AcceptedbytheAstrophysicalJournal PreprinttypesetusingLATEXstyleemulateapjv.12/16/11 DENSEGASINMOLECULARCORESASSOCIATEDWITHPLANCK GALACTICCOLDCLUMPS JinghuaYuan(袁敬华)1,YuefangWu2†,TieLiu3,TianweiZhang4,JinZengLi1,Hong-LiLiu1,FanyiMeng5, PingChen2,RunjieHu2,andKeWang6 1NationalAstronomicalObservatories,ChineseAcademyofSciences,20ADatunRoad,ChaoyangDistrict,Beijing100012,China; 2DepartmentofAstronomy,PekingUniversity,100871Beijing,China; 3KoreaAstronomyandSpaceScienceInstitute776,Daedeokdae-ro,Yuseong-gu,Daejeon,RepublicofKorea305-348; 4PekingUniversityHealthScienceCenter,XueyuanRoad38th,HaidianDistrict,Beijing100191,China; 5PhysikalischesInstitut,Universita¨tzuKo¨ln,Zu¨lpicherStr.77,50937,Germany;and 6EuropeanSouthernObservatory,Karl-Schwarzschild-Str.2,D-85748GarchingbeiMu¨nchen,Germany 6 DraftVersion,January20,2016 1 0 ABSTRACT 2 WepresentthefirstsurveyofdensegastowardsPlanckGalacticColdClumps(PGCCs). Observationsinthe n J =1−0transitionsofHCO+andHCNtowards621molecularcoresassociatedwithPGCCswereperformed a usingthePurpleMountainObservatory13.7-mtelescope. Amongthem,250sourceshavedetection,including J 230 cores detected in HCO+ and 158 in HCN. Spectra of the J = 1−0 transitions from 12CO, 13CO, and 9 C18Oatthecentersofthe250coreswereextractedfrompreviousmappingobservationstoconstructamulti- 1 linedataset. Thesignificantlylowdetectionrateofasymmetricdouble-peakedprofiles,togetherwiththewell consistence among central velocities of CO, HCO+, and HCN spectra, suggests that the CO-selected Planck ] coresaremorequiescentcomparedtoclassicalstar-formingregions. Thesmalldifferencebetweenlinewidths A ofC18OandHCNindicatesthattheinnerregionsofCO-selectedPlanckcoresarenotmoreturbulentthanthe G exterior. The velocity-integrated intensities and abundances of HCO+ are positively correlated with those of HCN, suggesting these two species are well coupled and chemically connected. The detected abundances of . h bothHCO+ andHCNaresignificantlylowerthanvaluesinotherlow-tohigh-massstar-formingregions. The p low abundances may be due to beam dilution. On the basis of the inspection of the parameters given in the - PGCC catalog, we suggest that there may be about 1 000 PGCC objects having sufficient reservoir of dense o gastoformstars. r t Keywords:ISM:abundances–ISM:clouds–ISM:kinematicsanddynamics–ISM:molecular–stars: forma- s a tion [ 1 1. INTRODUCTION limitsduetothepoorresolutionofPlanckbands. Amongthe v C3PO clumps, 915 early cold cores (ECCs) were identified Starsformindenseregionsofmolecularclouds. However, 3 based on criteria of S/N > 15 and T < 14 K where S/N is some physical and chemical properties of cold compact ob- 8 the signal-to-noise ratio. The ECC sample is delivered as a jects that breed stars are sill poorly understood. A key ap- 7 partofthe Planck EarlyReleaseCompactSourceCatalogue proachwouldbetostatisticallyinvestigatecolddenseclumps 4 (ERCSC,PlanckCollaborationetal.2011b).Asthehigh-S/N 0 intheGalaxy. Inthelastdecade,numerousdenseclumpsre- andcoldtailoftheC3POobjects,ECCsareexcellenttargets . vealed by surveys in continuum at (sub-)millimeter domain 1 for studying the earliest stages of star formation (Juvela et usingground-basedfacilitiesandtheHerschelSpaceObser- 0 al. 2010, 2011, 2012, 2015; Parikka et al. 2015). Recently, vatory(e.g.Schulleretal.2009;Andre´ etal.2010;Molinari 6 Planck Collaboration et al. (2015) released the Planck Cata- et al. 2010; Aguirre et al. 2011) have led to significant ad- 1 logofGalacticColdClumps(PGCCs). Asthefullversionof vancement.However,thesesurveysaremainlyconfinedinthe : theECCcatalog,thePGCCcatalogprovides13188Galactic v Galactic plane and several well-known nearby star-forming sources spreading across the whole sky. Containing sources i regions,anall-skysurveyatmulti-bandsiscrucialtotheun- X derstandingofstarformationindifferentenvironments. inverydifferentenvironments,thePGCCcatalogisusefulfor r ThePlancksatellite(Tauberetal.2010;PlanckCollabora- investigatingtheevolutionfrommolecularcloudstocores. a Torevealthemolecularcharacteristics,wecarriedoutsur- tionetal.2011a)carriedoutthefirstall-skysurveyatmulti- veyobservationsinthe J = 1−0transitionsof12CO,13CO, bandsinthesubmm-to-mmrangewithunprecedentedsensi- andC18Otowardsmorethan600Planckcoldclumpsselected tivityandprovidesaninventoryofcoldcondensationsofin- from the ECC catalog in both single-pointing and on-the-fly terstellar matter in the Galaxy. The first all-sky Cold Clump modes(Wuetal.2012;Liuetal.2012,2013a,2015;Menget CatalogueofPlanckObjects(C3PO)releasedinPlanckCol- al.2013). WefoundthatPlanckcoldclumpshavethesmall- laborationetal.