Planck Early Results: The Galactic Cold Core Population revealed by the first all-sky survey

3.1 MB·English

by  Planck Collaboration

Astronomy&Astrophysicsmanuscriptno.Planck˙Early˙Paper˙1˙v3.1 c ESO2011
(cid:13)
January12,2011
Planck Early Results: The Galactic Cold Core Population revealed
by the first all-sky survey
PlanckCollaboration:P.A.R.Ade68,N.Aghanim45,M.Arnaud55,M.Ashdown53,74,J.Aumont45,C.Baccigalupi66,A.Balbi27,
A.J.Banday72,6,60,R.B.Barreiro50,J.G.Bartlett3,51,E.Battaner76,K.Benabed46,A.Benoˆıt46,J.-P.Bernard72,6,M.Bersanelli25,40,R.Bhatia33,
J.J.Bock51,7,A.Bonaldi36,J.R.Bond5,J.Borrill59,69,F.R.Bouchet46,F.Boulanger45,M.Bucher3,C.Burigana39,P.Cabella27,
C.M.Cantalupo59,J.-F.Cardoso56,3,46,A.Catalano3,54,L.Cayo´n18,A.Challinor75,53,8,A.Chamballu43,R.-R.Chary44,L.-YChiang47,
P.R.Christensen63,28,D.L.Clements43,S.Colombi46,F.Couchot58,A.Coulais54,B.P.Crill51,64,F.Cuttaia39,L.Danese66,R.D.Davies52,
R.J.Davis52,P.deBernardis24,G.deGasperis27,A.deRosa39,G.deZotti36,66,J.Delabrouille3,J.-M.Delouis46,F.-X.De´sert42,C.Dickinson52,
K.Dobashi14,S.Donzelli40,48,O.Dore´51,7,U.Do¨rl60,M.Douspis45,X.Dupac32,G.Efstathiou75,T.A.Enßlin60,E.Falgarone54,F.Finelli39,
1
O.Forni72,6,M.Frailis38,E.Franceschi39,S.Galeotta38,K.Ganga3,44,M.Giard72,6,G.Giardino33,Y.Giraud-He´raud3,J.Gonza´lez-Nuevo66,
1
K.M.Go´rski51,78,S.Gratton53,75,A.Gregorio26,A.Gruppuso39,F.K.Hansen48,D.Harrison75,53,G.Helou7,S.Henrot-Versille´58,D.Herranz50,
0
2 S.R.Hildebrandt7,57,49,E.Hivon46,M.Hobson74,W.A.Holmes51,W.Hovest60,R.J.Hoyland49,K.M.Huffenberger77,A.H.Jaffe43,G.Joncas11,
W.C.Jones17,M.Juvela16,E.Keiha¨nen16,R.Keskitalo51,16,T.S.Kisner59,R.Kneissl31,4,L.Knox20,H.Kurki-Suonio16,34,G.Lagache45,
n
J.-M.Lamarre54,A.Lasenby74,53,R.J.Laureijs33,C.R.Lawrence51,S.Leach66,R.Leonardi32,33,21,C.Leroy45,72,6,M.Linden-Vørnle10,
a
J M.Lo´pez-Caniego50,P.M.Lubin21,J.F.Mac´ıas-Pe´rez57,C.J.MacTavish53,B.Maffei52,N.Mandolesi39,R.Mann67,M.Maris38,
D.J.Marshall72,6,P.Martin5,E.Mart´ınez-Gonza´lez50,G.Marton30,S.Masi24,S.Matarrese23,F.Matthai60,P.Mazzotta27,P.McGehee44,
1
A.Melchiorri24,L.Mendes32,A.Mennella25,38,S.Mitra51,M.-A.Miville-Descheˆnes45,5,A.Moneti46,L.Montier72,6⋆,G.Morgante39,
1
D.Mortlock43,D.Munshi68,75,A.Murphy62,P.Naselsky63,28,F.Nati24,P.Natoli27,2,39,C.B.Netterfield13,H.U.Nørgaard-Nielsen10,
] F.Noviello45,D.Novikov43,I.Novikov63,S.Osborne71,F.Pajot45,R.Paladini70,7,F.Pasian38,G.Patanchon3,T.J.Pearson7,44,V.-M.Pelkonen44,
A O.Perdereau58,L.Perotto57,F.Perrotta66,F.Piacentini24,M.Piat3,S.Plaszczynski58,E.Pointecouteau72,6,G.Polenta2,37,N.Ponthieu45,
G T.Poutanen34,16,1,G.Pre´zeau7,51,S.Prunet46,J.-L.Puget45,W.T.Reach73,R.Rebolo49,29,M.Reinecke60,C.Renault57,S.Ricciardi39,T.Riller60,
I.Ristorcelli72,6,G.Rocha51,7,C.Rosset3,M.Rowan-Robinson43,J.A.Rubin˜o-Mart´ın49,29,B.Rusholme44,M.Sandri39,D.Santos57,G.Savini65,
.
h D.Scott15,M.D.Seiffert51,7,G.F.Smoot19,59,3,J.-L.Starck55,9,F.Stivoli41,V.Stolyarov74,R.Sudiwala68,J.-F.Sygnet46,J.A.Tauber33,
p L.Terenzi39,L.Toffolatti12,M.Tomasi25,40,J.-P.Torre45,V.Toth30,M.Tristram58,J.Tuovinen61,G.Umana35,L.Valenziano39,P.Vielva50,
- F.Villa39,N.Vittorio27,L.A.Wade51,B.D.Wandelt46,22,N.Ysard16,D.Yvon9,A.Zacchei38,S.Zahorecz30,andA.Zonca21
o
r (Affiliationscanbefoundafterthereferences)
t
s
a Preprintonlineversion:January12,2011
[

1
ABSTRACT
v
5
WepresentthestatisticalpropertiesofthefirstversionoftheColdCoreCatalogueofPlanckObjects(C3PO),intermsoftheirspatialdistribution,
3
temperature, distance, mass, and morphology. We also describe the statistics of the Early Cold Core Catalogue (ECC) that is a subset of the
0
complete catalogue, and that contains only the 915 most reliable detections. ECC is delivered asa part of the EarlyRelease Compact Source
2
Catalogue(ERCSC).WehaveusedtheCoCoCoDeTalgorithmtoextractabout10thousandcoldsources.ThemethodusestheIRAS100µmdata
.
1 asawarmtemplatethatisextrapolatedtothePlanckbandsandsubtractedfromthesignal,leadingtoadetectionofthecoldresidualemission.
0 Wehaveusedcross-correlationwithancillarydatatoincreasethereliabilityofoursample,andtoderiveotherkeypropertiessuchasdistanceand
1 mass.
1 TemperatureanddustemissionspectralindexvaluesarederivedusingthefluxesintheIRAS100µmbandandthethreehighestfrequencyPlanck
: bands.Therangeoftemperaturesexploredbythecataloguespansfrom7Kto17K,andpeaksaround13K.Dataarenotconsistentwithaconstant
v
valueoftheassociatedspectralindexβoverthealltemperaturerange.βrangesfrom1.4to2.8withameanvaluearound2.1,andseveralpossible
i
X scenariosarepossible,includingβ(T)andtheeffectofmultipletemperaturecomponentsfoldedintothemeasurements.
