A Catalog of Local E+A(post-starburst) Galaxies selected from the Sloan Digital Sky Survey Data Release 5

0.31 MB·English

by  Tomotsugu Goto

Mon.Not.R.Astron.Soc.000,1–6(2006) Printed2February2008 (MNLATEXstylefilev2.2)
A Catalog of Local E+A(post-starburst) Galaxies selected from the
Sloan Digital Sky Survey Data Release 5
⋆
Tomotsugu Goto
InstituteofSpaceandAstronauticalScience,JapanAerospaceExplorationAgency,3-1-1Yoshinodai,Sagamihara,Kanagawa229-8510,Japan
8
0
0
2 2February2008;inoriginalform2007February1

n
a
J ABSTRACT
E+Agalaxieshavebeeninterpretedaspost-starburstgalaxiesbasedonthepresenceofstrong
7
Balmerabsorptionlinescombinedwiththeabsenceofmajoremissionlines([OII]norHα).

] Asapopulationofgalaxiesinthemidstofthetransformation,E+Agalaxieshasbeensubject
h to an intense research activity. It has been, however, difficult to investigate E+A galaxies
p statistically since E+A galaxies are an extremely rare population of galaxies (<1% of all
-
o galaxiesinthelocalUniverse).
r Here,wepresentalargecatalogof564E+A(post-starburst)galaxiescarefullyselected
t from half million spectra of the Sloan Digital Sky Survey Data Release 5. We define E+A
s
a galaxiesasthosewithHδequivalentwidth>5A˚ andnodetectableemissionin[OII]andHα.
[ Thecatalogcontains564E+Agalaxies,andthus,isoneofthelargestofthekindtodate.In

addition,wehaveincludedtheHαlineintheselectiontoremovedustystar-forminggalaxies,
1
v which could have contaminated previous [OII]-based samples of E+A galaxies up to 52%.
6 Thus, the catalog is one of the most homogeneous,let alone its size. The catalog presented
0 herecanbeusedforfollow-upobservationsandstatisticalanalysesofthisrarepopulationof
1 galaxies.
1
. Key words: galaxies: evolution, galaxies:interactions, galaxies:starburst, galaxies:peculiar,
1
galaxies:formation
0
8
0
:
v
i 1 INTRODUCTION 1990;Fabricant,McClintock,&Bautz1991;Abrahametal.1996;
X Dressleretal. 1999; Poggiantietal. 1999; Dressleretal. 2004;
Dressler&Gunn (1983, 1992) found galaxies with mysterious
r Gotoetal. 2003; Goto 2004, 2005b, 2007). However, the reason
a spectra while investigating high redshift cluster galaxies. The
whytheystartedstarburst,andwhytheyabruptlystoppedstarburst
galaxies had strong Balmer absorption lines with no emission
remainoneofthemysteriesingalaxyevolution.
in [OII]. These galaxies are called “E+A” galaxies since their
spectra looked like a superposition of that of elliptical galax-
ies (Mg5175, Fe5270 and Ca3934,3468 absorption lines) and that
of A-type stars (strong Balmer absorption lines) 1. The exis-
tence of strong Balmer absorption lines shows that these galax-
ies have experienced starburst recently (within a gigayear; Goto
2004). However, these galaxies do not show any sign of on-
going star formation as non-detection in the [OII] emission line
indicates. Therefore, E+A galaxies have been interpreted as a OneofthemajordifficultiesininvestigatingE+Agalaxieshas
post-starburst galaxy, that is, a galaxy which truncated starburst been their rarity: less than 1% galaxies are in the E+A phase in
suddenly (Dressler&Gunn 1983, 1992; Couch&Sharples 1987; thepresentUniverse(Gotoetal.2003).Inthiswork,wesolvethis
MacLaren,Ellis,&Couch 1988; Newberry,Boroson,&Kirshner problembycreatingalargecatalogof564E+Agalaxiescarefully
selectedfromthefifthpublicdatareleaseoftheSloanDigitalSky
Survey(SDSS;Adelman-McCarthyetal.2006).Inadditiontoits
size,duetotheavailabilityoftheinformationontheHαline,this
⋆ E-mail:[email protected]
catalogisoneofthemosthomogeneousofthekind,nicelyremov-
1 Since some of E+A galaxies are found to have disk-like morphol-
ing the contanimation from the Hα emitting galaxies. The cata-
ogy (Couchetal. 1994; Dressleretal. 1994; Caldwell&Rose 1997;
logispubliclyreleased2sothatworld-wideresearcherscanutilize
Dressleretal.1999),thesegalaxiesaresometimescalled“K+A”galaxies.
