Statistical study of OB stars in NGC 6334 and NGC 6357

21 Pages·2012·3.37 MB·English

A&A538,A142(2012) Astronomy
DOI:10.1051/0004-6361/201117299 &
(cid:2)c ESO2012 Astrophysics
(cid:2),(cid:2)(cid:2),(cid:2)(cid:2)(cid:2)
Statistical study of OB stars in NGC 6334 and NGC 6357
D.Russeil1,A.Zavagno1,C.Adami1,L.D.Anderson1,S.Bontemps2,F.Motte3,J.A.Rodon1,N.Schneider3,
A.Ilmane4,andK.J.Murphy5
1 Laboratoired’AstrophysiquedeMarseille–UMR6110,CNRS–UniversitédeProvence,13388MarseilleCedex13,France
e-mail:[email protected]
2 Laboratoired’AstrophysiquedeBordeaux,OASU–UMR5804,CNRS–UniversitédeBordeaux1,2ruedel’Observatoire,
BP89,33270Floirac,France
3 LaboratoireAIM,CEA/DSM–INSU/CNRS–UniversitéParisDiderot,IRFU/Serviced’Astrophysique,CEA-Saclay,
91191Gif-sur-YvetteCedex,France
4 LaboratoiredemathématiqueJeanAlexandreDieudonne,UMR6621CNRSUNSA–UniversitédeNice–SophiaAntipolis,
06108NiceCedex02,France
5 DepartmentofPhysicsandAstronomy,OhioUniversity,251BClippingerLab,Athens,OH45701,USA
Received20May2011/Accepted8December2011
ABSTRACT
Context.Star-formingcomplexesarelargestructuresexhibitingmassivestar-formationatdifferentstagesofevolution,fromdense
corestowell-developedHiiregions.Theyareveryinterestingforthestudyoftheformationandevolutionofstars.NGC6334and
NGC6357aretwoactiveandrelativelynearbystar-formingcomplexes.Fromtheextinctionmapandthesub-mmcolddustemission,
andbecausetheyhavesimilarvelocities,theseregionsaremostlikelyconnected.However,locatedinthedirectionoftheGalactic
centertheirradialvelocityisnotrepresentativeoftheirdistance.AnalternativeisthentodeterminethedistanceofNGC6334and
NGC6357fromtheirstellarcontent.
Aims.OuraimistoperformacensusofO-B3ionisingstarsinNGC6334andNGC6357,todeterminetheextinctioncoefficient,and
thedistanceofbothregions.AcensusofO-B3starsisanessentialbasisforestimatingthestatisticallifetimeoftheearliestmassive
star-formingphases.
Methods.WeperformedaU,B,V,andRphotometricsurveyofalargeareacoveringNGC6334andNGC6357withtheVIMOS
(ESO-VLT)andtheMOSAIC(CTIO)instruments.ThisallowsustohaveacompletecensusofOtoB3starsuptoV =22.6mag.The
OBstarsareselectedbasedontheirU−BandB−V colors.Themostrobustextinctioncoefficientisdeterminedfromcolor−color
plotsbeforecomputingthedistanceoftheOBstars.
Results.WefindahighervaluethantypicalofthediffuseinterstellarmediumforR of3.53±0.08and3.56±0.15forNGC6357and
V
NGC6334,respectively.AdoptingtheseR values,thedistancesofNGC6357andNGC6334are1.9±0.4kpcand1.7±0.3kpc.
V
Weconcludethat,withintheerrorbars,bothregionsarethusatthesamedistanceof1.75kpc(weightedmean).Weconfirmthatthe
valueofR islinkedtothelargedustgraincontent.Inparticular,wefoundthattherearemoreverysmallgrainsinNGC6357than
V
inNGC6334, suggesting thatNGC6357 couldbemoreevolvedthanNGC6334. PlacedintheGalacticcontext,theNGC6334-
NGC 6357 complex appears tobe locatedat the inner edge of theSagittarius-Carinaarm. Our census of O toB3starsleads toa
countof∼230,whichallowsustodeterminethestatisticallifetimeoftheearliestphasesofthemassivestars.Thestarlessandthe
protostellarphaseshaveameanstatisticallifetimeof∼1.5×104yrand∼2.2×105yr,respectively.
Keywords.Hiiregions–dust,extinction
1. Introduction a75to500μmSPIREandPACSimagingofacompletesample
ofmolecularcomplexesmoremassivethanOrionatdistanceof
Star-forming complexes are the main building blocks of the
lessthan3kpc.Thissamplehasbeenbuiltbasedondustextinc-
large-scale structure of galaxies and major sites to study how
tionimagesandCO surveysofthe Galacticplaneandincludes
massivestarsform.However,in contrastto low-massstars, the
NGC6334andNGC6357.
formationofhigh-massstarsremainspoorlyunderstood.Inthis
InourGalaxy,NGC6334andNGC6357aretwoveryactive
framework, the Herschel/HOBYS (PI: F. Motte, A. Zavagno,
star-formingcomplexesseenintheopticalastwoextensiveand
S. Bontemps) GuaranteedTime Key Programmeis performing intensely star-formingHii regions.The extinctionmap and the
morphologyofthe1.2mmcolddustemissionseemtoindicate
(cid:2) Based on observations made with the VIMOS instrument at the thatNGC 6334and NGC 6357are connectedby a filamentary
VLT-ESO.Basedonvisitingastronomerobservations,atCerroTololo
structure, suggesting that both regions belong to a single com-
Inter-AmericanObservatory,NationalOpticalAstronomyObservatory,
plex(Russeiletal.2010).
which is operated by the Association of Universities for Research in
The molecular emission associated with NGC 6334 has a
Astronomy,undercontractwiththeNationalScienceFoundation.
(cid:2)(cid:2) Appendicesareavailableinelectronicformat mean molecular velocity of VLSR = −4 km s−1 (Kraemer &
http://www.aanda.org Jackson 1999) similar to the value from radio recombination
(cid:2)(cid:2)(cid:2) FullTablesA.1andA.2areonlyavailableattheCDSviaanony- lines, VLSR = −3.6 km s−1, of NGC 6357 (Caswell & Haynes
mousftptocdsarc.u-strasbg.fr (130.79.128.5)orvia 1987).Thesesimilarvelocitiesforregionslocatedspatiallyclose
http://cdsarc.u-strasbg.fr/viz-bin/qcat?J/A+A/538/A142 to each other indicate that they are both at a similar distance.
ArticlepublishedbyEDPSciences A142,page1of21
A&A538,A142(2012)
Zone 2 Zone 1
354.000 G353.2+0.9
2.000
NGC6357
353.000 1.000 Zone 7
Zone 4
GUM 63 Zone 3
352.000 GUM 64b
GUM 62
-1.000 G351.2+0.5 GUM 61
NGC6334 Zone 5
351.000 GM24 Zone 6 Zone 8
Fig.1.a)Hαimage(UKSTHαsurvey,Parkeretal.2005)oftheHiiregionsNGC6334andNGC6357(Galacticcoordinates).b)Hαimagewith
overplottedthecoverageoftheVIMOSobservations(green)andCTIOobservations(red),aswellasthezones1to8(white)discussedinthetext.
Zones2,6,and8correspondrespectivelytoNGC6357,NGC6334,andGM24.
