsfor
UV spectral shape of a 100 Myr old population with constant
SFR, and the UV flux is multiplied by 2.2 to account for the
light emitted longwards of 3000 Å and shortwards of 1216 Å
bytheunobscuredstarsbelongingtotheyoungpopulation.
For all galaxies detected above the 83 μJy limit, we have
Figure1.Stellarmassvs.colordistributionofCOMBO-17selectedgalaxiesin
used the IR and UV to estimate the total SFR. For galaxies theECDFSandA901/2fieldwith0.4<z<0.8.Theverticallineshowsthe
undetected at 24 μm, or detected at less than 83 μJy, we use masslimitM(cid:5)>1010M(cid:6)usedtoselectoursample.Thismassselectedsample
insteadUV-onlySFRestimates. iscompleteexceptforredsequencegalaxiesatz>0.6.Thebluelineshows
thecutusedtoseparateredsequenceandbluecloudgalaxies.Redsymbols
2.3.1.IREmissionfromAGN-heatedDust denote24μmdetectedgalaxieswith>83μJy.
(Acolorversionofthisfigureisavailableintheonlinejournal.)
Possiblecontaminationofmid-IR-derivedSFRsfromactive
galactic nucleus (AGN)-heated dust is often addressed by
estimatingthefractionofstarformationheldinX-raydetected
sources. In our case <15% of the star-forming galaxy sample Despitethevariouslevelsofuncertainty,takingthedifferent
were detected in X-rays, in good agreement with the results lines of evidence together demonstrates that (cid:4)30% of the IR
foundby,i.e.,Silvaetal.(2004)orBelletal.(2005). luminosityinoursamplecomesfromsystemsthatmayhostan
Yet, there are two limitations of this estimate. First, this AGN, and that it is likely that <10% of the IR luminosity of
does not account for any contribution from X-ray undetected our sample is powered by accretion onto supermassive black
Compton-thick AGN, which could drive up the expected con- holes.Giventheotheruncertaintiesinouranalysis,wechoose
tribution from AGN in our sample. For example, applying an toneglectthissourceoferrorinwhatfollows.
m = 24 cut to the sample of Alonso-Herrero et al. (2006),
R
weestimatethefractionofX-rayundetectedAGNtobe∼30%,
whileRisalitietal.(1999)find∼50%oflocalAGNtobeComp- 3.SAMPLESELECTIONANDMETHOD
tonthick.Onthisbasis,itisconceivablethatupto30%of24μm
luminosityisfromgalaxieswithAGN.24 ThegoalofthispaperistoexploretheSFRinmajormergers
Second,eveningalaxieswithAGN,notalloftheIRemission between massive galaxies, from the pre-merger interaction to
will come from the AGN. Although the data do not currently after the coalescence of the nuclei. We chose a stellar mass-
existstoanswerthisquestionconclusively,itispossibletomake limited sample with M(cid:5) (cid:2) 1010M(cid:6) in the redshift slice of
aroughestimateoftheeffect.Inordertoestimatethefraction 0.4 < z (cid:3) 0.8 (see Figure 1). This roughly corresponds to
of mid-IR light that comes from the AGN (as opposed to star MV =−18.7forgalaxiesintheredsequenceandMV =−20.1
formationinthehost),wehavemadeuseoftheresultsofRamos for blue objects. We only included galaxies that fall into the
Almeidaetal.(2007),whoattemptedtostructurallydecompose footprintofboththeACSsurveysGEMSandSTAGESandof
mid-infrared imaging from Infrared Space Observatory for a existingSpitzerdata.Thesecriteriaresultedinafinalsampleof
sample of both Seyfert 1 and 2 AGN in the local universe, 2551galaxies.
some of which are very highly obscured in X-rays. Analyzing Given the flux limit mR (cid:4) 23.5 for which COMBO-17
the results in Tables 2 and 3 of Ramos Almeida et al. (2007), has reasonably complete redshifts (Wolf et al. 2004) we are
wehavefoundthatonlyasmallfractionoftheIRradiationat completeforM∗ >1010M(cid:6)bluecloudgalaxiesovertheentire
∼10μm(inthispaperweworkatrestframe13–17μm)comes redshift range 0.4 < z < 0.8. For red sequence galaxies,
fromthecentralpartsofthegalaxiesintheSeyfert2population, the sample becomes somewhat incomplete at z > 0.6, and
findingatotalcontributionof at z = 0.8, the limit is closer to 2 × 1010M(cid:6). We chose
to adopt a limit of 1010M(cid:6) in what follows, despite some
FAGN mild incompleteness in the red sequence, for two reasons.