However,Goto(2003)foundthattheirE+Asamplewithhighercomplete- forstatisticalanalysesandfollow-upobservationsinvariouswave-
nesshasearly-typemorphology.Followingthisdiscovery,wecallthemas length.Unlessotherwisestated,weadoptthebest-fitWMAPcos-
“E+A”throughoutthiswork. mology:(h,Ωm,ΩL)=(0.71,0.27,0.73)(Bennettetal.2003).
2 T. Goto
Figure1. AhistogramoftheHδequivalentwidth(A˚)ofallgalaxies.We
selectthosewithHδEW>5A˚ asE+Agalaxies.
Figure2. HαEW(emission)isplotted againstHδ EW(absorption) for
galaxieswithSN(r)> 10.E+Agalaxiesselectedinthispaperareshown
2 CATALOG withlargerdots.
2.1 Selection
shown inthe contours, these twofigures show that E+A galaxies
Our new catalog of 564 E+A galaxies is based on the fifth pub-
haveauniqueemission/absorptionlineproperties.
licdatareleaseoftheSDSS(Adelman-McCarthyetal.2006).This
WestressanadvantageinusingtheHαlineinselectingE+A
is the final data release of the SDSS I , and containts spectra of
galaxies. Previous samples of E+A galaxies were often selected
675,000galaxies.TheselectionalgorithmissimilartothatinGoto
based solely on [OII] emission and Balmer absorption lines ei-
(2005b), which we summarize below (also used in Goto 2004;
ther due to the high redshift of the samples or due to instrumen-
Yamauchi&Goto2005;Yagi&Goto2006;Yagi,Goto,&Hattori
talreasons.Gotoetal.(2003)showedthatsuchselectionsofE+A
2006).
galaxies without information on Hα line would suffer from 52%
Weonlyusethoseobjectsclassifiedasgalaxies(type=3,i.e.,
of contamination from Hα emitting galaxies, whose morphology
extendedobjects)withspectroscopicallyclassifiednottobeastar
andcolorareverydifferentfromthatofE+Agalaxies(Gotoetal.
(SpecClassisnot astar) and thespectroscopic signal-to-noise
2003).Tobackthisup,Blakeetal.(2004)selectedE+Agalaxies
>10perpixel(inthecontinuumofther-bandwavelengthrange).
fromthe2dFusingonlyBalmerand[OII]linestofindthatsome
Thesecriterianicelyremovecontaminationfromnearbystarsand
E+AgalaxiesintheirsamplehavetheHαlineinemission.When
star-forming regions. For these galaxies, we have measured Hδ,
Hαlineisnot availablefor the E+Aselection, sinceHβ and Hγ
[OII]andHαequivalentwidths(EWs)andobtainedtheirerrorsus-
absorptionfeaturesaresubjecttotheemissionfilling,Blakeetal.
ingthefluxsummingmethoddescribedinGotoetal.(2003).This
(2004)suggestedthatusingthreeBalmerabsorptionlines(Hδ,Hγ
fluxsummingtechniqueisadvantageousovertheGuassianfitting
andHβ)cansupressthecontaminationfromthegalaxieswithde-
methodinmeasuringreliablelinestrengthsfornoisyspectra.For
tectableHαemission.
the Hδ line, we only used the wider window of 4082−4122A˚ as
alineregiontoeasethecomparisonwithmodels(c.f. Gotoetal.