Independantdistance determinationsof each regionare notyet 2. Observationsanddataextraction
well-established. Owing to their Galactic coordinates close to
theGalacticcenterdirection,l, b = 351.16◦,+0.69◦ andl, b = 2.1.U,B,andVVIMOSdata
353.01◦,+0.89◦,respectively,forNGC6334andNGC6357,the We observed NGC 6334, NGC 6357, and GM 24 (GM 1-24,
kinematicdistanceisunreliable.Thebestwaytoderivetheirdis- Gyulbudaghian et al. 1977) in U, B, and V bands with the
tancesisfromthedeterminationofthedistancetotheirexciting VIMOS camera (ESO-VLT). VIMOS is a visible (360 to
stars. 1000 nm) wide field imager and multi-object spectrograph
mounted at the Nasmyth focus B of UT3 Melipal. We use
As reviewed by Persi & Tapia (2008), NGC 6334
VIMOS in its imaging mode. This led to the observation of
(=Sharpless 8 = RCW 127) contains the Hii regions (Fig. 1) 67 fields, each composed of 4 quadrants of 7(cid:5) × 8(cid:5) separated
GUM 61, GUM 62, GUM 63, and GUM 64 (Gum 1955) and by2(cid:5) gaps.ThecoverageispresentedinFig.1.Thetypicalob-
G351.2+0.5. Roslund (1966), Neckel (1978), Walborn (1982),
servingmethodwasthreeexposuresof3sinB,sevenexposures
and Persi & Tapia (2008) computedthe distance to NGC 6334
of 4 s in U, and four exposures of 3 s in V. The observa-
fromthedistancestothevisibleearly-typestarsintheseHiire-
tionsweretakeninservicemodeanddeliveredflatfielded,bias-
gions and obtained d = 1.45 kpc, 1.74 kpc, 2.30 kpc and
correctedandwith calculatedastrometricandphotometricdata
1.61kpc,respectively.Persi&Tapia(2008)underlinedthatthe (zeropoint,extinctioncoefficient).
discrepancies are due to the different adopted M calibrations
V We extracted stellar positions and magnitudes using
andtheadoptedextinctionlaw(R value).
V SExtractor (Bertin & Arnouts 1996). We kept objects a with
SExtractorflag≤3(suchaconditionensuresagoodphotometry
Intheoptical,NGC6357exhibitsseveralbubblesandshell-
like regions. The brightest Hii region (G353.2+0.9) shows a andexcludessaturatedobjects).Inaddition,weexcludeddataat
thebordersoftheCCDandindeadcolumns.Allobjectsbrighter
sharpboundaryfacingthemassiveopenclusterPismis24.The
than V = 10 mag were typically saturated. As we treated the
distance of NGC 6357 is usually established from the distance
imageindependently,weperformedacross-correlationbetween
of Pismis 24 (the exciting cluster of NGC 6357); its most re-
theextractedU, B,andV datainordertoestablishacatalogue
cent determination was that of Massey et al. (2001), who give
a value of 2.5 kpc. This distance differs from the 1.7 kpc ofstarswithU,B,andV magnitudes.
We measured the completeness as in Adami et al. (2006)
suggested by Neckel (1978) and Lortet et al. (1984) and the
1.1 kpc and 1.74 kpc derived by Conti & Vacca (1990) and andDurretetal.(2010).Thismethodaddsartificialstarsofdif-
ferent magnitudesand at different locations to the original im-
Van der Hucht (2001), respectively, for the Wolf-Rayet star
ages and then attempts to recover them by running SExtractor
HD157504(WR93).
again with the same parameters used for object detection and
ThedistanceofNGC6334andNGC6357hasthereforenot classification on the originalimages. We investigatedthe com-
yetbeenwell-established.However,sincethedistanceisanes- pletenessfromthreetypicalimageslocatedatdifferentpartsof
sential parameter for the determination of the mass, size, and thesurveyedareawithvaryinglevelsofbackgroundandsource
luminosity of the associated young objects (e.g. Russeil et al. crowding.The90%meancompletenesslevel(Fig.2)isrespec-
2011), it is essential to evaluate in an homogenous way the tively21.8,22.4,and22.6magforU,B,andV filters.Figure2,
distance of the early-type stars in the direction of NGC 6334 shows how the completeness levels vary from CCD to CCD
and NGC 6357. In this paper, we focus on the content of because of the quantum efficiency variations between individ-
ionizingOBstarsofthestar-formingcomplexesNGC6334and ualCCDsandbecauseoffluctuationsinthediffusebackground
NGC 6357 to more tightly constrain their distances, to place light.
them in the Galactic context, and to perform a census of their Fromastatisticalstudyofthezero-point,weestimateddis-
OBstars. persionsof0.09magforU,0.04magforB,and0.03magforV.
A142,page2of21
D.Russeiletal.:StatisticalstudyofOBstarsinNGC6334andNGC6357
center.ForthepurposeoflocatingOandBstars,wekeptstars
with magnitude values in the three (U, B, and V) bands. This
providedasampleof30388stars(U, B,andV magnitudesare
presentedinAppendix,TableA.1).
2.2.RbandCTIOimages
We obtained R images of NGC 6334, NGC 6357, GM 24,
and the more diffuse area between NGC 6334 and NGC 6357
withtheMOSAICcamera(Blanco-CTIO).Thisinstrumentisan
eightCCD mosaic imagerwith a pixelscale of0.27(cid:5)(cid:5) mounted
at the prime focus of the 4-m Blanco telescope at the Cerro
TololoInter-AmericanObservatory(CTIO).Theseobservations
(a)
consist of six fields composedof three exposuresof 10 s each.
The global coverage is presented Fig. 1. Each field consists of
eightquadrants(separatedby gapsof ∼13(cid:5)(cid:5) in rows and ∼9.5(cid:5)(cid:5)
incolumns)coveringa36(cid:5)×36(cid:5)fieldofview.Thebiasandthe
flat-fieldcorrectionswereappliedtoeachoftheeightCCDim-
ages independently, while the combination of the different ex-
posures, the astrometry, the reconstruction of the final image
and the construction of the weight images were done with the
modules MissFITS, SCAMP, and SWarp of TERAPIX (Bertin
et al. 2002). The photometric calibration was performed using
the observation of the Landolt calibration stars S107a, S107b,
S104, and S98 (Landolt2009), giving an extinction coefficient
of0.095±0.043magandazeropointof25.57±0.02mag.The
extractionofthestellarpositionsandmagnitudeswasdoneusing
SExtractorwiththeweightimagesproducedbytheSWarpmod-
(b) ule. We kept only objects with a SExtractor flag below 3, and
excludedstarsin thebordersoftheCCD andin deadcolumns.
This provided a sample of 53420 stars (the R magnitudes of
whicharepresentedinAppendix,TableA.2).Theestimatedsat-
uration limit is 11 mag and the typical magnitude uncertainty
(quadraticsumofthemagnitudeerrorsgivenbySExtractor,the
zero-pointuncertainty,andtheextinctioncoefficientuncertainty)
is0.05mag.
3. Otheravailabledatasets
3.1.Near-Infrareddatafrom2MASS
To complement our optical data, we used data from the Two
MicronAllSkySurvey(2MASS)pointsourcecatalog(Skrutsie
et al. 2006). The 2MASS survey scanned the entire sky uni-
(c) formly in three near-IR bands: J (1.25 μm), H (1.65 μm), and
K (2.17 μm). The 2MASS point source catalog1 (Skrutsie
S
Fig.2. U a), B b), and V c) completeness level in individual CCDs et al. 2006) consists of accurate positions (astrometric accu-
(eachdottedcurvecorrespondstoaCCDquadrant,andtheblackcurve racy rms better than 200 mas) and brightness information for
is the average of these dotted curves) for the fields centered respec- over400millionpointsources.Thepointsourcecatalogismore
tively at 17h20m10s.0; −35◦57(cid:5)18(cid:5).(cid:5)0 (NGC 6334, blue), 17h22m09s.0; than99%completeformagnitudesof J < 15.8,H < 15.1,and
−35◦03(cid:5)06(cid:5).(cid:5)0(green),and17h25m37s.0;−34◦13(cid:5)27(cid:5).(cid:5)3(NGC6357,red). K < 14.3 mag. The photometric signal-to-noise ratio is >10
S
(i.e.σJ,H,Ks ≤ 0.1mag)forsourcesbrighterthanthecomplete-
nesslimits.Thetypicalmagnitudeerroris0.05mag,0.06mag,
The delivered photometric calibration neglects the color term. and0.05mag,respectivelyinJ,H,andK .