FIRtotal =0.26±0.02. (3) First,adoptingacutof2×1010M(cid:6) acrossthewholeredshift
IR range reduces the sample size by a factor of 30%, leaving too
Thisresultshouldbeviewedasindicativeonly:obviously,the small a sample for the proposed experiment. Second, the vast
systems being studied will be different in detail from those in majority of the star-forming galaxies are blue cloud galaxies
oursample.Furthermore,the10μmluminositiesofthenuclei (83% of the star formation is occurring in blue galaxies),
willbepreferentiallyaffectedbysilicateabsorption,makingit makingthemodestincompletenessintheredsequenceofminor
possiblethatourvalueofFAGN/Ftotalisalowerlimit. importance.
IR IR
Later, we will use a subsample composed of star-forming
24 AlthoughnotethatinarecentinvestigationofX-rayundetectedIR-bright galaxies.Wewillreferto“starformers”asgalaxiesdefinedby
galaxiesintheCDFS,Lehmeretal.(2008)foundthatradio-derived(1.4GHz) havingeitherblueopticalcolorsorhavingbeendetectedinthe
SFRsagreewiththeUV+IR-derivedones.Thisimpliesthattherelative
MIPS24μmband.Weselectopticallybluegalaxiesadoptinga
strengthofanyAGNcomponentisnotdominantwhencomparedtothehost
galaxy. stellarmass-dependentcutinrest-frameU−Vcolor,following
328 ROBAINAETAL. Vol.704
Belletal.(2007;seeourFigure1)25: Thefirstcriterionisthatthestellar-massratiofallsbetween1:1
and1:4.W√efurtheronlyallowamaximumredshiftdifference
U −V (cid:5)1.06−0.352z+0.227(log10M∗−10). Δz=Δ= 2σz,whereσzistheerrorinredshiftoftheprimary
galaxy(seeEquation(1)),and,dependingonthecase,eitherthe
Weincludeallobjectsdetectedabovethe24μmlimitof83μJy primaryorbothgalaxiesinthepairhavetobestarformers(see
asstarforming. Section4).
Inordertotrackstarformationinveryclosepairs(<2(cid:4)(cid:4) and We can then study the possible enhancement of (specific)
hence unresolved by the ground-based COMBO-17 data) and SFRbymeansofaprojectedmarked(orweighted)correlation
merger remnants, we include only merging systems (from the function,whichcanbedefined
ACSdata)withM∗ >2×1010M(cid:6):i.e.,theminimumpossible
massforamergerbetweentwogalaxiesinoursample. E(r )= 1+W(rp)/Δ, (5)
p 1+w(r )/Δ
p
3.1.ProjectedCorrelationFunction
whereW(r )=Δ(PP/PP −1)and
p R
The correlation function formalism is a convenient and
(cid:7)
powerfultooltocharacterizepopulationsofgalaxypairs(e.g.,
PP(r )= P D ,
Davis & Peebles 1983; Beisbart & Kerscher 2000). Here, we P ij ij
useweightedprojectedtwo-pointcorrelationfunctionsbecause (cid:7)ij
redshift uncertainties (1%–3%) from COMBO-17 translate to PP (r )= P R .
R P ij ij
line-of-sight distance errors of ∼100 Mpc, necessitating the
ij
useofprojectedcorrelationfunctionstoexploretheproperties
P isthemark(orweight).WeadopttwodifferentweightsP
of close physical pairs of galaxies (Bell et al. 2006). For our ij ij
inwhatfollows,oneistheSFRofthepairofgalaxies
sampleathand,weestimatetheweighted(ormarked)two-point
correlationfunction(Boerneretal.1989;Beisbart&Kerscher P =S =SFR =SFR +SFR ,
ij ij ij i j
2000;Skibbaetal.2006;Skibba&Sheth2009),usingboththe
SFR and the specific SFR (SFR per unit stellar mass) as the andtheotheristhespecificSFRofthegalaxypair
weight.
