Interpreting the ionization sequence in AGN emission-line spectra

28 Pages·2013·5.91 MB·English

MNRAS437,2376–2403(2014) doi:10.1093/mnras/stt2056
AdvanceAccesspublication2013November26
Interpreting the ionization sequence in AGN emission-line spectra
Chris T. Richardson,1,2‹ James T. Allen,3‹ Jack A. Baldwin,1‹ Paul C. Hewett4‹
and Gary J. Ferland5‹
1Physics&AstronomyDepartment,MichiganStateUniversity,EastLansing,MI48864-1116,USA
2PhysicsDepartment,ElonUniversity,Elon,NC27244,USA
3SydneyInstituteforAstronomy,SchoolofPhysics,UniversityofSydney,NSW2006,Australia
4InstituteofAstronomy,UniversityofCambridge,MadingleyRoad,CambridgeCB30HA,UK
5PhysicsandAstronomyDepartment,UniversityofKentucky,Lexington,KY40506-0055,USA
D
o
w
n
Accepted2013October22.Received2013October11;inoriginalform2013April6 loa
d
e
d
fro
ABSTRACT m
h
Weinvestigatethephysicalcauseofthegreatrangeintheionizationlevelseeninthespectra ttp
s
ofnarrow-linedactivegalacticnuclei(AGN).Weusedarecentlydevelopedtechniquecalled ://a
meanfieldindependentcomponentanalysistoidentifyexamplesofindividualSloanDigital ca
d
Sky Survey galaxies whose spectra are not dominated by emission due to star formation em
(SF), which we therefore designate as AGN. We assembled high signal-to-noise ratio (S/N) ic.o
u
compositespectraofasequenceoftheseAGNdefinedbytheionizationleveloftheirnarrow- p
.c
line regions (NLR), and extending down to very low ionization cases. We then used a local o
m
optimally emitting cloud (LOC) model to fit emission-line ratios in this AGN sequence, /m
n
including the weak lines that can be measured only in the co-added spectra. These weak ra
s
line ratios provide consistency checks on the density, temperature, abundances and ionizing /a
continuum of Seyfert galaxies determined from strong-line ratios. After integrating over a rticle
widerangeofcloudsatdifferentradiianddensities,ourmodelsindicatethattheradialextent -a
b
s
oftheNLRisthemajorparameterindeterminingthepositionofhighertomoderateionization tra
c
AGNalongoursequence.Thisprovidesaphysicalinterpretationfortheirsystematicvariation. t/4
3
HigherionizationAGNcontainoptimallyemittingcloudsthataremoreconcentratedtowards 7
/3
thecentralcontinuumsourcethaninlowerionizationAGN.OurLOCmodelsindicatethatfor /2
3
thespecialsetofobjectsthatlieonourAGNsequence,theionizingluminosityisanti-correlated 76
withtheNLRionizationlevel,andhenceanticorrelatedwiththeradialconcentrationandactual /10
2
physicalextentoftheNLR.Apossibleinterpretationthatdeservesfurtherexplorationisthat 66
9
theionizationsequencemightbeanagesequencewherelowionizationobjectsareolderand 1
b
have systematically cleared out their central regions by radiation pressure. We consider the y g
u
alternativethatourAGNsequenceinsteadrepresentsamixingcurvecombiningSFandAGN e
s
spectraindifferentproportions,butarguethatwhilemanygalaxiesinfactdohavethistype t o
n
ofcompositespectra,ourAGNsequenceappearstobeaspecialsetofobjectswithnegligible 0
1
SFexcitation. A
p
Key words: galaxies: active–galaxies: evolution–galaxies: nuclei–galaxies: Seyfert– ril 2
0
1
galaxies:structure. 9
thesequenceofAGNgalaxiesdefinedbyMFICA,andfinallythe
1 INTRODUCTION
goalsofthispaper.
Thispaperexploresthenatureofthenarrow-lineregion(NLR)in
activegalacticnuclei(AGN)bycomparingplasmasimulationsto 1.1 Thenarrow-lineregion
theresultsobtainedfromarelativelynewandnoveltechnique,mean
The NLR of AGN is a region of ionized and neutral gas gener-
fieldindependentcomponentanalysis(MFICA).Thisintroduction
firstdescribesthepresentunderstandingoftheNLR,followedby allyclassifiedbystrong[NII]λ65841 and[OIII]λ5007emission.
(cid:2)
E-mail: [email protected] (CTR); [email protected] (JTA); 1PaperIdesignatedemissionlinesbytheirvacuumwavelengths,butherewe
[email protected](JAB);[email protected](PCH); switchtothemorecommonlyusedairwavelengthsfortransitionslongward
[email protected](GJF) of2000A.
(cid:2)C 2013TheAuthors
PublishedbyOxfordUniversityPressonbehalfoftheRoyalAstronomicalSociety
Interpretingtheionization sequenceinAGNspectra 2377
D
o
w
n
lo
a
d
e
d
fro
m
h
ttp
s
://a
c
a
d
e
m
Figure1. TheBPTdiagram,withoursampleofgalaxiesplottedasgreydots.ThesolidbluelineisthetheoreticalupperlimitforSFgalaxiespresentedby ic
Kewleyetal.(2001).ThedashedbluelineistheclassificationcurveusedbyKauffmanetal.(2003)asalowerlimitforfindingAGN.Thedottedblueline .o
u
showsthedivisionofAGNandLINERsfromKauffmanetal.(2003).ThesolidredcurveistheAGNlocusdiscussedinSection1.2,andthedashedredcurve p
.c
istheSFlocus.TheredsymbolsshowoursequencesofAGNpickedoutbyMFICAwitha4jrepresentingthehighestionizationobservations,andalsothe o
m
positionofthelow-ionizationSFsubsamples01thatisusedinSection5.4. /m
n
ra
s
Incontrasttotheverysmallbroad-lineregion(BLR),theNLRisof ies called ‘transition’ or ‘composite’ objects which merge down /a
theorderof103pcinsizeandcontainsrelativelylow-densitygas, intotheSFgalaxiesontheseline-ratiodiagrams(Ho,Filippenko rtic
withelectrondensitiesne∼104cm−3.However,thephysicalpicture &Sargent1993;Kauffmanetal.2003;Kewleyetal.2006).These le-a
oftypicalNLRsremainsunclear.Animportantrouteforadvancing lie between the dashed and solid blue lines in Fig. 1. They may bs
ourunderstandingofthisregionistocompareobservedemission- beamixofSF,AGNandcompositeSF−AGNcases,introducing tra
c
linestrengthstotheonespredictedbymodelssimulatinganentire considerableconfusionabouttheionizationmechanismoveranex- t/4
3
NLR.Diagramsconstructedfromemission-lineintensityratiosare tensive part of the full BPT diagram. Our study here includes an 7
/3
widelyusedforthissortofcomparison.Earlyempiricalworkby investigation of the relationship between these composite objects /2
3
Baldwin,Phillips&Terlevich(1981)pointedouttheimportanceof andtypicalhigher-ionizationTypeIIAGN. 7
6
the[OIII]λ5007/Hβversus[NII]λ6584/Hαdiagram(hereafterthe Wenowturntowhatisknownaboutthephysicalnatureofthe /10
BPTdiagram)forclassifyingemission-linegalaxiesaseitherAGN NLR.Therehavebeenmanystudiesusingtheobservedintensityra- 2
6
6
orstarforming(SF).Theopticaldiagnosticdiagramspresentedin tiostoidentifytheionizationsourceinnarrow-lineSeyfertsandSF 9
1
Veilleux & Osterbrock (1987; hereafter VO87) added further im- galaxies.Photoionizationmodelshavebeenremarkablysuccessful b
y
portant intensity ratios that are minimally affected by reddening. inreproducingtheobservedemission-lineratiosinhigh-ionization g
u
More recent work has extended this method into the ultraviolet AGN(Fergusonetal.1997,hereafterF97;Komossa&Schulz1997; e
s
(UV)andinfrared(IR)[Spinoglio&Malkan1992;Allen,Dopita G04b).Theexcitationmechanismforlow-ionizationNLRinthere- t o
n
&Tsvetanov1998;Sturmetal.2002;Groves,Dopita&Sutherland gionoverlappingwithSFgalaxiesremainsuncertainanditisnot 0
1
2004a,b(hereafterG04aandG04b,respectively)]. evenclearwhetherallsuchobjectscontainAGN.Photoionization A
p
herFeig(.se1esSheocwtisona1B.P2)Tpdloiatgterdamaswgirtehythdeotssa.mWpeledoisfcugsaslatxhieeslaursgeedr mhigohd-eelsxcoiftaAtiGonNahnadvelofwoc-euxsceidtaotniomnagtaclhaixnigesthbeutemnoinsseiohnasfraotmtemeipthteedr ril 20
1
symbolsandthelabelledredlinesbelow.Theunlabelledbluelines tosimultaneouslymatchboth.Modelsthatincorporateshockexci- 9
show commonly used boundaries to separate different classes of tation(Dopita&Sutherland1995,1996)canaccountforLINERS
objects.ThebluedashedcurveistheboundarybetweenSFgalax- butfailtofithigh-ionizationAGN.