(2011d)consistsof10,342coldsourcesthat estlinewidthscomparedtootherstar-formaingsamples(e.g., stand out against a warmer environment. The C3PO clumps IRDCs, Wangetal.2009,2014)indicatingquiescentfeatures are cold with dust temperatures ranging from 7 K to 19 K, andnatureofveryearlyevolutionarystages. with a distribution peaking around 13 K. A detailed anal- Although some kinematic properties have been revealed, ysis of ten C3PO clumps carried out by Planck Collabora- the low critical densities of CO and its isotopologues pre- tion et al. (2011c) revealed low column densities of N ∼ H2 vent us investigating the denser inner regions. Observations (0.1−1.6)×1022 cm−2 which can only be treated as lower in dense gas tracers are crucial to unveiling the physical and chemicalfeaturesoftheinterioroftheseobjects. †[email protected] 2 Yuanetal. Inthispaper, wereportonasurveyofthe J = 1−0tran- servedin2013. Thesystemtemperaturesareintherangeof sitionsofHCO+ andHCNtowardsthesmallersub-structures 133to363K,withameanvalueof190K.TheCLASSsoft- enclosed in the ECCs which have been previously mapped wareoftheGILDASpackage(Pety2005)wasusedtoreduce inCO(seesection2). CombiningspectraofCOanditsiso- thedata.Inordertoachieveahighersensitivity,wesmoothed topologuesextractedfrompreviousmappingobservations,we thespectratoalowervelocityresolutionof0.42kms−1. The havecarriedoutacomprehensivestudyofdensegasinthese finalreducedspectrahave1σnoisesrangingfrom0.02to0.23 sources. Thispaperisarrangedasfollows. Wepresentade- K,withameanvalueof0.07Kandamedianof0.06K. scriptionofthesampleandobservationsinSection2andthe resultsinSection3. InSection4,wetrytocomprehensively 3. RESULTS discuss the data. The main findings are summarized in Sec- Among the 621 observed molecular cores, 250 ones have tion5. validdetectionineitherHCO+ J = 1−0orHCN J = 1−0, including138coresdetectedinbothlines, 92coresonlyde- 2. THESAMPLEANDOBSERVATIONS tected in HCO+ J = 1 − 0, and 20 cores only detected in As mentioned in section 1, more than 600 Planck cold HCN J = 1−0. The designations and coordinates of these clumps have been mapped in the J = 1 − 0 transitions of 250sourcesaregiveninthefirstthreecolumnsofTable1. In 12CO, 13CO, and C18O. A 22(cid:48) × 22(cid:48) region for each clump Figure1,theyarerepresentedwithfilledcircles. Wenotethat wasmappedwithaspatialresolutionofabout52(cid:48)(cid:48). Onlythe thespatialdistributionofCO-selectedcoresisbiasedtowards central14(cid:48)×14(cid:48)portion,wherethesensitivityishighenough, nearbystar-formingregions(e.g.,Perseus,TaurusandOrion) havebeentakenintoaccount. Detailsaboutthemappingob- whiletheGalacticplaneisunder-represented.Thisisbecause servations are referred to Liu et al. (2012) and Meng et al. ofthehigherconfusionintheGalacticplane,sothattheS/N (2013).AsdemonstratedinMengetal.(2013),wehaveiden- tends to be lower there, but also because the dust tempera- tified sub-structures in regions defined by a contour of 60% turesintheplanetendtobehigherconflictingtotheselection ofthepeakintensityof13CO.Intotal, 621smallersubstruc- criteriaofECCs(PlanckCollaborationetal.2011d). Thedis- tureshavebeenidentified(Liuetal.2012;Mengetal.2013, tributionofthetargetsintheface-onMilkyWayisshownin T. Zhang et al. in prep; P. Chen et al. in prep; R. Hu et al. Figure 2. Most sources (∼ 85%) reside in the 10 kpc inner in prep), and are referred as CO-selected cores in this work regionfromtheGalacticcenterwithaconcentrationnearthe to be distinguished from the parental clumps. These CO- Sun. ExamplespectraofseveralsourcesarepresentedinFig- selectedcoreshavesource-averagedcolumndensitiesranging ure3. Mostsourceshaveonevelocitycomponentexceptfor from 7.6×1020 cm−2 to 3.7×1022 cm−2 with a mean value fouroneswhichshowtwovelocitycomponents.Inallthe250 of 7.2 × 1021 cm−2. The source-averaged volume densities sources,254velocitycomponentshavebeendetected. are in the range of (0.1−68.6)×103 cm−3 with a mean of 4.6×103 cm−3. ThelowmeanexcitationtemperatureofCO 3.1. ObservedParameters (∼ 11 K) suggests quiescent features. The distances to 164 CO-selectedcoreshavebeenadoptedfromthePGCCcatalog From previous mapping observations, we have extracted (PlanckCollaborationetal.2015). Thedistancestoother382 spectraofthe J = 1−0transitionsof12CO,13COandC18O sources, whose distances are not provided in the PGCC cat- atthepeakofeachcoretoconstructadatasetwithfivelines. alog, arefromWuetal.(2012). Notethatwehaveconfined Forspectraof12CO,13CO,C18OandHCO+,peakintensities, the Galactocentric distance to be smaller than 20 kpc while centroid velocities, and line widths of all the 250 cores have matchingthesamplewiththecatalogsofWuetal.(2012)and beenobtainedbyfittingGaussianprofiles. PlanckCollaborationetal.