Theprojectedcorrelationfunctionw(r )istheintegralalong SFR +SFR
P P =s =SpecificSFR = i j.
thelineofsightofthereal-spacecorrelationfunction ij ij ij M +M
(cid:5),i (cid:5),j
(cid:2)
∞ (cid:3)(cid:4) (cid:5) (cid:6) Then,theestimatorthatweuseforE(r )is
w(r )= ξ r2+π2 1/2 dπ, (4) p
p p
−∞ PP/DD
E(r )= , (6)
p (cid:11)P (cid:12)
wherer isthedistancebetweenthetwogalaxiesprojectedon ij
p
the plane of sky and π the line-of-sight separation. A simple
where (cid:11)P (cid:12) is the average value of the weight used (SFR,
estimator for this unweighted correlation function is w(r ) = ij
p or specific SFR) across the sample. This normalization is the
Δ(DD/RR−1),whereΔisthepathlengthbeingintegratedover,
averagevalueofpairSFRorSSFRfortheactualpairsamples
DD(r ) is the histogram of separations between real galaxies
P used in this analysis, out to a projected separation of 8 Mpc,
and RR(r ) is the histogram of separations between galaxies
P inordertoprobegalaxypairssamplingdifferentenvironments
in a randomly distributed catalog (this is the same estimator
tobuildarepresentativecosmic-averagedweight.TheSFRsor
usedinBelletal.2006).Basically,theaimistofindtheexcess
SSFRs of individual galaxies used to find the normalization
probability (compared to a random distribution) of finding a
are exactly the same as for the numerator, as described in
galaxy at a given distance of another galaxy. This estimator
Section 2.3. It is worth noting that with our definition of the
accomplishes that by subtracting the random probability of
enhancementgiveninEquation(5)therandomhistogramsRR
findingtwogalaxiesatagivenseparationfromtheprobability
and PP cancel in the process of obtaining the expression in
in the real data sample and normalizing to the probability in R
Equation (6), so they are not used in the computation of our
the random case. Other estimators (i.e., Δ[(DD −DR)/RR]
enhancement.
or Δ[(DD −2DR +RR)/RR]) for the two-point correlation
In the present work we perform two analyses: the cross-
functiongiveresultsdifferentby<5%(lessthanothersources
correlation of star-forming galaxies (as defined above) as pri-
ofuncertainty).Thus,
marygalaxieswithallgalaxiesassecondaries,andtheautocor-
(cid:7)
relationofstar-forminggalaxies.Wewillestimatetheerrorsin
DD(r )= D ,
P ij ourmarkbymeansofbootstrappingresampling.
ij
(cid:7)
3.2.VisualMorphologies
RR(r )= R ,
P ij
A particular challenge encountered when constructing a
ij
census of star formation in pairs and mergers is accounting
wherethesumisoverallnon-repeatedpairsinthesample,and for systems with separations of <2(cid:4)(cid:4) (which corresponds to
D (R )equals1onlyifthepair-selectioncriteriaaresatisfied
ij ij <15 kpc, the radius within which we can no longer separate
inthereal(random)galaxycatalog,andisequalto0otherwise.
two massive galaxies using COMBO-17; Bell et al. 2006). In
order to pick up the SF in all the stages of the interaction, we
25 Duetominormagnitudeandcolorcalibrationdifferencesbetweenthetwo
need to have an estimate of the SFR not only in galaxy pairs
fields,theredsequencecutisslightlyfielddependent,withtheinterceptat
1010M(cid:6)andz=0beingU−V =1.01and1.06fortheECDFSandthe withseparations>15kpcbutalsoinextremelyclosepairsand
A901/902fields. inrecentmergerremnants.Weconductourcensusofsuchclose
No.1,2009 GALAXYINTERACTIONSANDSTARFORMATION 329
physical pairs by including in the <15 kpc range sources that One practical issue is that of passband choice and shifting.