ies and AGN defined by Kauffman et al. (2003), while the blue ThestructureoftheNLRisanotheropenquestion.Threedifferent
solidlineisthetheoreticalupperlimitonSFlineratiosfoundby modelsthatcanreproducetheobservedemissionfromtheNLRof
Kewley et al. (2001). An additional class of objects is the low- Seyfertsare:(1)acombinationofmatter-andionization-bounded
ionizationgalaxiescalledLINERs(Heckmann1980).LINERsare clouds(Binetteetal.1996);(2)radiationpressuredominateddusty
characterized by very strong emission lines from neutral species, clouds(G04a)and(3)locallyoptimallyemittingclouds(F97;see
andarewellseparatedfrombothAGNandSFgalaxiesintheBPT alsoKomossa&Schulz1997).
and similar line-ratio diagrams. The blue dotted diagonal line in Themodelthatcombinesmatter-andionization-boundedclouds
Fig.1istheboundarybetweenSeyfertsandLINERSsuggestedby isparametrizedbytheratioofthesolidanglessubtendedbymatter-
Kauffman et al. (2003), although this definition was later refined andionization-boundedclouds.Thematter-boundedcomponentis
(Kewleyetal.2006).Finally,thereareotherlow-ionizationgalax- responsibleforhighionizationlines,namelyHeIIλ4686,andthe
2378 C.T.Richardsonetal.
ionization-boundedcomponentisresponsibleforlowtointermedi- Fig.2showsthespectraofthesefiveemission-lineMFICAcom-
ateionizationlines.Thedusty,radiation-pressure-dominatedmodel ponents.Thecomponentswerethenfittedtothelargersampleof
isbasedontheprinciplethatradiationpressureincidentongrainsis ∼104galaxies.Theeffectoffittingthecomponentsistoprojecteach
sufficienttomoderatethedensity,excitationandsurfacebrightness galaxy on to the five-dimensional space of component weights,
of high ionization regions. Finally, the locally optimally emitting which we designate W −W . The weights were normalized such
1 5
cloud(LOC)modelassumesthattheemissionlinescomefroma that they total unity for each galaxy. By performing this fit on a
vast sea of clouds distributed over a wide range of densities and large number of galaxies we were able to examine their distribu-
radialdistancesfromacentralsourceofionizingradiationandthat tioninthespaceofcomponentweights,whichthroughtheMFICA
theobservedemission-lineintensityratiosaretheresultofapow- componentsthemselvescorrespondsdirectlytotheirdistributionof
erful selection effect: emission lines are produced in clouds that observedphysicalproperties.
optimallyemitthem. Galaxies that lie below the SF classification line of Kauffman
Each of these models can successfully represent the NLR in etal.(2003)areexpectedtobedominatedbystarformation(SF);
at least some galaxies, but it is not clear whether any of them Stasin´skaetal.(2006)showedthatnomorethan3percentoftheir D
o
canexplainallNLRs.Thelow-ionizationAGN,whichweinclude emission-line flux arises from AGN. These galaxies form a tight w
n
in this present study, fall in a region of the BPT diagram that is ‘SF’sequenceintheBPTplane(the‘SFlocus’showninFig.1), lo
a
alsooccupiedbylowionizationSFgalaxies(cf.Tanaka2012a,b). andasexpectedthiscorrespondstoasimilarlytightsequenceinthe d
e
TorhinsootvoenrleapgewnietrhalSFtypceaseosfgmreoadtelyl ccaonnfeuxspeslaitnhethisesufeulolfsewqhueetnhceer MFIsIoClAatiwngeiaghstasmsppaleceo.f pure AGN presented a greater challenge, d from
from low-ionization non-SF objects to the much more luminous because the BPT diagram does not allow for easy identification h
high-ionizationobjectssuchasclassicalSeyfertgalaxies.Thelow- ofarepresentativesampleofAGNwithoutsignificantcontamina- ttp
s
ionization AGN may or may not include cases where the ioniza- tionfromSF.However,thegreaterdimensionalityoftheMFICA ://a
c
tion is due to a composite non-thermal and SF spectrum. What weights does allow a sample to be identified. Fig. 3 shows the a
d
isneededisanunbiasedsetofAGNthatspanthefullionization two-dimensionalprojectionsinvolvingthefirstthreeMFICAcom- em
range. ponents(additionalprojectionsinvolvingW andW areshownin ic
4 5 .o
fig. 15 of Paper I). Fig. 3 shows that the SF sequence (the blue u
p
line on the figure) has a significant contribution from component .c
1.2 Theactivegalacticnucleisequence o
2 for almost all of its length, excepting only the end at which it m
InniqAuelletnhaettiasl.a(b2le01t3o;sheepraeraaftteerAPGapNerfrIo)mweSdFevgealloapxeiedsawniethwmteucchh- ccoannvbeergeexspetocteWd3to=ha1v.e0.emHiesnscioen, glianleasxiwesithwviethryloliwttlevaclounetsribouftiWon2 /mnras
gasretahteorsepubraistyedthoannBisPTusduiaalglyraamcshioevrepdriwncitihpaollcdoemr mpoenthenotdsansaulcyh- afrlosomhSaFs.sItnrodnegedc,otnhterribeuitsioanssefqruoemncceomofpgoanleanxtises4wanitdh5lo,windWic2attihnagt /article
sis.ThisMFICAtechniqueallowedustoseparategalaxiesalonga emissionsourcesotherthanSF.Frominspectionofthedistribution -a
b
star-forming(SF)locusfromthosealonganactivegalaxy(AGN) ofMFICAweightswechoseW2≤0.05asourfirstAGNcriterion. stra
locusextendingdowntoevenquitelowionizationlevels.Weused Asecondcriterionisrequiredinordertodefinethelow-ionization c
asampleofabout104 low-redshift(0.1<z<0.12)SloanDigital limit of the AGN sequence (defined below), i.e. how close to the t/43
7
SkySurvey(SDSS)emission-linegalaxiestoisolatesequences,or SFsequenceitisallowedtoextend.Tothisend,werequiredthat /3
‘loci,’ of pure SF and of AGN cases. Full details of the MFICA W4+W5 ≥ 0.18. The value of the cut-off was selected by trying /23
7
analysisaregiveninPaperI;herewegiveasummary,focusingon a range of values, constructing an AGN locus for each one, and 6
/1
the derivation of the AGN locus and of the increasing ionization attempting to reconstruct the input sample in terms of summed 0
2
sequencethatliealongthatlocus. contributionsfromtheAGNandSFloci.Forcutoffvaluesgreater 66
9
Allemission-linegalaxiesusedinthisstudywereselectedfrom than 0.18, there was a population of galaxies that could not be 1
the redshift range 0.10 ≤ z < 0.12. Those used to generate the describedbyacombinationofthetwoloci,indicatingthattheAGN by
g
emission-linecomponents,asampleof727,wereselectedtohave sequencewasinsufficientlylonginthosecases.Whenacutoffvalue u
e
maHnαoddapenroadstiet[ilNvyeIhI]eigqλhu6irv5-a8bl4ae,nndtwwistihigdntSha/lN-(tWo≥-λn)o5fi.os0erferoaartcihothoe(1ffl6H.u0βx≤,m[OSea/NIsIIu]r<eλm520e30n.07ts), oo0ff.10c8.o1pn8rtrowidbauusctuieosdnensdo,ftrhfouemrstehthgeeralitamwxpoierlsoovccoei.muRledenbdte,uscdoiens0gc.1rtih8beewdcaubstyoafadfocvopamtleudbeifnboaertlitoohwne st on 01
of Hβ and [NII] λ6584. Their Hα line widths were in the range followinganalysis. Ap
1.9Å ≤ σHα < 3.0Å. The ∼104 galaxies used to define the Applyingtheabovecriterialeaves5519galaxiesdominatedby ril 2
SF and AGN loci were allowed a wider range in r-band S/N SF,and379dominatedbyanAGN,withtheremaining4221galax- 01
(15.0≤S/NR <30.0)andlinewidth(1.9Å≤σHα <3.5Å),but iesfromouroverallsamplefallingoutsidetheBPTSFregionbutnot 9
otherwisefollowedthesamecriteria.Thislattersampleisplotted classifiedasAGNbecauseeitherW2>0.05or(W4+W5)<0.18.
