Evidence for Expanding Superbubbles in a Galaxy at z=0.7443

12 Pages·0.36 MB·English

by  Nicholas A. Bond

DRAFTVERSIONFEBRUARY1,2008
PreprinttypesetusingLATEXstyleemulateapj
EVIDENCEFOREXPANDINGSUPERBUBBLESINAGALAXYATz=0.74431
NICHOLASA.BOND,CHRISTOPHERW.CHURCHILL,JANEC.CHARLTON2
DepartmentofAstronomyandAstrophysics
ThePennsylvaniaStateUniversity
UniversityPark,PA16802
bond,cwc,[email protected]
AND
STEVENS.VOGT
UCO/LickObservatories
BoardofStudiesinAstronomyandAstrophysics
1 UniversityofCalifornia,SantaCruz,CA96054
0 [email protected]
0 DraftversionFebruary1,2008
2
ABSTRACT
n
a Theinterveningz =0.7443MgIIabsorptionsysteminthespectrumofMC1331+170showsanunusualseries
J oflinepairs,eachwith∆v ∼30kms−1. Theselinescouldbeexplainedastheshellsofexpandingsuperbubbles

4 residing in the outer regions of an edge–on spiral galaxy visible in the optical image of the MC 1331+ 170
field. Thecolorandbrightnessofthisgalaxymakeitthemostlikelycandidatez = 0.7443absorber,thoughtwo

1 othergalaxiesinthequasarfieldcouldalsobecontributingtotheMgIIabsorptionprofile. Kinematicmodelsof
v absorptionfrom compactgroupsand galaxypairs produceprofileslargely inconsistentwith the observedMgII
9
spectrum.Superbubbleswouldnaturallygeneratemoreregularstructuressuchasthoseobserved.Photoionization
4
models of the superbubbleshell are consistentwith the observedprofile for many realistic physicalconditions.
0
1 In a pure superbubble model, the large velocity spread of the MgII absorption system is inconsistent with the
expectedspread of a quiescent, rotatingdisk. This requiresunusualkinematicswithin the host galaxy,perhaps
0
1 duetoarecentinteraction.
0 Subjectheadings:quasars:absorptionlines—interstellarmedium—superbubbles
/
h
p 1. INTRODUCTION HI map of face–on galaxies or as HI emission–line filaments
- inedge–ongalaxies.Singlestructuresofthistypeareobserved
o Strong MgII absorbers show complex kinematic structure withsizessimilartoGalacticbubblesandmassesbetween107
str crootnastiinstgendtiswkit(hLaanmzeitxtatu&reBofowraednia1l9i9n2fa;llCfhraormltotnhe&haClhouarnchdilal and108M⊙. Smallershells(evensingle–starstellarwindbub-
bles) are also thought to exist, but cannot be detected in ex-
a 1998; Churchill& Vogt2001). A largevarietyof phenomena
: ternal galaxies. Superbubbles larger than 1 kpc in radius are
v cancontributetotheprofiles,suchasHI regions,HII regions,
thoughttobeformedbyeitheramergingofseparateshellsor
i high–velocityhaloclouds,andsuperbubbles. StrongMgIIab-
X sorbersalmostalwayshaveL > 0.08L∗ galaxieswithin35 by a process of self–propagatingstar formation (e.g. Bomans
B B et al. 1996). The number of bubbles in a given galaxy varies
r kpc (Steidel, Dickinson, & Persson 1994; Steidel 1995). One
a dramaticallyfromalmostnonetoseveralhundred,perhapsde-
particular absorption system, at z = 0.7443, in the spectrum
pendingonthegalaxy’shistoryofinteractions.
of MC 1331+ 170 shows multiple pairs of lines, each with
∆v ∼ 30kms−1. Therearefourstrongopticalsourceswithin Stellar wind bubbles for a single early–type star were first
5′′ ofthequasar,oneofwhichisanedge–onspiral. Theseare modeledbyWeaveretal.(1977). MacLow&McCray(1988)
extended this concept to clusters of O and B stars, proposing
theonlydetectedopticalsourceswhichcouldhaveimpactpa-
Galactic superbubbles as the explanation for the HI filaments
rameterslessthan35kpc.
observedbyHeiles(1979).Accordingtothesemodels,asuper-
OB associations are capable of producing large shells of
bubble begins as an oversized stellar wind bubble in the ISM
swept up ISM using the energy input of stellar winds and
around an OB association. The swept up material from this
multiplesupernovae. Suchstructures,knownassuperbubbles,
cavityquicklycollapsesintoacold,denseshellwithadensity
have beenobservedin nearbygalaxies(Kamphuis1993; Sun-
greaterthanfourtimesthatof theambientISM.Theshellbe-
geon et al. 1999) as well as in our own galaxy (Heiles 1979;
ginsexpandingwiththeonsetofthefirstsupernovaeintheOB
McClure-Griffithsetal.2000).MgIIandFeIIabsorptionwould
association,whichcontributetheirenergyandmetalstothesu-
arise within the cold superbubble shell, appearing as pairs of
perbubble.Duringthis,theadiabaticphase,thesupernovaecan
linesfromtheoppositelyexpandingsidesoftheshellintersect-
beapproximatedasacontinuousenergysourceandthebubble
ingthelineofsight.
radiusincreasesast3/5. Theinteriorissparse,highlyionized,
Inourgalaxy,superbubblesaredetectedin HI asfilaments,
and isobaric during this phase. The continuing energy input
with inferred radii ranging from 100 pc to 2 kpc (e.g. Heiles
1979). Extragalacticsuperbubblesaremanifestasholesinthe maintainsthe expansionata rate of t1/2, in excessof the rate
1BasedinpartonobservationsobtainedattheW.M.KeckObservatory,whichisoperatedasascientificpartnershipamongCaltech,theUniversityofCalifornia,
andNASA.TheObservatorywasmadepossiblebythegenerousfinancialsupportoftheW.M.KeckFoundation.