For one third of the objects the distances are obtained using various methods such as the extinction signature, or the association with known
r
a molecularcomplexesorInfra-RedDarkClouds.Mostofthedetectionsarewithin2kpcintheSolarneighbourhood, butafewareatdistances
greaterthan4kpc.Thecoresaredistributedoverthewholerangeoflongitudeandlatitude,fromthedeepGalacticplane,despitetheconfusion,to
highlatitudes(>30 ).Theassociatedmassestimatesderivedfromdustemissionrangefrom1to105solarmasses.Usingtheirphysicalproperties
◦
suchastemperature,mass,luminosity, densityandsize,thesecoldsourcesareshowntobecoldclumps, definedastheintermediatecoldsub-
structuresbetweencloudsandcores.Thesecoldclumpsarenotisolatedbutmostlyorganizedinfilamentsassociatedwithmolecularclouds.The
ColdCoreCatalogueofPlanckObjects(C3PO)isthefirstunbiasedall-skycatalogueofcoldcompactobjectsandcontains10783objects.Itgives
anunprecedented statisticalviewtothepropertiesofthesepotentialpre-stellarclumpsandoffersauniquepossibilityfortheirclassificationin
termsoftheirintrinsicpropertiesandenvironment.
Keywords.ColdCores,Galaxy,Sourceextraction
1. Introduction
The main difficulty in understanding star formation lies in the
⋆ [email protected] vastrangeofscalesinvolvedintheprocess.Ifstarformationit-
2 PlanckCollaboration:TheGalacticColdCorePopulationrevealedbythefirstPlanckall-skysurvey
selfistheoutcomeofgravitationalinstabilityoccurringincold parts of the nearby star-forming clouds but cannot cover high
and dense structuresat sub-parsecscales, the characteristicsof Galacticlatitudeswherestarformationisknowntooccur.Inthis
thesestructures(usuallycalledpre-stellarcores)dependontheir endeavorthemainchallengeishowtolocatethecoresbecause,
large-scaleenvironment,uptoGalacticscalesbecausetheirfor- evenwith Herschel, detailedstudies mustbe limited to a small
mation and evolution is driven by a complex coupling of self- fractionofthewholesky.
gravitywith coolingprocesses,turbulenceand magneticfields, The Planck 1 satellite (Tauber et al. 2010) improves
to namea few. To progressin the understandingof star forma-
over the previous studies by providing an all-sky submillime-
tionpre-stellarcoresneedtobeobserved,inavarietyofenviron- tre/millimetresurveythathasboththesensitivityandresolution
ments.Moreimportantly,broadsurveysarerequiredtoaddress
neededforthedetectionofcompactsources.Theshortestwave-
statisticalissues,andprobetheoreticalpredictionsregardingthe
length channels of Planck cover the wavelengths around and
initial mass function (IMF) largely determined at the stage of
longwardsoftheintensitymaximumofthecolddustemission:
fragmentationofpre-stellarcores. ν2B (T =10K)peakscloseto300µmwhile,withatemperature
ν
Unfortunately,thepropertiesofthepre-stellarcoresarestill
of T 6K, the coldest dust inside the coreshas its maximum
poorlyknownmostly becauseof observationaldifficulties.The close∼to 500µm. Combined with far-infrared data such as the
total number of Galactic pre-stellar cores is estimated to be
IRASsurvey,thedataenableaccuratedeterminationofboththe
around3 105(Clemensetal.1991)butmostofthemhavesofar
dusttemperatureandthespectralindex.We usethePlanck ob-
×
escaped detection, simply because they are cold and immersed servationsto search forGalactic cold cores,i.e. compactcloud
inwarmer(thereforebrighter)environments.
coreswithcolourtemperaturesbelow14K.Becauseofthelim-
The thermal dust emission of nearby molecular clouds has itedresolution,wearelikelytodetectmainlylargerclumpsin-
beenmappedfromthegroundinthemillimeterandsubmillime- side which the cores are located. The cores will be pre-stellar
terrangeswithinstrumentssuchasSCUBA,MAMBO,SIMBA, objects before (or at the very initial stages) of the protostel-
andLaboca.Becauseoflimitedsensitivity,butalsothepresence lar collapse,orpossiblymoreevolvedsourcesthatstill contain
of the atmospheric fluctuations that call for beam-throw of at significant amounts of cold dust. The Cold Core Catalogue of
mostafewarcmin,thestudieshaveconcentratedonthebright- Planck Objects (C3PO) which will be made public at the end
estandmostcompactregionsthatarealreadyinanactivephase of the Planck proprietary period, will be the first all-sky cat-
ofstarformation.Thankstosub-arcminuteresolution,theseob- alogue of cold cloud cores and clumps. It will reveal the lo-
servations(togetherwithdedicatedmolecularlinestudies)have cations where the next generations of stars will be born and
been the main source of information also on the structure of will provide an opportunity to address a number of key ques-
the pre-stellar cores (Motte et al. 1998;Curtis & Richer 2010; tionsrelatedtoGalacticstarformation:Whatarethecharacter-
Hatchelletal.2005;Enochetal.2006;Kauffmannetal.2008). isticsofthissourcepopulation?Howdoesthedistributionofthe
Many compactclouds were detected as absorption features cores/clumps correlate with the current star formation activity
onphotographicplates.A newpopulationofthousandsofcold andthelocationofthemolecularcloudringsandthespiralarms?
dark clouds was discovered by observations of mid-infrared Howarethesourcesrelatedtolarge-scalestructuresliketheFIR
absorption towards the bright Galactic background (MSX and loops,bubbles,shells,andfilaments?Aretherepre-stellarcores
ISOGALsurveys;seeEganetal.1998;Peraultetal.1996).The at high latitudes? How much do the core propertiesdepend on
absorptionstudiesare,however,stronglybiasedtowardsthelow theirenvironment?Investigationssuchasthesewillhelpusun-
latitudes and do not directly provide information on the tem- derstandthe originof the pre-stellarcores,the instabilities that
perature of the detected sources. For a definitive study of the initiate the collapse, and the roles of turbulence and magnetic
coldcloudcores,onemustturntohighresolutionobservations fields.Thecataloguewillproveinvaluableforfollow-upstudies
inthesubmillimetreormillimetrerange(Andreetal.2000).The to investigate in detail the internal properties of the individual
Bolocam Galactic Plane Survey(BGPS) is producingmm data sources.
for the central part of the Galactic plane (Aguirre et al. 2010).
In this paper we describe the general properties of the
Thefirstresultssuggestthatatkpcdistances,evenwithahalfar-
current cold cores catalogue that is based on data that the
cminresolution,oneisdetectingmainlyclusterformingclumps
Planck satellite has gathered during its first two scans of the
ratherthan coresthatwouldproduce,at most,a small multiple
full sky. In particular, we will describe the statistics of the
system(Dunhametal.2010).