inFig.1. Oftheserejectedgalaxies,986(23.3percent)haveanAGNcontri-
Before analyzing the emission lines, an MFICA analysis was bution(i.e.W4+W5>0.18)butW2>0.05.
usedtosubtracttheunderlyingstellarcontinuumfromeachgalaxy Thedistributioninthefive-dimensionalspaceofMFICAweights
spectrum.Asampleof393galaxies,chosentolieintheSFregion wasthenparametrizedseparatelyfortheSFandAGN-dominated
oftheBPTdiagramasdefinedbyKauffmanetal.(2003),werethen subsamples, using the algorithm described in Newberg & Yanny
used to generate three emission-line MFICA components. These (1997,hereafterNY97).Thealgorithmdescribesadistributionof
threecomponentsaresufficienttodescribetheemission-linespectra pointsasalocusmadeofmanysegments.Eachsegmentisaprism
of SF galaxies, when combined linearly. The broader sample of withanellipticalcross-section,orahigher-dimensionalequivalent,
727emission-linegalaxiesofalltypeswasthenusedtogenerate surroundingacentral‘locuspoint’.Startingfromapairofendpoints
a further two components; the combined set of five components specified after visual inspection of the distribution, the algorithm
allowsthedescriptionofAGNspectra. addssegmentsonebyone,stoppingwhenthedistancebetweenthe
Interpretingtheionization sequenceinAGNspectra 2379
D
o
w
n
lo
a
d
e
d
fro
m
h
ttp
s
://a
c
a
d
e
m
ic
.o
u
p
.c
o
m
/m
n
ra
s
/a
rtic
le
-a
b
s
tra
c
t/4
3
7
/3
/2
3
7
6
/1
0
2
6
6
9
1
b
y
g
u
e
s
t o
n
0
1
A
p
ril 2
0
1
9
Figure2. Spectraforthefiveemission-linecomponentsgeneratedbyMFICA.Panel(a)showsthefullspectra.Theotherpanelsshowenlargementsthat
slightlyoverlapinwavelength.TheemissionlineslistedinTables2and3areidentifiedbytheirions.‘bl’indicatesablendedline.Allfluxvaluesareshown
inunitsofthepeakHβintensityintheparticularcomponent.
existing locus points is less than a fixed multiple, Nσ , of the weresparser.TheNY97algorithmallowsforamaximumdistance
spacing
widthofthelocus.WeusedNσ =3.0,thetypicalvalueusedin betweendatapointsandlocuspointstobeset,beyondwhichthe
spacing
NY97.Weaddedanadditionalcriterionthatanewlocussegment datapointsdonotcontributetothepositionofthelocuspointorto
would only be inserted between two existing segments if there thesizeofthecross-section.Forthedistributionsexaminedhere,
wereatleast20datapointsbetweenthetwoexistinglocuspoints. this criterion was found to make no significant difference to the
Thiscriterionimprovedstabilityinregionswherethedatapoints parametrizedlocus,soitwasnotapplied.
2380 C.T.Richardsonetal.
D
o
w
n
lo
a
Figure3. DistributionsoftheMFICAweightsforthefirstthreecomponents.InthisrepresentationtheAGNlocusappearsasthesolidredline,andtheSF de
d
lthoceutsexistathse‘wdainshges’d(baliu1ealnindea.iT2)hewoovualdlsfaollolonwgethaechoulotecrusboreupnrdeasreinetstohfetshceasteteorvoaflsdfaotartphoeinAtGssNtillloicduesn.tTifiheedloascaAtiGonNsoorfpthuereASGFN,asnudbtsheetssdeeqsucernibceedsdinesScericbtieodnin2 from
aremarkedbyblacksquares,andthelocationofthelow-ionizationSFsubsets01usedinSection5.4ismarkedbyablackcross. h
ttp
s
Aswellasdefiningthecentralspineoflocuspoints,theNY97 includesonlysomeoftheobjectsintheAGNandcompositeregions ://a
c
algorithmdefinesan(possiblyhigher-dimensional)ellipticalcross- oftheBPTdiagram.Thissequenceappearstohavespecialsignif- a
d
sectionaroundeachlocuspoint,whichdescribesthescatterofdata icance in a statistical way, so we want to understand its physical e
m
pointsinthatregion.Thealgorithmdefinestheaxesoftheellipseto interpretation. One end of the sequence contains high-ionization ic
.o
beequaltotherootmeansquare(rms)ofthedistancebetweenthe NLRs that are unambiguously classified as AGN, but the AGN u
p
spineandeachdatapoint.Inthiswork,wemultiplythermsvalues sequence descends down into the low-ionization transition object .c
o
by1.5toencompassagreaterfractionofthedatapointswithinthe regioninawaythatrejectsSFgalaxieswithsimilaremission-line m
/m
locus,andrefertotheseincreasedvaluesasthelocuswidth. spectra.SincetheAGNsequenceismainlyamonotonicprogres- n
Withthesetechniques,weidentifiedtwostatisticallyindependent sionalongasingleAGNlocus,weinvestigatethepossibilitythata ra
s
loci of observed galaxies, threading through the five-dimensional singlephysicalparameter,onethatismorefundamentalthanjustthe /a
component space. We called these the ‘SF’ (star-forming) and resultingionizationleveloftheNLR,isresponsiblefortheposition rticle
‘AGN’loci.ThoselociareshowninFig.3.Combinationsofspectra ofanobjectalongthesequence.Thisdefinesthegoalofthispaper: -a
b
fromtheSFandAGNlocicanreconstructthespectraofgalaxies weseektofindasingletunablephysicalparameterresponsiblefor stra
thatliebetweenthetwoloci.Thereisconsiderableseparationbe- thevariationinAGNrangingfromSeyfertsallthewaydowntothe c
tweentheloci,exceptattheverylowest-ionizationendoftheAGN lowest-ionizationAGN. t/43
7
locus.ThisdegreeofseparationshowsthatAGNgalaxiesfallalong As a check on this, we will also initially include objects that /3
asinglelocussuggestingthattheirvariationcanberepresentedby represent the observed scatter orthogonal to the AGN sequence. /2
