ebook img

A 2D multiwavelength study of the ionized gas and stellar population in the Giant HII Region NGC 588 PDF

0.74 MB·English
Save to my drive
Quick download
Download
Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.

Preview A 2D multiwavelength study of the ionized gas and stellar population in the Giant HII Region NGC 588

Mon.Not.R.Astron.Soc.000,1–??(2010) Printed11January2011 (MNLATEXstylefilev2.2) A 2D multiwavelength study of the ionized gas and stellar population in the Giant HII Region NGC 588 1 A. Monreal-Ibero1,2⋆†, M. Relan˜o3, C. Kehrig2, E. Pe´rez-Montero1, 1 0 1 2 2 2 J. M. V´ılchez , A. Kelz , M. M. Roth , O. Streicher 2 1InstitutodeAstrof´ısicadeAndaluc´ıa(CSIC),GlorietadelaAstronom´ıa,s/n,18008Granada,Spain n 2AstrophysikalischesInstitutPotsdam,innoFSPECPotsdam,AnderSternwarte16,D-14482,Potsdam,Germany a 3Dpto.deF´ısicaTeo´ricaydelCosmos,UniversidaddeGranada,CampusFuentenueva,Granada,Spain J 0 1 revisedversion ] O C ABSTRACT h. Giant HII regions(GHIIRs)innearbygalaxiesarealocalsampleinwhichwe canstudyin p detailprocessesin the interactionofgas, dust, andnewlyformedstars whichare analagous - to those which occurred in episodes of higher intensity in which much of the current stel- o lar population was born. Here, we present an analysis of NGC 588, a GHIIR in M33, r based on optical Integral Field Spectroscopy (IFS) data obtained with the PMAS instru- t s ment at the 3.5 m telescope of Calar Alto Observatory, CAHA, together with Spitzer in- a fraredimagesat8µmand24µm.Theextinctiondistributionmeasuredintheopticalshows [ complex structure, with three maxima which correlate in position with those of the emis- 1 sion at 24 µm and 8 µm. Furthermore, the Hα luminosity absorbed by the dust within v the HII region reproduces the structure observed in the 24 µm image, supporting the use 3 of the 24 µm band as a valid tracer of recent star formation. A velocity difference of 5 ∼50 km s−1 was measured between the areas of high and low surface brightness, which 8 would be expected if NGC 588 were an evolved GHIIR. We have carefully identified the 1 areas which contribute most to the line ratios measured in the integrated spectrum. Those . 1 line ratios which are used in diagnostic diagrams proposed by Baldwinetal. (1981) show 0 a larger range of variation in the low surface brightness areas. The ranges are ∼0.5 to 1 1.2 dex for [NII]λ6584/Hα, 0.7 to 1.7 dex for [SII]λλ6717,6731/Hα, and 0.3 to 0.5 dex 1 : for[OIII]λ5007/Hβ,withhighervaluesof[NII]λ6584/Hαand[SII]λλ6717,6731/Hα,and v lowervaluesof[OIII]λ5007/Hβ intheareasoflowersurfacebrightness.Ratioscorrespond- i ing to large ionization parameter (U) are found between the peak of the emission in Hβ X andthemainionizingsourcedecreasingradiallyoutwardswithintheregion.Differencesbe- r a tween the integrated and local values of the U tracers can be as high as ∼0.8 dex, notably whenusing[OIII]λλ4959,5007/[OII]λλ3726,3729andinthehighsurfacebrightnessspax- els. [OII]λλ3726,3729/Hβ and [OIII]λλ4959,5007/[OII]λλ3726,3729 yield similar local valuesfor theionizationparameter,whichare consistentwith those expectedfromthe inte- grated spectrumof an HII regionionized by a single star. The ratio [SII]λλ6717,6731/Hα departssignificantlyfromtherangepredictedbythisscenario,indicatingthecomplexioniza- tionstructureinGHIIRs.Thereisasignificantscatterinderivationsofthemetallicityusing stronglinetracersasafunctionofposition,causedbyvariationsinthedegreeofionization. ThescatterissmallerforN2O3whichpointstothistracerasabettermetallicitytracerthan N2. One interesting result emerges from our comparison between integrated and local line ratiovalues:measurementsofthelineratiosofGHIIRingalaxiesatdistances>∼25Mpcmay bedominatedbytheionizationconditionsintheirlowsurfacebrightnessareas. Key words: HII regions:individual:NGC 588 – galaxies: individual:M33 – stars: Wolf– Rayet–ISM:abundances–ISM:kinematicsanddynamics–dust,extinction. ⋆ E-mail:[email protected](AM-I) Center, Calar Alto, jointly operated by the Max-Planck-Institut fu¨r As- † Based on observations collected at the German-Spanish Astronomical tronomieHeidelbergandtheInstitutodeAstrof´ısicadeAndaluc´ıa(CSIC). 2 A. Monreal-Iberoet al. 1 INTRODUCTION ogy.Moreover, weevaluatedthereliabilityofdifferentlineratios as metallicity (Z) tracers. In a companion paper, we presented a Large areas of ionized gas known as Giant HII regions (GHI- novel approach tomodel these complex structures. There, were- IRs) constitute the most conspicuous places of star formation produced our observations by jointly fitting the radial profiles of in normal galaxies (see Shields 1990, for a review). Their di- different optical (i.e. Hα, [OII]λλ3726,3729, [OIII]λ5007) and ameters typically range between ∼100 pc to ∼800 pc (e.g. infrared(i.e.8µm,24µm)magnitudestoasetofCLOUDY-based Kennicutt1984;Alonso-Herreroetal.2002)whiletheirHαlumi- photoionizationmodels(Pe´rez-Monteroetal.2010). nosity range expands up to 3 orders of magnitudes (∼ 1038 − OurexperiencewithNGC595showstheimportanceofcarry- 1040 erg s−1, e.g. Kennicutt 1984; Rozasetal. 1996; Firpoetal. ingoutananalysisusingthecombinedinformationofopticaland 2005; Monreal-Iberoetal. 