2003).AhistogramofHδEWdistributionisshowninFigure1. 2.2 Catalogof564E+Agalaxies
Once these lines are measured, we select E+A galaxies as
thosewithHδEW>5.0A˚,andHαEW>−3.0A˚,and[OII]EW Among451759 galaxieswhichsatisfytheredshiftand S/Ncrite-
> −2.5A˚3. Although the criteria on emission lines allow small ria with measurable [OII], Hδ and Hα lines, we have found 564
E+A galaxies. Thisis one of the largest sample of E+A galaxies
amount of emission in the E+A sample, they are relatively small
to date. The 564 E+A galaxies span a redshift range of 0.03276
amountintermsofthestarformationrate(SFR).Wealsoexclude
z60.3421,withinwhichHαlineissecurelycoveredbytheSDSS
galaxies at 0.35 < z < 0.37 from our sample because of the
sky feature at 5577A˚ . Our criteria are more strict than previous spectrograph.
ones(e.g.,HδEW>4.0A˚ and[OII]EW> −2.5A˚),suppressing In Figure 4, we show four example spectra of E+A galax-
ieswithstrong Hδ EWof 8.62-9.27A˚. Thecorresponding g,r,i-
possiblecontaminationsfromotherpopulationsofgalaxies(Goto
compositeimagesareshowninFigure5.
2004). Toillustrateour selection criteria,weplot HαEW (emis-
The catalog is publicly available in the Table at
sion)againstHδEW(absorption)forgalaxieswithSN(r)> 10in
http://www.ir.isas.jaxa.jp/∼tomo/ea5 . The
Figure2.E+A galaxiesareshown inlargerdots. InFigure3,we
columns arename, z, RA,DEC,Hδ EW,errorsof Hδ EW,[OII]
show[OII]EW(emission)againstHδEW(absorption)forgalax-
EW,errorsof[OII]EW,HαEW,errorsofHαEWerr,Petrosian
ieswithSN(r)>10.Comparedwiththedistributionofallgalaxies
magnitudeinr,signal-to-noiseratioinr-bandwavelengthrange,
andalinktotheNASAskyviewRandK images.Ontheobject
2 http://www.ir.isas.jaxa.jp/∼tomo/ea5 names,thereisalsoalinktotheSDSSobjectexplorerwhereone
3 Absorptionlineshaveapositivesignthroughoutthispaper. canseeag,r,i-compositeimageandaGIFspectrum.
ACatalogofLocalE+A(post-starburst)GalaxiesselectedfromtheSloanDigitalSkySurveyDataRelease5 3
Figure4.Examplespectraof4E+AswithlargestHδEWof8.62-9.27A˚.Eachspectrumisshiftedtotherestframewavelengthandsmoothedusinga20A˚
box.ThecorrespondingimagesareshowninFig.5.(Theimageofthesamegalaxycanbefoundinthesamecolumn/rowpanelofFig.5)
3 DISCUSSION These nearby E+A galaxies may have remaining star-
formation activity outside of the 3” fiber. Nevertheless, it is also
3.1 RedshiftDistribution true that these galaxies have a post-starburst stellar population
within the 3” of radius. It is still important to investigate what
InFigure6,weshowtheredshiftdistributionoftheE+Agalaxies.
caused the central post-starburst in these galaxies, and thus, we
Themedianredshiftisz=0.138.
keep these nearby E+A galaxies in our sample. Due to their
Thereisapossiblepeakatthelowestredshiftbin(z ∼0.04).
larger apparent size, these E+A galaxies are suitable targets for
The reason for the peak is not certain. We have checked all the
spatially-resolvedobservations(e.g., Yagi,Goto,&Hattori2006;
imagesofz < 0.05E+Agalaxiesbyeyetofindthatallofthem
Yagi&Goto2006).
show strong Hδ absorption, confirming that our selection criteria
In Figure 8, we show four example spectra of E+A galax-
workproperly.We,however,suspecttheaperturebiasmaybeone ieswithalargeapparent sizeof Petrosian90% radiusof > 14′′.