3
7
asinglefreeparameter.Werefertothesequenceofgalaxiesthatlie However,themeasuredemission-lineratiosdescribedbelowshow 6
/1
atpointsalongtheAGNlocuswithincreasingvaluesofthatfree that there is little spread in properties perpendicular to the AGN 0
2
parameterasthe‘AGNsequence’. sequence, over the range in parameter space that the Newberg & 6
6
We measured emission-line intensity ratios for subsamples of Yanni(1997)algorithmhasidentifiedaslyingwithinthegroupof 91
galaxieslyingalongtheSFandAGNloci(Section2)andmapped AGNwithlittlecontributionfromSF. by
thelocibackontotheBPTdiagraminFig.1,wheretheyareshown The open squares in Fig. 1 define a sequence leading up from gu
e
as the two red curves labelled ‘SF locus’ and ‘AGN locus’. The the HII region into the AGN region of the BPT diagram. Differ- st o
squares in Figs 1 and 3 represent the positions of the individual encesingalaxies’locationsalongthisdirectionareoftenascribed n
galaxysubsetsthatlieontheAGNlocus,whilesmallertriangles todifferencesintheamountofmixingbetweenphotoionizationby 01
anddiamondsshowthepositionsoftwoparallelsequenceslyingat hot stars and photoionization by an AGN continuum. Given that Ap
eitherendofthemajoraxisoftheellipses,representingtheerror, theMFICAtechniquecantellthedifferencebetweenSFandAGN ril 2
as identified by the NY97 algorithm. We also show the position objects at points well down into what is called the composite re- 01
9
ofalow-ionizationsubsample(calleds01)ontheSFlocus,which gion,weexploreherethealternativethatallobjectsontheAGN
we will use later in the paper (Section 5.4). In Fig. 1 and other locushaveNLRsinwhichthephotoionizationisdominatedbyan
line-ratiodiagrams,theAGNlocusdoesnotpassthroughtherealm AGN-like ionizing spectrum, and try to determine if some other
of LINERs. Rather, it starts out below the LINER region in the underlying physical variable causes the range in properties along
area occupied by transition objects, and then angles up into the theAGNsequence.
AGN region. Fig. 1 clearly shows that the [OIII]/Hβ ratio is one InSection2wedescribecompositeobservedspectraformedat
parameterthatlocatesgalaxiesalongtheAGNlocusandtherefore fiveroughlyequallyspacedpointsalongtheMFICAAGNlocus,
thattheionizationparameterisakeyfactorinthedefinitionofthe andattwoadditionalpointstoeithersideoftheAGNsequenceat
AGNsequence. each of those positions on the sequence. These composites range
from very low ionization objects up through very highly ionized
1.3 Thegoalsofthispaper
TypeIIAGN,andhavesufficientlyhighS/Nthatmanyadditional
TheMFICAtechniquehaspickedoutadistinctsetofgalaxies−the observedlineratiosinvolvingweaklinescanbemeasuredascon-
AGNsequence−whicharedifferentfrompurelySFgalaxies,but sistency checks in addition to using the usual strong line ratios
Interpretingtheionization sequenceinAGNspectra 2381
whichhavebeenusedtoconstrainpreviousmodels.Noneofthese Selectedregionsoftheco-addedspectraofthefivecentral(j=1)
compositespectrashowsanyevidenceofanon-thermalcontinuum, subsetsareshowninFig.4.Theadditionalsubsets(j=0,2)arenot
except for their emission lines. Then in Section 3, following F97 showninFig.4duetotheirclosesimilaritytothecentralsubset.
weadoptanLOCmodelanduseittoreproducetheobservational We then measured emission-line strengths from the co-added
line-ratiodiagrams.Weinvestigatethesensitivityofthelineratios spectrumforeachsubset.Formostlines,wejustintegratedtheflux
tovariousphysicalparameters.Thephysicalinterpretationofour withinthelineprofileabovealocallyfittedcontinuum.Theaccuracy
results is discussed in Section 4. This will include a comparison wasabout±5percentforthestrongerlinesrangingto±20percent
of the results from our LOC models to those for a purely empir- fortheweakestlinesbasedontheS/Nintheadjacentcontinuum.We
ical ‘mixing model’ that combines spectra from SF galaxies and separatedthe[SII]doubletbysimplydividingtheblendedprofile
from a high-excitation AGN. Finally, Section 5 summarizes our atthelowestpointbetweenthetwolines.HeI5876isonthewing
conclusions.WewilladdresstheSFlocusinafuturepaper. of Na D, which is the one absorption feature in the underlying
galaxyspectrumthatobviouslydidnotsubtractoffproperlyalong
withtherestofthecontinuum.OurHeIλ5876measurementsare D
2 A COMPARISON SAMPLE AT based on fitting the HeI line with the profile of the Hβ emission ow
n
REPRESENTATIVE POINTS ALONG line,withatypicaluncertaintyofabout10percent.Table2liststhe lo
a
THE AGN LOCUS observed(reddened)value,whileTable3liststhedereddenedvalues d
e
Compositespectrawereformedfromsubsetsofgalaxieslyingalong caussruvmei(nCgarfdoerllsii,mCpllaiycittoyn, a&sMtanadthairsd1G9a8l9a)ctwicitRhvE=(B–3V.1)rcehdodseennintgo d from
theAGNsequenceandtoeithersideofit.Theyarenamedaij,where produce dereddened I(Hα)/I(Hβ) = 2.86, appropriate for Case B h
thefirstindexindicatesthepositionalongtheAGNlocusranging recombination at n = 102cm−3 and T = 104K (Osterbrock & ttp