(2015). Thedistancesrangefrom ForHCNJ =1−0,wehaveperformedhyperfinestructure 0.1to14.7kpcwithameanvalueof1.3kpc. fittingbyinitiatingtheintensityratiosofIF=1−1/IF=2−1 = 0.6 Single-pointing observations of the CO-selected cores in and IF=0−1/IF=2−1 = 0.2 using the HFS method in CLASS HCO+ J = 1−0(89.189GHz)andHCN J = 1−0(88.632 whichresultedinfourparametersp1,p2,p3,andp4.Here,p2 GHz) were carried out using the Purple Mountain Observa- andp3arethecentroidvelocityandlinewidth.Theintensities tory (PMO)13.7-m telescope withthe position-switch mode and optical depths of the main and satellite components can from June to July of 2013 and from May to June of 2014. beexpressedasfunctionsofp1andp4, The half power beam width (HPBW) and main beam effi- ciency at 89 GHz are about 59(cid:48)(cid:48) and 55%, respectively. The T = p1(cid:16)1−e−p4(cid:17), pointingandtrackingaccuracieswerebothbetterthan5(cid:48)(cid:48) de- main p4 termined by scanning solar planets. A double sideband SIS p1(cid:16) (cid:17) T = 1−e−R·p4 , (1) receiver was used to cover both lines in the upper sideband satellite p4 (USB)whichhasawidthof1GHzallocatedto16384chan- τ =p4, nels (Shan et al. 2012). The spectral resolution is about 61 main kHz, corresponding to a velocity resolution of 0.21 km s−1. τsatellite=R·p4. Inourobservations,theoffpositionforeachsourcewasorig- inallysettobe30(cid:48) westawayfromthetargetalongtheright Here, R is the intrinsic intensity ratio of the satellite compo- ascensiondirection.Wehavecarefullycheckedthetheoffpo- nent to the main component. The detailed procedure of the hyperfine fitting is referred to the online documents of the sitionswithdirectobservationstofindthatmostofthemare emission-free in both HCO+ and HCN. For the five sources, CLASSsoftware1.Theresultingobservationalparametersfor HCO+ J =1−0andHCNJ =1−0aregiveninTable1. whichshowanomalousabsorptionfeaturesinthespectrain- dicating possible line contamination, the off positions were reset to the opposite direction, i.e. 30(cid:48) east from the target. 3.2. DerivedParameters The on-source time was about 5 minutes for most sources, andlongerthan10minutesforasmallnumberofobjectsob- 1http://www.iram.fr/IRAMFR/GILDAS/doc/html/class-html DensegasinCOselectedcoresassociatedPGCCs 3 90 ◦ 60 ◦ 30 30 25 ◦ 20K k b 0◦ 15m s − 10 1 30 ◦ − 5 60 0 ◦ − 90 ◦ − Figure1. Spatialdistributionoftheobservedsources(blackcircles)andtheoneswithvaliddetection(bluecircles). ThebackgroundshowsCOJ = 1−0 emissionoftheType3mapfromPlanckCollaborationetal.(2014). themeasuredbrightnesstemperatures: 16 T (12CO) 1−exp(−τ ) r ≈ 12 . (3) T (13CO) 1−exp(−τ ) r 13 8 Here,τ12/τ13 =[12CO]/[13CO]. Theisotoperatioof12C/13C foreachsourcehasbeendeducedfromthefollowingequation (Pinedaetal.2013), ) c (kp 0 12C =4.7Rgal +25.05, (4) Y 13C kpc whereR istheGalactocentricdistanceinkpc. Withthede- −8 gal rivedopticaldepth,theexcitationtemperaturecanbereached fromEquation2basedontheintensityof12CO. Under LTE condition, the column density of a linear −16 moleculecanbeexpressedas (cid:16) (cid:17) 3k Q exp Eup −16 −8 0 8 16 N= rot kTex X (kpc) 8π3νµ2S gJ+1 J(Tex) (cid:90) 1 τ Figure2. Spatialdistributionofthesourceswith(bluecrosses)andwithout ×J(T )−J(T )1−exp(−τ) Tr dv. (5) (blackcircles)detectionintheface-onMilkyWay.Thespiralarmsarerepro- ex bg dΘouf0c1e=0dk2up2sc0i.nkgmthse−41-fiarttmedmtoodHeliiorfeHgioouns&. THhaenla(2rg0e14d)aswhietdhcRi0rcl=eh8a.5skaprcadainuds Htheere1,−µ0isttrhaenspietiromna,ntehnetldinipeoslteremnogmtheSnt=oftJh+e1m=ole1c,utlhee. Fdoe-r generacy gJ+1 = 2J +3 = 3, where J isthe2Jr+o3tatio3nalquan- tumnumberofthelowerstate. WehavefollowedMcDowell Forthesakeofrevealingphysicalfeaturesofthesemolec- (1988)towritethepartitionfunctionasQ = kT + 1 where ularcores,somefundamentalparametershavebeendeduced rot hB 3 B is the rotational constant. The molecular parameters have fromtheobservedspectraundertheassumptionoflocalther- beenadoptedfromtheCologneDatabaseforMolecularSpec- malequilibrium(LTE).Themeasuredbrightnesstemperature troscopy(CDMS)2. (T ) for a specific transition as a function of excitation tem- r Inthecalculation,theexcitationtemperaturesof13COand perature(T )canbeexpressedas ex C18O have been assigned to be equal to that of 12CO. This hν would be reasonable while considering the fact that 12CO, Tr= k [J(Tex)−J(Tbg)]×[1−exp(−τ)]f, (2) 13CO and C18O are well coupled in ECCs (Wu et al. 2012). The optical depths and column densities of 13CO have been whereJ(T)=[exp(hν/kT)−1]−1,Tbg =2.73Kisthetemper- obtainedfromEquations3and5. ForC18O,wederivedopti- atureofthecosmicbackgroundradiation,and f isthebeam- caldepthsandcolumndensitiesfromEquations2and5. fillingfactor.UnderLTEcondition,theopticaldepthsof12CO and13COcanbestraightforwardlyobtainedfromcomparing 2http://www.astro.uni-koeln.de/cdms/ 4 Yuanetal. Figure3. Examplespectra. For lacking observations of transitions from an