ofthecauses;theSDSSfiberspectrographonlysamplesthelight
These E+A galaxies are large enough for detailed morphologi-
withintheinner3”(Straussetal.2002).Thisdoesnotbringalarge
cal/substructure studies. The corresponding g,r,i-composite im-
bias into the distant E+A galaxies, whose sizes are smaller (the
agesareshowninFigure9.Thereisahintofpossibledynamical
median Petrosian 50% radius of our sample is 1.4”), but it does
disturbancesinallfourgalaxiesshownhere.
bringalargeaperturebiastonearbyE+Agalaxiesduetotheirlarge
apparentsize.Atz ∼ 0.04,themedianphysicalsize(takenfrom
the Petrosian50% light radius) of E+A galaxies is2.0”, whereas
3.2 E+AsinthehighredshiftUniverse
the 3”-fiber only collects the light from the inner 1.5” of radius.
InFigure7,weplottheratioofPetrosian50%lightradiustothe OurcatalogofE+Agalaxiesprovidesuswithausefulbenchmark
fiberradius(1.5”)asafunctionofredshift.Thefigureshowsthatat instudying theevolutionof E+Agalaxiesinhigher redshiftUni-
z∼0.04,therearesignificantnumberofE+Agalaxiesafewtimes verse. It has been known that cosmic star formation density was
larger thanthefiber size. Therefore, werecommend tousealow higheratz >1(Madauetal.1996).AccordingtoLeBorgneetal.
redshiftcut(e.g.,z>0.05;seealso Go´mezetal.2003;Gotoetal. (2006),thefractionsofthepost-starburstgalaxieswaslargerbya
2003)whenoneperformsstatisticalanalysisonthesample. factorof2atz∼1(alsosee Roseboometal.2006).Bycomparing
4 T. Goto
Figure5.Examplesofg,r,i-compositeimagesofE+Agalaxieswith4HδEW>7A˚.Theimagesaresortedfromlowtohighredshift.Only24lowestredshift
E+Asareshown.Thecorresponding spectrawithnameandredshiftarepresented inFig.4.(Thespectrumofthesamegalaxycanbefoundinthesame
column/rowpanelofFig.4.)
thefractionofE+Agalaxiesindifferentredshift,wecantracethe 1.1%,avaluelowerthanthatingalaxyclustersatcomparablered-
cosmicstarformationhistoryintermsofthepost-starburstgalax- shifts.4
ies. TakentogetherwiththemuchsmallerfractionofE+Agalax-
iesinthelocalfield,thesuddenincreaseofthefractionsindistant
Inhighredshiftclusterenvironments,E+Agalaxiesaremuch
clusters over all other environments suggests that there is some-
more numerous; pioneering work were done by Dressleretal.
thinguniquetothistimeandtheenvironmentthatgivesrisetoso
(1999); Poggiantietal. (1999), who found E+A galaxies (Hδ
manyE+Agalaxies.ItisalsoimportanttonotethatE+Agalaxies
EW>3A˚and not detectable emission in [OII]) are significantly
mayhaveheterogeneous origins; inthelocal Universe, E+Asare
more common in 10 clusters at 0.37< z <0.56 than in the field
preferentiallyfoundintherarefiedfieldregions,whereasathigher
(21±2%comparedwith6±3%).Later,Tranetal.(2003)found7-
redshift, E+Asarefound inthe cluster environment. Considering
13%ofE+Agalaxiesinthreehighredshiftclustersatz=0.33,0.58,
and0.83,claimingthat>30%ofE+S0membersmayhaveunder-
gonetheE+AphaseiftheeffectsofE+Adownsizingandincreas- 4 NotethatBaloghetal.(1999)foundsmallerE+Afractions of1.5%±
ingE+Afractionasafunctionofredshiftareconsidered(theirse- 0.8%inhighredshiftclusters(z ∼0.25).However,ithasbeencontrover-
lectioncriteriais HδEW+2HγEW > 4A˚ and[OII]EW>-5A˚).In sialwhetherthere isadiscrepany onthefraction ofE+Agalaxies inthe
theirsearchforfieldE+Agalaxiesamong800spectra,Tranetal. highredshiftclusters.SeeSection5.1ofDressleretal.(2004)fordetailed
(2004)measured theE+Afractionat0.3< z <1tobe2.7%± discussion.