are not resolved by COMBO-17, but appear to be interacting We choose to classify the F606W images of the GEMS and
pairsormergerremnantsonthebasisofvisualclassificationof STAGES fields (in STAGES because that is the only available
the∼0(cid:4).(cid:4)1resolutionACSimages.Wetrytorecovervisuallyall HST passband and in GEMS for consistency and because
<15kpcseparationpairsoftwoM∗ > 1010M(cid:6) galaxieswith F606WhashigherS/NthattheF850LPdata).Thiscorresponds
a mass ratio between 1:1 and 1:4 missed by COMBO-17. In to rest-frame ∼430(330)nm at redshift 0.4(0.8). In previous
additiontothoseextremelyclosepairs,wealsoaccountforthe papers (Wolf et al. 2005; Bell et al. 2005; Jogee et al. 2009),
SF in recent merger remnants M∗ > 2×1010M(cid:6) (two times we have assessed whether the morphological census derived
theminimummassofagalaxyinthesampleandtheminimum fromGEMS/STAGESwouldchangesignificantlyifcarriedout
possiblemassofagalaxypairasdefinedbefore). data a factor of 5 deeper from the GOODS project (testing
sensitivity to surface brightness limits), or if carried out at
3.2.1.DiscussionofVisualClassifications F850LP (always rest-frame optical at these redshifts). We
foundthatthepopulationdoesnotshowsignificantlydifferent
Our goal is to include very close pairs or already-coalesced
morphologiesbetweenour(comparatively)shallowF606Wdata
major merger remnants into the census of “mergers” in order
andthedeeper/redderimagingdatafromGOODS(seeFigure5
to account for any SF triggered by the merger/interaction
process.26 Wedosoonthebasisofvisualclassificationofthe inJogeeetal.2009).
sample.Themotivationforvisualclassificationisapragmatic
3.2.2.Method
one:whileanumberofautomatedmorphologicalclassification
systemshavebeendevelopedinthelast15years(i.e.,Abraham An independent visual inspection of the galaxy sample has
et al. 1996; Conselice et al. 2003; Lotz et al. 2004, etc.), it beencarriedoutbyfourclassifiers,A.R.R.,E.F.B.,R.E.S.,and
seemsthatthesensitivityof theobservables used(asymmetry, D.H.M.,inordertoidentifymorphologicalsignaturesofmajor
clumpiness,Ginicoefficient,second-ordermomentofthe20% gravitationalinteractions.Eachclassifierassignedeveryoneof
brightest pixels) is insufficient for matching the performance the∼2500samplememberstooneofthethreefollowinggroups.
ofvisualclassificationincurrentintermediateredshiftgalaxies 1. Non-majorinteractions:thebulkofgalaxiesinthisbinshow
with the same level of precision that they display in the local no signatures of gravitational interactions. Asymmetric,
universe samples used for their calibration (Conselice et al. irregular galaxies with patchy star formation triggered by
2003;Lisker2008;Jogeeetal.2009). internal processes lie in this category. A small fraction of
Yet,thereisadegreeofsubjectivitytowhatonedeemstobe galaxiesinthisbinshowaclearlyrecognizablemorphology
a major merger remnant. Many factors shape the morphology (e.g.,spiralstructure)butalsosignaturesofaninteraction
ofagalaxymergerthatarebeyondthecontroloftheclassifier. (such as tidal tails, or warped, thick or lopsided disks)
Bulge-to-total(B/T)massratioshaveanstrongeffectonboth but have no clear interaction companion; note that these
the intensity of the SFR enhancement and the time at which objectscouldbeinteractingsystemswherethecompanion
theintensitypeakshalloccur(e.g.,Mihos&Hernquist1996). isnowreasonablydistantand/orfaintandmoredifficultto
Orbital parameters strongly shape the development of easily identify. The tidal enhancement of SF from such systems
recognizabletidaltailsandbridges(coplanarornot,retrograde will not be missed by putting them in this bin; rather, it
versusprograde,etc.).Priordustandgascontentoftheparent will be measured statistically and robustly from the two
galaxies (“dry” versus “wet” mergers) will make a difference point correlation function analysis. Minor mergers and
to the appearance of the final object during the coalescence. interactions(interactionswherethesecondaryisbelieved,
Furthermore, merging timescales will depend on whether the onthebasisofluminosityratio,tobelessthan1/4ofthe
galaxies are undergoing a first passage or are in the final massoftheprimary)alsobelongtothiscategory.