S
Thus we estimated the typical error caused by this assumption
about the mean color coefficient estimated for the VIMOS in-
strumentandthemeancolorofOBstars.Thisleadstoatypical 3.2.DENISdata
error of 0.017 mag in U and B and 0.006 mag in V. We then
The DENIS2 imaging survey (Epchtein et al. 1997) gives a
estimatedtheerrorinthemagnitudeofastartobethequadratic
nearly complete overview of the southern sky in three NIR
sumofthemagnitudeerrorsgivenbySExtractor,thezero-point bands: Gunn-I (0.85 μm), J (1.25 μm), and K (2.15 μm). The
uncertainty, and the color term uncertainty. This led to typical S
imageshavebeentakensimultaneouslyinallthreebands,which
errorsof0.04mag,0.09mag,and0.23mag,respectively,inV,
B,andU. 1 http://irsa.ipac.caltech.edu/
Finally, as some fields overlap, we identified the objects 2 Pointsourcecatalogavailablefromhttp://irsa.ipac.caltech.
counted twice and retained those located closer to the field edu/
A142,page3of21
A&A538,A142(2012)
in the DENIS, 2MASS, and USNO catalogues using a 5(cid:5)(cid:5) cir-
cular cone search radius. We identified 2148 OB stars with a
2MASScounterpart.
From the O-B3 star sample we discuss, in the nextsection,
the extinction law (Sects. 5.1 and 5.2) in connection with the
dustproperties(Sect.5.3),theGalacticstructure(Sect.5.4)and
theOBstarscensus(Sect.5.5).
5. Discussion
5.1.Extinctionlaw
The radiation from massive OB stars can modify the interstel-
Fig.3.Color−colordiagramforstarswithU,B,andVmagnitudes.The lar grains present in their immediate vicinity, hence affect the
redcurve isthetheoretical unreddened main-sequence andthelineis global interstellar extinction law (e.g. Chini & Wargau 1990;
thestandardreddeningvector,uptoA =5mag,foraB3Vstar(from Pandeyet al. 2000). Any changein the interstellar law will af-
V
Schimdt-Kaler1983). fect the local extinction and thus the photometric distance de-
termination.Manyinvestigations(e.g.Wegner1994;Whittet&
leads to a veryhigh accuracy in the colorsof the objects inde- vanBreda 1980; Rieke & Lebofsky 1985) have shown that the
pendentofphotometricerrors.TheDENISdatahaveapositional normalextinctionlaw(R = 3.1)isvalidnearlyeverywherein
accuracybetterthan2(cid:5)(cid:5) andthemagnitudelimitsare18.5mag, ourGalaxy,althoughsomVeexceptionsarenotedespeciallyinthe
16.5 mag, and 14 mag in I, J, and KS, respectively. The typi- directionofstar-formingregionsandopenclusters,wheretheRV
calmagnitudeuncertaintyis0.15mag,0.11magand0.10mag value is higher (e.g. O’Dell & Wen 1992). Different R values
V
inJ,KS,andI,respectively. relate to different environmental conditions and thus different
grain-size distributions and compositions, since the extinction
3.3.USNO-B1.0data dependsontheopticalpropertiesofthedustgrains(e.g.Whittet
USNO-B is an all-sky survey obtained from scans of 7435 2003;Mazzei&Barbaro2011).Inparticularlow-RV valuesare
Schmidt plates taken for the various optical sky surveys dur- expectedarisealongthelinesofsighttosmallergrains.
ing the past 50 years. The USNO-B1.0 is complete down to Therearefourmainmethodsfordeterminingtheextinction
V = 21mag,hasanastrometricaccuracyof0.2(cid:5)(cid:5),andaphoto- andtheextinctionlawinagivendirection:
metricaccuracyof0.3mag(Monetetal.2003).Fromthepoint
1. Thespectro-photometricmethod,whichisbasedonthecom-
sourcecatalog3weextractedRandImagnitudes.
parisonoftheobservedenergydistributionofastarofknown
spectraltypewithitsnon-reddeneddistribution(e.g.Cardelli
4. TheO-B3starsample
1989).
We selected the youngand massive stars (O to B3 stars) based 2. ThePaschen-Balmerlinesratiomethod,whichusesthecom-
ontheirU −Band B−V colors.Theinterstellarreddeningaf- parisonbetweenthepredictedandobservedPaschen-Balmer
fectstheintrinsicU−BandB−Vcolorsfollowingthereddening linepairratioemittedbytheHiiregions(e.g.Greve2010).
line thathas the form: EU−B/EB−V = X +YEB−V. In this equa- 3. The color excess method, which uses two color-excess di-
tion,EU−B andEB−V arethecolorexcesses,Xistheslopeofthe agrams of the form E(λ−V) versus E(B−V) (where λ is
reddeningline,andY isthecurvatureterm.Weusedtheredden- oneofthewavelengthsofabroadbandfilter)ofagroupof
ingpath EU−B/EB−V = 0.72+0.05EB−V (Schmidt-Kaler1983). starsandrequiresthespectraltype,hencetheintrinsiccolor,
ToselectOtoB3starsfromtheU −B/B−V color−colordia- ofobservedstarstobeknown(“methodA”ofPandeyetal.
gram, we selected stars above the reddening path of a B3 star 2003).
(Fig. 3). Following Turner (1996), this corresponds to a nor- 4. The photometric method, which uses two color diagrams
mal extinction law (R = 3.1), while larger R (as expected (TCDs)oftheform(λ−V)versus(B−V)(whereλisone
V V
forstar-formingregions)seemtoinduceashallowerslope.This ofthewavelengthsofabroad-bandfilter)ofagroupofstars
does not impact our selection process as stars following a red- andassumesthattheobservedstarsaremainsequencestars
dening law with a shallower slope will be above the adopted (“methodB”ofPandeyetal.2003).Intheseplots,thered-
reddeningpath.Inadditionthereddeningpathdoesnotdepend dening path coincides with the unreddened main sequence
strongly on the adopted intrinsic colors. To test this aspect we causing the stars of different spectral types and extinctions
plotEU−BversusEB−Vforstarswithknownspectra(Table2)and to form a linear relation parallel to the reddeningline. The
fittedtheEU−B/EB−V = X+YEB−V relationshipfortheSchmidt- slope of this distribution, relative to the expected slope for
Kaler(1983)andforthemostrecentMartinsetal.(2006)color the normal reddening path, allows one to derive the value
calibrationsofOstars.Ifoursmallsampledoesnotallowusto ofR .
V
constrain reliably X and Y, the fit results are similar for both
calibrations (with a smaller than 1% difference in X and Y). AsweareperformingastatisticalstudyofOBstarsinthedirec-
tionofNGC6334-NGC6357andhavenospectralinformation
However,Turner(1989)showedthattheslopeoftheUBVred-
aboutthestars,thelastmethodisthemostappropriate.Toinves-
deninglinevariesfromoneregionoftheMilky-Waytoanother
overarangeofatleast X = 0.62to0.80.Iftheslopeissteeper tigatethevalueofRV inthedirectionofNGC6334–NGC6357,
we used TCDs with λ as one of the wavelengths of the broad
than0.72,thiscancauseustomisssomeB3stars.
band filters (R, I, J, H, K). These TCDs provide an effective
Thisprovidedasampleof2394candidateOBstars.Wethen
methodforseparatingtheinfluenceofthenormalextinctionpro-
searchedforobjectsassociatedwiththe2394candidateOBstars
ducedbythediffuseinterstellarmediumfromthatoftheabnor-
3 Availablefromhttp://irsa.ipac.caltech.edu/ malextinction.In these diagrams,the slope of the distribution,
A142,page4of21
D.Russeiletal.:StatisticalstudyofOBstarsinNGC6334andNGC6357
Table1.R results.