ACatalogofLocalE+A(post-starburst)GalaxiesselectedfromtheSloanDigitalSkySurveyDataRelease5 5
Figure8.Examplespectraof4E+AswithlargestPetrosian90%lightradii(> 14′′).Eachspectrumisshiftedtotherestframewavelengthandsmoothed
usinga20A˚ box.ThecorrespondingimagesareshowninFig.9.(Theimageofthesamegalaxycanbefoundinthesamecolumn/rowpanelofFig.9)
thelargedifferenceintimeandtheenvironment,E+Asinthelocal 4 SUMMARY
fieldandhigh-zclustersmayhavebeencreatedbydifferentphys-
Wehaveconstructedoneofthelargestcatalogsof564localE+A
icalmechanisms.Forexample,Goto(2005b)hasshownthatlocal
(post-starburst) galaxies carefully selected from the SDSS DR5.
(z∼0.1) E+A galaxies have more close companion galaxies than
The sample provides us with a useful tool to investigate statisti-
average galaxies, showing that the dynamical merger/interaction
cal properties and/or for follow-up observations of this rare, but
couldbethephysicaloriginofE+Agalaxies.However,mostE+A
importantpopulationofgalaxies.
galaxies in the high redshift clusters are known not to be major
mergers(Dressleretal.1999).
ACKNOWLEDGMENTS
Weare grateful to the referee, Prof.A.Dressler, for many insight-
3.3 SelectioncriteriaofE+Agalaxies
fulcommentsthatimprovedthepapersignificantly.WethankDr.
We used the selection criteria of HδEW > 5A˚ and Hα EW < M.Yagi,andC.Yamauchiforusefuldiscussion.
−3.0A˚,[OII] EW< −2.5A˚ to select 564 E+A galaxies (Section TheresearchwasfinanciallysupportedbytheSasakawaSci-
2.2).However,wecautionreadersthatthenumber(orfraction)of entificResearchGrantfromTheJapanScienceSociety.
the selected E+A galaxies strongly depends on the selection cri- ThisresearchwaspartiallysupportedbytheJapanSocietyfor
teria.Forexample, ifweloosen ourcriteriatoHδ EW> 4A˚,the thePromotionof SciencethroughGrant-in-AidforScientificRe-
numberofE+Agalaxiesincreasesto1062(almosttwiceasnumer- search18840047.
ous).WithHδEW>3A˚,2298E+Agalaxiesareselected(4times FundingforthecreationanddistributionoftheSDSSArchive
morenumerous).Therefore,whenonecomparessamplesofE+A hasbeenprovidedbytheAlfredP.SloanFoundation,thePartici-
galaxiesselectedfromdifferentdatasets,itisimportanttosynchro- patingInstitutions, theNational Aeronauticsand SpaceAdminis-
nizetheselectioncriteria.Especially,onecomparestoasampleat tration, theNational Science Foundation, theU.S.Department of
higherredshift,inconsistencyinselectioncriteriacanproducespu- Energy,theJapaneseMonbukagakusho,andtheMaxPlanckSoci-
riousevolutionaryeffect. ety.TheSDSSWebsiteishttp://www.sdss.org/.