Early Cold Cores Catalogue (ECC) that is part of the recently
Balloon borne experiments have provided larger blind sur-
published Planck Early Release Compact Source Catalogue
veysofhigherlatitudes.PRONAOS discoveredcoldcondensa-
(ERCSC Planck Collaboration 2011c). ECC forms a subset of
tionsalsoincirrus-typeclouds(Bernardetal.1999;Dupacetal.
thefullC3POandcontainsonlythemostsecuredetectionsofall
2003)Similarly,Archeops(De´sertetal.2008)detectedhundreds
thesourceswithcolourtemperaturesbelow14K.Thefinalver-
of sources with temperatures down to 7K. The latest addition
sionofC3POwillbepublishedin2013.Forhistoricalreasons,
to the balloon borne surveys is the BLAST experiment which
weuse”ColdCores”todesignatetheentriesintheC3POandin
haslocatedseveralhundredsubmillimetresourcesinVulpecula
the ECC, andsimilarly in muchofthis paper.However,asthis
(Chapinetal.2008)andVela(Netterfieldetal.2009;Olmietal.
paper and the companion paper (Planck Collaboration 2011r,
2009),includinganumberofcoldandprobablypre-stellarcores.
hereafterPaperII)demonstrate,mostofthesearemorecorrectly
Since its launchin May 2009,the Herschelsatellite has al-
described as ”cold clumps”, intermediate in their structure and
readyprovidedhundredsofnewdetectionsofbothstarlessand
protostellar cores (Andre´ et al. 2010; Bontemps et al. 2010;
Ko¨nyvesetal.2010;Molinarietal.2010;Ward-Thompsonetal.
1 Planck (http://www.esa.int/Planck) is a project of the European
2010). There is an intriguing similarity between the core mass
SpaceAgency(ESA)withinstrumentsprovidedbytwoscientificcon-
function(CMF)derivedfromthesedata,andtheIMFthatneed
sortia funded by ESA member states (in particular the lead countries
to be investigated in different environments, towards the inner FranceandItaly),withcontributionsfromNASA(USA)andtelescope
Galaxyin particular.TheHerschelstudieswill eventuallycover reflectorsprovidedbyacollaborationbetweenESAandascientificcon-
a significant fraction of the Galactic mid-plane and the central sortiumledandfundedbyDenmark.
PlanckCollaboration:TheGalacticColdCorePopulationrevealedbythefirstPlanckall-skysurvey 3
physical scale between a true pre-stellar core and a molecular The noise in the channel maps is essentially white with
cloud. a mean standard deviation of 1.4 10 3, 4.1 10 3, 1.4
− −
Planck (Tauberet al. 2010;Planck Collaboration 2011a)is 10 3MJy/srat353,545and857GH×zrespectivel×y(PlanckHF×I
−
thethirdgenerationspacemissiontomeasuretheanisotropyof CoreTeam2011b).Thephotometriccalibrationisperformedei-
the cosmic microwave background(CMB). It observesthe sky ther at the ring level using the CMB dipole, for the lower fre-
inninefrequencybandscovering30–857GHzwithhighsensi- quencychannels,oratthemaplevelusingFIRASdata,forthe
tivityandangularresolutionfrom31 to5.TheLowFrequency higher frequency channels at 545 and 857GHz. The absolute
′ ′
InstrumentLFI; (Mandolesiet al. 2010;Bersanelli et al. 2010; gaincalibrationofHFIPlanckmapsisknowntobetterthan2%
Mennellaetal.2011)coversthe30,44,and70GHzbandswith at353GHzand7%at545and857GHz(seeTable2inPlanck
amplifierscooledto20K.TheHighFrequencyInstrument(HFI; HFICoreTeam2011b).
Lamarreet al. 2010;Planck HFI Core Team 2011a)coversthe Thedetectionalgorithmrequirestheuseofancillarydatato
100, 143, 217, 353, 545, and 857GHz bands with bolometers tracethewarmcomponentofthegas.ThuswecombinePlanck
cooled to 0.1K. Polarization is measured in all but the highest data with the IRIS all-sky data (Miville-Descheˆnes& Lagache
two bands(Leahy et al. 2010;Rosset et al. 2010).A combina- 2005).ThechoiceoftheIRIS100µm asthe warm templateis
tionofradiativecoolingandthreemechanicalcoolersproduces motivatedbythefollowing:(i)100µmisveryclosetothepeak
the temperatures needed for the detectors and optics (Planck frequencyofablackbodyat20K,andtracesthewarmcompo-
Collaboration 2011b). Two Data Processing Centers (DPCs) nentoftheGalaxy;(ii)thefractionofsmallgrainsatthiswave-
checkandcalibratethe dataandmakemapsofthesky(Planck lengthremainsverysmallanddoesnotsignificantlytheestimate
HFI Core Team 2011b;Zaccheiet al. 2011).Planck’ssensitiv- of the emission fromlarge grainsthat is extrapolatedto longer
ity, angularresolution, and frequencycoveragemake it a pow- wavelengths;(iii)theIRASsurveycoversalmosttheentiresky
erful instrument for galactic and extragalactic astrophysics as (only 2 bands of 2% of the whole sky are missing); (iv) the
wellascosmology.EarlyastrophysicsresultsaregiveninPlanck resolutionoftheIR∼ISmapsissimilartotheresolutionofPlanck
Collaboration,2011h–z. in the highfrequencybands,i.e. around4.5. Using the mapat
′
100µmasthewarmtemplateis,ofcourse,notperfect,because
a non-negligiblefractionof the cold emission is still presentat
2. SourceExtraction
thisfrequency.ThislowerstheintensityinthePlanckbandsaf-
terremovaloftheextrapolatedbackground.Wewilldescribein
2.1.DataSet
detail,especiallyinSect.2.3,howwedealwiththisissueforthe
As cold cores are traced by their cold dust emission in the photometryofthedetectedcores.
submillimetric bands, we use Planck channel maps of the HFI AllPlanck andIRISmapshavebeensmoothedatthesame
at 3 frequencies : 353, 545 and 857 GHz as described in de- resolution4.5 beforesourceextractionandphotometryprocess-
′
tail in Planck HFI Core Team (2011b). The temperature maps ing.
at these frequencies are based on the first two sky surveys
of Planck, provided in Healpix format (Go´rski et al. 2005) at
nside=2048.We givehereaverybriefsummaryofthedatare- 2.2.SourceExtractionMethod
duction, cf Planck HFI Core Team (2011b) for further details.
Rawdataarefirstprocessedtoproducecleanedtimelines(TOI) We have applied the detection method described in Montier et
and associated flags identifyingvarioussystematic effects. The al.2010,knownasCoCoCoDeT(standingforColdCoreColour
data analysis includes application of a low-pass filter, removal DetectionTool),onthecombinedIRISplusPlanckdatasetde-
and correction of glitches, conversion to absorbed power and scribedinSect.2.1.Thisalgorithmusesthecolourpropertiesof
decorrelationofthermalstagefluctuations.Forthecoldcorede- theobjectstobedetectedtoseparatethemfromthebackground.