2CenterforGravitationalPhysicsandGeometry
1
2 EXPANDINGSUPERBUBBLESINQ1331+170
expected for a pure snowplow solution (Mac Low & McCray ofDLAs(100−200K)occurwhentheylieinspiralgalaxies
1988). (e.g.Kanekaretal.2000). Evenifwearelookingthroughthe
Inthispaper,weexplorethehypothesisthatthelinesplittings diskofaspiralgalaxy(aswesuspect),the21–cmdatadonotre-
inthez = 0.7443absorberarisefromexpandingsuperbubbles strictthenatureoftheabsorber. However,theMgIIabsorption
inthevicinityofanobservededge–onspiralgalaxy(Churchill arisingfromknownDLAshasthecharacteristicofbeingtotally
et al. 1995). In § 2, we present the HIRES/Keck absorption saturated across the entire profile (Churchillet al. 2000b)and
profiles of detected low ionization transitions and an archival typically exhibits a total velocity spread of 50−100 kms−1.
HST/WFPC2 image of the quasar field. In § 3, we constrain Thesefeaturesarenotpresentinthissystem,suggestingthatit
theredshiftsandluminositiesofthefourgalaxiesinthequasar isnotaDLA.
field using the galaxies’ colors and magnitudes. We consider
the possibility that galaxy pairs and compact groups give rise 2.2. HST/PC2Image
to the observed MgII absorption in § 4. In § 5, we provide a Images of the MC 1331 + 170 field obtained through the
description of the photoionization models used in testing the
F702W (R) and F814W (I) filters were retrieved from the
superbubblehypothesis.Wediscussourresultsandsummarize
archive(Program5351,LeBrunetal.1997).Pipeline–reduced
in§6.
images for each filter were averaged and cleaned using the
IRAFtaskGCOMBINE.ThetaskCOSMICRAYSwasusedto
2. DATAANDANALYSIS removehotpixels.Toincreasethesignal–to–noiseratioforpre-
2.1. HIRES/KeckSpectra sentation purposes, the R and I band images were combined.
For our analysis, we use the photometry as presented in Sec-
The spectrum of the z = 2.084, V = 16.7 quasar
tion2,Table8ofLeBrunetal.1997. Tworealizationsofthis
MC1331+170wasobtainedwiththeW.M.KeckTelescope’s
HST/WFPC2 image, with the quasar centered in the PC, are
HIRES/Keckspectrograph(Vogtetal.1994)onthenightof19
presentedinFigure2.Theleft–handpanelhasalargerdynamic
March1993.Theslitwidthwas0.861′′(yieldingR=45,000). rangeandshowsthreeresolvedobjectswithin5′′ofthequasar,
A2×2pixelbinning(inboththecross–dispersionanddisper-
labeledinorderofangulardistancefromthequasar. Anaddi-
siondirections)wasusedduringreadout.Theresultingresolu- tionalsourceat≃0.8′′isapparentuponpoint–spreadfunction
tionisR = 34,000(v = 8.8kms−1 perresolutionelement).
(PSF) subtraction of the quasar (Figure 16, of Le Brun et al.
The wavelengthcoverageis 4263to 6733A˚, with gapsabove
1997).Intheright–handpanel,G5isemphasized.
5100A˚,becausetheCCDdidnotcapturethefreespectralrange
oftheechelleordersabovethiswavelength. Thespectrumwas 3. THEMC1331+170FIELD
reduced in the usual manner using IRAF3 v2.10. All wave-
3.1. Procedure
lengthsarevacuumandheliocentriccorrected.
TherearefourMgIIsystemsinthespectrumofMC1331+ ThephotometryofLebrunetal.(1997)givesalimitingmag-
170 at z = 0.7443, z = 1.3280, z = 1.776, and z = nitude of 26 in R, which implies that unresolved sources at
1.786. Due to its unusual structure, our focus is on the z = z = 0.7443 brighter than ∼ 0.04 L∗ should be seen in the
B
0.7443 system. The detected transitions (FeII λ2587, λ2600, image. Because strong MgII absorbers nearly always have
MgII λλ2796,2803, MgI λ2853) for the z = 0.7443 system LB > 0.08 L∗B counterparts within 35 kpc (Steidel, Dickin-
areshowninFigure1. Theprofilefits(solidcurves)wereob- son, & Persson 1994; Steidel 1995), we expect to detect the
tainedusingVoigtprofiledecomposition. Thisyieldsthenum- galaxygivingrisetothez =0.7443absorption.
berofcomponents,theirline–of–sightvelocities,columnden- Therearefourgalaxieswhichcouldbewithin35kpcofthe
sities, andDopplerbparameters. WeusedMINFIT,aχ2 min- quasarintheHST/WFPC2image(Figure2,G2–G5,numbered
imizationcodeof ourowndesign(Churchill1997). Theticks asinLeBrunetal.1997). Inspectionofthesourcesshowsan
abovethenormalizedcontinuainFigure1givecomponentve- edge–onspiral morphologyfor G5 and an elliptical shape for
locitycentroids. Table1liststheresultingparametersfromthe G3 (also see Le Brun et al. 1997). G4 shows some structure
Voigtprofilefitting. (possiblyirregular)andthelocationofG2infrontofthequasar
The equivalent width of the z = 0.7443 system is makesmorphologydifficulttodetermine.