isotopo- wardlyderivedfromEquation2usingopticaldepthsobtained logue and a higher excitation level, the excitation tempera- fromHFSfitting(seesection3.1andTable1). Forthederiva- tures of HCO+ cannot be reliably derived. Additionally, op- tion of column densities of HCN, Equation 5 was used by ticallythinfeatureformostsources(seeTable2)furtherpre- multiplying a factor of 1.8 which comes from the intensity ventedusobtainingestimatesusingEquation2underanopti- ratiosamongthemainandsatellitecomponents. callythickassumption. Inthiswork,weassumedtheexcita- ThecolumndensitiesofCOforthe189coreswithdetection tiontemperaturesofHCO+ J = 1−0tobeitsupperenergy inC18OwerecalculatedfromthatofC18Owithaconsidera- inKelvin(E /k = 4.28K).Then theopticaldepthsandcol- tionofisotoperatioof16O/18OasafunctionofGalactocentric u umndensitieswerederivedfromEquations2and5. Wenote distance(Wilson&Rood1994), thatthecolumndensitiesobtainedherearelowerlimits(e.g., 16O R Hatchelletal.1998;Miettinen2012,2014)duetotheuseof =58.8 gal +37.1. (6) upperenergyastheexcitationtemperatures. 18O kpc For HCN, the excitation temperatures were straightfor- ForsourceswithoutdetectioninC18O J =1−0,theCOcol- DensegasinCOselectedcoresassociatedPGCCs 5 The differences may be due to the sources in this work are 50 relativelydenseregionsinclumpsstudiedbyPlanckCollabo- rationetal.(2011d)andWuetal.(2012).Themeanexcitation temperatureisabout12.54Kapproximatelyequaltothemean 40 dust temperature (13 K, Planck Collaboration et al. 2011d) andslightlyhigherthanthevalueinWuetal.(2012). Theex- er30 citationtemperaturesofthesesourcewithHCO+and/orHCN b m detection are significantly higher than that of CO-selected u N coresintheTaurus,Perseus,andCaliforniacomplexes(Meng 20 et al. 2013), but similar to that of CO-selected cores in the Orioncomplex(Liuetal.2012). ThecolumndensitiesofH 2 covertherangeof(1.5−70.9)×1021cm−2withameanvalue 10 of 1.3×1022 cm−2, much higher than those of CO-selected coresintheOrion,Taurus,Perseus,andCaliforniacomplexes (Liu et al. 2012; Meng et al. 2013). This is a natural result 0.5 1.0 1.5 2.0 2.5 becausethesourceswithdetectioninHCO+ and/orHCNare NH182/NH132 relativedenserones. Figure4. Distribution of the N18/N13 ratios for the 189 sources with 4. DISCUSSIONS H2 H2 C18Odetection. In order to assure the detection of HCO+ and/or HCN is not purely distance and sensitivity limited, we have plotted umn densities were obtained from that of 13CO by adopting histograms of distances and RMS noises of HCO+ for CO- isotoperatiogiveninEquation4. ThentheH columndensi- 2 selectedcoreswithandwithoutdetection.AsshowninFigure tieswerederivedwiththeconsiderationofCOabundancesas 5(a),thedistancesforsourceswithandwithoutdetectionhave afunctionofGalactocentricdistances(Fontanietal.2012), asimilardistribution.Thesourceswithdetectionindensegas (cid:32) (cid:33) R tracers have distances ranging from 0.1 to 14.7 kpc with a XCO =8.5×10−5exp 1.105−0.13 gal . (7) meanof1.4kpcandamedianof0.8kpc. Forsourceswith- kpc out detection, the values range from 0.1 to 10.2 kpc with a Inordertotesttheconsistencybetweenthetwomethods,we meanof1.2kpcandamedianvalueof0.9kpc. Thesimilar have obtained two sets of H column densities of the 189 meanandmedianvaluesforsourcewithandwithoutdetection 2 sources from C18O and 13CO and denoted them as N18 and suggest that distances have not induced biases to the detec- H2 tionofdensegastracersintheCO-selectedcores. Asshown N13. As shown in Figure 4, the N18/N13 ratios are ranging H2 H2 H2 in Figure 5(b), the distributions of RMS noises of sources from0.3to2.6withameanof0.9andastandardderivation withandwithoutdetectionaresimilarwithanexcessfornon- of 0.3, indicating that the H column densities derived from 2 detection sources at the high noise end (rms > 0.12 K). The 13COandC18Oareconsistenttoeachother. fractionsofhighnoisesare0.18and0.07forsourceswithand OneshouldnotethatthederivedH columndensitieswould 2 withoutdetection. Nevertheless,morethan80%CO-selected suffer from large uncertainties due to the abundance varia- cores without detection have RMS noises as low as the ones tionsofCO.AsnotedintheliteraturetheGalacticgradients with detection. To further check the influence of the large of abundances (Equations 4, 6, and 7) would be valid in the range of sensitivity on the detection, we have separated the Galactocentricdistancesbetween3kpcand10kpcwhereCO observed sources into high- and low-sensitivity groups us- hassignificantemission(Wilson&Rood1994;Pinedaetal. ing the median RMS noise (∼ 0.06 K) as a threshold. The 2013). As shown in Figure 2, most sources reside inside 10 detection rates for the high- and low-sensitivity groups are kpc from the Galactic center. For these targets inside the 10 38.7% and 40.5%, respectively. Such consistency, together kpc circle, abundance ratios of [12C]/[13C], [16O]/[18O] and with the similarity between distributions of RMS noises of [CO]/[H ] have been obtained from Equations 4, 6, and 7. 