tection, and more generally for source detection, Solar System In the case of cold cores, the method selects compact sources
objects (SSO) are identified in the TOI data using the publicly colder than the surrounding envelope and the diffuse Galactic
availableHorizonephemeridesandanSSOflagiscreatedtoen- background, that is at about 17K (Boulanger et al. 1996) but
surethattheyarenotprojectedontothesky. canlargelyvaryfromoneplacetotheotheracrosstheGalactic
Focalplanereconstructionandbeam-shapeestimatesareob- planeorathigherlatitudes.ThisWarmBackgroundSubtraction
tained using observations of Mars. Beams are described by an method is applied on each one of the three Planck maps, and
ellipticalGaussianparameterisationleadingtoFWHMθ given consistsof6steps:
S
in Table 2 of Planck HFI Core Team (2011b). The attitude of
the satellite as a function of time is provided by the two star 1. foreachpixel,thebackgroundcolourisestimatedastheme-
trackersinstalledonthePlanckspacecraft.Thepointingforeach dian value of the Planck map divided by the 100µm map
bolometeriscomputedbycombiningtheattitudewiththeloca- withinadiscofradius15 aroundthecentralpixel;
′
tionofthebolometerinthefocalplanereconstructedfromMars 2. thewarmcomponentinapixelatthePlanckfrequencyisob-
observations. tainedbymultiplyingtheestimateofthebackgroundcolour
FromthecleanedTOIandthepointing,channelmapshave withthevalueofthepixelinthe100µmmap;
beenmadeusingbolometersatagivenfrequency.Thepathfrom 3. the cold residual map is computedby subtractingthe warm
TOItomapsintheHFIDPCisschematicallydividedintothree componentfromthePlanckmap;
steps, ring-making, destriping and map-making. The first step 4. the local standard deviation around each pixel in the cold
averages circles within a pointing period to make rings with residual map is estimated in a radius of 30 using the so-
′
highersignal-to-noiseratiotakingadvantageoftheredundancy called Median Absolute Deviation that ensures robustness
of observations provided by the Planck scanning strategy. The against a high confusionlevel of the backgroundand pres-
lowamplitude1/f componentisaccountedforinasecondstep enceofotherpointsourceswithinthesamearea;
usingadestripingtechnique.Finally,cleanedmapsareproduced 5. athresholdingdetectionmethodisappliedinthecoldresid-
usingasimpleco-additionoftherings. ualmaptodetectsourcesatasignal-to-noiseratioSNR>4;
4 PlanckCollaboration:TheGalacticColdCorePopulationrevealedbythefirstPlanckall-skysurvey
6. finaldetectionsaredefinedaslocalmaximaoftheSNRcon- 2.3.2. Step2:100µmphotometry
strained so that there is a minimum distance of 5 between
′ The photometryon the 100 µm map is obtained by surface fit-
them.
ting,performedonlocalmapsof1 1 centeredoneachcan-
◦ ◦
×
didate.Allcomponentsofthemaparefittedasawhole:apoly-
This processis performedat each Planck bandyieldingin-
nomialsurface of an order between three and six for the back-
dividual catalogues at 857 GHz, 545 GHz and 353 GHz. The
ground;asetofellipticalGaussianswhenotherpointsourcesare
laststepofthesourceextractionconsistsinmergingthesethree
detected inside the local map; and a central elliptical Gaussian
independentcataloguesrequiringa detection in all three bands
correspondingtothecoldcorecandidateforwhichtheelliptical
at SNR>4. Thisstep rejects spuriousdetectionsthatare due to
shapeissetbytheparametersobtainedduringstep1.Whenthe
mapartifactsassociatedwith a singlefrequency(e.g.stripesor
fitofthebackgroundispoor,i.e.acleardegeneracyisobserved
under-sampledfeatures).Itincreasestherobustnessofthefinal
betweenthepolynomialfitandthecentralGaussian,weswitch
catalogue,whichcontains10783objects.
toasimpleaperturephotometryonthelocalmap.Notethatthe
We stressthatnoanyothera-prioriconstraintsareimposed aperturephotometryisperformedtakingintoaccounttheellipti-
on the size of the expectedsources,other than the limited area calshapeofthecoldcoreprovidedbystep1.Insuchcases(140
on which the background colour is estimated. Thus the maxi- sources), the flag Bad Sfit 100µm is set to on. Occasionally no
mumscaleoftheC3POobjectsisabout12′.Notealsothatthis counterpartatallisobservedat100µm,whenthecoldcorecan-
Warm Background Subtraction method uses local estimates of didateistoofaintorverycold,ortheconfusionoftheGalactic
the colour, identifyinga relative rather than an absolute colour backgroundis too high.Insuch case, we are notable to derive
excess.Thuscoldcondensationsembeddedin coldregionscan anyreliableestimateofthe100µmfluxofthecore,soonlyan
be missed,while in hotregionscondensationsmaybe detected upper-limitcanbeprovided.Thisupperlimitisdefinedasthree
thatarenotactuallycold.Amoredetailedanalysisintempera- times the standard deviation of the cold residual map within a
tureisrequiredtoassessthenatureoftheobjects. 25 radiuscircle, and the flag Upper100µm is set to on. There
′
are 2356 objects for which only an upper limit of the temper-
ature is derived. This population represents a very interesting
2.3.Photometry
sub-sampleofthewholecatalogue,probablythecoldestobjects,
Wehavedevelopedadedicatedalgorithmtoderivethephotom- butwedonothaveconfidenceinthephysicalpropertiesderived
etryof theclumpitself. The fluxesare estimatedfromthecold from the Planck data and so it is excluded from the physical
residualmaps,insteadofworkingontheinitialmapswherethe analysis.
clumps are embedded in their warm surrounding envelope. As
alreadystressedabove,themainissueistoperformthephotom-
2.3.3. Step3:100µmcorrection
etryontheIRIS100µmmapsthatalsoincludeafractionofthe
coldemission.The fluxof the sourceat100µm hasto bewell Once an estimate of the flux at 100 µm has been provided by
determinedfortworeasons:(1)anaccurateestimateoftheflux steps1and2,the warm templateat100µm iscorrectedbyre-
atthisfrequencyisrequiredbecauseitisconstrainssignificantly moving an elliptical Gaussian correspondingto the flux of the
therestoftheanalysis(intermsofspectraldensitydistribution centralclump.Thisnewwarmtemplateisthenextrapolatedand
(SED)andtemperature);(2)anincorrectestimateofthefluxat subtractedfromthePlanckmapstobuildthecoldresidualmaps.
100µm will propagatethroughthe Planck bandsafter removal Whenonlyanupperlimithasbeenobtainedat100µm,thewarm
oftheextrapolatedwarmcomponent.Themainstepsofthepho- templateisnotchanged.
tometry processing are described in the following subsections.
AnillustrationofthisprocessisprovidedinFig.B.5oftheasso-
2.3.4. Step4:Planckbandsphotometry
ciatedPlanckEarlyPaperonColdClumpsdescribingindetaila
sampleof10sources(PlanckCollaboration2011r). Aperture photometryis performedon local cold residual maps
centered on each candidate in the Planck bands, at 857 GHz,
545GHzand353GHz.Thisaperturephotometrytakesintoac-
2.3.1. Step1:EllipticalGaussianfit
count the real extension of each object by integrating the sig-
nalinsidetheellipticalGaussianconstrainedbytheparameters
AnellipticalGaussianfitisperformedonthe1 1 colourmap
◦ ◦
× obtained at step 1. The background is estimated by taking the
857GHzdividedby100µmcenteredoneachC3POobject.This
medianvalueonanannulusaroundthesource.Nevertheless,in
results in estimates of three parameters: major axis extension
229cases,nopositiveestimateofthefluxhasbeenobtained,be-
σ , minor axis extension σ and position angle ψ. The re-
Maj Min
causeofthepresenceofcoldpointsourcesthataretoocloseor
lationbetweentheextensionσandtheFWHMθofaGaussian
because the backgroundis highlyconfused.These sources(for
isgivenby:
whichtheflagPSNegissettoon)aresimplyremovedfromthe
physicalanalysisdescribedinthispaper.