2007; Relan˜o&Kennicutt 2009). Re- infrareddatatogetherwithmodelling.However,giventheirdiver- gardingtheirmorphologies, someof thempresentacompact dis- sity,thesampleofGHIIRsstudiedbymeansofthismethodology tributionwithhighsurfacebrightnesswhileothershaveamoredif- cannot be reduced to only one example. Instead, studies of other fuseemission.Also,theycanpresentmultiplecoresand/orshells regionssamplingadifferentrangeintheparameterspacewouldbe or ring-like features. In the same manner, they present very dif- desirable.Wepresenthere,theanalysisofaregionwithadifferent ferent content of gas, varying between ∼ 103 M⊙ to almost morphology,andrelativelylowermetallicityandhighHαluminos- 107 M⊙ whiletheyusually have∼ 102 −105 M⊙ instars(e.g. ity:NGC588.ThisregionislocatedintheoutskirtsofM33,atthe Castellanosetal. 2002b). Finally, GHIIRs are small-scale exam- endofaspiralarm,ataradiusof∼14′ (i.e.∼ 3.42kpcforadis- ples of the extreme events of star formation occurring in star- tancetoM33of840kpc;Freedmanetal.1991).Thisareapresents burstgalaxies(e.g.Alonso-Herreroetal.2009;Garc´ıa-Mar´ınetal. HIemission(Verleyetal.2010;Gratieretal.2010).However,no 2009).Thus,agoodknowledgeoftheseobjectsishighlyvaluable local H (or CO) emission has been detected towards NGC 588 2 forabetterunderstandingofthesemoreviolentphenomena. (Verleyetal.2010;Israeletal.1990).Withasizeof∼30′′×50′′ ThevarietyofpropertiesinGHIIRsimpliesacomplexstruc- (i.e.∼120pc×200pcatourassumeddistance),NGC588hasbeen ture,farfromthetextbook-likeStro¨mgrensphere.Moreover,mod- classifiedwithinthering-likeclass(Sabbadinetal.1980).Itsstel- elling of star-forming regions showed that in case of multiple larcontenthasbeenthoroughlystudied(Jametetal.2004;Pellerin ionizing sources, geometrical effects affect the physical proper- 2006;U´beda&Drissen2009).Thedifferentestimatesforitstotal ties(i.eelectrontemperatureandionizationstructure)ofthesere- stellarmassrangebetween∼1.3×103and∼5.6×103M⊙with gions(Ercolanoetal.2007;Jamet&Morisset2008).Theinhomo- anagefortheburstof∼ 3.5−4.2Myr andlowmetallicity(i.e. geneities of the Interstellar Medium (ISM) have been taken into Z ∼ 0.4Z⊙),consistentwithdirectmeasurementsusinglong-slit account within the HII regions using the filling factor. This de- (12+log(O/H)=8.30,V´ılchezetal.1988).NGC588containstwo scribesthefractionofthetotalvolumeoftheHIIregionwithhigh Wolf-Rayet(WR)stars.OneofthemwasclassifiedasWNLwhile densegaswhiletheremainingvolumeisconsideredtobeofnegli- the other as Ofpe/WN9, an intermediate object between Of and gibledensity(Osterbrock&Flather1959).Recently,detailedmod- WNstars(Masseyetal.1996).Fromthekinematicpointofview, elsshow thatthedensityvariationsassumingopticallythickhigh TAURUS-2 Fabry-Pe`rot data show how NGC 588 seems a rela- densitygasclumpsgiverisetoinhomogeneitiesinthetemperature tivelyevolvedsystem,withitsionizedgaskinematicsdominatedby andionizationparameter (Giammancoetal.2004,2005).Thus,a acollectionoflargestellarwindshells(Mun˜oz-Tun˜onetal.1996). singlevalueperphysicalmagnitude,usuallyextractedfromaspe- Finally,arelativelyfaint andpoint-likeX-rayemittingsourceas- cificareaoftheregion,isnotnecessarilyrepresentativeofitsphys- sociatedwiththisregionhasbeendetected(Plucinskyetal.2008). icalconditions. Inthiswork, wecombine 2D optical spectroscopic observa- ItisinverynearbyGHIIRs(i.e.atD<∼6Mpc),whereground tionswiththePotsdamMulti-ApertureSpectrophotometer(PMAS, basedopticaltelescopescanachievehighlinearspatialresolution Rothetal.2005)andinfraredimagingwithSpitzer.Thesewillcon- (i.e.<∼40 pc arcsec−1), enough to resolve the different elements stitutethefirstpublished resultsfromdataobtained withthenew (i.e.starclusters,ionizedgas,dust,etc.),playingaroleintheinter- PMAS’s CCD. A modelling of NGC 588 intending to reproduce actionbetweenthemassivestarsanditssurroundingenvironment. theobserved magnitudes will bepresented inacompanion paper Also,itistherewherethevariationsofthephysicalandchemical (Pe´rez-Monteroetal.inpreparation).Wedescribetheobservations propertiesoftheionizedgascanbeproperlysampled. anddataprocessingtocreatethemapsoftherelevantmagnitudesin From the observational point of view, the study of these re- Section2.ThemainobservationalresultsarepresentedinSection gionswouldbenefitfromhighqualityspectroscopicdatathatmap 3.Finally,wesummarizeourmainconclusionsinSection4. in an un-biased way the surface of the GHIIR. Nowadays, the technique of integral field spectroscopy (IFS), able to record si- multaneously the spectra of an extended continuous field, offers thepossibility of performing such amapping. At present, studies 2 OBSERVATIONSANDDATAREDUCTION of GHIIRsbased onIFSare stillscarce. Anexample isprovided 2.1 Observations by Garc´ıa-Benitoetal. (2010), who analyzed one of the bright- estGHIIRsinNGC6946.Also,Lo´pez-Sa´nchezetal.(2010)pre- The IFS data of NGC 588 were obtained on October 9-10, 2009 sentedadetailedstudyofastar-formingregionatthelowerlimit duringthecommissioningrunofthenewPotsdamMulti-Aperture ofGHIIRsintermsofHαluminosityandsizeinIC10,ourclos- Spectrophotometer (PMAS, Rothetal. 