2 sources with and without detection, suggests that the detec- The[12C]/[13C]rangesfrom39to72,[16O]/[18O]from213 tionofHCO+and/orHCNisnotsensitivitylimited. to 625, and [CO]/[H ] from 0.7 × 10−4 to 1.7 × 10−4. For 2 sourceswithGalactocentricdistanceslargerthan10kpc,the 4.1. Kinematics abundance ratios at R = 10 kpc have been adopted. We gal Wehavecarefullyinspectedthespectraofsourceswithde- comparedthe[12C]/[13C]obtainedfromEquation4ratiosto tection in HCO+ and/or HCN to find that all targets show that from Wilson & Rood (1994) and found that the differ- single-peaked profiles except for seven ones showing self- encesaresmallerthan25%.Weestimatethatthevariationsof absorption features (see Figure 3). Among the seven excep- abundanceratioswouldintroduceuncertaintiessmallerthana tions,twoareredprofiles(i.e.lineswithhigherpeaksskewed factorofoneforsourceswithRgal <10kpc.Theuncertainties to the red sides; G084.79-01.11A1 and G120.16+03.09A1) forsourceswithRgal >10kpcwouldbelarger. andfourblueprofiles(G114.56+14.72A1,G126.49-01.30A2, Theresultingopticaldepthsof13CO,C18OandHCO+,ex- G172.85+02.27A1, and G226.36-0-0.50A1). The other one citationtemperaturesofCO,columndensitiesof13CO,C18O, showssymmetricdouble-peakedprofilewithadipattheline HCO+,andHCNaregivenincolumns4–11ofTable2. The center(G120.98+02.66A1). Asymmetricdouble-peakedpro- statisticsofsomederivedparametersarepresentedinTable3. filesarealwaysregardedaseffectivetracersofkinematicsin The excitation temperatures in the CO-selected cores with star-forming cores with blue profiles linked to inward mo- HCO+and/orHCNdetectionrangefrom7.6to34.4K,signif- tions (e.g., collapse or infall, Zhou et al. 1993; Wu & Evans icantlyhigherthanthevaluesinWuetal.(2012)andthedust 2003;Evansetal.2005;Wuetal.2005)andredprofilesfor temperatures (7–19 K, Planck Collaboration et al. 2011d). outward motions (e.g., outflow or expansion Wu et al. 2007; 6 Yuanetal. 120 120 (a) detection (b) detection no detection no detection 100 100 80 80 r r e e b b m 60 m 60 u u N N 40 40 20 20 10-1 100 101 0.0 0.1 0.2 Distance (kpc) rmsHCO+ (K) Figure5. Distributionsofdistances(a)andRMSnoiseofHCO+(b)forsourceswithdetection(cyan)inetherHCO+orHCN,andsourceswithoutdetection (yellow)inneitherHCO+norHCN.Thegreenmeansanoverlaybetweenthetwogroups. rareness of bulk-motions in these dense CO-selected cores, 1.2 suggestingtheyarerelativelyquiescentcomparedtoclassical Gaussian fit star-formingregions. Anotherindicatorofkinematicsthatcanbeextractedfrom single-pointobservationscouldbethelinewidths. Themean 0.9 widths of 12CO, 13CO, and C18O of the CO-selected cores y withvaliddetectioninHCO+and/orHCNare2.95,1.67,and c n 1.08 km s−1, relatively larger than that in Wu et al. (2012) e in which they are 2.0, 1.3, and 0.8 km s−1. The spectra are u q 0.6 also slightly wider than those of CO-selected cores in Orion e (0.9 km s−1 for 13CO), Taurus (1.8 km s−1 for 12CO and 1.1 Fr km s−1 for 13CO), and California (2.2. km s−1 for 12CO and 1.4 km s−1 for 13CO) molecular complexes (Liu et al. 2012; 0.3 Meng et al. 2013). Slightly larger line widths detected in this work may be attributed to that the cores with detection in HCO+ and/or HCN are significantly denser than those in Taurus,CaliforniaandOrioncomplexes(seesection3.2)and higherrampressuresmakeadditionalcontributiontotheline −2 −1 0 1 2 widths. Another possibility could be that the CO-selected cores with detection in HCO+ and/or HCN are more turbu- ∆V ∆V (km s 1) C18O− HCN − lent. Usingdataofmultiplelinesfromdistinctspecieswithsig- Figure6. Distributions between differences of line widths of C18O and nificantly different critical densities, we can statistically ex- HCN. aminethedifferenceofkinematicsinseparateregions. Here, Yuan et al. 2013). The detection rates of blue and red pro- wemainlycomparethelinewidthsofC18OandHCNwhich, files are 1.2% and 0.8%, significantly smaller than the val- in general, are optically thin. The typical source-averaged ues in surveys toward low- and high-mass star-forming re- volume density is 4.6 × 103 cm−3 (see section 2), indicat- gions. Basedonresultsintheliterature, Evans(2003)found ingthattheremaybeasignificantvolumewithdensityabove the detection rates of blue (red) profiles in Class -I, 0, and I 104 cm−3 in these sources. With a critical density no larger low-massstar-formingregionsare35%(6%),33%(5%),and than 104 cm−3 , C18O J = 1−0 mainly originates from the 50% (19%), respectively. Among the 48 high-mass clumps exterior of a core. Meanwhile, HCN J = 1 − 0 is a good mappedinHCN J = 3−2andanopticallythinline(H13CN traceroftheinnerregionduetoitssignificantlyhighercriti- J = 3−2orC34S J = 5−4), Wuetal.