σ=θ/ 8ln(2) (1)
p 2.4.Monte-CarloQualityAssessment
If the elliptical Gaussian fit is indeterminate, a symmetrical
Gaussian is assumed with a FWHM fixed to θ = 4.5, and the To assess the quality of our photometry algorithm, we have
′
flagAperForcedissettoon.Inthesecases,thesourcefluxesare performeda Monte-Carlo analysis. A total of 10000 simulated
severelyunderestimatedatallfrequencies.Thisflaggedpopula- sourcesarerandomlydistributedoverthewholeskyintheIRIS
tion contains978 sourceswhich are rejected from the physical andPlanckmaps.Thesourcesareassumedtofollowtheemis-
analysis of Sect. 4, but not from the entire catalogue, which is sionofamodifiedblackbodywith atemperaturerandomly,T,
usedtoassesstheassociationwithancillarydata(cfSect.3)and distributedbetween6Kand20K,andanassociatedspectralin-
tostudymorphologyatlargescale(cfSect.5). dex given by β = 11.5 T 0.66 within a 20% error bar, based
−
×
PlanckCollaboration:TheGalacticColdCorePopulationrevealedbythefirstPlanckall-skysurvey 5
Normal BadSfit100µm AperForced Upper100µm
Quantity Bias(%) 1σ(%) Bias(%) 1σ(%) Bias(%) 1σ(%) Bias(%) 1σ(%)
Fluxat100µm 1.4 31.7 1.0 4.7 -58.1 14.1 117.1 190.0
Fluxat857GHz -5.0 6.2 3.9 3.2 -56.3 13.8 -11.0 6.0
Fluxat545GHz -3.6 6.4 3.7 3.7 -55.8 14.4 -9.0 6.0
Fluxat353GHz -5.0 7.3 2.4 4.7 -58.9 14.8 -10.0 6.7
FWHM -0.6 16.2 30.9 27.7 -25.2 16.3 -6.7 15.3
Ellipticity 0.0 8.2 0.0 9.5 - - 0.0 9.0
T -4.2 5.2 -4.1 1.6 -6.5 3.8 0.4 16.0
β 9.8 7.3 10.5 2.4 11.2 6.7 2.7 18.7
Table 1. Statistics of the Monte-Carlo analysis performedto estimate the robustness of the photometryalgorithm. The bias (ex-
pressed in %) is defined as the relative error between the median of the outputdistribution of the photometryalgorithm and the
injectedinput.The1σ(expressedin%)representsthediscrepancyaroundthemostprobablevalueoftheoutputdistribution.Those
quantitiesaregiveninthevariouscasescorrespondingtotheoutputflagsprovidedbythealgorithm.Statisticsofthetemperature
andspectralindexisalsogivenheretoshowtheimpactoftheobservederroronfluxes.
Normal BadSfit100µm AperForced Upper100µm
Quantity Bias(%) 1σ(%) Bias(%) 1σ(%) Bias(%) 1σ(%) Bias(%) 1σ(%)
Fluxat100µm 11.5 44.3 0.8 8.4 -51.6 21.1 204.5 278.2
Fluxat857GHz -4.0 8.1 2.1 4.7 -58.3 20.1 -10.4 7.1
Fluxat545GHz -2.5 8.0 2.4 4.9 -57.4 21.3 -7.8 7.0
Fluxat353GHz -3.4 8.7 1.9 5.5 -59.3 21.3 -8.7 7.4
FWHM 0.0 18.1 31.0 31.1 -24.4 16.9 -5.2 17.6
Ellipticity 0.0 9.3 -0.5 9.2 - - 0.1 10.4
T -2.1 6.3 -3.2 1.8 -4.4 6.2 6.8 20.6
β 7.1 8.2 9.3 2.6 5.6 12.3 -4.9 20.4
Table2.SameasTable1intheGalacticplane(b <25 ).
◦
| |
ontheworkdoneonArcheopsdatabyDe´sertetal.(2008).The astrongsourceisembeddedinafaintbackground(e.g.athigh
FHWM ofthe simulatedsourcesspansfrom4.5 to 7 with an latitude),introducingadegeneracybetweenthefitofthecentral
′ ′
ellipticityrangingfrom0to0.87.Thefluxat857GHzistaken ellipticalGaussianandthepolynomialfitofthebackgroundsur-
from10to500Jy followingalogarithmicrandomdistribution. face at100µm. Althoughbiasand1-σvaluesare smallerthan
The derived fluxes in all IRAS and Planck bands take into ac- inthenormalcaseduetothestrongsignalofthesesources,we
countthe colour correction.We apply our complete process of reject this population from the physical analysis, because they
photometry on this set of simulated data, and retrieve an esti- couldintroducewrongestimatesofthephysicalpropertiesbased
mateofallquantities(fluxes,FWHM,ellipticity)inthevarious onahighlybiasedextension.
casesdescribedbytheflagslistedbefore(cfFig.A.1).Statistical
If we focus now on the normal case, when the photometry
bias and 1σ errors are derivedfor all quantitiesand cases, and
algorithm has performedwell, we first observe a slight bias of
arelistedinTable1and2forall-skyand b < 25 respectively.
◦ all fluxes estimates. The bias at 100 µm becomes larger when
| |
Wethetemperatureandspectralindexestimatesrecoveredatthe
looking into the Galactic plane (11.5% for b < 25 compared
endoftheprocessingarealsolistedtoillustratetheimpactofthe | |
to1.4%overthewholesky).ThefluxesPlanckbands,however,
errorsonthefluxes.
arelessunder-estimatedwhenlookinginsidetheGalacticplane,
with biases spanning from 2.5% to 5%. The associated 1σ er-
This Monte-Carlo analysis confirms, firstly, why sources
rorsareabout6to7%onall-skyand8-9%intheGalacticplane.
withAperforcedsettoonshouldberejectedfromthephysical
The impact of such a biased estimate of the fluxes will be dis-
study, since for these sources fluxes are systematically under-
cussedtogetherwiththe studyonthe calibrationuncertaintyin
estimatedbyabout60%.SourceswithUpper100µmsettoon,
Sect. 4.1.On the otherhand,theFWHM estimateare typically
for which only an upperlimit at 100µm hasbeen providedby
biasedbylessthan1%andhaveanaccuracyof 18%,whenthe
the algorithm, the flux at 100 µm is over-estimated by a factor
∼
ellipticitypresentsnobiasandanaccuracyof 9%.Finallythe
oftwo,withanassociateddiscrepancythatcanreachafactorof
∼
temperatureandspectralindexarederivedusingthemethodde-
threetimestheinputvalueinregionsclosetotheGalacticplane.
scribedinSect.4.1.Whereasthetemperatureisslightlyunder-
MoreoverthefluxesinthePlanckbandsaresignificantlybiased
estimated ( 2% in the Galactic plane), the associated spectral
tolowervalues,withabiasgreaterthanthe1σdiscrepancy.The
∼
index is over-estimated by 7%. The statistical 1-σ uncertain-
resulting temperature estimate is, as expected, greater than the
∼
tiesareabout6%and8%forT andβrespectively.Theseresults
injectedvalueandtheuncertaintiesinthetemperatureandspec-
willbetakenintoaccountindetailwhendiscussingthephysical
tralindexarearound20%.Thisillustratesthelimitationsonany
propertiesofthesecoldsourcesinSect.4.1.