2005) CCD at the 3.5 m eststarburst.Recently,wepresenteda2Dspectroscopicanalysisof telescopeattheCalarAltoObservatory(Spain).ThenewPMAS4k thesecondbrightest HIIregioninM33,theTriangulumGalaxy: x4kCCDisreadoutinfourquadrantswhichhaveslightlydiffer- NGC595(Relan˜oetal.2010).There,weshowedhowtheoptical entgains(Rothetal.2010).DataweretakenusingtheLensArray extinctionmapandtheabsorbedHαluminosityarespatiallycor- Mode(LARR)configurationwhichismadeoutofa16×16arrayof relatedwiththe24µmemissionfromSpitzer,andhowtheioniza- microlensescoupledwithfibres(hereafterspaxels).Weusedthe1′′ tionstructure of theregion nicely follows theHα shell morphol- magnificationwhichprovidesafieldofview(FoV)of16′′×16′′. NGC 588withPMAS 3 the spectra were traced on a continuum-lamp exposure obtained before each target exposure. We extracted the target spectra by addingthesignalfromthecentraltracedpixelsantitsfourneigh- bours.Thespectrawerewavelengthcalibratedusingtheexposures of HgNe arc lamps obtained immediately after the science expo- sures.Wechecked theaccuracyof thewavelengthcalibrationus- ing the [OI]λ5577 A˚ sky line, and found standard deviations of <0.14A˚,which allowedustodetermine thecentroidof linewith 44" anaccuracyof∼7kms−1at∼5000A˚. Fibres have different transmissions that may depend on the wavelength. The continuum-lamp exposures were used to deter- mine the differences in the fibre-to-fibre transmission and to ob- tainanormalizedfibre-flatimage,includingthewavelengthdepen- 30" 100 pc dence.Thisstepwascarriedoutbyrunningthefiber flat.pl scriptfromtheR3Dpackage(Sa´nchez2006).Inordertohomoge- nizetheresponseofallthefibres,wedividedourwavelengthcal- ibrated science images by the normalized fibre-flat. To estimate Figure1.MosaictomapNGC588overplottedonacontinuum-subtracted the accuracy of the fibre-to-fibre response correction, we fitted a Hαdirect imagefromNOAOScience Archive (Masseyetal.2007).The orientationisnorthupandeasttotheleft.TheHαimageisshowninlog- Gaussiantofouremissionlinesdistributedalongthewholespectral arithmicstretchtobetterenhanceallthemorphologicalfeaturesoftheHII rangeinanextracted,wavelengthcalibratedandflat-fieldcorrected regionandcoversarangeof3.4dex. arcexposure.Weusedtheratiobetweenthestandarddeviationand themeanfluxineachlineasaproxyfortheaccuracyofthefibre-to- fibreresponsecorrection.Forthoselinesinthecentralpartofour We used the V600 grating and the 2×2 binning mode achieving spectralrange,thiscorrectionwasverygood,withratiosof∼2%. aneffectivedispersionof 1.59A˚ pix−1,and a∼3.4A˚ fullwidth Intheblueandrededgestheseratiosreachedvaluesof∼12%,due halfmaximumspectralresolution.WiththenewPMAS’sCCD,the tothecontributionoffibresaffectedbyvignetting. coveredspectralrangewasfrom3620to6800A˚ formostofthe Inthenextstep,thethreeexposurestakenforthesamepoint- spaxels.Thispermitsustoobservethemainemissionlinesinthe ing were combined in order to remove cosmic rays, using the optical from [OII]λλ3726,3729 to [SII]λλ6717,6731. However, imcombineroutineinIRAF.1Fluxcalibrationwasperformedus- for ∼40 spaxels, always at the edge of the LARR, this spectral ingtheIRAFtasksstandard,sensfuncandcalibrate.We range was slightly reduced due to the vignetting associated with co-addedthespectraofthecentralfibresofthestandardstarexpo- the3.5mtelescope. Thisprevents us fromobtaining information suretocreateaonedimensionalspectrumthatwasusedtoobtain for the [OII]λλ3726,3729, [SII]λλ6717,6731, and Hαemission thesensitivity function. For wavelengths larger than 4000 A˚, the linesinsomespecificareas(seealsosection2.2). uncertaintyinthefluxcalibrationis∼1%,whileacrossthebluer We made a mosaic of 6 tiles to map most of the surface of spectralrange(i.e.<4000A˚),theassociatederrorcanreach∼5%. NGC588.ThedistributionofthedifferenttilesisshowninFig.1 Giventhelargesizeof thenew PMAS’sCCD,thefour cor- overplottedontheHαemission-lineimagefromNationalOptical nersofeachexposuresufferfromtelescopevignetting(Rothetal. Astronomy Observatory (NOAO) Science Archive (Masseyetal. 2010).DuetothewaythefibresoftheLARRarearrangedatthe 2007). Contiguous tiles had a 2′.′0 overlapping, to make easier a entranceofthespectrograph,amaximumoftwocolumnsoffibres commonrelativefluxcalibrationofthedata.Intotal,wecovereda attheeastandwestsidesoftheLARRwereaffectedbythis.The fieldof30′.′0×44′.′0whichatthedistanceofNGC588corresponds wavelengthrangeaffectedwaslargerforthosefibreslocatedmore to∼120pc×180pc. towardstheedgeofthespectrographandcouldbeupto∼3898 A˚ We obtained three exposures of 400 s per tile. Atmospheric intheblueendandfrom∼6498A˚ intheredend.Thesepartsof conditionsduringtheobservationswerenon-photometricandtypi- thespectraweremaskedandthenweusedtheoffsetscommanded calseeingrangedbetween1′.′2and1′.′6.Theskytransparencydur- to the telescope and the PMAS acquisition images toconstruct a ingtheobservingnightspresentedvariationsof<∼15%.Allthedata mosaicdatacube. weretakenatairmasses<∼1.1inordertopreventstrongeffectsdue After creating this datacube, maps for the different ob- todifferentialatmosphericrefraction. servables were derived following the methodology presented in In addition to the science frames, continuum and HgNe arc Relan˜oetal. (2010). Basically, we performed a Gaussian fit to lampexposuresinordertominimizetheeffectsduetoinstrument the emission lines using the IDL-based routine mpfitexpr flexures.Also,anearbyskybackgroundframewasobtainedduring (Markwardt 2009) and derived the quantities of interest for each thesecondnightbymovingtheIFUoff-target. individual spaxel. Then, weused thesetogether withtheposition Finally, exposures of the spectrophotometric standard star of the spaxels within the datacube to create an image suitable to BD+28D4211wereobtainedinordertocorrectfortheinstrument bemanipulatedwithstandardastronomicalsoftware.Hereafter,we responseandperformarelativefluxcalibration. willrefertothiswithbothterms:mapandimage. Evenafterhavingallowedfor2′′ofoverlapbetweentiles,vi- 2.2 Datareductionandmapcreation The first steps of the data reduction were done through the P3d 1 IRAF is distributed by the National Optical Astronomical Observato- toolthatisdesignedtobeusedwithfibre-fedintegral-fieldspectro- ries, which are operated by the Association of Universities for Research graphs (Sandinetal. 