V
Zone: 1 2 3 4 5 6 7 8
NGC63357 NGC6334 GM24
R : 3.32±0.12 3.53±0.08 3.44±0.07 3.37±0.03 3.46±0.06 3.56±0.15 3.16±0.08 3.89±0.09
V
Table2.StellardataavailableinliteratureforNGC6357(Pismis24)andNGC6334.
Name Equatorialcoordinates V B−V U−B Spectraltype V−K J−H H−K R d A
S V V
J2000(hms◦(cid:5)(cid:5)(cid:5)) mag mag mag mag mag mag kpc mag
NGC6357
HDE319718=Pis24-1 172443.41−341156.5 10.43 1.45 0.40 O3If* 4.538 0.551 0.283 3.37* 1.65 5.75
HD157504=WR93 172508.79−341112.1 11.46 1.15 WC7(+abs?) 5.595 0.503 0.668 − − −
Pis24-17 172444.70−341202 11.84 1.49 0.28 O3III(f*) 4.865 0.535 0.306 3.52* 2.03 6.39
Pis24-2 172443.20−341243.5 11.95 1.41 0.32 O5.5V((f)) 4.260 0.475 0.307 3.27 1.90 5.99
Pis24-15* 172428.86−341450.3 12.44 1.18 0.26 O8V 3.974 0.450 0.235 3.38 2.71 5.01
Pis24-13* 172445.68−340939.2 12.32 1.33 0.22 O6.5V((f)) 4.784 0.574 0.289 3.47 2.08 5.74
Pis24-3* 172442.21−341321.0 12.77 1.44 0.35 O8V 4.279 0.524 0.298 3.30 1.92 6.09
Pis24-8* 172438.81−341458.2 12.95 1.55 0.42 4.460 − 0.208 − − −
Pis24-10 172435.94−341359.9 12.96 1.50 0.38 O9V 4.272 0.487 0.276 3.31 1.66 6.33
Pis24-16 172444.30−341200 13.02 1.60 O7.5V 5.649 0.592 0.490 3.76 − −
Pis24-7 172447.81−341516.5 13.46 1.68 0.58 5.036 0.608 0.327 − − −
Pis24-12 172442.22−341141.1 13.88 1.47 0.38 B1V 4.097 0.508 0.208 3.07 1.51 6.19
Pis24-4 172440.39−341205.9 13.93 1.43 0.53 4.792 0.595 0.315 − − −
Pis24-18 172443.20−341142 13.97 1.48 B0.5V: 4.998 0.683 0.341 3.61 − −
Pis24-9 172439.29−341526.4 14.26 1.40 0.40 4.258 0.490 0.288 − − −
Pis24-11 172434.68−341317.1 14.53 1.57 0.30 4.741 0.660 0.296 − − −
Pis24-19 172443.50−341141 14.43 1.39 B1V 4.818 0.592 0.467 3.71 − −
NGC6334
CD-3511146 171907.5−353746 11.44 0.76 −0.04 B2IV 2.05 0.24 0.09 2.98 1.59 3.48
HD319701 171916.0−355408 10.10 1.08 0.04 B1Ib 3.14 0.37 0.21 3.19 1.73 4.65
HD39699 171930.4−354236 9.63 0.77 −0.24 O5V((f)) 2.29 0.22 0.15 3.37 1.75 3.78
CD-3511477 172005.0−355638 11.11 0.90 −0.03 B0.5V 2.46 0.27 0.09 3.05 1.4 4.12
HD319697 172024.8−354232 10.33 0.66 −0.24 B1V 1.62 0.18 0.04 2.84 1.10 3.28
CD-3511482 172026.5−354407 10.70 0.73 −0.25 B0.5Ve 3.42 0.38 0.43 4.58 1.41 3.61
CD-3511483 172033.6−360626 11.65 1.03 −0.10 B1e 2.53 0.30 0.19 2.80 − −
HD319703B 171945.0−360547 11.20 1.25 0.04 O6.5V((f)) 3.61 0.28 0.28 − 1.26 5.64
HD319703A 171946.2−360552 10.71 1.14 0.04 O7III 3.66 0.38 0.27 3.58 1.78 5.12
CD-3511484 172049.8−355221 11.33 0.95 0.04 B1V 2.73 0.27 0.18 3.17 1.10 4.27
HD319702 172050.6−355146 10.13 0.90 0.04 O8III((f)) 2.71 0.31 0.17 3.43 2.11 4.03
HD156738 172052.7−360421 9.37 0.86 −0.14 O6.5III((f)) 2.58 0.26 0.16 3.55 1.58 4.05
Notes. Spectral types are from Massey et al. (2001) for NGC 6357 and fromPersi & Tapia (2008) and Pinheiro et al. (2010) for NGC 6334.
AsteriskfollowingthenameinCol.1meansthatU,BandV dataarefromthispaper.AsteriskfollowingvalueintheR columnmeansthedata
V
arefromBohigasetal.(2004).JHK dataarefrom2MASS.DistanceandA werecalculatedfromtheR of3.56and3.53forstarsinNGC6334
S V V
andNGC6357,respectively.
m ,iscomparedtothetheoreticalslope,m (Pandeyetal. 2. WekeptonlystarswithV <21mag(toensurecompletness)
obs normal
2003).ToderivethevalueofR ,weusedtherelation(seeSamal andwithB−V uncertainties<0.15mag.
V
etal.2007;Pandeyetal.2000) 3. For CTIO data we retained stars with V − R uncertainties
≤0.15mag.
R =(m /m )×R , (1) 4. Forthe2MASSandDENISdata,wekeptstarswithmagni-
V obs normal normal
tude uncertainties≤0.15mag and magnitudesbrighterthan
whereR =3.1. themagnitudelimits(seeSects.3.1and3.2).
normal
ToinvestigatetheR valueinNGC6334-NGC6357,wedi-
V
vided the area into eight zones based on the individual region These criteria helpedto minimizethe incompletenessproblem.
(see Fig. 1) and plotted the TCDs for each zone. Zones 2, 6, For USNO data, we adopted an uncertainty of 0.3 mag for R
and 8 encompass NGC 6357, NGC 6334, and GM24 respec- andI.
tively. Zone 3 is the area between NGC 6357 and NGC 6334. IntheTCDs,thedatawerefitusinga1σ-clippinglinearre-
Zones4and1andzones5and7areontheeasternandwestern gression(seeinAppendixB,Figs.B.1toB.8),andtheresultsare
bordersofNGC6357andNGC6334,respectively. giveninTable1(thedetailsofthelinearregressionsandtheR
V
BeforeplottingtheTCDs, weadoptthefollowingselection determinationaregiveninAppendixB).Toillustrateourresults,
criteriaforoursample: we present in Fig. 4 some representativecolor−colordiagrams
(alltheplotsarepresentedinAppendix).Intheseplots,thema-
1. WeremovedIR-excessobjectsbasedontheir2MASSJ−H jority of the stars follow a linear distribution.Few stars are lo-
and H −K colors.FollowingHansonetal.(1997),wekept catedabovethegeneraltrend.Metallicityoragevariationscan-
starswith(J−H)−(1.83×(H−K)+0.15)≥0. notexplainthisresult,becauseametallicityeffectcorrespondsto
A142,page5of21
A&A538,A142(2012)
mostlikelyexplanationisthemis-associationwithnear-infrared
data. As expected, this effect is more importantin crowded ar-
eas such as zone 4, which is the closest to the Galactic plane.