W (2796) = 1.81 A˚. An HST/STIS spectrum that covers the Unfortunately,redshiftsforG2–G5havenotbeenobtained.
r
CIVλλ1548,1550doubletisavailableinthearchive(Program Wemust,therefore,determinehowlikelyitisthateachcandi-
7271). Unfortunately, the signal to noise is ∼ 1 at the posi- date galaxyis associated with the absorberat eachof the red-
tionoftheCIVabsorption,yieldingnousefulequivalentwidth shifts. Thiswillallowustoassesswhichgalaxy/galaxiescould
limit. The other systems are also strong4. Damped Lyα ab- give rise to the z = 0.7443 absorption. We do this by com-
sorption (DLA), N(HI) ≥ 2×1020 cm−2 is observedin the paring the apparent magnitudes and colors of each galaxy to
z =1.776system(Chaffeeetal.1988). redshifted model spectral energy distributions (SED) for dif-
RaoandTurnshek(2000)foundthat50%ofMgIIabsorbers ferentmorphologicaltypes. However,weappreciatethatthere
with W (2796) > 0.5 A˚ and W (2600) > 0.5 A˚ are DLAs may not be a one–to–one correspondence between absorbers
r r
(also see Boisse´ et al. 1998). This criterion is satisfied by the andgalaxies. Twoormoreoftheobservedgalaxiescouldgive
z =0.7443system,buttheLyαofthissystemisbelowtheLy- risetoabsorptionandlieatthesameredshift(Bowen,Blades,
manlimitbreakofthez =1.776DLAandcannotbemeasured. &Pettini1995). Additionally,DLAabsorptioncanarisefrom
Neutralhydrogen21–cmmeasurementsofthez =0.7443sys- relativelylowluminosityand/orlowsurfacebrightnessgalaxies
tem (Lane 2000) give N(HI) < 1.5×1018hTsi cm−2, where which are not seen in the image (e.g., Rao & Turnshek 2000;
hT iisthespintemperature. Theminimumspintemperatures Bouche`etal.2000;alsoseeBowen,Tripp,&Jenkins2000).
s
3IRAFisdistributedbytheNationalOpticalAstronomyObservatories,whichareoperatedbyAURA,Inc.,undercontracttotheNSF.
4TheMgIIatz=1.3284isblendedwiththeFeIIfromthehigherredshiftsystems(Steidel&Sargent1992)andapreciseequivalentwidthmeasurementisnot
possible,butweestimateWr(2796)=0.8A˚.
BONDETAL. 3
In order to obtain the L values expected for each galaxy ThemoststrikinggalaxyintheMC1331+170fieldisG5,an
B
at each redshift, we first obtained B and R magnitudes from edge–onspiral. Thisisourfavoredcandidateforaz = 0.7443
model SEDs for a set of standard morphologies (Bruzual & absorber,withaluminosityofL ∼ 1L∗. Thelowluminosi-
B B
Charlot 1993; Kinney et al. 1996). Models for E, S0, Sab, tiesofG3andG4makethemmarginalcasesforproducingsuch
Sbc, and Scd galaxies (µ = 0.95,0.30,0.20,0.10,0.01, re- strongMgIIabsorption.Additionally,theluminosityofG5be-
spectively5) were modelled with a 16 Gyr stellar population comes larger for the higher absorber redshifts (∼ 2−3 L∗);
B
with exponentially decreasing star formation. A redder ellip- thus,it∼10timeslesslikelytobeatthehigherredshifts. The
tical galaxy(E2) was modeledusinga composite spectrumof colorofG5,however,doesnotconstrainitsredshiftandallows
observed ellipticals from Kinney et al. 1996. For an irregu- ittobeastar–forminggalaxyatanyofthehigherredshifts. If
lar galaxy(Im), the observedSED ofNGC 4449was adopted thisis so, G3 or G4 wouldbe responsiblefor the z = 0.7443
(Bruzual&Charlot1993).Twostarburstmodelswerealsocon- system, but their small luminosities at this redshift make this
sidered, both with a constant star formation rate, assuming a unlikely.
SalpeterIMF.Theagesofthestarburstgalaxies(Starburst1and Inconclusion,G5isthe galaxymostlikelytobe producing
Starburst2,Bruzual&Charlaut1993)are106and107years. z =0.7443absorption,butG3andG4couldalsobecontribut-
TheK–correctionswereperformedbyredshiftingthemodel ing to the absorption at that redshift. If so, they would have
SEDs to the absorber redshifts and convolving them with the small luminosities that make them border–line candidates for
F702W and F814W filter response curves. A zero–pointnor- strongMgIIabsorption.
malization was performed using the measured R band appar-
ent magnitudeslisted in Table 2 (also see Table 8 in Le Brun 4. KINEMATICMODELSIMULATIONSOFGROUPSANDPAIRS
et al. 1997). Finally, the MB values were calculated using Though G5 is the best candidate for the z = 0.7443 ab-
H0 = 50 kms−1Mpc−1 and q0 = 0.05. Figure 3 shows sorber,G3andG4alsohavecolorsandluminositiesconsistent
R−I colorvs.redshiftforthestandardmorphologicaltypes. with this redshift. Furthermore,the bluecolorsof G4 and G5
The dotted lines on the figure correspond to the redshifts of seemtoindicaterecentstarformation,whichcouldeasilyhave
theMgIIabsorptionssystemsintheMC1331+170spectrum beentriggeredbyaninteractionbetweentwoormoregalaxies.
andthedashedlinestothecolorsofthefourgalaxies,G2–G5. Assuch,two compellingpossibilitiesarethatthe lineofsight
Intersections of the lines represent permitted locations on the passesthroughseveralmembersofa compactgroupofgalax-
diagram of the galaxies in the field and allow us to constrain ies or througha galaxy pair, giving rise to the observedMgII
theirtypeateachpossibleredshift. absorptionprofile.