stages of the merger. Finally, there is a degeneracy between 2. Majorcloseinteractions:closepairsresolvedinHSTimag-
all these parameters and the relative masses of the galaxies ing but not in ground-based COMBO-17 data, consisting
undergoingtheinteraction,whichmakesdifficultinsomecases oftwogalaxieswithmassratiosbetween1:1and1:4based
to distinguish the morphological signatures of a major merger onrelativeluminosity,andclearsignaturesoftidalinterac-
fromthoseofaminormerger. tion such as tidal tails, bridges or common envelope (see
Someofthesefactors(e.g.,gasfraction,B/Tratios,etc.)will Figure2).Fromnowonweshallrefertoobjectsclassified
also affect the enhancement of the SFR during the interaction inthisgroupas“veryclosepairs.”
(e.g.,diMatteoetal.2007,2008;Coxetal.2008).Whilethere 3. Majormergerremnants:objectsthatarebelievedtobethe
isconsiderablemerger-to-mergerscatter,encountersoftwogas- coalesced product of a recent major merger between two
richdiskgalaxieswithparallelspinstendtodevelop,onaverage, individual galaxies. Signposts of major merger remnants
the strongest morphological features, but at the same time are include a highly disturbed “train wreck” morphology,
morelikelytothrowoutlargeamountsofcoldgasintidaltails, double nuclei of similar luminosity, tidal tails of similar
preventingthefunnelingofthisgastothecentralregions.Thus, length,orspheroidalremnantswithlarge-scaletidaldebris
samples selected to have the strongest morphological features (see Figure 3). Galaxies with clear signs of past merging
mayhaveanaverageSFRenhancementdifferentfromtheactual but a prominent disk (e.g., highly asymmetric spiral arms
meanenhancement.27 oronetidaltail)weredeemedtobeminormergerremnants
and were assigned into the Group 1. Naturally, there
26 Notethataconsistentcomparisonwiththeprojectedcorrelationfunction
is some uncertainty and subjectivity in the assignment
samplerequirestheinclusionofallnon-interactingpairsthatarephysically
associated(inthesamecluster,filament,etc.),areseentobecloseprojected of this class, in particular; such uncertainty is taken into
pairsonthesky,butmaybeseparatedbyasmuchasafewMpcalongtheline account in our analysis by the Monte Carlo sampling of
ofsight.
all four classifications in order to properly estimate the
27 Thisbiasmightalsobepresentinthecaseofstudieslookingforsignsof
dispersionintheopinionsoftheindividualclassifiers(see
interactionsinthehostgalaxiesofAGNs,attemptingtoassesswhetherthe
AGNactivityisprecededbyamerger. below).
330 ROBAINAETAL. Vol.704
Figure2.ObjectsclassifiedinGroup2:majorcloseinteractions.Thepresenceoftwogalaxiesandsignsofinteractionarerequired.Theclassifierbelievesthemass
ratioisbetween1:1and1:4.Atthisstageoftheinteraction,drymergersarestillrecognizableasseeninpanelsattopcenter,bottomcenterandbottomright.The
blackbaratthebottomofeverypanelshowsaproperdistanceof20kpcattheredshiftoftheobject.Someoftheobjectsclassifiedinthisgroupwerealsoseparated
astwogalaxiesintheground-basedcatalogandtreatedinconsequence.
We then assign the objects in the Groups 2 and 3 (very included in the two point correlation function analysis, and
close pairs with morphological signatures of interaction and any SF triggered by major interactions or early-stage major
merger remnants, respectively) to a small projected separation merging is accounted for by that method. As we have four
andtreateveryoneofthemasagalaxypairinordertocombine different classifications for every object (one given by each
themwiththecorrelationfunctionanalysisresultforpairswith human classifier), we randomly assign one of them, calculate
separations>2(cid:4)(cid:4).AllobjectsinGroup2(extremelyclosepairs the average value of the weight we are using and repeat the
with projected separations <15 kpc as measured by centroids processanumberoftimes.AsbydefinitionobjectsinGroups
in HST imaging) are assigned to a separation of 10 kpc and 2 and 3 are considered to be a galaxy pair by themselves, we
all objects in Group 3 (merger remnants) are assigned to a removeineveryMonteCarlorealizationtheobjectsassignedto
separation of 0 kpc. We have checked for duplicate pairs in those groups before we run the weighted correlation function.