(2010)reportedde- caldensityof>105 cm−3. ThelinewidthsofC18Oareinthe tection rates of 44% and 31% for blue and red profiles. We rangeof0.17–3.39kms−1 withameanvalueof1.08kms−1, note that the difference between detection rates of asymmet- while those of HCN ranging from 0.69 to 2.70 km s−1 with ric profiles in this work and that in the literature are not due a mean of 1.06. The typical optical depths are 0.2 and 0.3 toadifferenceinS/Noftheobservations,becausethedetec- for C18O J = 1 − 0 and HCN J = 1 − 0, respectively. tion rates are still significantly low (< 3%) even if we only Such small optical depths can only broaden the line widths consider sources with S/N > 10. The small detection rates by a factor of < 10% which has been estimated using Equa- of asymmetric double-peaked profiles in this work indicate tion 3 of Phillips et al. (1979). We note that the minimum DensegasinCOselectedcoresassociatedPGCCs 7 value for HCN is only an upper limit due to the poor spec- 4 tralresolutionofourobservations(seesection2). Themean (a) ) widthsofC18OandHCNarealmostequaltoeachother. The 1− distribution of difference between line widths of C18O and m s 2 HCN, which can be fitted using a Gaussian with a µ = 0.01 (k km s−1 and a σ = 0.51 km s−1, is shown in Figure 6. In- +O 0 C triguingly, the width difference is very small with statistical VH significance. ThisiscontrarytotheresultsinWuetal.(2010) −−2 O inwhichlargerwidthsoftransitionswithhighercriticalden- V13C sities are treated as an indication of larger turbulence in the −4 innerregionofdensecores. Thealmostequivalencebetween −60 −40 −20 0 20 widths of C18O and HCN suggests that the inner regions of CO-selected cores are not more turbulent than the exterior, VLSR(13CO) (km s−1) different from the situation in high-mass star-forming cores 4 wheretheinteriorisoflargerturbulence(e.g.,Wuetal.2010; (b) Garay et al. 2015). Compared to the two optically thin lines 1)− (i.e., C18O J = 1 − 0 and HCN J = 1 − 0), the lines of m s 2 13CO J = 1−0 are relatively broader with widths ranging k from 0.59 to 6.25 km s−1 with a mean value of 1.67 . The (N 0 C typical difference between line widths of 13CO and HCN is VH 0.65kms−1,andthevalueis0.42kms−1evenaccountingfor − O −2 thebroadeningattributedtoopticaldepth. Thisindicatesthe C V13 surroundingcloudsaremoreturbulent. −4 4.2. CouplingofDifferentSpecies −60 −40 −20 0 20 In the survey toward 674 Planck cold clumps, Wu et al. VLSR(13CO) (km s−1) (2012) found that the central velocities of transitions from CO and its isotopologues are strikingly consistent with each 4 (c) other. Thisfeaturecouldbereasonableinquiescentregions, 1)− forCOanditsisotoploguesshouldbewellcoupledwitheach m s 2 other without active disturbance. In this work correlations k aamreosnhgowcenntirnalFviegluorceiti7e,shoafv1e3CbOee,nHeCxOam+inaendd.HTChNe, dwihffiecrh- (HCN 0 V encesofcentralvelocitiesbetween13COandHCO+ (V − − VHCO+), 13CO and HCN (V13CO − VHCN), HCO+ and13HCOCN +CO −2 (VHCO+ −VHCN)havemeanvaluesof0.006±0.4,0.05±0.4, VH 0.002 ± 0.6 km s−1. Intriguingly, the differences are strik- −4 inglysmall,whichmaybeanindicationofquiescentfeatures −60 −40 −20 0 20 for these CO-selected cores. Noticeably, the deviations of VLSR(HCO+) (km s−1) (V13CO −VHCO+), (V13CO −VHCN), and (VHCO+ −VHCN) from zero seems to be consistent with the critical densities of the Figure7. Correlations among velocity centroids of different species. (a) cJm=−31),−a0ntdraHnsCitNion(s∼of513×C1O05(∼cm10−33)c.m−If3)H,HCCNO,+H(C∼O7+×, a1n0d4 V(HLCSRN()HvCs.OV+L)SvRs(.HVCLSOR+()1.3TChOe)l;o(wbe)rVpLaSnRel(sHsChNow)svsth.eVvLeSlRoc(i1t3yCdOiff);e(rce)ncVeLsS.R 13CO trace gases from the interior to the exterior, the devia- ofHCO+ areintherangeof(0.07−24.16)×1012 cm−2 with tionofvelocitydifferencefromzerocanalsoreflectthespatial ameanvalueof1.69×1012 cm−2 ,whilethatofHCNrang- correlation among different species. For instance, the small ingfrom0.24×1012 cm−2 to16.15×1012 cm−2 withamean deviationof(VHCO+−VHCN)mayindicatethatHCO+isbetter of1.57×1012 cm−2. TheabundancesofHCO+ (HCN)range coupledwithHCNthanCO. from3.0×10−12(1.3×10−11)to1.5×10−9(5.6×10−10)witha Shown in Figure 8 are relations of velocity integrated in- meanvalueof1.5×10−10(1.4×10−10)whicharesignificantly tensitiesof12CO,13CO,C18O,andHCO+ withthatofHCN. lower than values in low- to high-mass star-forming regions The intensities of CO and its isotopologues seem not corre- where the abundances of both HCO+ and HCN are in the lated with that of HCN. However, the relationship between rangeof10−10−10−8 (Dutreyetal.1997;Hirotaetal.1998; HCO+andHCNfollowsapowerlawofWHCO+ =2.3×WH0.C8N