fromi=0atthelow-ionizationendtoi=4atthehigh-ionization Ferland 2006, hereeafter AGN3). The deereddened values are the s://a
end.Thesecondindexisthepositioninadirectionorthogonalto c
onesusedthroughouttheremainderofthepaper.TheS/Nobtained a
tSsohenFeqtusAheeeGqncucNoeenn.lvocFeceinugats.ni,o1dwnsjaiht=lhoBwj2P=sTcwo0drhicreaeorgsreprraeotmhsnped.soiNennodgptiotenoignth‘ttwsaot(itonshhngeo’t‘whfwuenirBntahgPse’Ttrcrfidlaroionasmgegslretatshtmo)efttSahhFleel fc[srNuaosmmIeI]so,cλfpo5tr-h7oae5vd5id[dS,iinHnIgI]gesλiIpmλλe4pc60to6r6ra7t88aa,n4lalt0onc7wdo6sn[dOswoisueItIab]eknlλecl7tyi)3n,c2e[hAs5esrtcuoIkVcsb]heoλan4ms7o[e1uSa1rsI,Ium][rNλoed4dI]0eil7λns05.m2(0toh0set, demic.oup.c
lowestionizationsubsets(a0j)fallveryclosetothelinerepresenting om
the Kauffman et al. (2003) lower boundary for finding AGN and /m
n
arewellbelowthelinerepresentingtheKewleyetal.(2001)upper 3 DUST-FREE LOC MODELS ras
limitforSFgalaxies.Table1listssomegeneralpropertiesofthese /a
1w5hoasijesMubFsIeCtsA,ewacehigohftswmhiocshtccolonstaeilnysmaastacmhepdlethoeftsheele5c0tedgaplaoxinietss 3.1 Method rticle-a
along the AGN locus. The E(B–V) values were determined from We used the LOC model, which treats the NLR as the sum of a bs
the Hα/Hβ intensity ratio as described below. We then list the large number of separate gas clouds spread out around a central tra
c
observedemission-lineluminositiesL(Hβ)andL([OIII]λ5007)and ionizing source (Baldwin et al. 1995). Once the incident spectral t/4
tbhyetchoenctionrureusmpolunmdiinngosdietyreadtdλe5n0ed07v,aLluλ(eλs5L0c0(H7)β.T),hLecs(e[OarIeIIf]oλll5o0w0e7d) emnaejrogryfdreisetrpibauratimonet(eSrsEDin)tahnedLthOeCchmemodiceallaarbeutnhdeandcisetsriabruetisoent,sthoef 37/3/2
and Lcλ(λ5007), and the observed equivalent widths Wλ of Hβ theindividualcloudsintheirradialdistancesrfromthesourceof 376
and[OIII].Finally,thetableliststherelativeweightingsofeachof ionizingradiation,andintheirgasdensitiesnH. /10
theMFICAemission-linecomponentswhenfittedtotheco-added Theindividualcloudsweremodelledusingversion10.0ofthe 26
6
spectrumrepresentingeachindividualsubsetalongtheAGNlocus. plasmasimulationcodeCLOUDY(Ferlandetal.1998).Forsimplicity, 9
1
b
y
Table1. Propertiesoftheaijsubsets. g
u
e
s
Subset a00 a01 a02 a10 a11 a12 a20 A21 a20 a30 a31 a32 a40 a41 a42 t o
n
0
E(B–V) 0.80 0.83 0.68 0.43 0.49 0.56 0.40 0.43 0.45 0.30 0.33 0.39 0.20 0.20 0.21 1
ObservedL(Hβ)1 1.8 2.1 2.4 2.0 2.2 2.6 3.1 2.9 3.2 3.1 2.4 2.8 2.9 2.9 2.8 Ap
ObservedL([OIII]5007)1 0.95 0.85 0.92 3.1 3.6 3.7 9.2 9.6 9.1 16. 12. 13. 24. 25. 24. ril 2
ObservedLλ(λ5007)2 0.47 0.49 0.52 0.45 0.52 0.58 0.50 0.67 0.60 0.60 0.49 0.53 0.52 0.50 0.50 01
DereddenedLc(Hβ)1 29. 38. 25. 9.0 12. 18. 12. 13. 15. 8.8 7.7 11. 5.9 5.9 5.7 9
DereddenedLc([OIII]5007)1 14. 14. 8.8 13. 18. 24. 35. 40. 41. 43. 35. 47 47. 49. 48.
DereddenedLcλ(λ5007)2 6.97 8.14 4.94 1.88 2.60 3.77 1.89 2.80 2.70 1.62 1.42 1.91 1.02 0.98 0.99
Wλ(Hβ)(Å) 3.81 4.25 4.64 4.47 4.23 4.48 6.23 4.32 5.35 5.13 4.94 5.30 5.55 5.78 5.65
Wλ([OIII]5007)(Å) 2.01 1.72 1.78 6.93 6.92 6.37 18.5 14.3 15.2 26.5 24.7 24.6 45.9 49.8 48.4
Componentweights
Component1 0.03 0.02 0.01 0.14 0.13 0.11 0.25 0.25 0.25 0.40 0.40 0.40 0.54 0.57 0.59
Component2 0.02 0.01 0.01 0.01 0.01 0.02 0.02 0.01 0.01 0.00 0.01 0.01 0.01 0.00 0.00
Component3 0.73 0.74 0.75 0.52 0.54 0.55 0.34 0.34 0.34 0.13 0.16 0.18 0.01 0.01 0.01
Component4 0.09 0.17 0.24 0.16 0.34 0.32 0.17 0.26 0.35 0.22 0.30 0.38 0.22 0.26 0.29
Component5 0.13 0.07 0.00 0.17 0.09 0.00 0.22 0.14 0.05 0.24 0.14 0.04 0.23 0.16 0.11
1Emission-lineluminositiesareinunitsof1040ergs−1.
2Continuumluminositiesareinunitsof1040ergs−1Å−1.
2382 C.T.Richardsonetal.
D
o
w
n
lo
a
d
e
d
fro
m
h
ttp
s
://a
c
a
d
e
m
ic
.o
u
p
.c
o
m
/m
n
ra
s
/a
rtic
le
-a
b
s
tra
c
t/4
3
7
/3
/2
3
7
6
/1
0
2
6
6
9
1
b
y
g
u
e
s
t o
n
0
1
A
p
ril 2
0
1
9
Figure4. Observed,co-addedspectraforthefivesubsetsthatfalldirectlyalongthecentral AGNlocus.Panel(a)showsthefullspectra.Theotherpanels
showenlargementsthatslightlyoverlapinwavelength.TheemissionlineslistedinTables2and3areidentifiedbytheirions.‘bl’indicatesablendedline.All
fluxvaluesareshowninunitsofthepeakHβintensityintheparticularspectrum.
weassumedthatthereisnoISMattenuationoftheAGNcontin- mainlyimportantforIRlineswhichformincoolatomicgasand
uumradiationincidentontheclouds.Theconsequencesofthisare whicharenotconsideredhere.Forsolarabundances,weusedthose
describedinF97.Weusedconstantdensitymodels;theresultsof summarized in G04a, taken from a series of papers by Asplund
Pellegrini et al. (2007) show that constant-density and constant- and his collaborators (Asplund 2000; Asplund et al. 2000, 2004;
pressuremodelsgiveverynearlythesameresultforopticallines, AllendePrietoetal.2001,2002).Abundances foradditionalele-
whichforminthewarmgaswithintheH+zone.Thedistinctionis mentsnotgivenbyG04aaretakenfromF97.FollowingF97,we
Interpretingtheionization sequenceinAGNspectra 2383
Table2. Measuredemission-linestrengthsfortheAGNlocus,relativetoHβ.