physical conclusions that could be drawn from this population
ofsources.Whena badfitofthe 100µm backgroundhasbeen The Monte-Carlo simulations described here demonstrate
obtained,Bad Sfit 100 µm flag set to on, the main error comes the robustness of our photometryalgorithm,and justify the re-
fromthehighlybiasedestimateoftheFWHM( 31%),leading jectionofentirecategoriesofobjectsusingthephotometryflags,
∼
toanover-estimateofthefluxesinallbands.Thishappenswhen such as the Aper Forced, PS Neg and Bad Sfit 100µm. The re-
6 PlanckCollaboration:TheGalacticColdCorePopulationrevealedbythefirstPlanckall-skysurvey
Fig.1.Colour-Colourdiagramofthecatalogue.Theover-plotted Fig.2. Signal-to-noise ratio (SNR) of new sources (dash line)
symbolsstandforthepositivecross-matcheswithnonISMob- overlaidontheSNRofallsources(solidline).
jects.Theredcontoursgivethedomainofthediagramfilledby
Archeopscoldcoresassumedtofollowagrey-bodylaw,witha Simbadtype C3PO <MC>
temperaturerangingfrom6K < T < 25K,andaspectralindex [%] [%]
βgivenbyDe´sertetal.(2008). ISM 49.0 21.7
Star 2.3 4.9
Gal 2.1 7.4
mainingsample consists of 9465objects, dividedinto two cat- Radio 5.3 7.7
egories: 1840 objects have only an upper limit estimate of the QSO 0.1 0.3
Others 0.3 0.2
fluxat100µmand7625havewelldefinedphotometryinIRAS
Newdetections 40.9 57.8
and Planck bands. We will focus on this last category of 7625
sources for the rest of the analysis on the physical properties. Table3.CrossmatchwithSimbaddatabaseforC3POandsimu-
BasedonthisMonte-Carloanalysis,wewilladoptthefollowing latedcatalogues,foreachcategoryofSimbadtype.The<MC>
estimateofthe1σuncertaintyonfluxes:40%onIRAS100µm, columngivesanestimateoftheprobabilityofchancealignment
and8%onPlanck bands.Thiserrorismuchlargerthanthein- foreachSimbadtype.
trinsicpixelnoiseandsoinstrumentalerrorsareneglected.
2.5.Cross-Correlationwithexistingcatalogues
contoursof this figure show the domainfilled by dusty objects
As one step of the validation of our detections, we have per- assumingagrey-bodyemissionlaw,with6K < T < 25K,and
formed an astrometric search on the Simbad database2 for all a spectralindexβgivenbyDe´sertetal. (2008).Thematchbe-
known sources within a 5 radius of our sources. There are a tweenPlanckdetectionsandthiscolour-colourdomainisstrong.
′
large number of objects in the Simbad database which raises Only a few objects (17) show the colour-colour properties of
the question of chance alignments. This is especially true for radio emitters, located in the top-right corner, indicating real
extragalacticobjectswhichhaveareasonablyisotropicskydis- matches with extragalactic objects. For the rest of the sample,
tribution. To judge the number of chance alignments that can theprobabilityofchancealignmentishigh.Concerningtheas-
be expected by performing this kind of search, we have also sociationwithstars,exceptforafewX-rayemitters,mostlyall
conducted a Simbad cross check on the positions of a set of Simbadmatchesseemassociatedwithdustyemission,andthus
100Monte-CarlosimulatedcataloguespresentedinSect.5.1.1. representchancealignment.
TheseMonte-Carlorealizationsreproducetheobjectdensityof Wefinallyrejectonlytheobviousextragalacticmatches,lo-
the Planck catalogue per bin of longitudeand latitude. The re- catedinthetop-rightcornerofthecolour-colourdiagram,lead-
sultspresentedinTable3showthatthenumberofcoincidences ingto7608objects.
in the ISM category is greater in the C3PO catalogue than the Out of the 7608 sources in the photometric reliable cata-
probabilityofchancealignmentestimatedfromtheMonte-Carlo logue,40%havenocounterpartintheSimbaddatabase.Inad-
simulations. On the contrary, the fraction of contaminants (i.e. dition, these newdetectionshave a similar SNR distributionas
Galaxies,QSO,RadioSources,stars) isalwayslowerinC3PO theentirecatalogueasshowninFig.2,andcanbeconsideredas
thanintheMonte-Carlorealizations.Thusextragalacticobjects reliableastheentirecatalogue.
and Galactic non-dusty objects are mostly rejected by the de-
tection algorithm, whereas actual ISM structures are preferen-
3. SpatialDistribution
tiallydetected.AmoredetailedcomparisonbetweenC3POand
IRDCscataloguesispresentedinSect.7.1. 3.1.AssociationwithGalacticstructures
Neverthelessthe association with probablecontaminantsin
C3POisquitehigh( 10%)andnotallarenecessarilytheresult Theall-skydistributionofthe10783C3POsourcesispresented
ofchancealignments∼.Todisentanglebetweenchancealignment intheupperpanelofFig.3.MostlyconcentratedintheGalactic
and real matches, we use colour-colour information as shown plane, the distribution clearly follows Galactic structures be-
in Fig. 1. Mostly objects are distributed in the bottom-leftcor- tweenlatitudesof 20◦and+20◦.Afewdetectionsareobserved
−
nerofthediagram,typicalofdust-dominatedemitters.Thered at high Galactic latitude (b > 30◦) and after cross-correlation
| |
with external catalogues have been confirmed not to be extra-
2 http://simbad.u-strasbg.fr/simbad/ galacticobjects(seeSect.2.5).
PlanckCollaboration:TheGalacticColdCorePopulationrevealedbythefirstPlanckall-skysurvey 7
ColdCoreDensityMap
COcontoursonColdCoreDensityMap
AvcontoursonColdCoreDensityMap
Fig.3. Upper panel: All-sky map of the number of C3PO Planck cold clumps per sky area, smoothed at 3 . Middle panel: CO
◦
contoursareover-plottedontheC3POdensitymapwhichissetto0whereCOmapisnotdefined.Lowerpanel:Avcontoursare
over-plottedontheC3POdensitymapwhichissetto0whereAvmapislowerthan0.1Av.
8 PlanckCollaboration:TheGalacticColdCorePopulationrevealedbythefirstPlanckall-skysurvey
In themiddlepanelofFig.3, contoursoftheintegratedin- Name Lon Lat Area Distance Nb
tensitymapoftheCOJ1-0lineareoverlaidonthePlanckcold [deg] [deg] [deg2] [pc]
clumps density all-sky map. This CO map is a combinationof AquilaSerpens 3 28 30 260 59
COdatafromDameetal.(2001)andNANTENdata(Fukuietal. PolarisFlare 24 123 134 150 55
1999;Matsunagaetal.2001;Mizuno&Fukui2004),asdefined Camelopardalis 20 148 159 240 11
UrsaMajor 35 148 44 240 13
in Planck Collaboration (2011o). The correlation between CO
Taurus -15 170 883 140 393
and C3PO Cold Clumps is quite impressive and demonstrates
TaurusPerseus -15 170 883 350 227
onceagaintherobustnessofthedetectionprocessandthecon-
λOri -13 196 113 400 66
sistency of the physical nature of these Planck cold objects. A
Orion -9 212 443 450 353
detailed analysis shows that more than 95% of the clumps are Chamaeleon -16 300 27 150 114
associatedwithCOstructures. Ophiuchus 17 355 422 150 311
Hercules 9 45 35 300 16
The lower panelshowsthe same kindof spatial correlation
withtheall-skyAvmap(Dobashi2011inpreparation).TheAv Table 4. Molecular complexes used to associate C3PO cold
map traces more diffuse regions of the Galaxy and extents to clumpstoGalacticwell-knownstructures,forwhichanestimate
higherlatitude,wherecoldclumpsarealsopresent.About75% ofthedistanceisavailable.
oftheC3POobjectsareassociatedwithanAvsignaturegreater
than1.