In the distribution of the stars, we can delineate overdensities
(whichareclearlyillustratedinFig.4a).Theseprovideinforma-
tionaboutthenumberoftheextinctionlayerspresentalongthe
lineofsight.IntheTCDofzone4,thelowestGalacticlatitude
zone, we can distinguish up to four extinction layers centered
on (B − V) ∼ 0.7, 1, 1.3, and 1.75. Stars are mainly located
around(B−V) = 1.5and(B−V) = 1,respectively,inzones2
(NGC6357)and6(NGC6334),underliningthesmallextinction
differencebetweenbothregions.Finally,inzones2,6,and7we
notethesmallnumberofstarsaroundandbelow(B−V) = 0.5
thatdepartfromthegeneraltrend.Thesestarsareprobablyvery
nearbyforegroundstars.
WenotethatR variesfromzonetozone.Fromourdatawe,
V
(a)
confirmahighervalueofR inNGC6357andNGC6334than
V
normalextinction.ThevalueofR isrespectively3.53±0.08and
V
3.56±0.15forNGC6357andNGC6334,whichisverysimilar
for both regions. Neckel & Chini (1981) found a mean value
ofR of3.8forNGC6357-NGC6334usingasimilarmethod.
V
Thisvaluecomesfrom(R−V)versus(vs.)(B−V)and(I−V)
vs.(B−V)plots(givingR = 4.08and3.64,respectively)only
V
andusingstarsinbothNGC6357andNGC6334.Considering
thesameplots,theNeckel&Chini(1981)resultsareconsistent
with ours (see Table B.3). Bohigas et al. (2004) found that for
Pismis 24 (NGC 6357) R = 3.51, and Pinheiro et al. (2010)
V
foundthat(cid:8)R (cid:9)=3.5forstarsinNGC6334.
V
The largest value, R = 3.89 ± 0.09, is found for zone 8
V
correspondingtoGM24.ThezonesoutsidestheHiiregionsbut
closertotheGalacticplane(zones3,4,and5)exhibitasimilar
R ,around3.43,whichis,despitethemorediffuseaspectofthe
V
interstellarmediumexpectedinthese zones,andis higherthan
thenormalvalue.ThezonesathigherGalacticlatitude(zones1
(b) and7)exhibitacloser-to-normalR withameanvalueof3.25.
V
5.2.Starswithknownspectra
TheextinctionparametersR andA canalsobedeterminedfor
V V
starsofknownspectraltype.Thiscanthenbecomparedwithour
purephotometricapproach.WecompiledinTable2fromthelit-
eraturealltheOandBstarsinNGC6334andNGC6357with
a knownspectraltype.Inthistable,Cols. 1to3 and7givethe
name, the coordinates, and the spectral type of the stars, while
Cols.4to6and8to10givetheUBVJHKphotometricinforma-
tion. Column 11 gives the R value deduced from Eq. (2) (see
V
below),whilethelasttwocolumnsgivethedistanceandtheex-
tinction for each star calculated with the photometricR value
V
determinedintheprevioussection.
Moffat & Vogt (1973) obtained photometry of 15 stars in
Pismis24,forwhich12aremembers.Masseyetal.(2001)ob-
tainedspectroscopyof11stars,fourofwhichwerenotinthelist
(c)
ofMoffat&Vogt(1973).Theyfoundthattwooftheclusterstars
Fig.4. V − K versus B−V for zone 4 a), zone 2 = NGC 6357 b), are of type O3, one of which is a supergiant (HDE 319718 =
andzone6=NGC6334c).Thestarsidentifiedspectroscopically(see Pis 24-1) and the other appears to be a giant (Pis 24-17). On
Sect.5.2)arealsooverplotted(filledtriangles).Thesmallthicksegment the basis of the 10 stars with reliably determined luminosity
onthebottomleftoftheplotsisthelocusoftheunreddenedO-B3stars classes, Massey et al. (2001) inferred a distance modulus of
(fromMartinsetal.2006;Koornneefetal.1983;andWegner1994). 12.03±0.14mag.WhentheyusedonlythesixOdwarfstode-
rivethedistance,theycomputed11.99±0.05mag.Theyadopted
a distance modulusof12.0magcorrespondingto a distanceof
acolorvariationsmallerthan0.1mag(DeGrijsetal.2001)and 2.5 kpc, which is larger than the 1.7 kpc suggested by Neckel
anagevariationcanberepresentedasisochronesparalleltothe (1978) and Lortet et al. (1984). The cluster also contains the
reddeningvectorincolor−colordiagrams(Leitherer1999).The Wolf-Rayet star HD 157504 (WR 93), which is of type WC7.
A142,page6of21
D.Russeiletal.:StatisticalstudyofOBstarsinNGC6334andNGC6357
Conti&Vacca(1990)describedthisstarasa“WCE+abs”and sizedistribution.Seab&Shull(1983)foundsignificantgrainde-
derivedadistanceof1.1kpc. structiononlyforshockvelocitiesabove∼40kms−1.Theypre-
In contrast to NGC 6357, NGC 6334 is ionized by a dictedthatlargegrains,withsizeslargerthan500Å(thegrains
small number of lightly reddened OB stars (Table 2) that are that produce most of the 100 μm radiation), are preferentially
spread about the whole nebula. Persi & Tapia (2008) sum- affected.
marized the available stellar data for early-type stars in the To investigate this finding in terms of the value of R , we
V
region. NGC 6334 is a grouping of the well-known HII re- evaluated the average infrared (12, 25, 60, and 100 μm) fluxes
gions GUM 61, GUM 62, GUM 63, and GUM 64. The ex- (backgroundsubtracted)fromIRIS4 inthe differentzones.The
citing stars of GUM 61 (HD319703A and HD319703B) and MIPSGAL24and70μm images(Careyetal.2009)cannotbe
GUM 62 (HD156538) can be clearly distinguished, while the used because most of the HII regions are saturated. However,
stars HD319702and CD-35 11484 appear as possible exciting IRAS data facilitates the comparison with dust models similar
starsofGUM64. tothoseofDésertetal.(1990).TheIRASmeanfluxesandtheir
In GM24, no optical exciting star has been identified. uncertaintyaremeasuredfromthecalibratedimageswithineach
However, a small cluster dominated by a few young massive areausingtheds9-“Funtools”tool.Thefluxesarecorrectedfrom
starsliesatitscore(Tapiaetal.1991).Onthenear-infraredpho- abackgroundvaluetakenatthesamelocation,onthenorth-west
tometryoftheclusterstarsandextinctionconsiderations,Tapia border (to minimize the contribution of the Galactic plane), in
etal.(1991)deducedadistanceof2kpc. every12to100μmimage.
FollowingSamaletal.(2007),onewaytoestimatethevalue As expected from Hα emission and OB star content, and
of RV is to use stars of known spectral type (Table 2). The RV based on a large I(60)/I(100) and low I(12)/I(25) ratio corre-
value toward a star can be evaluated using the empirical rela- sponding to an increase in the intensity of the radiation field
tionshipestablishedbyWhittetetal.(1976) (Boulanger 1988), zones 2, 6, and 8 (corresponding, respec-
R (cid:10)1.1E(V−K)/E(B−V). (2) tively, to the HII regions NGC 6357, NGC 6334, and GM24)
V
clearlyexhibitastrongerradiationfieldthanotherzones(Fig.5).