Ion λ a00 A01 a02 a10 a11 a12 a20 a21 a22 a30 a31 a32 a40 a41 a42
[OII] 3727 1.47 1.43 1.18 1.48 1.30 1.31 1.78 1.58 1.48 2.05 1.82 1.71 2.48 2.37 2.35
HI 3798 <0.03 <0.04 <0.04 0.01 0.03 0.04 0.01 0.03 0.05 0.03 0.01 0.01 0.05 0.04 0.04
HI 3835 <0.01 <0.02 <0.05 0.02 0.02 0.03 0.01 0.03 0.04 0.03 0.01 0.02 0.08 0.07 0.08
[NeIII] 3869 0.10 0.02 0.08 0.16 0.17 0.18 0.41 0.42 0.38 0.57 0.54 0.56 0.78 0.79 0.81
HI 3889 0.04 0.03 0.08 0.11 0.09 0.11 0.14 0.14 0.14 0.13 0.13 0.14 0.16 0.16 0.16
[SII] 4070 0.12 0.06 0.04 0.07 0.03 0.01 0.10 0.13 0.11 0.11 0.10 0.08 0.12 0.11 0.11
HI 4102 0.17 0.14 0.18 0.18 0.17 0.20 0.21 0.20 0.20 0.19 0.20 0.19 0.22 0.21 0.21
[FeV]? 4229 0.04 0.10 0.09 0.05 0.02 0.04 0.04 0.04 0.04 0.04 0.05 0.06 0.03 0.01 0.01
HI 4340 0.33 0.32 0.35 0.42 0.40 0.44 0.42 0.44 0.44 0.38 0.39 0.41 0.42 0.42 0.43
[OIII] 4363 <0.07 <0.04 <0.04 0.05 0.04 0.08 0.07 0.07 0.07 0.10 0.11 0.13 0.11 0.12 0.12 D
HeII 4686 <0.05 <0.01 <0.03 0.05 0.04 0.03 0.09 0.09 0.09 0.14 0.15 0.14 0.23 0.23 0.24 ow
[ArIV] 4711 <0.02 <0.02 <0.04 0.03 <0.02 <0.03 0.02 0.02 0.01 0.04 0.04 0.05 0.06 0.06 0.06 n
lo
HI 4861 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 a
d
[OIII] 4959 0.17 0.18 0.14 0.60 0.60 0.50 1.31 1.23 1.11 1.85 1.81 1.71 2.76 2.86 2.92 ed
[OIII] 5007 0.53 0.40 0.39 1.52 1.60 1.39 2.94 3.25 2.88 5.11 4.76 4.62 8.34 8.64 8.76 fro
[NI] 5200 0.05 0.07 0.08 0.04 0.08 0.08 0.06 0.08 0.07 0.06 0.07 0.07 0.07 0.07 0.07 m
H[NeIII] 55787565 <00..1047 <00..0078 <00..0074 <00..1016 <00..0079 <00..0069 00..0132 00..0151 00..0141 00..0133 00..0132 00..0141 00..0122 00..0122 00..0122 https
FevII 6087 <0.00 <0.01 <0.00 <0.01 <0.01 <0.01 0.02 0.03 0.03 0.04 0.05 0.07 0.08 0.08 0.08 ://a
[OI] 6300 0.24 0.25 0.16 0.23 0.23 0.22 0.33 0.35 0.31 0.35 0.31 0.33 0.40 0.40 0.40 ca
d
[OI] 6363 0.10 0.06 0.04 0.07 0.06 0.06 0.10 0.10 0.10 0.11 0.13 0.14 0.11 0.11 0.11 e
m
[NII] 6548 1.05 1.35 1.30 0.91 1.15 1.40 0.97 1.20 1.35 1.03 1.19 1.34 0.94 0.95 0.98 ic
HI 6563 6.58 6.79 5.82 4.49 4.73 5.11 4.35 4.45 4.58 3.92 4.03 4.29 3.51 3.52 3.54 .ou
[NII] 6584 3.36 4.13 3.78 2.85 3.34 3.92 2.92 3.44 3.55 2.88 3.22 3.45 2.58 2.54 2.64 p.c
HeI 6678 0.04 0.03 0.02 0.02 0.02 0.04 0.04 0.03 0.03 0.04 0.04 0.03 0.04 0.04 0.04 o
m
[SII] 6716 1.25 1.21 0.89 0.90 0.86 0.86 0.93 0.93 0.88 0.97 0.94 0.92 1.01 0.96 0.96 /m
[[ASrII]III] 67713315 00..9088 10..0110 00..7066 00..7069 00..7069 00..7059 00..8104 00..8134 00..7184 00..8159 00..8148 00..8128 00..8255 00..8215 00..8215 nras
[OII] 7325 0.13 0.12 0.07 0.08 0.13 0.17 0.16 0.19 0.13 0.16 0.15 0.16 0.15 0.16 0.16 /artic
le
Table3. Dereddenedemission-linestrengthsfortheAGNlocus,relativetoHβ. -ab
s
tra
Ion λ a00 a01 a02 a10 a11 a12 a20 a21 a22 a30 a31 a32 a40 a41 a42 c
t/4
3
[HOIII] 33772978 <30..6007 <30..6029 <20..5047 20..4001 20..2045 20..4057 20..7092 20..5045 20..4068 20..8085 20..6032 20..6042 30..1006 20..9076 20..9066 7/3/2
HI 3835 <0.03 <0.04 <0.10 0.03 0.04 0.06 0.01 0.05 0.06 0.05 0.02 0.04 0.09 0.09 0.10 37
6
[NeIII] 3869 0.22 0.04 0.17 0.25 0.28 0.31 0.62 0.64 0.60 0.77 0.76 0.83 0.96 0.97 1.00 /1
0
HI 3889 0.08 0.07 0.15 0.17 0.15 0.19 0.22 0.22 0.22 0.17 0.18 0.21 0.20 0.19 0.20 2
6
[SII] 4070 0.23 0.12 0.07 0.10 0.05 0.02 0.14 0.18 0.16 0.14 0.13 0.11 0.14 0.13 0.13 69
HI 4102 0.32 0.27 0.31 0.25 0.25 0.31 0.29 0.28 0.29 0.24 0.25 0.26 0.25 0.25 0.25 1 b
[FeV]? 4229 0.06 0.18 0.15 0.06 0.03 0.06 0.05 0.05 0.06 0.04 0.07 0.07 0.03 0.02 0.01 y
g
HI 4340 0.50 0.51 0.51 0.53 0.52 0.59 0.53 0.56 0.56 0.45 0.46 0.51 0.47 0.47 0.48 u
e
H[OeIIIII] 44638663 <<00..0160 <<00..0016 <<00..0036 00..0056 00..0045 00..0130 00..1009 00..1009 00..1009 00..1152 00..1163 00..1156 00..2142 00..2143 00..2143 st on
[ArIV] 4711 <0.02 <0.02 <0.05 0.04 <0.03 <0.04 0.03 0.02 0.01 0.04 0.04 0.05 0.06 0.06 0.06 01
HI 4861 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 A
p
[[OOIIIIII]] 45905097 00..1468 00..1366 00..1335 01..5483 01..5570 01..4279 12..2768 13..1087 12..0771 14..8901 14..7566 14..6359 28..7112 28..8411 28..8573 ril 20
[NI] 5200 0.04 0.05 0.06 0.04 0.07 0.07 0.06 0.07 0.06 0.05 0.07 0.06 0.06 0.06 0.07 19
[NII] 5755 <0.04 <0.04 <0.02 <0.04 <0.05 0.04 0.03 0.04 0.03 0.03 0.03 0.03 0.02 0.02 0.02
HeI 5876 0.08 0.04 0.04 0.08 0.06 0.06 0.09 0.08 0.08 0.10 0.10 0.09 0.11 0.10 0.10
FeVII 6087 <0.01 0.01 <0.01 0.01 <0.01 <0.01 0.02 0.02 0.02 0.03 0.05 0.07 0.07 0.07 0.08
[OI] 6300 0.11 0.12 0.09 0.15 0.15 0.13 0.23 0.23 0.21 0.26 0.23 0.23 0.33 0.34 0.33
[OI] 6363 0.05 0.03 0.02 0.05 0.04 0.03 0.07 0.07 0.07 0.08 0.09 0.10 0.09 0.09 0.09
[NII] 6548 0.46 0.57 0.64 0.58 0.70 0.79 0.64 0.77 0.85 0.75 0.85 0.90 0.77 0.77 0.79
HI 6563 2.86 2.86 2.86 2.86 2.86 2.86 2.86 2.86 2.86 2.86 2.86 2.86 2.86 2.86 2.86
[NII] 6584 0.11 1.72 1.84 1.81 2.01 2.18 1.91 2.20 2.21 2.10 2.28 2.29 2.09 2.05 2.13
HeI 6678 0.05 0.01 0.01 0.01 0.01 0.02 0.02 0.02 0.02 0.03 0.03 0.02 0.03 0.03 0.03
[SII] 6716 0.46 0.48 0.41 0.56 0.50 0.46 0.60 0.58 0.53 0.69 0.65 0.59 0.81 0.76 0.77
[SII] 6731 2.86 0.40 0.35 0.46 0.44 0.40 0.51 0.52 0.47 0.60 0.58 0.53 0.68 0.64 0.65
[ArIII] 7135 0.03 0.03 0.02 0.05 0.05 0.05 0.08 0.08 0.08 0.13 0.12 0.11 0.19 0.19 0.19
[OII] 7325 0.04 0.04 0.03 0.05 0.06 0.08 0.09 0.10 0.07 0.10 0.10 0.09 0.11 0.11 0.11
2384 C.T.Richardsonetal.
usedanionizingluminosityL =1043.5ergs−1,typicalofSeyfert gas with 2log(r) + log(n) < 41.5, which roughly corresponds to
ion
galaxies.Oursimulationsproceededuntiltheelectrontemperature log(U) > 0.4, was not included because it is so highly ionized
fellbelow4000Korabove105K.Gasbelow4000Kdoesnotsignif- thatthegasthatproducestheobservedemissionlinesisessentially
icantlycontributetoopticalemissionlines,althoughitdoesproduce transparent to ionizing radiation. This is the justification for the
IR lines, and gas above 105K contributes principally to X-Rays. integrationlimitsusedbyF97.