3.2.3. Distancesfromextinctionsignature
3.2.DistanceEstimation
Genetic forward modelling (using the PIKAIA code
Distanceestimatesareessentialtoproperlyanalysethepopula-
Charbonneau 1995) is used along with the Two Micron
tionofdetectedcoldclumps.Wehaveusedfourdifferentmeth-
All Sky Survey (Skrutskie et al. 2006) and the Besanc¸on
ods:associationwithIRDCs,associationwithknownmolecular
Galactic model (Robin et al. 2003) to deduce the three di-
complexes,three dimensionalextinctionmethodusing2MASS
mensional distribution of interstellar extinction towards the
data,andextinctionmethodusingSDSSdata.
cold clump detections. The derived dust distribution can then
be used to determine the distance and mass of the sources,
independently of kinematic models of the Milky Way. Along
3.2.1. DistancestoIRDCs
a line of sight that crosses a cold clump, the extinction is seen
torise sharplyatthedistanceofthe cloud.Themethodisfully
Simonetal.(2006b)andJacksonetal.(2008)providekinematic
explainedinMarshalletal.(2006)andMarshalletal.(2009).
distance estimates for a total of 497 IRDCs extracted from the
Thedistance,asdeterminedbythistechnique,providesline
MSXcatalogue(Simonetal.2006a)thatconsistsof10931ob-
of sight information on the dust distribution. However, it does
jects. Kinematic distances are obtained via the observed radial
not have sufficient angular resolution to perform morphologi-
velocity of gas tracers in the plane of the Galaxy. By assum-
cal matches on the cold clumps. To ensure that the extinction
ing that the Galactic gas follows circular orbits and a Galactic
risedetectedalongthelineofsightisindeedrelatedtotheinner
rotation curve,an observedradialvelocity at a given longitude
structureweperformaconsistencycheckonthecolumndensity
corresponds to a unique Galactocentric radius. Of course, this
derivedfromthe extinctionandfromthe sourceflux,corrected
means that in the inner Galaxy, two heliocentric distances are
foritstemperature.Onlydetectionswherethetwocolumnden-
possible. Thistechniqueis onlyapplicablein the plane andre-
sitiesareinagreementwithinafactoroftwoareretained.This
quires the availability of appropriate molecular data. We find
leadstodistanceestimatesfor978objectsoftheentireandpho-
127 Planck cold clumps, over the complete catalogue, associ-
tometricreliablecatalogue.
atedwithIRDCsthatalreadyhaveakinematicdistanceestimate.
Thisnumberdecreasesto32associationsoverthe7608objects
ofthephotometricreliableC3POcatalogue. 3.2.4. DistancesfromSDSS
AmorerecentworkbyMarshalletal.(2009)usesanextinc-
Distancestocoldclumpswithin1kpcareobtainedbyanalysis
tionmethod,detailedinSect.3.2.3,onthesameMSXcatalogue
ofdistance-reddeningrelationsforlatespectraltypestarswithin
ofIRDCstoderivethedistanceof1259objects.Thisyields188
thelineofsighttoeachsource(McGehee2011inpreparation).
associations with C3PO clumps over the entire catalogue, and
Specifically,weuseSloanDigitalSkySurveyphotometryofM1
47overthephotometricreliableC3POcatalogue.
toM5dwarfscolour-selectedbythereddening-invariantindex
E(g r)
3.2.2. Distancestoknownmolecularcomplexes Qgri =(g r) − (r i). (2)
− − Er i −
−
The all-sky distribution of cold clumps follows known molec- TheupdatedugrizreddeningcoefficientsofSchlaflyetal.(2010)
ular complexes. Many of these have distances estimates in the areused.ThemedianstellarlocusofCoveyetal.(2007)forms
litterature. To assign the distance of a complex to a particular the basis of a calibration between Q and the intrinsic g i
gri
−
cold clump we use the CO map of Dame et al. (2001)to trace colour.Afterdereddening,thedistancetoeachstarisdetermined
thestructureofthemolecularcloudaboveagiventhreshold,and includingcorrectionsforGalacticmetallicityvariationfollowing
test for the presence of cold clumpsinside this region.The as- Bochanskietal.(2010).
sociation has been performed on 14 molecular complexes (see Thedistance-reddeningprofileisconstructedbycomputing
Table4),leadingto1152distanceestimatesovertheentirecat- the median reddeningfor stars within a circularpatch centered
alogueand947onthephotometricallyreliablecatalogue.cata- on the core location for 25 pc wide distance bins spanning 0
logue. to 2000 pc. We fit the observedreddeningprofile to the model
PlanckCollaboration:TheGalacticColdCorePopulationrevealedbythefirstPlanckall-skysurvey 9
Method EntireC3PO ReducedC3PO
(10783) (7608)
IRDCs(Kinematic) 127 32
IRDCs(Extinction) 188 47
2MASSExtinction 978 978
SDSSExtinction 1452 1004
MolecularComplexes 1152 947
Total 3411 2619
Table 5. Number of distance estimates available of the C3PO
sources for each method. Notice that the total numbersare not
equaltothesumofallmethods,duetooverlapbetweenthem.
Fig.4.DistributionofC3POcoldclumpsasseenfromtheNorth
Galactic Pole. Colours stand for methodsused to estimate dis-
tance: Molecular Complex association (green), SDSS extinc-
tion(lightblue),2MASSextinction(darkblue),IRDCsextinc-
tion(orange)andIRDCskinematic(red).Thereddashedcircle
showsthe 1kpc radiusaroundthe sun. Black dashedlines rep- Fig.5. Distance distribution of the MSX IRDCs (Simon et al.
resent the spiral arms and local bar. The black circles give the 2006b) (solid line) and of the subset associated to the cold
limitsofthemolecularring. clumps of the entire C3PO catalogue (dot-dash-dash line) and
thephotometricreliablesubsetofC3PO(dottedline).
definedbyconvolutionofthenear-fieldplussinglecloudprofile
withaGaussian(indistancemodulus),thisfunctionis:
E(B V) =a+c x−x0 1 exp −t2 dt (3)
− obs Z−∞ √2πσ2 2σ2! Thenumberofobjectsforwhichwehaveadistanceestimate
wherexistheindependentvariable(distancemodulus),x isthe is 2619 out of a total of 7608 objects in our photometric reli-
0
locationofthesinglecloud,aisthenear-fieldreddening,cisthe able subset, i.e. 34%. The distances of the cold clumps span
∼
reddeningassociated with the cloud, and σ is the width of the from 0.1 to 7kpc, but they mainly concentrated in the nearby
Gaussian.Thefittedσvaluesaretypically0.4to0.5magnitudes Solar neighbourhood as shown on Fig. 4. This type of distri-
in m M, as expected from the standard deviation of the (r bution has been already demonstrated using simulations, see
z,M )−usedtoassignabsolutemagnitudes. − Fig.10ofMontieretal.(2010).Thelackofdetectionsatlarge
r
Analysisofcalibrationfieldscontainingwell-studiedmolec- distances is mainly caused by the effects of confusion within
ular cloud complexes, e.g. the Orion B Cloud, reveal that the the Galactic plane, from which suffers the detection method.