Equation(2)isbasedonthefindingthattheratioAV/E(V−K)(cid:10) However, NGC 6334 appears to have a higher radiation field
1.1doesnotchangeappreciablywiththeRV valueandacrossthe than NGC 6357, while NGC 6357 contains more O-type stars
Galaxy(Whittet&vanBreda1978,1980).ToapplyEq.(2),we (Table 2) and has a higher radio flux than NGC 6334 (see
transformedtheJHKSmagnitudesofthestarsfromthe2MASS Sect.4.6).
systemtotheKoornneefsystem(Koornneef1983)usingthere- To characterise the dust properties, we determined the ra-
lationgivenbyCarpenter(2001).Intrinsic(V−K)0colorswere tios of mid-infrared to far-infrared emission (defined as X1 =
takenfromKoornneef(1983)andintrinsic(B−V)0colorsfrom νI(12 μm) / [νI(60 μm) + νI(100 μm)] and X2 = νI(25 μm) /
Schmidt-Kaler (1983). The results are listed in Table 2. From [νI(60μm)+νI(100μm)]).Theseratiosaretoolsforprobingthe
these results, we note that RV varies from place to place on a contentof small grains such as PAHs (polycyclicaromatic hy-
typical spatial scale that is smaller than our zone size. In this drocarbons)andVSGs(verysmallgrains)relativetobiggrains
way,Pismis 24 (NGC6357)exhibitsa meanRV = 3.43,while (Désertet al. 1990).In particular,X1 dependsprimarilyonthe
RV =3.42forstarsassociatedwithNGC6334.Theseresultsare abundanceofsmallparticlessuchasthePAHs.Figure5presents
slightlysmallerthan,butinagreementwith,thoseobtainedfrom thevariationinX1andX2withR .BothX1andX2havevalues
V
thephotometricapproach(RV = 3.53±0.08and3.56±0.15in similartothoseinDésertetal.(1990)forHIIregions.
thedirectionofNGC6357andNGC6334respectively). Globally,X1showsnotrendrelativetoR ,whichisconsis-
V
From our photometric determination of RV, we found for tentwith RV dependingonbig grains,butalso onthe radiation
NGC6357andNGC6334ameandistanceof1.93±0.36kpcand field. When we assumed that zone 3 is the most representative
1.72±0.26kpc,respectively,andameanAV of5.93±0.49mag of the typical dust properties, as it is the least affected by HII
(Bohigas et al. 2004, give for Pismis 24 AV = 6.37 mag) and regionsandstarformation,wefoundthatthezones1,4,6,and8
4.52±0.68mag,respectively.Fangetal.(2011),fromisochrone are deficient in small particles (PAH and VSG) while zones 7
fittingofOstarsinPismis24,foundadistanceof1.7±0.2kpc and2areoverabundant.NGC6357(zone2)appearstohavethe
andamedianextinctionof5.3mag.We canthenconcludethat highest X1 value suggesting that shocks can act in addition to
NGC6334andNGC6357areatanaveragedistanceof1.75kpc thephotonfluxwithinNGC6357.Thisisconsistentwith mor-
(weightedmean). phology(filamentarystructures)ofNGC6357asseeninHα.
In the X2 versus R plot, a clear split is observed between
V
HIIregions(zones2,6,and8)andotherregions.TheseHIIre-
5.3.Generaldustproperties
gionsexhibitahigherX2,whilezones1,3,4,5,and7havesimi-
Small values of RV are generally assumed to relate to a preva- larX2valuesofaround0.10±0.02.Thisrelativeexcessemission
lenceofsmalldustgrains,whichaffectstheextinctioncurveat at25μmcouldmeanthatthereisanexcessofVSGs,thatthebig
ultravioletto opticalwavelengths(Fitzpatrick 2004). A change grain emissivity has been modified (Désert et al. 1990), and/or
in the size distributionnaturallyexplainsthe variationof RV in thatthebiggrainemissioncontributessignificantlytothe25μm
differentinterstellarmedia.Inparticular,anychangeinthedis- bandflux.
tributionofbiggrains,withasizeofbetween15nmand100nm,
is expected to be directly related to a change in R . The size 4 IRIS(ImprovedReprocessingoftheIRASSurvey)dataaretheim-
V
proved version of thelatest IRASimages, followong the second gen-
distributionof biggrainscanbe alteredbyshockwaves(Jones
etal.2005).High-velocityshocksaffectgrainsthroughsputter- eration processing (or IRAS Sky Survey Atlas, ISSA). These maps
havebeenreprocessedtoimprovesensitivityandabsolutecalibration.
ing,whichreducesthenumberofsmallparticles,whileinshocks
ComparedtothelatestversionoftheIRASdata,thesenewimageshave
withlowervelocities,grain-graincollisionsalterthesizedistri-
higherqualityzodiacallightsubtraction,calibration,andzerolevelad-
butionbyincreasingthe small-to-largegrainsize ratio(Mazzei justmentstomatchtheDIRBEdataonlargeangular scales.TheIRIS
& Barbaro 2008). Heiles et al. (1988) proposed that shock ve- data have a resolution near 4(cid:5) in the four wavelengths (12, 25, 60,
locitiesofat least ∼30km s−1 are requiredto modifythe grain and100μm).
A142,page7of21
A&A538,A142(2012)
5.4.Galacticstructure
The relatively large number of OB stars (Table 3) in our sam-
pleallowedustoperformastatisticalapproachtomeasuringthe
stellar distance. The zones exhibit similar star-number surface
densities, excepts for zone 4, which has a larger value. This is
naturallyexplainedbyitslocationpointingtowardstheGalactic
plane,whichnaturallyprobesdenserstellarregions.Sincefrom
photometricdataonlyitisimpossibletodeterminetheluminos-
ity class, when determining the distance we assumed that the
starsaremain-sequencestarsandifthisisnotthecasethisleads
ustounderestimatethedistancebyuptoafactorof0.6respec-
tively to class III stars. Taking as a starting point the observed
colorsU−BandB−V,theintrinsiccolors(U−B) and(B−V)
0 0
andtheextinctionAV =RV×EB−V aredeterminedfollowingthe
(a) reddeninglaw and the unreddenedmain sequence of Schmidt-
Kaler(1983).Theintrinsiccolorsareusedtodeterminethespec-
traltypeusingthemain-sequencecalibrationtable ofSchmidt-
Kaler (1983). The distance is then calculated by adopting the
M -spectraltypecalibrationfromRusseil(2003).
V
The distance uncertainty is evaluated from the photometric
uncertaintiesasσ /d =(ln(10)/5)×[σ +R ×σ +R ×σ +
d V V B V V
σRV×EB−V].Todeterminethedistance,weadoptedtheRV value
oftheareawherethestarsarelying(Fig.7a).
We considered the U-band completeness level (see
Sect. 2.1), assuming a typical diffuse extinction (A =
V
0.7 mag kpc−1, Marshall et al. 2006) and a mean extinction of
A =5magatthedistanceofNGC6334–NGC6357,andtak-
V
ingtheabsoluteU magnitudefora B3V(Schmidt-Kaler1983)
andaO9.5V(Martins&Plez2006)starwecanestimatethatwe
havea90%meancompletenesslevelforB3VandO9.5Voutto
3300 pc and 5000 pc, respectively. At these distances, we can
(b) probethenearestspiralarmstructure.
In the model of Russeil et al. (2007), the line of sight in
the direction of NGC 6334 – NGC 6357, is expected to cross
the Sagittarius-Carina arm at ∼1.5 kpc, the Scutum-Crux arm
at ∼3 kpc, the Norma arm at ∼4.3 kpc, and a small feature
around6kpc.FortheHouetal.(2009)model,thedistancetothe
armsareslightlydifferent:0.82kpc,2.9kpcand4.7kpcforthe
Sagittarius-Carina, Scutum-Crux, and Norma arms. Figure 7a
showspeaksat1kpc,1.8kpc,2.6kpc,and4.6kpc.Everypeak
canbeassociatedwithaspiralarm.
Toestimatethepossibleimpactofourassumptionsaboutthe
main-sequencetypeofthestars,wenote(Table2)that65%are
main sequence stars, 17% are type III stars, and 9% are type I
stars.Wethenrepeatedourdeterminationofthestellardistance
distribution (Fig. 7b) considering these proportions of class V
and III and adopting an average factor of 1.5 for the distance
of class III stars, respectively, to class V stars. We can discern
(c) that the featuresare globallyrecoveredwith peaksat distances
ofaround1.4,2.6,and4.6kpc.