Thesemodelsdidnotincludeanycosmicrays.TheCLOUDYmod- Here we present LOC integrations within the limits
elsallterminatebeforereachingmolecularregionswherecosmic 2.0 < log(νH) < 8.0 and 17.48 < log(ρ) < 22.0. Although our
ray ionization from the Galactic background becomes important. gridsofmodelsextendoutto1023cm,themajorityoftheequiv-
Komossa&Schulz(1997)studiedtheeffectofaddingthestandard alent widths considered in this work peak at much smaller radii.
Galacticcosmicraybackgroundtosimilarmodels,andfoundthat Ouradoptedlimitof1022cmisthesameoneusedbyF97.Inaddi-
theemission-lineratioschangedbyonly1percent. tion,the3arcsecdiameterSDSSfibretranslatestoapproximately
Aseriesofgridsofindividualcloudswerecomputedoverarange 1022cm for galaxies observed at z ∼ 0.1.We experimented with
ofdensitiesandradiithatrepresentplausiblevaluesforgasinthe changing the integration limits to include a larger range, but this D
o
NLR.Eachgridconsistedof7171cloudmodels.Theradialdistance onlyproducessmallchangestotheintegratedspectrum.Changes w
fromtheionizing sourcetotheindividual cloud, r,was variedin intheabundances,theSED,γ orβ producedmorenoticeableef- nlo
a
0.1 dex steps from 1016to 1023cm and the hydrogen density, nH, fects. de
twhaast tvhaerieiodniinzin0g.1flduexxwstaespsisfortormopi1c00sototh1a0t10acmdi−ff3e.reWnceeaisnsurmadeidi tioWnsebcyoumsipnagrelidneth-reatoiobsdeiravgerdaminsteagsrianteBdalpdrwedinicetitoanl.s(t1o98th1e),pVreOd8ic7-, d from
reflectsadifferenceinfluxproportionaltor−2.DifferencesinL and elsewhere. We discuss these diagrams below. We will exclu- h
Twhoeultdoataplpheyardarosgaernesdceanlisnitgyowftahsekraedpitaclodnissttaannctefsobrye(aLcihon/i1n0d4iv3.i5d)u1i/oa2nl. saidvderleyssdethsecrfiabcetdthuastt-ofurereduLsOtyCmgordideslsudnotinloSteficttisoonm3e.5ofwthheerAeGwNe ttps://a
c
cloudbutthemolecularandelectrondensitiesweredeterminedself- subsetsaswellasourdust-freemodels. a
d
e
consistentlyandsovarywithdepth.Theresultsfromthesegridsare m
shownbelowintermsoftheequivalentwidthofvariousemission 3.2 SEDoptimization ic.o
lines,whichindicatestheefficiencyofeachcloudinreprocessing u
p
the continuum into a particular emission line. These equivalent TheintrinsicionizingSEDinAGNisuncertain.Thevariousforms .co
m
widthsarerelativetothecontinuumlevelat4860Åinthemodels, oftheincidentcontinuumassumedinpreviousworkincludesimple /m
whichincludesonlytheionizingAGNcontinuum,aquantitythat powerlaws,fν ∝να(Binetteetal.1996,G04a),acombinationof n
isnotdirectlymeasurableintheobservations. power-lawandblackbodycontinua(Komossa&Schulz1997)and ras
/a
Wecomputedbothdust-freeanddustygridsofclouds.By‘dusty’, multi-componentdistributions(F97).Inaddition,theSEDviewed rtic
wemeanLOCgridsthatincludedustintheindividualcloudsinparts by the BLR (Korista et al. 1997) may be different from the SED le
of the NLR where sublimation is not expected to have occurred, viewedbytheNLR,andpartsoftheSED,suchasemissionoriginat- -a
b
s
while‘dustfree’referstoLOCgridswithnodustanywhere.This ingfromadustytorus,areonlyvisibleforAGNatcertainviewing tra
currentsectiondescribesthedust-freemodels. angles.Weadoptedamulti-componentapproachasinF97andused c
t/4
TheLOCmodelisbasedonthefactthatdifferentemissionlines HeI/HIlineintensityratiostoinfertheSED. 3
7
areoptimallyemittedindifferentindividualcloudsaccordingtothe WeconsideredthethreedifferentSEDsshowninFig.5.Theyall /3
/2
correct density and incident flux for efficient production of each havethegeneralform 3
7
particular line, while we observe only the integrated spectrum of 6
the ensemble of clouds. This gives rise to a powerful selection fv ∝vαuve−hv/kTcut−kTIR/hv+Cvαx (2) /10
2
effect in which the overall radial and density distributions of the 66
9
clouds largely determine the measured spectrum. As in F97 (see 1
b
alsoBaldwinetal.1995),wemodelledthetotalemittedspectrum y
g
integratedoverradialdistanceanddensitydistributionsdefinedas u
e
f(r) ∝ rγ and g(n) ∝ nβH, respectively, where γ and β are free st o
parameters.Anaveragespectrumovermanygalaxiesaswehave n
0
producedhereismorelikelytobecorrectlydescribedbyabroad 1
A
power-law distribution than might be the case for an individual p
galaxy.Thetotalluminosityofthelineisthen ril 2
(cid:2)(cid:2) 0
1
L ∝ r2F(r,n )rγnβdn dr (1) 9
line H H H
whereF(r,n )isthefluxofthelineandrγ nβ isthespatialdistri-
H H
butionfunction(Bottorffetal.2002).Throughouttheremainderof
thepaper,whenthelogofdensitiesandradiiaregiven,theunits
arecm−3 andcm, respectively. We chose the integration limits to
includetherangeofparameterspacethatisphysicallyrelevantfor
theNLR.Gaswithdensitieslog(ν )>8isabovethemaximumcrit-
icaldensityforopticalforbiddenHlines.Gasatradiilog(ρ)<17.48 Figure5. TheSEDsforF97,MF87,ourbestmodel(optimizeddust-free)
discussedinSection3.4.2andalsotheSEDforourdustymodel(Section4).
would either be much too hot to produce strong optical emission
Ouroptimizeddust-freemodelsresultfromanSEDthathasashiftedUV
lines, or would have high density which would suppress the for- ‘BigBump’inordertomatchHeIIλ4686,awell-knownindicatorofthe
biddenlines.Gaswithdensitieslog(nH)<2.0doesnotoptimally SED.ThelabelledpartsoftheSEDcorrespondtothevariouscomponents
emitmanylines,thenotableexceptionbeing[SII]λ6720.Finally, forouroptimizeddust-freemodel.
Interpretingtheionization sequenceinAGNspectra 2385
where α is the low-energy slope of the ‘big bump’, T is the subsequentline-ratiodiagramsshowonlytheobservedpointsfor
uv cut
UV temperature cut-off, T is the IR temperature cut-off, C is a thecentralAGNlocus.