recovereddistancemoduliareunderestimatedby0.2to0.3mag- Nevertheless, when comparing the distance distribution of the
nitudes,consistentwiththebiasexpectedfromtheMdwarfmul- C3POcoldclumpsassociatedtoMSXIRDCswiththetotalsam-
tiplicityfraction. pleofSimonetal.(2006b)inFig.5,wenoticethatthefraction
Thisprocessingleadsto1452distanceestimatesovertheen- ofC3PO-IRDCsmatchesdoesnotdependondistanceandex-
tirecatalogueand1004overthephotometricreliableone. tendsto8kpc.
BecausethesubsetofC3POcoldclumpswithadistancees-
timatehasbeenobtainedusingdifferentmethods,exploringvar-
3.2.5. Combinedresults
iousregionsanddistancesoverthesky,thissampleappearshet-
The numberof sourcesfor whichdistancescould be recovered erogeneous.Thecompletenessofthecataloguewithdistancesis
depends on the method used (cf Table 5). There is some over- quite difficultto assess. Thuswe define two subsets forfurther
lapbuteachmethodhasitsdistinctadvantagesaccordingtothe analysis, especially when looking at number counts, for which
distancerangebeingconsidered.The2MASSextinctionmethod weknowthatthesampleismorehomogeneous:thefirstsubset
isnotverysensitive nearby(D<1kpc),asthereare notenough (1790objects)dealswiththelocalobjects(D< 1kpc)anduses
stars to determine accurately the line of sight information. In onlyestimatesfrommolecularcomplexesassociationandSDSS
contrast, the extinction method using SDSS is especially de- extinction;thesecondsubset(674objects)focusesondistantob-
signedfornearbyobjects.Forobjectswith1kpc,wehaveused jects(D>1kpc)andusesonly2MASSextinctionestimatesand
SDSSdistanceswhenavailbleormolecularcomplexdistances. IRDCsassociations.
10 PlanckCollaboration:TheGalacticColdCorePopulationrevealedbythefirstPlanckall-skysurvey
Fig.7. Reduced χ2 obtained in the case β = 2 as a functionof
the temperature obtained with β free. When T becomes lower,
theχ2becomeslarger.
line),ofthewarmenvelopeT (redline),andofthelocalback-
env
groundT (reddotdashline)distributionspeakrespectivelyat
bkg
13.9K, 15.1K and 16.1K. The uncertainty on the temperature
estimatesisabout7%.Theseresultsareingoodagreementwith
theexpectedvaluesofcoldcores(e.g. Bergin&Tafalla 2007)
and consistentwith the results of our Monte-Carlosimulations
demonstratingthat our source extractionmethodaccuratelyre-
Fig.6.Distributionofthetemperatureofthecoldclumps(blue),
coversthe cold source parametersin the presence of a warmer
of the warm envelope (red) and of the total (green) estimated
background.
inside the elliptical Gaussian of the clump itself. The averaged
temperature of the local background is plotted in red dot-dash Inasecondanalysis,weperformedathreeparameter(A,T
line. and β) χ2 fit leading to the temperature distributions shown in
thelowerpanelofFig.6.Theχ2 fitisperformedonagridtak-
ingintoaccountthecolourcorrectionasdefinedinPlanckHFI
4. PhysicalProperties
Core Team (2011b) and gives the exact minimum of the χ2 in
4.1.Temperature the(A,T,β)spaceandprovidingtheassociated1-σuncertainty.
Evaluatingtheχ2obtainedwithβ=2asafunctionofthebestfit
ThetemperatureofthesourcesisestimatedfromSEDsusing4 temperatureobtained from the full three parameter fits, we see
bands:theIRAS100µmandthethreehighestfrequencyPlanck that a modelβ = 2 is reasonable for temperaturesin the range
bands857GHz,545GHzand353GHz.Theassumedemission 10K < T < 18K (forwhichthe χ2 < 1),butdoesnotprovide
modelisamodifiedblack-bodylaw,definedas: a good fit at lower temperature T < 10K (see Fig. 7). In fact,
the lower the temperature, the worse the fit. Using β as a free
S = AB (T)νβ, (4)
ν ν parameter, the temperature distributions peak at 13K, 13.9K,
15.5Kand17KforT ,T ,T andT respectively,withan
where S is the flux integrated over the solid angle C tot env bkg
ν
error of about7%. The associated spectral index β varies from
Ω = πσ σ , A is the amplitude, T is the temperature,
C Maj Min
1.5to3,withanuncertaintyof21%andameanvalueof2.1for
βisthespectralindexandB isthePlanckfunction.
ν
coldclumpsand1.8forthetotalemission,consistentwithother
Foreachsource,asetoffourtemperaturesismeasured:(1)
studiesbased on Planck data (PlanckCollaboration2011o,t,u).
the temperature of the clump T is defined as the temperature
C Thetemperatureofthecoldclumpsspantherange7Kto17K.
basedontheSEDsofthecoldresidualasdescribedinSect.2.3;
(2)thetemperatureofthewarmenvelopeT isobtainedfrom Thebiasandtheuncertaintyofthetemperatureandspectral
env
aperturephotometryoverthesameregionbutperformedonthe indexhavetobeadjusted,takingintoaccounttheMonte-Carlo
warmcomponent;(3)thetotaltemperatureT isdefinedasthe analysisofthephotometryalgorithm(seeSect.2.4),andtheim-
tot
temperatureofthesourceintheinitialmap,i.e.withoutremov- pactofthecalibrationuncertaintydetailedinSect.B.Werecall
inganywarmcomponent;(4)thetemperatureofthelocalback- thatabiasof -2%onTand 7%onβisinducedbythepho-
∼ ∼
groundT isdefinedasthetemperatureoftheaveragesurface tometryitself. Ontheotherhand,thecalibrationuncertaintyof
bkg
brightnessaroundthesource. fluxesdoesnotintroducedanybiasonT orβ,butgeneratesan
We have first fixed the spectral index to β = 2 (Boulanger errorof 8%onβandfrom3%to5%onT,thatshouldbeadded
etal.1996).Aχ2 fitisperformedontheSEDstoderivealles- quadrati∼callytotheuncertaintyduetostatisticalerrors.Allthese
timatesoftemperaturesandassociated1-σerrors.Thedistribu- considerationsleadtoafinalrangeoftemperaturespanningthe
tionofthesetemperaturesisshownontheupperpanelofFig.6. range7Kto17Kwithanuncertaintyofabout9%,andaspec-
ThetemperatureofthecoresT (blueline)peaksat13.4Kand tralindexβvaryingfrom1.4to2.8withanuncertaintyofabout
C
spansfrom9Kto16K.ThetemperatureofthetotalT (green 23%.
tot