Fig.5. Far infrared dust emission of the 8 zones. The plots are TheaveragedistanceoftheNGC6334-NGC6357complex
log(I(60)/I(100))versuslog(I(12)/I(25))a),X1versusRV b)andX2 (1.75 kpc) places it in the Sagittarius-Carina arm, although it
versusR c).Labels2,6,and8correspondtotheregionsNGC6357,
V appears to be located slightly farther away than the first stel-
NGC6334,andGM24,respectively.
lar peak. Such a configuration was already noted for the other
Sagittarius-Carina arm HII region RCW108 (Georgelin et al.
1996)atadistanceof1.4kpc,whilethestellarpeakofthefield
Wesuggestthatthereisaglobaldifferencebetweenthedust starsisat1.1kpc.Inparallel,Tovmassianetal.(1996),studying
sizedistributionsofNGC6334andNGC6357.ThesimilarR anareaofabout12◦2 centeredonl = 297◦,b = −1◦ findatthe
V
in both regions suggests that their big grain content is similar, distanceoftheSagittarius-Carinaarmtwoseparatestellarasso-
whileNGC6357containsmoreverysmallgrains.TheOBstars ciationsat1.2and1.5kpc.Thissuggeststhatthereisapossible
inNGC6357seemthentohavealreadyre-processedpartofthe branch of the arm or a possible age gradient(with HII regions
dust, which could be interpreted as a more evolved status for andfieldstarslocatedatthefarandnearedgeofthearm).Itis
NGC6357thanforNGC6334. most probable that there is a non-zero age gradient across the
A142,page8of21
D.Russeiletal.:StatisticalstudyofOBstarsinNGC6334andNGC6357
Table3.Meanstarnumberandsurfacedensity.
Zone: 1 2 3 4 5 6 7 8
Number: 327 458 409 595 113 199 194 99
Numberper◦2: 346.4 394.8 281.9 1075.9 465.0 314.9 570.6 436.1
NGC6357
354.000 GP3i5s3m.2is+ 204.9
AH03J1725
2.000
353.000 1.500
1.000
0.500 GUM 63
352.000 GUM 64b Bochum 13
-0.500 GGU35M1. 26+20.5
GUM 61
-1.500 -1.000 351.000 NGC6334 GM24 (a)
-2.000
Fig.6.Hαimage(UKSTHαsurvey,Parkeretal.2005)oftheHiire-
gionsNGC6334andNGC6357(Galacticcoordinates). Thedifferent
regionsarelabeled.OverplottedaretheO-B3starsinthedistancerange
of1.5to1.95kpc(redcircles)andstellarclusters/groups(yellowsym-
bolsandnames).Clusters/groupswithoutnameareclustersfromBica
etal.(2003)andFeigelsonetal.(2009).
Sagittarius-Carinaarm,asMel’niketal.(1998)show,becausea
stratification of stellar ages exists across the Sagittarius-Carina
arm,withyoungobjectsbeinglocatedneartheinneredgeofthe
arm,whileoldgroupingslieclosertotheouteredge.
(b)
5.5.TheOBstarscensus
Uncertainty about the first phase of high-mass star formation Fig.7.a)DistributionofdistancestoOBstars(blackhistogram).Only
concerns the relative lengths of the lifetimes of high-mass starswithdistanceuncertaintiessmallerthan400pchavebeenconsid-
IR-quiet and high-luminosity protostars as well as prestellar ered.ThebluedashedhistogramistheBstardistribution,whilethered
dashedhistogramistheOstardistribution.b)Distributionofdistances
sourcesdetectedinsubmmsurveys(Motteetal.2007).Tohelp
toOBstarsassumingaproportionofclassVandclassIIIstarsof65%
resolvethisuncertainty,weperformedacensusofmassivestars
and17%.
(earlierthanB3)intheNGC6334-NGC6357complex,because
the statistical lifetime is measuredrelativeto the knownage of
OBstars. (Diaset al. 2002). For Pismis 24, Wang etal. (2007) identified
The 2.7 GHz radio continuum flux of NGC 6334 and 34O-B3starsfromX-raystudies.Theinformationontheclus-
NGC 6357 is 439.4 ± 13.5 Jy and 924.1 ± 29.0 Jy, respec- ter AH03J1726-34.4 is very sparse. AH03J1726-34.4 contains
tively(fluxesestimatedfromPaladinietal.2003).Thesefluxes atleast4OBstars(Neckel1984;Damkeetal.2006).We were
correspond to ionising photons fluxes (on a logarithmic scale) abletocountatleast38OBstarsassociatedwithNGC6357.
of50.04s−1 and50.37s−1. FromTable2,we evaluatethestel- ForNGC6334(Fig.6),Neckel(1978)found14O-B3stars
lar ionising photon flux for both regions using Panagia (1973) from optical photometry, and Bica et al. (2003) listed 7 em-
andSternbergetal.(2003).Thisledto50.04s−1 and50.32s−1 beddedclusters/groupsassociatedwithradiosources.Feigelson
forNGC6334andNGC6357,respectively,ingoodagreement et al. (2009) show (from a X-ray census of stellar sources in
with the radio values and suggests that the main ionizing stars NGC6334)acomplicatedspatialpatternwith ∼10distinctstar
are clearly identified. However, to have a complete census of clusters(theheavilyobscuredclustersaremostlyassociatedwith
OB stars we must also consider the OB stars that are not opti- known far-infrared sources and radio HII regions). They also
callyvisible. founddozensoflikelyOBstars,bothinclusters,anddispersed
In NGC 6357, two open clusters (Fig. 6) were previously throughouttheregion,suggestingthatstarformationinthecom-
identified, the well-known Pismis 24 and AH03J1726-34.4 plex has proceeded over millions of years. Tapia et al. (1996)
A142,page9of21
A&A538,A142(2012)
deduced that the young cluster embedded in NGC 6334I has Acknowledgements. Part of this work was supported by the ANR (Agence
around 93 stars members brighter than K = 16 and that the Nationalepourlarecherche)project“PROBES”,numberANR-08-BLAN-0241.
majorityofthe observedmembersare ZAMSstars earlierthan PartofthisworkwascompletedthankstoSIMBADandCDS.Theauthorsac-
knowledgeD.CloweandM.P.Ulmer,respectively PIandCo-IoftheNOAO
B3-B4.Tapiaetal.(1996)found12faintsources(detectedonly
observingtime(NOAO2009A-0137),andL.Guennoufortheirimplicationin
in K band) centered on the HII region NGC 6334E. They es- theNOAOobservations.
timated that the ionization of NGC 6334E requires at least an
O7.5ZAMSstaror12-13B0-B0.5ZAMSstars.Theythenpro-
posed that a cluster of B stars is responsible for the ionization References
of NGC 6334E.Bochum 13, a cluster at the north-westborder
Adami,C.,Picat,J.P.,Savine,C.,etal.2006,A&A,451,1159
of NGC 6334, and at the same distance, contains5 O-B3 stars
Ahumada,A.,Claría,J.,&Bica,E.2007,A&A,473,437
(McSwain2005) andaclusterdominatedbyafewyoungmas- Bertelli,G.,Bressan,A.,Chiosi,C.,etal.1994,A&AS,106,275
sivestars(∼10)liesinthecoreofGM24(Tapiaetal.1991).We Bertin,E.,&Arnouts,S.1996,A&AS,117,393
thencountatleast150OBstarsassociatedwithNGC6334. Bertin,E.,Mellier,Y.,&Radovich,M.2002,ASPC,281,228
Bica,E.,Dutra,C.,&Barbuy,B.2003,A&A,397,177
OurUBVdataprovideanestimateof40additionaloptically
Bohigas,J.,Tapia,M.,Roth,M.,etal.2004,AJ,127,2826
visibleO-B3starsatthedistanceofNGC6334-NGC6357com- Boulanger,F.,Beuchman,C.,Désert,F.,etal.1988,ApJ,332,328
plex (Fig. 6), giving a total count of at least 228 OB stars. We Cardelli,J.A.,Clayton,G.C.,&Mathis,J.S.1989,ApJ,345,245
canthusestimatethat230massivestars(earlierthanB3)belong Carey,S.,Noriega-Crespo,A.,&Mizuno,D.2009,PASP,121,76
Carpenter,J.2001,AJ,121,2851
totheentirecomplex.