Vasyuninaetal.2011;Sanhuezaetal.2012;Liuetal.2013b; with a coefficient of determination of R2 = 0.7 indicating Gerneretal.2014;Miettinen2014). Thelowabundancesin about 70% of the data points can be well represented by the thisworkcouldbeattributedtopoorspatialresolutionsofour fitted linear relation in the log-log space. Such prominent observationswhichhaveledtosevereeffectofbeamdilution. correlationgivesadditionalevidencetothecouplingbetween The CO-selected cores with valid detection in HCO+ and/or HCO+andHCN.StrongcorrelationbetweenHCO+andHCN HCN have distances ranging from 0.1 pc to 14.7 kpc with a was also detected in Liu et al. (2013b) and interpreted as an meanof1.4kpc. Thecorrespondingphysicalscalescovered indicatorofatightchemicalconnection. by one beam range from 0.03 pc to 4.3 pc with a mean of 0.35 pc. However the region where the gas is dense enough 4.3. AbundancesofHCO+andHCN for HCO+ and HCN to be detectable would be smaller than Thestatisticsofcolumndensitiesandfractionalabundances 0.1 pc. Thus the measured column densities of HCO+ and ofHCO+andHCNaregiveninTable3. Thecolumndensities HCN have been severely diluted by a factor of tens on aver- 8 Yuanetal. (a) (b) 102 1m s)− 1m s)− k k K K 101 (WCO (W13CO 101 10-2 10-1 100 101 10-2 10-1 100 101 W (K km s 1) W (K km s 1) HCN − HCN − 101 (c) (d) 101 1)− 1)− m s m s K k 100 K k 100 (O (+ W18C WHCO 10-1 log(y)=(0.96 0.01)log(x)+(0.54 0.01) 10-1 ± ± 10-2 10-1 100 101 10-2 10-1 100 101 W (K km s 1) W (K km s 1) HCN − HCN − Figure8. VelocityintegratedintensitiesofCO(a),13CO(b),C18O(c),andHCO+(d)asfunctionsofthatofHCN. age,evenbyafactorof103 forthefarthestsources. Wenote sitiesof13COandexcitationtemperaturesofCOforsources that the H column densities, which have been used to ob- detectedinHCO+and/orHCNaresystematicallyhigherthan 2 tainfractionalabundancesofHCO+andHCN,wereestimated that of sources without detection. For H column densities, 2 basedonCOobservations. Asaforementioned,thevariation thetwogroupsofsourcesshareasimilardistributionatlow- of[12C]/[13C],[16O]/[18O]and[CO]/[H ]ratioswouldlead to intermediate densities. There is a prominent excess at H 2 2 tolargeuncertaintiestotheresults. Furtherdustobservations column densities higher than 2×1022 cm−2 for sources de- with sufficient resolution are needed to better constrain the tected in dense gas tracers. In general, sources detected in abundances. HCO+and/orHCNwouldhavehigherCOexcitationtemper- aturesand,probably,highercolumndensitiescomparedtothe 4.4. PGCCswithDenseGas oneswithoutdetection. Inordertopinpointwhichparameter(s)giveninthePGCC As the first survey of dense gas towards Planck cold catalog can point to the presence of dense gas, we have re- clumps,thestatisticalresultsinthisworkmayleadustodeter- visited H column densities, dust temperatures, masses and minewhetherthereareanyhints,inthemanyphysicalparam- 2 luminosities of all PGCCs (black) and the ones consisting etersgiveninthePGCCcatalog,thatcanpointtothepresence CO-selected cores with (red) and without (blue) detection in ofdensegas. HCO+and/orHCN.Thehistogramsandstatisticsofthesepa- Forsourcesdetectedindensegastracers,whatmakesthem rameters are presented in Figure 10 ahd Table 4. In general, stand out of the rest? To address this question, we have ex- the PGCCs containing CO-selected cores (for both with and aminedsomeparametersfromCOobservationsi.e.,peakin- withoutdetectionindensegastracers)aredenserandcolder tensities of 13CO, line widths of 13CO, H column densities, 2 comparedtothewholePGCCsample.Thereisnosurprisefor and excitation temperatures of CO. Shown in Figure 9 are thispointbecausetheCO-selectedcoreswereidentifiedfrom histograms of these inspected parameters. The CO-selected coreswithandwithoutdetectioninHCO+ and/orHCNshare ECCs which constitute the high-S/N and cold tail of PGCC objects. PGCCs containing CO-selected cores detected in commondistributionsforlinewidthsof13CO.Thepeakinten- DensegasinCOselectedcoresassociatedPGCCs 9 80 (a) detection 80 (b) detection no detection no detection 60 60 r r e e b b m m u40 u40 N N 20 20 0 2 4 6 8 10 12 0 1 2 3 4 5 Tmb(13CO) (K) ∆V13CO (km s−1) 80 (c) detection 80 (d) detection no detection no detection 60 60 r r e e b b m m u40 u40 N N 20 20 1021 1022 1023 5 10 15 20 25 NH2 (cm−2) Tex (CO) (K) Fwiigthurdeet9e.ctDioinstr(icbyuatnio)nisnoefthveerloHcCityO-+intoergHraCteNd,inantednssiotuiersceosfw13iCthOou(at)d,eltiencetiwonid(thyselolofw13)CinOn(ebi)t,hHer2HcoClOum+nndoernHsiCtiNes.aTnhdeegxrceietnatimoneatnesmapneroavtuerrelasyobfeCtwOefeonrsthoeurtcweos groups. HCO+ and/orHCNareslightlydensercomparedtotheones Wehavecarriedoutsingle-pointingobservationsintheJ = withnodetection. Thereisnosignificantdifferencebetween 1−0transitionsofHCO+andHCNtowards621CO-selected