Charlton and Churchill (1998)foundthat the kinematicsof
3.2. Results strong MgII absorbers can be described by assuming that the
absorbingcloudsarelocatedbothinthedisksandhalosofnor-
Table 2 summarizesthe permitted morphologicaltypes and
malgalaxies. Inordertosynthesizeanabsorptionlineprofile,
their calculated L (in units of L∗, using M∗ = −20.9) at
B B B we run a line of sight through a model galaxy with a given
each absorber redshift (excepting z = 1.786 because of its
orientation, impact parameter, and size (where the size of the
proximitytoz =1.776).
galaxy is defined by its luminosity). We used the luminosity
TheerrorsonthecolorofG2andmagnitudemightbequite
andimpactparameterfromTable2astheinputforeachgalaxy
large due to the PSF subtraction of the quasar. The color of
model.TheorientationofgalaxyG5wassettobeedge–onbut,
G2 makes it unlikely to be at z = 0.7443, and it must be an
whentheothergalaxieswereused,theirorientationswerecho-
elliptical if it is at z = 1.3284 or z ∼ 1.8. If only its color
sen randomly. Each galaxymodelwas chosenfrom75D/25H
is considered, it could be an E or S0 at z ∼ 1.8, but the de-
Hybrid2inCharlton&Churchill(1998),with75%diskclouds
rivedM valuesgiveL ∼ 15L∗ foranEandL ∼ 5L∗
B B B B B and25%haloclouds,bynumber. Inordertosimulateabsorp-
foranS0. Suchluminousgalaxiesareexceedinglyrareandwe
tionfromeithercompactgroupsorgalaxypairs,theindividual
concludethatG2isprobablyatz =1.3284.
syntheticabsorptionprofileswere shiftedin velocitybasedon
ThemorphologyandcolorofG3seemtoindicatethatitisan
observedcompactgrouporgalaxypairkinematics. Wegener-
early–typegalaxy,butitsredshiftisdifficulttoconstrain. Ifat
ated1000realizationsofsyntheticspectraforeach.
z = 0.7443,G3wouldhaveL ∼ 0.1L∗ andcouldbeanE,
B B
S0,orearly–typespiral. Duetothelowluminosity,itisuncer-
4.1. CompactGroups
tainwhetherG3wouldbegivingrisetostrongMgIIabsorption
at this redshift. G3 could be an S0 at either z = 1.3284with The median velocity dispersion of compact groups,
L ∼ 1L∗ orz ∼ 1.8withL ∼ 4.3L∗. Ifitisthelatter, 200kms−1 (Hicksonetal. 1992),resultsinprofilesthathave
B B B B
thegalaxywouldbeunusuallyluminous. Therefore,itismost velocity spreads larger than in the z = 0.7443 system. A
likelyanEorS0atz =0.7443orz =1.3284. smallervelocitydispersionwasrequiredsothatourmodelsdid
G4 isveryblueandwouldhaveto beactivelystar–forming notproducesyntheticprofileswithtotalvelocityspreadsincon-
to be atz = 0.7443. Just as with G3, however,it may notbe sistent with the data. Therefore, we chose the galaxy relative
luminousenough(L ∼0.1L∗)atthisredshifttobeexpected velocitiesfromHGC22,thecompactgroupwiththeminimum
B B
to giverise to strongMgII absorption. Theirregularstructure velocity spread (100 kms−1, Hickson et al. 1992). Thus, we
observed in the image may indicate an Im galaxy. The color havealreadyrestrictedourselvestoalimitedregionofparame-
and luminosity (L ∼ 0.1 L∗) are also consistent with such terspace;thisisnota“typical”compactgroup.
B B
a morphology at either z = 1.3284 or z ∼ 1.8. The color– Twenty–five randomly selected synthetic spectra from our
redshiftcurvesconvergetowardR−I ∼0athighredshiftfor three–galaxy compact group models (using G3, G4, and G5)
thespiraltypes,allowinganScdorSbcidentificationforG4at areshownintheupperpanelofFigure5. Clearly,thevelocity
z ∼1.8. spreadsofthecloudsinthesimulatedspectratendtobegreater
5µisthefractionofthemassinthestarsafteroneGyr(Bruzual1983)
4 EXPANDINGSUPERBUBBLESINQ1331+170
thanthatinthez = 0.7443system,evenwiththebiastowards dominatesatageslessthan∼ 107 years. Mostofthe models,
a small dispersion in galactic velocities. Also, the saturation however,aredominatedbytheHaardt–Madauspectrum.
levelsinthesimulatedspectratendtobehigher,implyingthat WeassumetheISMtohaveaconstantinitialatomicdensity
ourparticularlineofsightpassesthroughlessmaterialthanin ofn0 cm−3. AnOB associationresidinginthismediumgen-
an average model compact group. Finally, the regular struc- erates a total luminosity of L38 (in 1038 ergs cm−3) through
tureofpairsinabsorption,separatedby∼ 30kms−1,isnota a series of supernovaexplosions. If each supernova produces
characteristic feature of lines of sight throughmodel compact ∼ 1051 ergs and the OB association lasts 50 million years,
groups. the time–averaged luminosity is related to the number of su-
4.2. GalaxyPairs spwerenpotvuapesbhyelNliSsNn≃(e5x8p0reLss3e8d. Tinhuenaittosmoficcdme−n3si)tyanidnstihdeemofettahle-
e
Inordertoreducethetotalvelocityspreadandsaturation,we licity is Z, in solar units. The energy input from supernovae
alsoconsideredmodelsofgalaxypairs,assumingthateitherG3 drives the bubble radially outward into the ISM. The similar-
or G4 is at z = 0.7443 along with G5. Again, G5 was taken itysolution(MacLow&McCray1988)givestheradiusofthe
tohaveanedge–onorientation,andtheothergalaxywastaken bubble,
to be randomlyoriented. The velocitydifferencebetween the
galaxypairistunedto∼50kms−1. Infact,thisisarelatively R=0.27 (n0/L38)−0.2 t70.6 kpc, (1)
typicalvalueforfieldgalaxypairs.