2010). After trimmimg, combining the four inAstronomy,Inc.,undercooperativeagreementwiththeNationalScience quadrantsandsubtractingthebiaslevel,theexpectedlocationsof Foundation. 4 A. Monreal-Iberoet al. sityoveralocallyfittedcontinuum.AswasshowninRelan˜oetal. Table1.Observedandextinction-corrected emissionlinefluxes,normal- izedtoI(Hβ)=1fromtheintegratedspectrumofNGC588.f(λ)isthered- (2010),bothsplotandmpfitexprgivesimilarresultsforthe deningfunctionnormalisedtoHβfromFluksetal.(1994). high signal-to-noise integrated spectrum. We derived the redden- ing coefficient, c(Hβ), from the F(Hδ)/F(Hβ), F(Hγ)/F(Hβ) and F(Hα)/F(Hβ) line ratios. We performed a linear fit (by minimiz- EmissionLine λobs f(λ) I(λ)/I(Hβ) I(λ)/I(Hβ)corr ingthechi-squareerrorstatistic)tothedifferencebetweenthethe- 3727[OII] 3725.52 0.255 1.922±0.048 2.317±0.091 oretical and observed Balmer decrements vs. the reddening law 3868[NeIII] 3866.01 0.227 0.389±0.013 0.461±0.020 (Fluksetal. 1994), while simultaneously solving for the effects 3889H8+HeI 3886.21 0.223 0.423±0.014 0.501±0.021 3970Hǫ+[NeIII] 3966.03 0.208 0.152±0.011 0.179±0.013 of underlying Balmer absorption with equivalent width, EWabs. 4101Hδ 4099.02 0.180 0.199±0.013 0.274±0.016 We assumed that the EW is the same for all Balmer lines abs 4340Hγ 4338.19 0.133 0.397±0.010 0.451±0.020 (e.g.Kobulnickyetal.1999).ThetheoreticalBalmerlineintensi- 4861Hβ 4859.46 0.000 1.000±0.016 1.000±0.021 4959[OIII] 4957.02 -0.027 1.500±0.021 1.524±0.046 tieswereobtainedfromStorey&Hummer(1995)assumingCase 50508776[OHeIIII] 55080742..9400 --00..024117 40..410504±±00..004094 40..412480±±00..103017 B,Te=104K,ne=100cm−3(typicalvaluesfoundinHIIregions, 6563Hα 6559.34 -0.314 3.508±0.034 2.938±0.114 Osterbrock&Ferland2006). 6584[NII] 6579.40 -0.316 0.246±0.004 0.210±0.008 WederivedameanextinctionfortheregionofAV = 0.548. 6678HeI 6674.29 -0.329 0.036±0.002 0.031±0.002 GiventhattheGalacticextinctioninthelineofsighttoNGC588 6717[SII] 6712.40 -0.332 0.210±0.004 0.178±0.007 6731[SII] 6726.76 -0.335 0.147±0.003 0.124±0.005 is AGVal = 0.146 (Schlegeletal. 1998) most of it is intrinsic to NGC588 andagrees withintheuncertaintieswithprevious mea- surementonitsbrightestparts(A =0.49−0.81,Melnick1979; V Viallefond&Goss1986;Melnicketal.1987). Table2.Maindiagnosticemissionlineratiosandintegratedphysicalprop- Table 1presents both the measured and extinction-corrected erties ofNGC588. Allthe physical properties were estimated usingour fluxes. Errors were estimated using the formula presented in extinctioncorrectedintegratedfluxes. Castellanosetal.(2002a). EW Parameter Valuecorr σline=σcont×N1/2(1+ N∆λ)1/2 (1) log([NII]λ6584/Hα) -1.146±0.077 whereσ isthestandarddeviationinacontinuumclosetothe log([SII]λλ6717,6731/Hα) -0.988±0.078 cont lineofinterest,N isthenumberofpixelssamplingtheline,EW log([OIII]λ5007/Hβ) 0.646±0.050 isitsequivalent widthand ∆λ isthe dispersion in A˚ pix−1. The log([OIII]λλ4959,5007/[OII]λ3727) 0.406±0.069 quoteduncertaintiesofthereddeningcorrectedlinefluxestakeinto logR23 0.915±0.053 N2O3 -1.792±0.127 accountthemeasurementandreddeningerrors. [SII]λ6717/[SII]λ6731 1.435±0.143 Toinvestigateifthenebularpropertiesderivedfromlong-slit log([NII]λ6584/[OII]λλ3726,3729) -1.043±0.077 arerepresentativeofthewholeHIIregion,wecomparedourflux log([NII]λ6584/[SII]λλ6717,6731) -0.271±0.078 measurements from the integrated spectrum of NGC 588 to the values presented by Jametetal. (2005) and V´ılchezetal. (1988) c(Hβ) 0.26±0.03 ne(cm−3) <100 whousedlongslitatpositionangle(P.A.)∼ −45◦ includingthe qeff (cms−1) 3.9×107 mainionizing clusterinNGC588 (e.g.seeFig.1in Jametetal. Te(K) 11140±180(∗) 2005).Mostofourmeasurementspresentadifferencewithrespect 12+log(O/H) 8.16±0.02 toHβwhencomparingwithpreviouslyreportedmeasurementsthat canrangebetween∼8%(e.g. [OIII]λ5007andHeIλ6678)upto (∗) Valuederivedfromthe[OIII]λ4363/λ5007lineratiobyJametetal. ∼64% in the case of H8+HeIλ3889. This indicates that it is not trivialhowthelineratiosmeasuredatthebrightestknotstracethose (2005). fortheHIIregionasawhole.Thisissuewillbeexploredinmore detailinSec.3.6. gnetting prevents us from deriving information for some lines in Table2containstheprincipaldiagnosticemission-lineratios thecentralcolumnsofourmosaic.Inthosemapswhere(atleast) measuredfromtheintegratedspectrumaswellasthephysicalpa- one line affected by vignetting was involved, those spaxels were rametersderivedfromthem.Giventheblue-shiftforM33andthe maskedandinterpolatedusingfixpixwithinIRAF,forpresenta- presenceintheskysubstractedspectrumofstrongresidualsforthe tionpurposes. HgIλ4358sky-line,itwasnotpossibletomeasurethe[OIII]λ4363 nebular line and thus determine the electron temperature (T ) by e means of the [OIII] λ4363/λ5007 line ratio. Another possibility 3 RESULTS would have been using [SII]λ4067 and/or [NII]λ5755 together with [SII]λλ6717,6731 and/or [NII]λ6584. To search for these 3.1 Integratedproperties lines,inadditiontothetotalspectrum, wecreatedaspectrum by Here we analyze the integrated spectrum for NGC 588, obtained co-addingthespectrainanapertureof5×9spaxelscentredatthe afterco-addingthesignalof∼1000spaxels.Theskysubstracted peakofemissionfortheionizedgas.Inthisway,weincreasedthe spectrumisshowninFig.2,whereweusedtwodifferentnormal- signal-to-noise ratio – and thus, our detection limit – since only ization factors in order to better display all the observed emis- thebrightestspectrawereincluded.However,thesefeatureswere sion lines. The positions of the detected nebular emission lines detectedneitherinthetotalspectrumnorintheoneinvolvingthe were marked with labels. In addition, several sky line residuals brightestspaxels.Thus,weassumedaT of11140K,asderivedby e are clearly visible. Flux for the main emission lines was mea- Jametetal.