both the visually selected sample and the COMBO-17 catalog This approach presents two clear advantages: (1) the resultant
inordertoavoidrepeatedpairs.GalaxiesinGroup1arealready bootstrappingerrornotonlyrepresentsthestatisticaldispersion
No.1,2009 GALAXYINTERACTIONSANDSTARFORMATION 331
Figure3.ObjectsclassifiedinGroup3:majormergerremnants.Theblackbaratthebottomofeverypanelshowsaproperdistanceof20kpcattheredshiftofthe
object.
butalsothedifferentcriteriaofthefourhumanclassifiers;and 1010M(cid:6) and the pair has a stellar-mass ratio between 1:1 and
(2) the morphology of every object is weighted with the four 1:4.Ourprimaryanalysisisbasedonamarkedcross-correlation
classifications given. This means that objects with discrepant between star-forming galaxies, as defined in Section 3, and
classifications are not just assigned to one category when we all galaxies in the sample. For morphologically selected very
calculate the SFR (or specific SFR) enhancement; rather, any close pairs or interactions (unresolved by COMBO-17), we
dispersion in classifications is naturally accounted for (e.g., also require them to be blue or detected by Spitzer to be
minor/major criteria). The numbers of such systems and their considered as part of the star-forming sample,28 with a mass
uncertainties,estimatedfromtheclassifier-to-classifierscatter, ofM∗ >2×1010M(cid:6).
aregiveninTable1. Weperformtwoanalysesinthispaper:thecross-correlation
of star formers as primary galaxies with all galaxies as secon-
4.RESULTS daries(ourdefaultcase),andtheautocorrelationofstar-forming
We are now in a position to quantify the triggering of star
formation in galaxy interactions and mergers in the redshift 28 Allgalaxies,irrespectiveoftheircolororIRflux,wereclassified;the
interval0.4<z<0.8,inthecaseswhereeachgalaxyhasM∗ > star-forminggalaxiesaresimplyasubsampleofthislargersample.
332 ROBAINAETAL. Vol.704
Figure4.PairspecificSFRenhancementasfunctionoftheprojectedseparationbetweentwogalaxies.Thetwosmallestradiibinsarederivedfrommorphologically
selectedveryclosepairs(shownwithrP ∼10kpc)andmergerremnants(shownwithrP =0);enhancementsatlargerradiiaredeterminedusingweightedtwo-point
correlationfunctions.Astatisticallysignificantenhancementispresentingalaxypairsandmergersbelow40kpcinboththecross–correlationbetweenstar-forming
galaxiesasprimariesandallgalaxiesassecondaries(blackfilledsymbols)andtheautocorrelationofstar–forminggalaxies(emptydiamonds).Errorbarshavebeen
calculatedbybootstrapping.
Table1 starformation,ratherthanbeingduetoacorrelationwithsome
ResultsfromtheMorphologicalClassification otherunidentifiedquantity:(1)itiswellknownfromsimulations
(Mihos&Hernquist1996;diMatteoetal.2007;Coxetal.2008)
LowerMassLimit SampleSize Group1 Group2 Group3
thataburstofstarformationisexpectedinthecollisionsofgas-
1010M(cid:6) 2551 2380±37±49 106±7±10 72±7±8
rich galaxies; and (2) the observed effect is in the opposite
2×1010M(cid:6) 1749 1640±32±40 69±6±8 44±5±7
sense of the usual SFR–density relation (e.g., Balogh et al.
2002), which says that galaxies in dense environments (where
Notes. Galaxy and interaction sample. Group 1: isolated objects and minor
interactions.Group2:extremelyclosepairs(rP <15kpc).Group3:merger preferentially close galaxy pairs tend to be found, as shown
remnants.Thefirsterrorbarrepresentsclassifier-to-classifierscatter,whilethe inBartonetal.2007)have,onaverage,weakerstarformation
secondonerepresentsPoissonnoise. activitythangalaxiesinlessdenseregions.