IR
normalizingconstantandα istheslopeoftheX-raycomponent. We explored the entire range of β, γ and SED presented in
x
ThefactorCsetsthescalingbetweenthebigbumpandtheX-ray Fig.6whenfittingthesubsetswithinthecentrallocus.Ourgoal,
powerlaw,andinCLOUDYisdeterminedbyspecifyingαox,whichis as presented in the Introduction, is to find a single tunable free
givenby parameter responsible for the systematic variation of AGN along
theBPTdiagram.AsisdiscussedinSection3.4,varyingonlythe
fv(2keV) =403.3αox (3) radialdistributionparameterγ ofcloudswhileholdingthedensity
fv(2500Å) weightingparameterβconstantprovidesareasonablefittotheAGN
locus.Thisbecameapparentatanearlystageintheinvestigation
(Koristaetal.1997).TheX-raycomponent(lasttermofequation andsoweusedthisapproachtooptimizetheSED.
2)wasassumedtofalloffasν−2above100keV. Itcanbeseenthatthehigher-ionizationobservedpointsinthe
A commonly used AGN SED is given by Mathews & Ferland uppertwopanelsofFig.6areingoodagreementwiththecalculated D
o
(1987, hereafter MF87), and is shown as a dotted line in Fig. 5. curves.Todeterminethebestsetoffreeparameters,wevariedthe w
n
Acombinationofdirectobservations,withinferencesbasedonthe radialdistributionofclouds,startingwiththetopofthelocus,trying lo
a
emission-line spectrum, determined the components of this SED. toreproducethehighestionizationsubsettowithinafactorof2, d
e
AThGeNX(-Zraaymsolroapneioefttahle. 1S9E8D1,).αAxs∼w−e0.d7e,sicsrtiybpeicbaellofwor,rtahdisioh-laosuda bweefoardeopprtoacefaecdtionrgotfo2thaesnoeuxrtcsuribtesreito.nHfeoreraanndacthcreoputagbhloeufittthbeeptwapeeern, d from
significant effect on the equivalent width contours but minimally the predicted and observed line ratios. This is principally due to h
affectstheLOCintegrations.F97usedamodifiedSED,shownin uncertaintiesinthemodelsaftertakingintoaccounttheuncertainty ttp
s
Fig.5asthedashedline.ThisSEDdiffersfromthatofMF87in intheSEDandabundances,andthefactthattheLOCmodelsjust ://a
c
thatitlacksIRemissionfromadustytorus,includesasoftX-ray use simplified power laws to describe the distributions of r and a
d
component,andhasasofterX-rayslopeαx ∼−1.0thatistypical nH.AsmentionedinSection2,themeasurementofanemission- em
ofradio-quietquasarsandAGN(Elvisetal.1994). line ratio containing two weak emission lines would possess an ic
TheHeIIλ4686/Hβ intensityratioisanimportanttestofthese uncertaintyof∼30percent.Thus,theobservationaluncertaintyis .oup
possibleSEDs,because(1)HeIIandHβarerecombinationlinesand lessthanthesystematicuncertaintyinourmodels. .co
thereforelesssensitivetoT thancollisionallyexcitedlines(AGN3); For the F97 SED (Panel (a) of Fig. 6), the models producing m
(2)theratioisrelativelyinseensitivetolargechangesintheionization integratedspectrawithmoderate[OIII]λ5007/Hβratios(moderate /mn
parameter(Binette,Courvoisier&Robinson1988).Photoionization ionization) overpredict the HeII/ Hβ ratios by almost a factor of ras
m19o8d6e)l.sofAGNoftenunderpredictthisratio(Ferland&Osterbrock f3a.cItnorcoofn2tr.aTsth,itshiescMloFs8e7lySrEelDatepdrotoduthcees‘SantoyHmeIeI/tHhoβd’raotfiodewteitrhminina- /article
UsingtheF97andMF87SEDs,wecomputedtwoseparategrids ingtheSED(AGN3)andweadjustitshardnesstoreproducethe -a
b
ofCLOUDYmodelsasoutlinedinSection3.1,employingasolarcom- ratio. stra
positionforthispreliminarystep,andthenranLOCintegrations. We optimized the SED to produce a third continuum shape, c
Thepredicted[OIII]λ5007/HβversusHeIIλ4686/Hβlineratiodi- shown in Fig. 5 as the solid line labelled ‘optimized, dust-free’, t/43
7
agramsforourdust-freeLOCmodelsusingeachoftheseSEDsare to try to produce a better fit to the measurements from our /3
shownintheuppertwopanelsofFig.6ascolouredlinesconnecting galaxysample.InadditiontoaSEDsimilartoradio-loudquasars /23
7
aseriesofcircles.EachlinerepresentstheLOCresultsforadiffer- (MF87), we wanted one that was typical of AGN (F97) but that 6
entdensityweightingintherange−1.8≤β≤−0.6inincrements still matched the moderate ionization HeII/Hβ ratio. To achieve /10
2
of0.4withβ=−1.8[blue],β=−1.4[green],β=−1.0[red]and this, we modified the F97 SED by shifting the ‘big bump’ until 66
β=−0.6[cyan].Alongeachindividuallinetheradialweightingis the HeII/Hβ ratio matched the moderate to high ionization sub- 91
variedovertherange−2.0≤γ ≤2.0,andthecirclesalongtheline sets. This modified ‘bump’ turned out to match the one in the by
representdifferentLOCmodelsrunatincrementsinγ of0.25.The MF97SED.Wealsokepttheαξ toprovidecloseragreementwith gu
e
lcalroguedsst,sγol=idc−ir2c.l0e.sTinhdeicciartcelethsewmitohswtnheitgeactievnetrreasdiraelpdreissternibtumtioodneolsf t(hZeammoeraanniveatluael.o1f9−811;.0Stfeoffuenndeitnarla.d2i0o0-q6u);ietthiqsumasaakrsesanthdeAoGptNi- st on
0
usingourbest-fittingsetoftheremainingfreeparameters,asde- mizedSEDsignificantlydifferentfromtheMF87SEDintheX-ray 1
A
scribedinSection3.4.Theblackdiamonds,trianglesandsquares region. p
representtheratiosmeasuredfromtheAGNobservations,witheach Our optimized SED is characterized by Tcut ∼ 4.2 × 105 K, ril 2
sequenceintheobservationsindicatedbyadifferentlinestyleand αUV=−0.3,αox=−1.35,αx=−1.0,andkTIR=0.14eV.Thebot- 019
symbol.ThelargestofthesesymbolsrepresentsthetopoftheAGN tompanelofFig.6showstheresultingpredictedline-ratiodiagrams
locusforthatsequence.Thelargestsquareshouldbecomparedto andverifiesthefittothemoderateandhigh-ionizationHeII/Hβra-
thelargestwhite-filledcircle,whichrepresentsourbest-matching tio.Furtheradjustmentsweremadetomatchthelowerionization
setoffreeparametersforthehighionizationendoftheobserved HeII/Hβ ratiosbutwewereunabletosimultaneouslyaccountfor
sequence. boththelowerandupperendofthelocus.
Fig.6showsallthreeobservedAGNsequences(thecentralai1 Fig.7showstheWλ distributionsforseveralemissionlinesfor
locusandthetwo‘wings’correspondingtoai0andai2).Thethree each of the three SEDs discussed so far. We note that the peak
observed sequences are very similar in Fig. 6. In fact, even in emissivityofsomestronglinesoccursattwolocationsontheLOC
the full set of line ratio diagrams described below the difference planefortheMF87AGNcontinuum,buttheF97continuumonly
between each sequence is so slight that we cannot determine a createsasinglemaximum.Wefoundthatthisiscausedbythediffer-
physicallymeaningfulinterpretationforthewings;theyappearto entassumptionsintheformoftheX-raycontinuum.Thesesecond
just represent object-to-object scatter about the central locus. For optimallyemittingregionsminimallyaffectedtheLOCintegrations.
thisreason,wecompareourmodelsonlytothecentrallocusandon Furthermore,thedensitiesatwhichthesecondoptimallyemitting