(2005)fromthe[OIII]λ4363/λ5007lineratiotoesti- suredusing splotwithinIRAF,whichintegratesthelineinten- matetheelectrondensity(n )andusedthetasktemden,basedon e NGC 588withPMAS 5 Integrated spectrum [NeIII]+HeI+H8 [NeIII]+Hε HeI 3 [OII] Hδ Hγ Hβ [OIII] HeI Hα+[NII] [SII] a.u.) 2 x ( u Fl 1 0.4 u.) 0.3 a. x ( u 0.2 Fl 0.1 4000 4500 5000 5500 6000 6500 λ(Å) Figure2.IntegratedspectrumofNGC588obtainedbyco-addingthesignalofallthespaxelsinthefieldofviewwithtwodifferentnormalizationstobetter visualizethedifferentobservedemissionlines.Fluxesareinarbitraryunits.Theskyspectrumwascreatedbycombiningthespectracorrespondingtothe vigneting-freefibresofthebackgroundframe. thefivelprogram(Shaw&Dufour1995)includedintheIRAF per or the lower branch. Instead, the metallicity of NGC 588 packagenebular.Thederivedn wasconsistentwithbeingbe- falls in the knee region of the Z − R23 relation, where un- e lowthelowdensitylimit. certainties can be as high as 0.7 dex. Alternatively, one can AnotherquantityquotedinTable2istheionizationparameter, usetheN2O3=log(([NII]λ6584/Hα)/([OIII]λ5007/Hβ)).Wede- definedas: rived a metallicity of 12 + log(O/H) = 8.15 and 8.18 us- ing the parametrization proposed by Pettini&Pagel (2004) and Q(H ) qeff = 4πR2on (2) Pe´rez-Montero&Contini (2009), respectively. Thus for the pur- s e pose of this section, wewill consider as thecharacteristic metal- whereQ(Ho)isthenumberofionizingphotonspersecondemitted licity of the region, the mean of those derived from the N2 and bythestars,RsistheStro¨mgrenradiusoftheHIIregionandneis N2O3parameters: 8.16±0.02. Thisvalue agreewithintheuncer- theelectrondensity.Weestimatedthenumberofionizingphotons taintieswiththe12+log(O/H) = 8.17 metallicityreportedby usingtheexpressionprovidedbyKennicutt(1998): Jametetal. (2005), is slightly lower than the expected value as- sumingthatthisregionfollowsthemetallicitygrandientforM33 Q(Ho)(s−1)=7.31×1011L(Hα)(ergs−1) (3) (8.28±0.08, Rosolowsky&Simon 2008) and ∼0.15 dex lower and the extinction corrected Hα luminosity reported by thanthevaluereportedbyV´ılchezetal.(1988). Relan˜o&Kennicutt(2009).Sincethesulfurlineratioisconsistent withbeingbelowthelowdensitylimitregime,weassumedaface valueofn = 20cm−3.Also,weutilizedaStro¨mgrenradiusof e 3.2 Structureoftheionizedgasandthestellarcomponent 80pcwhichisanapproximatedvalueinferredfromtheHαimage. Table2alsoincludesseveralmetallicitysensitivelineratios. In Figure 3, we present the Hβ and [OIII]λ5007 flux maps for ThemostwidelyusedisprobablytheR23=([OII]λλ3726,3729+ NGC 588 derived from our PMAS data. Contours reproducing [OIII]λλ4959,5007)/Hβ index(Pageletal.1979).However,the archivecontinuumimagesintheredandbluespectralbandshave Z −R23 relation is two valued and thus, independent metallic- been overplotted for reference (see caption of Fig. 3 for details). ity tracers are needed to determine which of the two branches OurPMASdatacover thewholesouthern part of theregion plus is appropiate for NGC 588. One possibility would be the most of the northern one. This GHIIR is dominated by emission N2 = log([NII]λ6584/Hα) line ratio. According to the empiri- fromabrokenelongatedring-likestructurewithmajorandminor calparametrizationproposedbyPe´rez-Montero&Contini(2009), axesof∼40′′and∼25′′(i.e.∼160pcand∼100pc),respectively we derived a metallicity of 12 + log(O/H) = 8.16. This re- andatP.A.∼10◦ and∼100◦.Inaddition,thereisabridgeofion- sult does not point in a conclusive manner to either the up- izedgasemissionjoiningthetwohalvesofthering-likestructure 6 A. Monreal-Iberoet al. Flux(Hβ) Flux([OIII]λ5007) 15 +2.30 15 +3.30 10 10 5 +1.69 5 +2.47 + + 0 0 ) ) c c e e s −5 +1.09 s −5 +1.64 c c r r a a ( ( ∆δ −10 ∆δ −10 −15 −15 +0.48 +0.81 −20 −20 −25 −25 −0.12 −0.02 5 0 −5 −10 −15 −20 5 0 −5 −10 −15 −20 ∆α (arcsec) ∆α (arcsec) Figure3.Left:MapoftheobservedHβfluxderivedfromourPMASdata.Eachspaxelhasa1′′×1′′size.Alogarithmicstretchcovering2.4dexwasusedto betterenhanceallthemorphologicalfeatures.Unitsarearbitrary.ContourscorrespondtotheRbanddirectimagefromNOAOScienceArchive(Masseyetal. 2007)andshowthelocationoftheionizingstarswithintheregion.Theorientation isnorthupandeasttotheleft.Thepeakinthiscontinuumimage,at coordinates RA(J2000):1h32m45.7s,Dec.(J2000):+30d38m55.1s,markstheoriginofourcoordinate systemandwillappearasacrossinthefollowing figuresforreference.Right:Similarmapfortheobserved[OIII]λ5007flux.Thelogarithmicstretchcovers3.3dexandcontoursrepresenttheHST-WFPC2 imagewiththeF336W filter(program5384;P.I.W.William),convolvedwitha0′.′5Gaussianfilter. from∼[1′.′0,−7′.′0]to∼[−7′.′0,−3′.′0]2 atP.A.∼ −70◦.Themor- 10′′×8′′.However,itwasnotdetected,andthusimpossibletobe phologies of the Hβ and the [OIII]λ5007 maps are very similar fittedforindividualspaxels. but show some rather subtle differences which indicate the com- The reddening map was created assuming E(B − V) = plexionizationstructure,whichwillbeexploredinmoredetailin A /3.1(Rieke&Lebofsky1985)andisdisplayedintheleftpanel V section3.6.Themainionizing cluster,asdepicted bythearchive ofFig.4withthe24µmimagefromSpitzeroverplottedwithcon- continuumimages,isnotatthecentreoftheringlikestructure,but tours.Thismapshowshowirregulartheextinctiondistributionis at ∼2′.