Twenty–fiverandomlyselectedgalaxypairsyntheticspectra wheret7 is theage inunitsof107 years. Theshellexpansion
aredisplayedinthebottompanelofFigure5. Comparedtothe speedthenfollowsas
compactgroupsyntheticprofiles,alargerpercentageofthepair
pwriothfiltehsehoabvseervveeldocMitygIsIprperaodfislea.nHdoswateuvraetri,otnhelerveegluslacronabsissoterpn-t R˙ =16 (n0/L38)−0.2 t7−0.4 kms−1, (2)
tion line pairswe observein the spectrumofMC 1331+170
areonlyrarelyproduced,eveninthissimplegalaxypairmodel. whichisconstrainedbytheobservedvelocitysplitting,∆v,of
theabsorptionlinesaccordingto
4.3. KinematicModelResults
The z = 0.7443 absorber toward MC 1331+170 was se- R˙ =0.5∆v 1−z2/R2 −1/2, (3)
lectedforthisstudybecauseofitsuniquelyregularproperties. (cid:0) (cid:1)
Assuch,thisisabiasedstudy.Nonetheless,wehaveseenthat,
in 1000Monte–Carlosimulationsof galaxypairsand in 1000 where z is the impact parameter (to the center of the bubble)
simulations of compact groups, we can only rarely produce a of the line of sight. This equationis not very sensitive to im-
profile with such low levels of saturation and regular velocity pact parameter unless z/R ∼ 1. From the observed profile,
splittings. However,itisdifficulttoquantifytheshapesofthe ∆v ∼30kms−1;thus,wesetR˙ =15kms−1.
profilesinsuchawaythatwecanmakedirectstatisticalcom- If we choose L38 and t7, we can calculate R and Ntot for
parisons. We remindthereaderthatweusedacompactgroup input to the Cloudy model, where N is the total hydrogen
tot
with an unusuallysmall velocitydispersion. Therefore,based column density. The radius is found by combining equations
on this and on a visual inspection, we conclude that the ob- (1)and(2),giving
servedprofileisunlikelytoarisefromeitheracompactgroup
orgalaxypairenvironment. R=0.017R˙ t7 kpc. (4)
5. SUPERBUBBLESIMULATIONS
Thespectrumofthez = 0.7443absorptionsystemappears Fromequation(1),n0isgivenby
to be kinematicallyconsistent with a line of sight intersecting
theoppositelyexpandingsidesofthe“cold”shellsofmultiple 5
superbubbles (Churchill et al. 1995). In order to test this hy- n0 =L38 (cid:16)16t7−0.4/R˙(cid:17) cm−3. (5)
pothesis,wecalculatedwhatthecolumndensitiesofMgIIand
FeIIshouldbeinthecontextofsuchamodel. Finally,N is
tot
5.1. TheModel
The global properties of our model superbubbles are gov- Ntot =1021Rn0 cm−2. (6)
ernedbythesimilaritysolutionsofMacLow&McCray(1988)
(alsoseeWeaveretal.1977).Modelsofthecoldshellofthesu- Theluminosity,L38 (0.01,0.1,1,10),waschosenforOBas-
perbubbleswere performedusing Cloudy, version94.00 (Fer- sociations containing between 5 and 5,000 stars (e.g. McKee
land 1996), taking into account chemical and ionization con- &Williams1997;Bresolinetal.1998).Theage,t7(0.5,1,2,4),
ditions. The photoionizationsource consists of the ultraviolet was chosen to represent different points in the evolution of a
background from quasars (Haardt & Madau 1996) and the O bubble expanding for 50 million years (Mac Low & McCray
and B stars drivingthe expansionofthe superbubble. The in- 1988). The Cloudy models of the shell required that we also
tensity of the Haardt–Madau background is set to a constant input logn (−2.5 to 2, with intervals of 0.25). Cloudy was
e
hydrogen–ionizingphotonnumberdensityof10−5.2 cm−3 (as thenrunforeverycombinationofthechosenparameters(L38,
appropriate for z ∼ 1). The stellar contribution depends on t7, ne, and Z). The metallicity range (logZ ∼< 0) was based
the luminosity of the parent OB association and the radius of uponobservationalconstraints(see§5.2).Eachsimulationwas
the superbubble. Based upon our models, we find that for lu- haltedwhenthehydrogencolumndensity(N )calculatedin
tot
minosities greater than ∼ 1038 ergs s−1, the stellar spectrum Equation6wasreached.
BONDETAL. 5
5.2. SimulationResults on the shell density, n , but we can constrain it for particular
e
For each L38 and Z, Cloudy grids were generated. In Fig- combinations of L38 and Z. As can be seen in Figure 4, in-
ure4,weshowaplotofoutputgridsfromCloudyphotoioniza- creasing L38 raises the grid and brings higher shell densities
tionmodelsforthelogZ = −1andlogZ = −0.5cases. The intotheallowedparameterspace. Forexample,theL38 =0.1,
logZ = −1 grid has only −1.5 > logn > 1 cm−3 lying
thicknessoftheplaneparallelsupernovashellisplottedversus e
withintheallowedparameterspace.
its MgII columndensity. Eachpanelrepresentsa differentlu-
minosityoftheparentOBassociation(givenbyL38).Thesolid
linesrepresenttheagesofthemodelsuperbubbles,correspond- 6. CONCLUSIONANDDISCUSSION
ing (from left to right) to t7 = 4, 2, 1, and 0.5. The dashed The line pairs in the z = 0.7443 MgII absorption system
linesrepresentthehydrogennumberdensityoftheshell,which
in the spectrum of MC 1331+170are stronglysuggestiveof
rangesfrom(toptobottom)logn = −3tologn = 2cm−3
e e shellsofexpandingsuperbubbles.Theidentityofthehostofthe
in logarithmicintervalsof 0.25. We picked the regionsof pa-
z = 0.7443absorptionsystemisnotimmediatelyobvious,but
rameterspacewhichareintheobservedMgII columndensity
is likely to be the spiral galaxy(G5) seen in the HST/WFPC2
range(seeTable1)andareconsistentwithobservedsuperbub-
image (Figure 2). G3 and G4 are also candidates for being
bles (Heiles 1979). These regions are enclosed by the dotted
at z = 0.7443 and may form a compact group or pair with
boxes.