6 T. Goto
Figure9.Examplesofg,r,i-compositeimagesofE+AgalaxieswithlargestPetrosian90%lightradii(> 14′′).Thecorrespondingspectrawithnameand
redshiftarepresentedinFig.8.(Thespectrumofthesamegalaxycanbefoundinthesamecolumn/rowpanelofFig.8)
The SDSS is managed by the Astrophysical Research Con- BennettC.L.,etal.,2003,ApJS,148,1
sortium (ARC) for the Participating Institutions. The Participat- BlakeC.,etal.,2004,MNRAS,355,713
ingInstitutionsareTheUniversityofChicago,Fermilab,theInsti- CaldwellN.,RoseJ.A.,1997,AJ,113,492
tuteforAdvancedStudy,theJapanParticipationGroup,TheJohns CouchW.J.,SharplesR.M.,1987,MNRAS,229,423
Hopkins University, Los Alamos National Laboratory, the Max- CouchW.J.,EllisR.S.,SharplesR.M.,SmailI.,1994,ApJ,430,
Planck-InstituteforAstronomy(MPIA),theMax-Planck-Institute 121
forAstrophysics(MPA),NewMexicoStateUniversity,University Dressler A.,Oemler A. J., Poggianti B. M., Smail I.,Trager S.,
ofPittsburgh,PrincetonUniversity,theUnitedStatesNavalObser- ShectmanS.A.,CouchW.J.,EllisR.S.,2004,ApJ,617,867
vatory,andtheUniversityofWashington.
DresslerA.,SmailI.,PoggiantiB.M.,ButcherH.,CouchW.J.,
EllisR.S.,OemlerA.J.,1999,ApJS,122,51
DresslerA.,GunnJ.E.,1992,ApJS,78,1
DresslerA.,OemlerA.J.,ButcherH.R.,GunnJ.E.,1994,ApJ,
REFERENCES
430,107
AbrahamR.G.,etal.,1996,ApJ,471,694 DresslerA.,GunnJ.E.,1983,ApJ,270,7
Adelman-McCarthyJ.K.,etal.,2006,ApJS,162,38 FabricantD.G.,McClintockJ.E.,BautzM.W.,1991,ApJ,381,
BaloghM.L.,MorrisS.L.,YeeH.K.C.,CarlbergR.G.,Elling- 33
sonE.,1999,ApJ,527,54 Go´mezP.L.,etal.,2003,ApJ,584,210
ACatalogofLocalE+A(post-starburst)GalaxiesselectedfromtheSloanDigitalSkySurveyDataRelease5 7
Figure3. [OII]EW(emission)isplottedagainstHδEW(absorption).E+A Figure7. Theratioofthephysicalgalaxysize(takenfromthepetrosian
galaxiesselectedinthispaperareshownwithlargerdots. 50%lightradius)totheSDSSfibersizeisshownasafunctionofredshift
fortheE+Agalaxies.
sonD.D.,MageeD.,2004,ApJ,609,683
Tran K.-V. H., Franx M., Illingworth G., Kelson D. D., van
DokkumP.,2003,ApJ,599,865
YagiM.,GotoT.,HattoriT.,2006,ApJ,642,152
YagiM.,GotoT.,2006,AJ,131,2050
YamauchiC.,GotoT.,2005,MNRAS,359,1557
Figure6. RedshiftdistributionoftheE+Agalaxies.
GotoT.,2007,MNRAS,377,1222
GotoT.,2005b,MNRAS,357,937
GotoT.,2004,A&A,427,125
GotoT.,YamauchiC.,FujitaY.,OkamuraS.,SekiguchiM.,Smail
I.,BernardiM.,GomezP.L.,2003,MNRAS,346,601
GotoT.,etal.,2003,PASJ,55,771
GotoT.,2003,PhD.Thesis,TheUniversityofTokyo
LeBorgneD.,etal.,2006,ApJ,642,48
MacLarenI.,EllisR.S.,CouchW.J.,1988,MNRAS,230,249
MadauP.,FergusonH.C.,DickinsonM.E.,GiavaliscoM.,Stei-
delC.C.,FruchterA.,1996,MNRAS,283,1388
NewberryM.V.,BorosonT.A.,KirshnerR.P.,1990,ApJ,350,
585
PoggiantiB.M.,SmailI.,DresslerA.,CouchW.J.,BargerA.J.,
ButcherH.,EllisR.S.,OemlerA.J.,1999,ApJ,518,576
RoseboomI.G.,etal.,2006,MNRAS,373,349
StraussM.A.,etal.,2002,AJ,124,1810
TranK.-V.H.,FranxM.,IllingworthG.D.,vanDokkumP.,Kel-