′0fromthepeak of emission inHβ.Finally, therearealso withlowvaluesofreddening(∼ 0.00−0.25).Thisstronglycon- several secondary peaks of emission, most of them in the north- trastswiththefindingsofJametetal.(2004)who,usinglong-slit, ernhalfoftheregion,whichareassociatedwithverymassive(i.e. reportedanalmostconstantextinctionofE(B−V)=0.11±0.02 30−45M⊙)individualionizingstars(Jametetal.2004). andreinforcestheneedof2Dunbiasedspectralmappingtochar- acterizethephysicalpropertiesofGHIIRs.Theopticalreddening mappresentsthreemaximathatspatiallycorrelateverywellwith 3.3 Extinctiondistributionanddust themaximaofdustemissionintheSpitzer24µmand8µmbands (seeFig. 4). Thisisconsistent with theidea of extinction caused ThedistributionoftheextinctionwasderivedbymeansoftheHα by absorption of dust associated with the GHIIR. Other dust-gas andHβemissionlinemaps.WeassumedanintrinsicBalmeremis- configurationswouldhavecausedadifferentsetofmaps.Forex- sionlineratioofHα/Hβ=2.86(Osterbrock&Ferland2006)fora ample,ifdust werebehindtheregion,therewouldnothavebeen caseB aproximation andTe = 11150 K and usedthe extinction a counterpart in the E(B −V) map to the peaks in the map at curveofFluksetal.(1994).Weincludeda1A˚ correctiontotake 24 µm. In the right panel of Fig. 4 we present the absorbed Hα intoaccounttheHβabsorptionlineduetheunderlyingstellarpop- luminosity-definedasthedifferencebetweenthetotalextinction- ulation. This absorption feature was clearlyvisible ina co-added correctedHαluminosityandtheHαluminositycorrectedforthe spectrumextractedinanrectangularareawithlowsurfacebright- foregroundGalacticextinction,E(B−V)=0.044,(Schlegeletal. nessintheemissionlineslocatedat ∼ [−12′.′0,−17′.′0]ofabout 1998)-obtainedusingourderivedreddeningmapwiththe8µm contoursoverlaid.Acomparisonofthesetwopanelsshowshowthe 24µmemissionpresentsamorecompactdistributiontowardsthe 2 Hereafter,thereportedpositionswillrefertherelativecoordinatestothe centreoftheregionandcorrelatesbetterwiththeabsorbedHαlu- mainionizingcluster. minositymapthanthe8µmemission.ThiswasseeninotherHII NGC 588withPMAS 7 Figure4.Left:ReddeningmapforNGC588obtainedassuminganintrinsicHα/Hβ=2.86andacorrectionof1A˚ inabsorptionforHβ.Theextinctionlaw ofFluksetal.(1994)andE(B−V)=AV/3.1(Rieke&Lebofsky1985)wereutilized.ThemaphasbeenconvolvedwithaGaussianfilterwithσ=1′′to bettertracetheextinctionstructure.The24µmemissionwasoverplottedwithcontours.Theintensitycontoursareat2,5,10,20,40,60,80,95percentof themaximumintensitywithintheregion.A1%contourlevelcorrespondstoa3σvalue.Right:AbsorbedHαluminosityofNGC588with8µmemission contoursoverplotted.Thecontourlevelscorrespondtothesamepercentageasforthe24µmcase.Here,a1%contourlevelcorrespondstoa1σvalue. regions like NGC 604 (Relan˜o&Kennicutt 2009) and NGC 595 themainionizingclusters.Here,weapplythesamemethodology (Relan˜oetal.2010).The24µmemissionhasprovedtobeagood toNGC588. traceroftherecentstarformationforalargerangeofHαluminosi- There are two known WR stars in NGC 588 that have been ties, ranging from moderate HII regions to dusty powerful star- widely studied in the past. The first one, named UIT-011 by bursts(e.g.Riekeetal.2009;Calzettietal.2010).Thespatialcor- Masseyetal.(1996)-MC3byDrissenetal.(2008)-wasdetected relationbetweentheHαluminosityand24µmmapsindicatesthe for the first time using narrow band imaging by Conti&Massey infrared band as astar-formation tracer also inlow dust environ- (1981) and was spectroscopically confirmed by Massey&Conti ments,suchasNGC588.Duetothelowspatialresolutionofthe (1983)lateron.ItwasclassifiedasWNLwithM = −7.9mag. V 70µmand160µmSpitzerbandswewerenotabletocomparethe The second one, named UIT-008 is a transition Of/WN9 star. emission at these wavelengths withour reddening map. Herschel Both stars were modelled using multi-band photometry with the data,whichwillbeavailablesoontothescientificcommunity,will HSTby U´beda&Drissen (2009). They derived effective temper- permitapropercomparisontobemadeinthefuture. atures of 57000 K and 32000 K and bolometric luminosities of log(L/L⊙) = 6.48 and 5.97, for UIT-011 and UIT-008 respec- tively. 3.4 WRstellarpopulation Fig.5presentsthecontinuumsubstractedbluebumpmapaf- Relan˜oetal.(2010)presentedanovelandsimpletechniquetode- ter convolving with a 1′′-Gaussian. We identify two main peaks tectWRstarsinaswiftwayandcompareditsresultstoclassical, of emission whose positions agree well withthose previously re- moretime-consuming techniques. Usingthesamesetofobserva- portedforUIT-011andUIT-008(Drissenetal.2008).Noaditional tions (i.e. the datacube), one can identify the WR candidates by WRinNGC588wasfound.Theextractedspectraforthetwostars simulatingtheactionofnarrowfiltersandcreatingcontinuumsub- ispresentedinFig.6andshowclearlyboththeblueandredbump. tractedmapsattheemissionbumpat4700A˚ (thebluebump)andat ThecasesofNGC595andNGC588,inHIIregions,aswell 5700A˚ (theredbump),characteristicofWRstellaremission,and asexistingonesforstarburstgalaxies(e.g.theAnntenae,IIZw70, localizingthepeaksofemission,afterwards.Then,thecandidates Mrk996,NGC5253,IC10,Bastianetal.2006;Kehrigetal.2008; canbeconfirmedbyextractingthespectraoftheassociatedspax- Jamesetal.2009;Monreal-Iberoetal.2010;Lo´pez-Sa´nchezetal. els.Followingthismethodology,weconfirmedpreviouscataloged 2010)illustratetheeffectivenessofIFSinthefindingandcharac- WRstarsinNGC595anddiscoveredanewonefurtherawayfrom terization of the WR population. At this stage, the possibility of 8 A. Monreal-Iberoet al. 465 − 475 nm 15 +0.08 70 60 UIT−011 a.u.) 50 10 Flux ( 3400 20 10 5 +0.01 4000 4500 5000 5500 6000 λ(Å) UIT−008 UIT−011 + 0 30 UIT−008 u.) 25 sec) −5 −0.06 Flux (a. 1250 c 10 r a 5 ∆δ (−10 4000 4500 5λ0(Å00) 5500 6000 Figure6.ExtractedspectrafortwodetectedpeakofemissioninthePMAS −15 continuumsubtracted4650−4750A˚.Verticalarrowsmarkthepositionof −0.13 theexpectedblueandredbumps. −20 v (km s−1) 15 Hβ −110.00 −25 −0.20 5 0 −5 −10 −15 −20 10 ∆α (arcsec) 5 −130.00 Figure5.4650−4750A˚ mapderivedfromthePMASdataafterconvolv- + 0 ingwithaGaussianofσ=1′.