G5. However, kinematic models of galaxypairs and compact
Themodelsproducevaluesforne/n0overalargerange,the groupsseemtoyieldvelocityspreadslargerthantheobserved
maximum being ∼ 1000. We only plot modelsfor which the ∼ 200kms−1 andrarelyreproducethe symmetryseen in the
shellismoredensethantheISM(ne > n0,wheren0 isfound MC 1331+170 system. This evidence seems to suggest that
fromEquation5).Thiswillprobablybethecaseunlessthereis
the absorptionarises in a single galaxy,either fromsuperbub-
an extremeISM density gradient(Mac Low, McCray, & Nor-
bles or random cloud distributions. Photoionization models,
man1989)orwearepassingthroughacloudymedium(Silich
consideringtheconditionsoftheparentOBassociationandthe
etal.1996).Realsuperbubbleshaveshelldensitiesgreaterthan
surroundingmedium,supportthesuperbubblehypothesiswith-
fourtimestheISM densities, perhapsgreaterthanten timesif
outrequiringfine–tuningofthespecificconditionsfromwhich
theshellsareoverstable(Ryu&Vishniac1988).
thesuperbubblesarise.
Itmayseemcounter–intuitivethatMgII columndensityde-
The models of superbubbles do, in fact, produce the ob-
creasesastimeincreases,butthesuccessivet7gridlinesdonot served MgII column densities and velocity splittings of the
represent an evolutionary sequence. The decrease is a conse-
z = 0.7443 system, but there are still some discrepencies.
quence of holding shell speed constant for all models. In or- The ∼ 250 kms−1 spread of our profile is the most prob-
dertoproducethesamespeedatalatertime,theISMdensity
lematic feature in the context of disk superbubbles. A typi-
muchbe lower (providingless resistance). Thus, the different cal L∗ galaxy has a rotation velocity (V0) of ∼ 220 kms−1.
t7linesrepresentvaryinginitialconditions(seeEquation2). If If one assumes that the superbubblesfollow the disk rotation,
you have a smaller ISM density, then you will have a smaller
the maximum spread (V ) over which they could occur is
amountofmaterialsweptupintotheshell,resultinginasmaller max
MgIIcolumndensity. Vmax =V0(1−D/Rg),whereDistheimpactparameterand
R isthemaximumgalacticradiusatwhichsuperbubblescan
Increasing the luminosity of the clusters corresponds to a g
occur. The parameter R is likely to be approximately equal
shift of the grid toward higher MgII column densities and g
totheopticalradiusofthegalaxy. InM101,superbubblesare
thicker shells. Increasingmetallicity causes a shiftin the grid
seenoutto30kpc(Kamphuis1993).IfweassumethatR can
towardslowerMgIIcolumndensities. Wehavesetconstraints beaslargeas∼40kpc,wegetonlyV ∼100kms−1g. For
on the value of L38, so we can now constrain Z to fit ob- G5atz =0.7443,theimpactparametemraisx20h−1−1kpc.
servational parameter space (the dotted box). The three left–
There are severalpossible explanationsfor the large spread
handpanelsofFigure4showlogZ = −1,whichisconsistent
with the observed MgII column densities for L38 = 0.01 or in the z = 0.7443 MgII absorber: 1) A merger is inducing
globular cluster formation in the halo (Ashman & Zepf 1992;
L38 = 0.1. Agridwithanincreasedmetallicitywouldrequire
Zepf& Ashman1993)andthe clustersare creatingsuperbub-
a lower L38 (as apparent in the three right–hand panels). If
bles. 2) Superbubbles are forming within tidal debris from a
logZ >0,themodelsrequirealuminositythatistoolowforan
recentinteraction(Kniermannetal.2000;Gallagher2001). A
OBassociation(L38 <0.01).ThelargestOBassociationscon-
lineofsightthroughtidaldebriscancreatevelocityspreadsas
tain∼7000supernovaprogenitors(McKee&Williams1997),
largeas1000kms−1 (Gallagher2001). 3)Someofthesuper-
correspondingtoL38 ∼ 10(notshowninFigure4). Evenun-
bubblesareoutliersinvelocityspace,possiblylocatedwithina
der these extreme conditions, we cannot place a useful lower
companiondwarfgalaxy(MacLow&Ferrara1999). 4)Some
limitonthemetallicityintheshell.
of the absorption lines are not from superbubblesand lie out-
We can not obtain general lower and upper limits on t7
side of thedisk. Thisabsorptioncouldarise in halocloudsor
(though t7 < 5 is a physical limit in our models, see § 5.1).
indwarfgalaxies.
The ages are limited, however, for particular combinations of
Womble, Junkkarinen, and Burbidge (1991) point out that
L38 andZ. For example,the L38 = 0.1, logZ = −1gridin
largevelocityspreadsinedge–ongalaxiesarebetterexplained
Figure4hasonly1<t7 <4lyingwithintheallowedparame-
bychimneys,worms,orsuperbubblesblowingoutintothehalo.
terspace(dottedbox).Presumably,themultiplebubblesgiving
These are processesknownto occurat the end ofa superbub-
risetoourspectrumcouldarisefromthesameburstofstarfor-
ble’slifetimewhentheshellacceleratesthroughadensitygra-
mation and have similar ages and metallicities. For this to be
dientintothehalo. TheshellthenfragmentsduetoRayleigh–
true, a logarithmicspreadof 1.5isrequiredin L38 in orderto
Taylorinstabilitiesandthebubblebeginsexpellingmaterialout
reproducetheentirerangeofMgIIcolumndensities.
of the disk. This process is quite messy, however, and is not
Aswitht7,wecannotobtaingenerallowerandupperlimits
likelytogiverisetosuchevenlysplitlines.