distributions of dust temperatures for these two subsamples. molecularcoresassociatedwithPlanckcoldclumps. Among AsshowninFigure10(c),theclumpmassesofthefullPGCC them,250sourceshavevaliddetectionineitherHCO+and/or sample and the PGCCs without detection in dense gas have HCN, including 138 cores detected in both lines, 92 cores similar distributions, while that of the the PGCCs with de- only in HCO+, and 20 cores only in HCN. Spectra of the tection in HCO+ and/or HCN slightly deviating to the high J = 1−0 transitions of 12CO, 13CO, and C18O for the 250 mass end with larger median values. For luminosities, the cores were extracted from previous mapping observations to fullPGCCsampleandtheothertwosubsamplessharesimi- construct a multi-line data set to reveal some kinematic and lardistributionsandthemedianvaluesarealsoconsistentina chemicalproperties. factorofone(seeFigure10andTable4). Althoughthereare SpectraofHCO+ J =1−0andHCNJ =1−0ofallsources nosignificantdifferencesforparametersamongdistinctsam- show single-peaked profiles except for seven ones showing ples,westilltentativelyfoundthatPGCCsourcescontaining self-absorption features. Among the seven exceptions, two densegaswouldbedenserandmoremassivecomparedtothe are red profiles, four blue profiles. The other one shows a remainingpool. OnthebasisofthestatisticspresentedinTa- symmetric double-peaked profile with a dip at the line cen- ble 4, we tentatively estimate that there may be about 1000 ter. The low detection rate of asymmetric profiles suggests PGCCobjectshavingsufficientreservoirofdensegastoform theCO-selectedcoresaremorequiescentcomparedtoclassi- stars. calstar-formingregions. The difference between line widths of C18O and HCN are close to zero with a small standard deviation, indicating that 5. CONCLUSIONS the inner regions of CO-selected cores are not more turbu- 10 Yuanetal. 150 (a) All PGCC/10 100 (b) All PGCC/10 detection detection 120 no detection no detection 80 r 90 r be be 60 m m u u N N 60 40 30 20 1019 1020 1021 1022 1023 5 10 15 20 25 NH2 (cm−2) Tdust (K) 40 (c) All PGCC/10 50 (d) All PGCC/10 detection detection no detection no detection 40 30 r r be be 30 m m u 20 u N N 20 10 10 10-2 10-1 100 101 102 103 104 105 10-2 10-1 100 101 102 103 104 105 M (M ) L (L ) clump clump ⊙ ⊙ Fwiigthur(ree1d0).anDdiwstritihbuotuito(nbsluoef)Hd2etceocltuiomnnindHenCsiOti+esa,nddu/sotrtHemCpNe.rWatuerneso,temtahsasteasllatnhdelpuamrainmoestietiressaoreffarlolmPGthCeCPsG(CblCaccka)taalnodgt(hPelaonnceksCcoolnlsaibsotirnagtioCnOe-tsealle.c2t0e1d5c)ores lent than the exterior. This is different from the situation in HCN would have higher CO excitation temperatures and, high-mass star-forming cores where the interior is of larger probably,highercolumndensitiescomparedtotheoneswith- turbulence(Wuetal.2010;Garayetal.2015). outdetection. AmongtheparametersgiveninthePGCCcat- The central velocities of HCO+ are more consistent with alog,wefoundthattheH columndensitywouldserveasan 2 thatofHCNthan13CO.Thevelocity-integratedintensitiesof indicatorofthepresenceofdensegasinPGCCs.Onthebasis HCO+ are well correlated with that of HCN. These features of the inspection of the parameters given in the PGCC cata- indicatethatgaseousHCO+ isbettercoupledwithHCNthan log,wesuggestthatthereareabout1000PGCCobjectsmay COanditsisotopologues. havesufficientreservoirofdensegastoformstars. The H column densities of sources with detection in 2 HCO+ or HCN have been deduced from spectra of C18O or 13CO,andareintherangeof(1.5−70.9)×1021 cm−2 witha This work is supported by the National Natural Science meanvalueof1.3×1022cm−2,significantlyhigherthanthose FoundationofChinathroughgrantsof11503035,11573036, 11373009, 11433008 and 11403040, the China Ministry of ofCO-selectedcoresinOrion, Taurus, Perseus, andCalifor- Science and Technology under State Key Development Pro- niacomplexes. The estimated fractional abundances of HCO+ (HCN) gramforBasicResearchthroughthegrantof2012CB821800, rangefrom3.0×10−12(1.3×10−11)to1.5×10−9(5.6×10−10) theInternationalS&TCooperationProgramofChinathrough withameanvalueof1.5×10−10 (1.4×10−10)whicharesig- the grand of 2010DFA02710, the Beijing Natural Science Foundationthroughthegrantof1144015. andtheYoungRe- nificantlylowerthanvaluesinlow-tohigh-massstar-forming searcherGrantofNationalAstronomicalObservatories,Chi- regions. The low abundances detected in this work may be neseAcademyofSciences. KWacknowledgessupportfrom attributedtobeamdilution. theESOfellowshipandgrantWA3628-1/1throughtheDFG The inspection of some parameters from CO observations suggests that CO-selected cores detected in HCO+ and/or priorityprogramme1573‘PhysicsoftheInterstellarMedium’ (http://www.ism-spp.de). We give our thanks to the staff at