′0.Continuumwassubtractedbyaveragingthe spectralrangesof4490−4540A˚ and4755−4805A˚.Alogarithmicstretch ) c covering∼0.3dexwasusedtobetterenhancethepeaksofemission.Labels se −5 −150.00 withthedetectedWRareincludedandwhitecontoursdelineatethespaxels c r a utilizedtocreatetheextractedspectra.Contourscorrespondtothecontin- ( uumsubtractedRbroad-banddirectimagefromNOAOScienceArchive ∆δ −10 (Masseyetal.2007).Theorientation isnorthupandeasttotheleft.The mainionizingcluster,atcoordinatesRA(J2000):1h32m45.7s,Dec.(J2000): +30d38m55.1s,markstheoriginofourcoordinatessystem. −15 −170.00 usingthistechniqueroutinelyshouldbetakenintoaccount.Inpar- −20 ticular,itwouldsuitperfectlyinthecaseofWRfindingingalaxies atlargerdistancesand,moreimportant,withlargegradientsintheir −25 velocityfields.Here,thetraditionaltechniqueofsearchforcandi- −190.00 datesviaimagingfirst,andspectroscopicconfirmationafterwards, 5 0 −5 −10 −15 −20 might well miss some of the WR populations since the blue/red ∆α (arcsec) bumpmightmoveoutsidethespectralrangeofthenarrowfilter.On thecontrary,themethodologypresentedherecanbeeasilymodi- fiedandimplementedtodefinewhatcanbecalledsynthetictunable Figure7.VelocityfieldderivedfromtheHβemissionline.Contourscor- filtersthattakeintoaccountthemovementsofthegalaxyandthus respond to the continuum subtracted Hα direct image from NOAO Sci- preventingtheselosses.AnadditionaladvantageofusinganInte- enceArchive(Masseyetal.2007).Theorientationisnorthupandeastto gralFieldUnit,especiallyifithasalargefieldofviewisthedetec- theleft.Themainionizingcluster,atcoordinatesRA(J2000):1h32m45.7s, tionofrunawayWRstarsejectedbythecentralstarclusterofthe Dec.(J2000):+30d38m55.1s,markstheoriginofourcoordinatessystem. region(Drayetal.2005).Inparticular,IFS-basedinstrumentswith relativelylargefieldofviewlikePPak/PMAS(Kelzetal.2006),or MUSE (Baconetal. 2010) are(or will be soon) under operation. ityfieldmapsfromthestrongestemissionlines.InFig.7weshow Theyopenthepossibilityofcarryingoutsurveysoflargesamples themapcorrespondingtoHβ.Norelevantdifferenceswerefound ofgalaxieswherethiskindofsimpletechniquescouldbeparticu- fromthemapderivedusing[OIII]λ5007.Thevelocityfieldhasa larlyuseful. complexstructurewithvaluesrangingbetween−190kms−1and −110kms−1.Thenorth-westpartoftheregionseemstobemore redshiftedthanthesouth-east,whichistheregionwithhigherHβ 3.5 Kinematicsoftheionizedgas surfacebrightness(seeFig.3).Inthesurroundingsofthelocation Thesuperiorspectralresolutionofthepresentobservations,twice ofthestellarcluster, markedinFig.7asablackcross, thereisa thatoftheobservationsofNGC595,allowedustoderivetheveloc- pronouncedvelocitygradient:thenorth-westparthasvelocitiesof NGC 588withPMAS 9 −120kms−1whilethesouth-easthasvelocitiesof−170kms−1. 3.6.2 LineratiosintheBPTdiagnosticdiagrams Thevelocityseparationof ∼25kms−1 between thesezones and Diagnostic diagrams, where different areas of a given diagram the location of the stellar cluster, corresponding to a velocity of ∼30kms−1 inthegalaxyplane(inclinationofM33,i = 56deg areoccupied by gasexcited via different mechanisms, have been widelyused tostudy theionization conditions of theISM.Inthe vandenBergh2000),andthesymmetryofthevelocityfieldsug- opticalspectralrange,themostpopularareprobablytheBPTdia- gesttheexistenceofashellexpandingintheinterioroftheregion. Theshell expansion velocity, ∼30km s−1, isslightlylower than grams, firstproposed by Baldwinetal. (1981) andlater reviewed by Veilleux&Osterbrock (1987). Their wide use to study the thevaluesobservedinhighluminosityHIIregionsofasetofspiral galaxies(v ∼ 40−90kms−1,Relan˜oetal.2005)andalsoin ionization conditions in star-forming and starburst galaxies (e.g. exp NGC604(v ∼40kms−1,Yangetal.1996).However,basedon Alonso-Herreroetal. 2010) is due to the fact that they involve exp emission lines that are relatively strong and line ratios that have akinematicstudyofNGC588andNGC604,Mun˜oz-Tun˜onetal. (almost)nodependenceontheextinction. (1996)suggestthatNGC588ismoreevolvedthanNGC604and With the present 2D unbiased mapping, one can determine therefore,wewouldexpectlowervelocitiesfortheshellsinthefirst theselineratioslocally.Inthisway,itispossibletoevaluatetheir regionthaninNGC604,consistentwiththeresultfoundhere. dependenceonthepositionwithintheGHIIRandrelativesurface Inorder tocheckwhether thewindscomingfromthestellar brightnessoftheareaunderstudy.Moreover,onecanmakeacom- population withinthe HII regioncould produce theexpansion of parisonbetweenintegratedandlocalvalues. theobservedshell,wehavemadeacrudeestimationofthekinetic energy involved in the shell and compared with the input kinetic We present the maps for the three available line ra- energyfromthestars.UsingtheHαluminosityoftheregionfrom tios involved in the BPT diagrams - namely [NII]λ6584/Hα, Relan˜o&Kennicutt(2009),wepredictanemissionmeasure(EM) [SII]λλ6717,6731/Hα, and [OIII]λ5007/Hβ - in Fig. 8. These of4000(pccm−6)foranHIIregionradiusof140pc,correspond- maps show that the ionization structure in NGC 588 is complex. ingtotheapertureradiususedtoobtainedtheHαluminosity.The The [NII]λ6584/Hα and [SII]λλ6717,6731/Hα maps present a EMisthenusedtoderivea< n > of5cm−3 andintegrat- rathersimilarstructure.Inbothcases,theminimumislocatednei- e rms ingovertheHIIregionvolumewederiveatotalionizedmassof theratthepeakofemissioninHβ(i.e.ionizedgas)norattheone ∼6×105M⊙.Assuming,asanupperlimitthatthewholemasshas forthecontinuum(i.e.stars)butinthemiddlepointbetweenthem. beensweptupbytheshellweobtainakineticenergyfortheshell Then,lineratiosincreaseoutwards,followingtheringstructureof of3.8×1051erg.Starburst99models(Leithereretal.1999)using theregion. rangesofvaluesforthestellarmassof1−6×103 M⊙ andage The [OIII]λ5007/Hβ map is roughly similar to the of 3.5−4.2 Myr, aSalpeter initialmassfunction, and metallici- [NII]λ6584/Hα and [SII]λλ6717,6731/Hα maps but its value tiesofZ=0.004and0.008givearangeofkineticenergyinputof varies in the opposite sense. Moreover, there are two main 5.9−64.2×1051 erg,atleasttwiceashighastheupperlimitof differences. The peak of [OIII]λ5007/Hβ (maximum for this the kinetic energy of the shell. This crude calculation shows that line ratio) is broader than the peak for [NII]λ6584/Hα and thewindsfromthestellarclusterwithinNGC588areabletopro- [SII]λλ6717,6731/Hα(minimumfortheseones). Morerelevant, ducetheobservedshellintheHIIregionandcreatetheobserved there is an area of ∼ 5′′ × 13′′ centred at ∼ [7′.′0,−8′.′0] of Hβ morphology withholesandfilamentsshowninFig.3.More- elevated [OIII]λ5007/Hβ values that do not show specially low over,thelowerenergyassociatedwiththeshell,30kms−1,com- [NII]λ6584/Hαand[SII]λλ6717,6731/Hα. paredwithvaluescloserto100kms−1 foryounger, lessevolved Inordertobetterassesshowtheionizationconditionschange regions(Relan˜oetal.2005),isconsistentwithwiththefindingsby in different parts of NGC 588, we divided our data in three lu- Mun˜oz-Tun˜onetal.(1996). minosity bins, which sample the low, medium and high surface brightness areas of this GHIIR. In Fig. 9, we present the posi- tionofeachindividual spaxelintheBPTdiagnosticdiagramsto- gether with the borders that separate HII region-like ionization fromionizationbyothermechanismsaccordingtoseveralauthors (Veilleux&Osterbrock1987;Kewleyetal.2001;Kauffmannetal. 3.6 Characterizationoftheionizedgas 2003;Stasin´skaetal.2006).Asexpected,alllineratiosarewithin thetypicalvaluesexpectedforanHIIregion-likeionization. 3.6.1 Densitystructure However, therearedifferencesbetweenthediagramsassoci- Electron density can be determined from the ratio between two ated with the different luminosity bins. Firstly, the range of ob- lines of the same ion emitted by different levels with similar ex- servedvaluesvariesfrom∼0.3to∼0.6dexfor[OIII]λ5007/Hβ, citationenergies.Weusedthe[SII]λ6717/[SII]λ6731ratioinour from ∼0.6 to ∼1.1 dex for [NII]λ6584/Hα and from ∼0.5 to analysis.AsisshowninthelowerrightmapofFig.8,nostructure ∼1.1 dex for [SII]λλ6717,6731/Hα, with larger ranges in those forthen wasfound.Wemeasuredamean(±standarddeviation) areas with lower surface brightness Secondly, as indicated by e [SII]λ6717/[SII]λ6731valueof1.20(±0.17),whichisconsistent the position of the red triangles, higher [NII]λ6584/Hα and withthevaluederivedfromtheintegratedspectrum(seeTable2). [SII]λλ6717,6731/Hαratiosandlower[OIII]λ5007/Hβratiosare For the assumed temperature, this implies a n of ∼250 cm−3 detectedintheareasoflowersurfacebrightness.Thisimpliesthat e and agrees within the uncertainties with the values reported by thedegreeofionizationgetssmallerwithincreasingdistancefrom V´ılchezetal. (1988) and Jametetal. (2005). This value for the theionizingsource. electron density corresponds to the density of the clumps within HowdotheseresultscomparewithourfindingsforNGC595? theregionanddiffersfromther.m.s.electrondensityderivedinthe Ingeneral,thetendenciesinthedifferencesbetweentheintegrated previoussectionusingtheHαsurfacebrightnessoftheregion.The valuesandthosefortheindividualspaxelsinNGC588aresimilar ratioofbothdensityestimatesisameasureofthevolumefraction tothosefound inNGC595. However, therangeof observed val- occupiedbydenseclumps. uesinNGC588issmaller.Thisisseenin[SII]λλ6717,6731/Hα 10 A. Monreal-Iberoet al. log([NII]/Hα) log([SII]/Hα) 15 −0.60 15 −0.40 10 10 5 −0.95 5 −0.73 + + 0 0 ) ) c c e e s −5 −1.30 s −5 −1.05 c c r r a a ( ( ∆δ −10 ∆δ −10 −15 −15 −1.65 −1.38 −20 −20 −25 −25 −2.00 −1.70 5 0 −5 −10 −15 −20 5 0 −5 −10 −15 −20 ∆α (arcsec) ∆α (arcsec) log([OIII]/Hβ) [SII]λ6717/[SII]λ6731 15 +0.90 15 +2.20 10 10 5 +0.77 5 +1.78 + + 0 0 ) ) c c e e s −5 +0.65 s −5 +1.35 c c r r a a ( ( ∆δ −10 ∆δ −10 −15 −15 +0.52 +0.93 −20 −20 −25 −25 +0.40 +0.50 5 0 −5 −10 −15 −20 5 0 −5 −10 −15 −20 ∆α (arcsec) ∆α (arcsec) Figure8.Mapoftheobserved[NII]λ6584/Hα(upperleft),[SII]λλ6717,6731/Hα(upperright),[OIII]λ5007/Hβ(bottomleft)and[SII]λ6717/[SII]λ6731 (bottomright)lineratios.ContourscorrespondtothecontinuumsubtractedHαdirectimagefromNOAOScienceArchive(Masseyetal.2007).Theorientation is northupandeast totheleft. Themain ionizing cluster, atcoordinates RA(J2000): 1h32m45.7s,Dec.(J2000): +30d38m55.1s, marks theorigin ofour coordinatesystem.

See more

The list of books you might like

Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.