Evenwhenweconsiderourmorphologicalclassificationand
galaxies. While the first analysis is a rather more direct attack furtherMonteCarloresamplingmethodtobeveryrobust,po-
onthequestionofinterest,weshowresultsfromtheautocorre- tentialclassificationerrorscouldactintwodifferentdirections.
lationofstar-forminggalaxiestoillustratetheeffectsofmaking Interactingsystemsmisidentifiedasnon-interactingwillbedi-
differentsamplechoicesonthefinalresults. lutedintothebackgroundstarformationassinglegalaxiescon-
tributingtopairsatrandomseparations.WhilethisSFshouldbe
4.1.EnhancementintheStarFormationActivity
losttotheinteractingbin,theeffectontheaverageSFRwould
Our main results are shown in Figure 4, which shows the beminimal.Ontheotherhand,isolatedgalaxiesmisidentifiedas
enhancementofthespecificstarformationrate(SSFR)inpairs interactingsystemsbecauseofinternalinstabilitiesorstochastic
as a function of their projected separation. As explained in starformationwouldacttoreducetheenhancement.
Section 2.3 we use UV+IR SFRs for the objects detected in As mentioned before, the SFR for the objects undetected at
24μmandonlyUVSFRsforthoseundetected.Forthewhole 24μmhasbeencalculatedbasedonlyontheUV.Inthe24μm
sample, 38% of the galaxies where detected by Spitzer above detectedobjects,wehavefoundnocleartrendinboththeUV
the 83 μm limit, while if we restrict to the Groups 2 and 3 versusUV+TIRSFRsandintheTIR/TUVversusopticaldust
in our morphological classification we find a detected fraction attenuationbutfoundinsteadaconstantcorrectionfactorwitha
of 60%. Figure 4 shows a clear enhancement in the SSFR for largescatter(4.1 ± 2.4asestimatedfromtherelationbetween
projectedpairseparationsr <40kpc.Itcouldbearguedthat TIR/TUV versus optical attenuation.) We have checked the
P
the SSFR is a better measure of the SF enhancement than the effectsofsuchadust-correctionoftheUV-onlySFRs:theresults
SFR-weightedestimator,becausethestrongscalingofSFRwith differinallbinsby<10%,comparabletoorsmallerthanother
galaxy mass is factored out. The figure shows both the cross- sourcesofsystematicuncertainty.
correlationbetweenstar-formingprimariesandallsecondaries Yet, in order to understand the degree of obscuration in
(SF–All,solidline)andthestar-forminggalaxyautocorrelation galaxy interactions we have repeated our analysis including
(SF–SF,dottedline).Thetwobinsatr (cid:3)15kpcarecalculated only UV–derived SFRs, this is, excluding the TIR compo-
p
from morphologically selected very close pairs (r = 10 kpc) nent in Equation (2) for 24 μm detections. The result of this
p
andmergerremnants(r =0).AlltheerrorsinE havebeen analysis is shown in Figure 5. The enhancement in the un-
P SSFR
computedbybootstrapresampling.Thisapproachallowsusto obscured SSFR measured for close pairs (r < 40 kpc) in
P
treat both the morphologically selected objects and the galaxy thiscaseisdramaticallysmallerthantheenhancementinclud-
pairs exactly in the same way, having as a result a coherent ingthedust-obscured(IR-derived)SFR.Thisismoreapparent
displayoftheerrorbars. in the very close pairs and merger remnants, where the ex-
TherearetworeasonswhythisexcessinE inclosepairs cessintheSSFRevendisappearscompletelyinthecaseofthe
SSFR
andremnantsislikelyasignthatinteractionsinduceadditional SF–SFautocorrelation(E(r <15kpc)(cid:7)1).Thisimpliesthat
P
No.1,2009 GALAXYINTERACTIONSANDSTARFORMATION 333
Figure 5. Same as Figure 4 but tracing only unobscured (UV-derived) star
formation. The unobscured SSFR enhancement found in galaxy pairs with Figure 7. SFR enhancement measured after randomizing the SFR between
rseesppaercattiotonsthrePca<se4i0nkwphcicahndthmeeorbgsecrusreredmsntaarnftsorimsadtriaomnaistictaaklleynriendtuocaecdcowuintht goafl∼ax1i6es0okfpsci.mWilearsshtoewllatrhmeapsosi.nAtsmcoilrdreesnphoanndcienmgetnottihsefoSuFn–dSoFuatutotosceoprareralatitoionns
(Figure4). atdistances>15kpc,wherenomorphologicalinformationisused.
the SFR (rather than in the SSFR) is a better quantity to
consider.