6 EXPANDINGSUPERBUBBLESINQ1331+170
Our spectrum shows no more than six superbubbles within with respect to most of the ISM material. This separation in
the line of sight, each ∼ 10 pc thick (Mac Low, McCray, & velocityspacemayariseifthebubblesarelocalizedinacertain
Norman 1989); thus, there must be a substantial path length partofthediskorexistinastar–formingtidaltailnearby. The
throughthe ISM in between the superbubbles. Why don’twe linesat∼ −100kms−1 inFigure1showlesssymmetrythan
see dominating absorption from the ISM between the super- theothersandcouldbearisingfromISMcloudsinthedisk.
bubbles? The simplest explanationis that the filling factor of
the superbubbles is so large that the path length through the
ISMmakesanegligiblecontributiontotheabsorption. Obser- We acknowledge support from the NSF/REU program and
vations of HI holes in nearby galaxies (Kamphuis 1993) give NASA/LTSA through grant NAG5-6399. We would like to
maximum covering factors of ∼ 0.3; however, they are only thank Mordecai-Mark Mac Low for his insightful comments.
sensitive down to R ∼ 0.5 kpc. Another possible explana- WethankJanetGeoffroyfortheuseofherspectralintegration
tionisthatOBassociationsresideinregionsofenhancedden- codewhichweusedtocalculategalaxycolors. Wealsobene-
sityandthesuperbubblessweepthedensestmaterialintotheir fitfromusefulconversationswithMattBershadyNielBrandt,
shellbeforereachingthesparse, interveningmaterial. Finally, Robin Ciardullo, Jenn Donley, Rajib Ganguly, Dick McCray,
theparentOBassociationsthemselvesareatunusualvelocities JaneRigby,KenSembach,andSteinnSigurdsson.
REFERENCES
Ashman,K.M.&Zepf,S.E.1992,ApJ,384,50 Kanekar,N.,Chengalur,J.N.,Subrahmanyan,R.,&Petitjean,P.2000,A&A,
Boisse´,P.,LeBrun,V.,Bergeron,J.,&Deharveng,J.-M.1998,A&A,333,841 inpress(astro-ph/0012079)
Bomans,D.J.,deBoer,K.S.,Koornneef,J.,&Grebel,E.K.1995,A&A,313, Kniermann, K. A., Hunsberger, S. D., Gallagher, S. C., et. al., 2000, ApJ,
101 submitted
Bouche´,N.,Lowenthal,J.D.,Charlton,J.C.,Bershady,M.A.,Churchill,C. Kinney,A.L.,Calzetti,D.,Bohlin,R.C.,McQuade,K.,Storchi–Bergmann,T.,
W.,&Steidel,C.C.2000,ApJ,submitted &Schmitt,H.R.1996,ApJ,467,38
Bowen,D.V.,Blades,J.C.,&Pettini,M.1995,ApJ,448,634 Lane,W.M.2000,Ph.D.Thesis,RijksuniversiteitGroningen
Bowen, D. V., Tripp, T. M., & Jenkins, E. B. 2000, AJ, submitted (astro- Lanzetta,K.M.&Bowen,D.V.1992,ApJ,391,48
ph/0011134) LeBrun,V.,Bergeron,J.,Boisse´,P.,&Deharveng,J.M.1997,A&A,321,733
Bresonlin, F.,Kennicutt, R. C.,Ferrarese, L.,Gibson, B.K.,Graham, J.A., MacLow,M.-M.,&McCray,R.1988,ApJ,324,776
Macri,L.M.,Phelps,R.L.,Rawson,D.M.,Sakai,S.,Silbermann,N.A., MacLow,M.-M.,McCray,R.,&Norman,M.L.1989,ApJ,337,141
Stetson,P.B.,&Turner,A.M.1998,AJ,116,119 MacLow,M.-M.&Ferrara,A.1999,ApJ,513,142
Bruzual,A.G.,&Charlot,S.1993,ApJ,405,538 McClure-Griffiths,N.M.,Dickey,J.M.,Gaensler,B.M.,Green,A.J.,Haynes,
Charlton,J.C.,&Churchill,C.W.1998,ApJ,499,181 R.F.,&Wieringa,M.H.2000,AJ,119,2828
Chaffee,F.H.,Black,J.H.,&Foltz,C.B.1988,ApJ,335,584 McKee,C.F.&Williams,J.P.1997,ApJ,476,144
Churchill,C.W.,Vogt,S.S.,Steidel,C.C.,1995,Frommetal-lineabsorption Rao,S.M.&Turnshek,D.A.2000,ApJS,inpress
profiles to halo kinematics. In: Meylan G. (ed.) QSO absorption lines, Ryu,D.,&Vishniac,E.T.1988,ApJ,331,350
Springer,Berlin,p.153 Silich,S.A.,Franco,J.,Palousˇ,J.,&Tenorio-Tagle,G.1996,ApJ,468,722
Churchill,C.W.1997,Ph.D.Thesis,UniversityofCalifornia,SantaCruz Steidel,C.C.&Sargent,W.L.W.1992,ApJS,80,1
Churchill,C.W.,Mellon,R.R.,Charlton,J.C.,Jannuzi,B.T.,Kirhakos,S., Steidel,C.C.,Dickinson,M.&Persson,E.1994,ApJ,437,L75
Steidel,C.C.,&Schneider,D.P.2000a,ApJS,130,91 Steidel, C. C. 1995, in QSO Absorption Lines, ed. G. Meylan (Garching:
Churchill,C.W.,Mellon,R.R.,Charlton,J.C.,Jannuzi,B.T.,Kirhakos,S., Springer–Verlag),139
Steidel,C.C.,&Schneider,D.P.2000b,ApJ,543,577 Sungeon,K.,Dopita,M.A.,Staveley-Smith, L.,&Bessell, M.S.1999,AJ,
Churchill,C.W.&Vogt,S.S.2001,ApJ,submitted 118,2797
Ferland,G.1996,Hazy,UniversityofKentucky,InternalReport Vogt,S.S.,etal.1994,inProc.SPIEConf.2198,InstrumentationinAstronomy
Gallagher, S. C., Charlton, C. C., Hunsberger, S. D., & Zaritsky, D. 2001, VII,editedbyD.L.CrawfordandE.R.Craine(SPIE,Bellingham,WA),p.