Caswell,J.,&Haynes,R.1987,A&A,171,261
In Russeil et al. (2010), we presented a census of high- Chini,R.,&Kruegel,E.1983,A&A,117,289
mass young stellar objects at various evolutionary stages: we Chini,R.,&Wargau,W.1990,A&A,227,213
identified 1 starless clump, 6 IR-quiet protostars, and 9 high- Chini,R.,Elsasser,H.,&Neckel,T.1980,A&A,91,186
Conti,P.,&Vacca,W.1990,AJ,100,431
luminosity IR protostars in NGC 6334-NGC 6357. Since the
Curti,R.,etal.2003,2MASSAll-SkyPointSources(Pasadena:NASA/IPAC),
timescaleofthehigh-massyoungobjectscannotbederiveddi- http://iras.ipac.caltech.edu/applications/Gator
rectly, a statistical method was used. Assuming a steady star Damke,G.,Barba,R.,&Morrell,N.2006,Rev.Mex.Astron.Astrofis.,26,180
formation, the ratio of the number of objects at each stage is deGrijs,R.,O’Connell,R.W.,&Gallagher,J.S.2001,AJ,121,768
Désert,F.,Boulanger,F.,&Puget,J.L.1990,A&A,237,215
proportional to the stage timescale. The unit of time was esti-
Dias,W.,Alessi,B.,Moitinho,A.,etal.2002,A&A,389,871
mated from the number and age of OB stars. The adopted age
Durret,F.,Laganá,T.F.,Adami,C.,&Bertin,E.2010,A&A,517,A94
of OB stars in NGC 6334-NGC 6357 comes from the age of Epchtein,N.,deBatz,B.,Capoani,L.,etal.1997,Msngr,87,27
theclusterPismis24,whichisbetween1.73Myr(Masseyetal. Fang,X.,Storey,P.J.,&Liu,X.-W.2011,A&A,530,A18
2001) and 5± 3 Myr (Ahumada et al. 2007). This means that Feigelson,E.,Martin,A.,McNeill,C.,etal.2009,AJ,138,227
from our number of ∼230 OB stars, one high-mass young ob- Fitzpatrick,E.L.2004,ASPC,309,33
Georgelin,Y.M.,Russeil,D.,Marcelin,M.,etal.1996,A&AS,120,41
jectcorrespondstoastatisticallifetimeofbetween∼7.5×103yr Greve,A.2010,A&A,518,A62
and∼2.2×104yr.Wecanthen,inNGC6334-NGC6357,esti- Gum,C.S.1955,MmRAS,67,155
matethatthestarlessandtheprotostellarphaseshaveastatistical Gyulbudaghian,A.,&Maghakian,T.1977,PAZh,3,113
lifetimeof7.5×103−2.2×104 yr and1.1×105−3.3×105 yr, Hanson,M.,Howarth,I.,&Conti,P.1997,ApJ,489,698
Heiles,C.,Reach,W.,&Koo,B.1988,ApJ,332,313
respectively.
Hou,L.G.,Han,J.L.,&Shi,W.B.2009,A&A,499,473
Jones,A.2005,Proc.ThedustyandMolecularUniverse,ESAP,577,239
Koornneef,J.1983,A&A,128,84
6. Conclusions Kraemer,K.,&Jackson,J.1999,ApJS,124,439
Kroupa,P.2001,MNRAS,322,231
Landolt,A.U.2009,AJ,137,4186
The results that we have presented in this paper stress the ef-
Leitherer, C. 1999, Spectrophotometric Dating of Stars and Galaxies, ed. I.
fectivenessofmulticoloropticalphotometryinthestudyofthe
Hubeny,S.Heap,&R.Cornett,ASPConf.Proc.,192,3
extinction and distance towards HII regions. Our large stellar Lortet,M.,Testor,G.,&Niemela,V.1984,A&A,140,24
samplehasallowedustocarryoutastatisticalstudyofthelocal Marshall,D.J.,Robin,A.C.,Reylé,C.,etal.2006,A&A,453,635
structureoftheMilky-Way,aswellasthecensusofO-B3starsin Martins,F.,&Plez,B.2006,A&A,457,637
Massey,P.,DeGioia-Eastwood,K.,&Waterhouse,E.2001,AJ,121,1050
theNGC6334-NGC6357star-formingcomplex.Spectroscopic
Mazzei,P.,&Barbaro,G.2008,MNRAS,390,706
informationofourlargeO-B3stellarsampleisnowrequiredto Mazzei,P.,&Barbaro,G.2011,A&A,527,A34
confirmthephotometricspectraltypesandimproveourdistance Mel’nik,A.,Sitnik,T.,Dambis,A.,etal.1998,AstL,24,594
determinations. Moffat,A.,&Vogt,N.1973,A&AS,10,135
Asexpected,thetotal-to-selectiveextinctioncoefficient,R , Monet,F.,Levine,S.E.,Canzian,B.,etal.2003,AJ,125,98
V Motte,F.,Bontemps,S.,Schilke,P.,etal.2007,A&A,476,1243
ishigherinastar-formingregionsuchasNGC6334-NGC6357 Neckel,T.1978,A&A,69,51
than in the diffuse interstellar medium. Futhermore, this coef- Neckel,Th.1984,A&A,137,58
ficient varies throughoutthe HII regions, suggesting that there Neckel,Th.,&Chini,R.1981,A&AS,45,451
O’Dell,C.R.,&Wen,Z.1992,ApJ,387,229
is a variation in the dust properties on the same scale. The R
V Paladini,R.,Burigana,C.,&Davies,R.2003,A&A,397,213
valuesweobtainedforNGC6334andNGC6357(3.56±0.15
Panagia,N.1973,AJ,78,929
and3.53±0.08,respectively)haveallowedustodeterminemore Pandey,A.,Ogura,K.,&Sekiguchi,K.2000,PASJ,52,847
preciselythe distanceofthese regions.We haveconfirmedthat Pandey,A.,Upadhyay,K.,Nakada,Y.,etal.2003,A&A,397,191
theyarelocatedatanaveragedistanceof1.75kpc.Ourcensus Parker,Q.,Phillipps,S.,Pierce,M.,etal.2005,MNRAS,362,689
Persi,P.,&Tapia,M.2008,HandbookofStarFormingRegions,Vol.II:The
of O-B3 stars in the NGC 6334-NGC 6357 star-forming com-
SouthernSkyASPMonographPublications,ed.B.Reipurth,5,456
plexhasidentifiedatotalof300O-B3stars.Thiscensusisim- Pinheiro,M.,Copetti,M.,&Oliviera,V.2010,A&A,521,A26
portant to help us to determine the statistical lifetimes of the Rieke,G.,&Lebofsky,M.1985,ApJ,288,618
different stages of the high-massstar formation in NGC 6334- Roslund,C.1966,Arkiv.Astr.,4,101
Russeil,D.2003,A&A,397,133
NGC 6357 that we observe in the infrared with the Herschel
Russeil,D.,Adami,C.,&Georgelin,Y.M.2007,A&A,470,161
satellite(Russeiletal.,inprep.).
Russeil,D.,Zavagno,A.,Motte,F.,etal.2010,A&A,515,A55
A142,page10of21