WeshowinFigure6theenhancementintheSFR(E (r ))
SFR p
as a function of the projected pair separation. For the cross–
correlation function (our default case) the enhancement in the
SFR is similar to the one found in the SSFR at all separations
except for the merger remnants (r = 0), in which the excess
p
abovethewholepopulationis∼50%lower.TheSFR–weighted
autocorrelation of star-forming galaxies matches that of the
SSFR–weighted one for r < 40 kpc but differs beyond:
p
E = 1.25 for 40 < r < 180 kpc. While most of these
SFR P
points in Figure 6 are individually compatible with the error
bars shown in Figure 4, taken together they represent a ∼2σ
significant difference between E and E for the entire
SSFR SFR
region40<r <180kpc.
p
Figure6.SFRenhancementingalaxyinteractions.Thetwosmallestradiibins A potential driver of the SFR enhancement in the regime
aarneddmereivrgeedrfrreommnmanotrsph(sohloowgincawlliythserlPect=ed0e)x;treenmhaenlycecmloesnetpsaairtsla(rrgPe∼rra1d0iikparce) 40 < rP < 180 kpc is the fact that more massive galaxies
determined using weighted two-point correlation functions. There is a clear tend to be both more clustered and have higher SFR (Noeske
enhancement at rP < 40 kpc for the cross-correlation analysis (black-filled etal.2007);thiscouldtranslateintoaweakenhancementinthe
symbols) which is compatible with ESSFR (Figure 4) except for the merger SFR in galaxy pairs living in dense environments (see Barton
remnants,wheretheexcessis∼50%lower.Theautocorrelationofstar-forming
et al. 2007, for a thorough discussion on the relation between
galaxies(emptysymbols)presentsanunexpectedbehavior,showingaverymild
galaxypairsandenvironment)whichwillnotbepresentinthe
enhancementatrP <180kpc.
SSFR,becausethenormalizationbygalaxymassfactorsoutthis
dependence.Totesttherelevanceofthissystematiceffect,we
mostofthedirectlytriggeredstarformationisdustobscured,in randomizedtheSFRsamonggalaxiesofsimilarmass500times
goodagreementwiththeexpectationsfromMihos&Hernquist inthesampleandrepeatedtheanalysis.Weshowtheresultsof
(1994,1996),diMatteoetal.(2007),Coxetal.(2008),andthe thisexerciseinFigure7,wherewecanseeatailofenhancement
detailed models by Jonsson et al. (2006). In these simulations withabehaviorsimilartotheoneseeninFigure6.Webelieve
mostofthestarformationistriggeredinthecentralregionsof thatacombinationofthedensity–mass–SFRrelationplusnoise
thegalaxyafterthecoldgashasbeenfunneledtotheinnerkpc. isdrivingESFR >1(autocorrelation)between40and180kpc.
This scenario is also supported by our measurement of Accordingly, we consider only the enhancement at rP <
the mean ratio between the total SFR and the UV-derived, 40 kpc as produced by major merging in what follows, and
which gives an idea of the degree of dust-obscuration use the differences between the SFR and SSFR enhancement
(SFR /SFR ). We find 6.64 ± 0.66 in the case of the on <40 kpc scales as a measure of systematic uncertainty.
IR+UV UV
mergerremnants(Group3inSection3.2)and6.63 ± 0.64inthe Under those assumptions, we find a weak enhancement of
caseoftheveryclosepairs(Group2),comparedto3.15 ± 0.53 star formation at rP < 40 kpc of (cid:8) = 1.50 ± 0.25 in the
forallobjectsinthesample. SF–SF autocorrelation and (cid:8) = 1.80 ± 0.30 in the SF–All
cross-correlation. These values have been computed as the
average of the enhancement in the bins r < 40 kpc together
4.1.1.StarFormationRateversusSpecificStarFormationRate P
in E and E . These (conservative) error bars include
SSFR SFR
To study the fraction of the global star formation directly boththestatisticaluncertaintiesandthesystematicsdrivingthe
triggered by galaxy–galaxy interactions the enhancement in differencesbetweentheSFRandtheSSFR.