submitted 362
Haardt,F.&Madau,P.1996,ApJ,461,20 Weaver,R.,McCray,R.,Castor,J.,Shapiro,P.,&Moore,R.1977,ApJ,218,
Heiles,C.1979,ApJ,229,533 377
Hickson,P.,deOliveira,C.M.,Huchra,J.P.,&Palumbo,G.G.C.1992,ApJ, Whitmore,B.,Zhang,Q.,Leitherer,C.,Fall,S.M.,Schweizer,F.,&Miller,B.
399,353 1999,AJ,118,1551
Kamphuis,J.1993,Ph.D.Thesis,UniversityofGroningen Womble,D.S.,Junkkarinen,V.T.,&Burbidge,E.M.1991,ApJ,388,55
Zepf,S.E.&Ashman,K.M.1993,MNRAS,264,611
BONDETAL. 7
TABLE 1
CLOUD PROPERTIES FORQ1331+170
Cloud MgII FeII MgI
hvi logN b logN b logN b
[kms−1] [cm−2] [kms−1] [cm−2] [kms−1] [cm−2] [kms−1]
−140.7 11.81±0.08 5.01±1.68 <11.89 ··· <10.80 ···
−117.9 13.15±0.01 9.09±0.16 12.82±0.03 8.33±0.93 <10.78 ···
−90.2 13.05±0.01 4.70±0.12 12.61±0.04 4.81±0.85 11.02±1.15 ∼9.5
−71.4 12.63±0.01 7.13±0.37 12.07±0.19 2.19±1.57 <10.80 ···
−47.9 12.62±0.08 6.29±0.59 11.62±0.06 0.31±0.08 <10.84 ···
−32.8 13.22±0.02 13.22±0.68 12.87±0.04 13.54±1.79 11.21±0.22 4.03±2.87
−6.9 13.26±0.02 4.03±0.20 13.25±0.04 3.57±0.26 11.81±0.04 5.50±0.80
2.1 13.03±0.04 10.33±0.58 12.54±0.09 3.87±1.09 <10.75 ···
37.6 12.01±0.06 5.88±1.19 <11.82 ··· <10.80 ···
63.0 12.21±0.04 8.65±1.10 12.52±0.22 39.46±22.98 <10.80 ···
103.4 13.08±0.02 13.33±0.56 12.83±0.06 12.21±1.71 11.21±0.28 12.81±8.38
121.4 13.08±0.02 7.77±0.23 12.45±0.08 6.15±1.24 11.09±0.30 ∼2.6
148.3 13.47±0.01 9.07±0.12 13.31±0.01 9.75±0.38 11.43±0.07 7.59±1.58
TABLE 2
Q1331+170FIELDANALYSIS SCENARIOS
Galaxy D R I 0.7443 1.3284 1.776
D Type Luminosity D Type Luminosity D Type Luminosity
[′′] [kpc] [L∗] [kpc] [L∗] [kpc] [L∗]
2 0.75 24.9 23.8 3.61 ··· ··· 4.36 E 3.57 4.63 E 14.5
S0 5.14
3 1.58 25.1 24.4 7.61 S0 0.08 9.18 S0 1.09 9.76 S0 4.28
E 0.08
4 2.86 24.2 24.2 13.77 SB 0.08 16.62 SB 0.163 17.67 SB 0.208
Im 0.12 Im 0.498 Im 0.822
5 3.86 21.40 21.47 18.58 SB 1.16 22.43 SB 2.12 23.84 SB 2.71
8 EXPANDINGSUPERBUBBLESINQ1331+170
| | | | || || | | | | |
1
0
| | | | || || | | | | |
1
0
| | | | || || | | | | |
1
0
| | | | || || | | | | |
1
0
| | | | || || | | | | |
1
0
-200 -100 0 100 200
FIG. 1.— The HIRES/Keck spectrum of the z=0.7443 system toward Q1331+170showing transitions detected at > 5σ aligned
in rest frame velocity space. The solid lines representa model from a simultaneousVoigt profile fit to the spectra, with the ticks
markingindividualcomponents.
BONDETAL. 9
FIG.2.— Twoimages,eachwithadifferentdynamicrange,oftheMC1331+170fieldcombiningtheR(F702W)andI (F814W)
filters. Galaxies are numbered G3 through G5, in order of distance from the quasar (at center). G2 is not labeled because of its
proximitytothequasarandcanonlyberesolvedviaPSFsubtraction.Theleft–handpanelemphasizesG3andG4andtheright–hand
panelemphasizesG5(forpresentationpurposesonly).
10 EXPANDINGSUPERBUBBLESINQ1331+170
1
.5
0
0 .2 .4 .6 .8 1 1.2 1.4 1.6 1.8 2
FIG. 3.— A color-redshift diagram plotted of model spectral energy distributions representing a variety of morphologicaltypes
(Bruzual & Charlot 1993; Kinney et al. 1996). The dotted vertical lines denote the redshifts of the absorption systems in the
spectrumofQ1331+170andthedashedhorizontallinesdenotethecolorsofthesourcesintheopticalfield.