Table Of Content

Inpreparation.RevisedVersion:February2,2016 PreprinttypesetusingLATEXstyleemulateapjv.04/20/08 INTERACTIONSOFTHEINFRAREDBUBBLEN4WITHTHESURROUNDINGS Hong-LiLiu1,2,8†,Jin-ZengLi1,YuefangWu3,Jing-HuaYuan1,TieLiu4,G.Dubner5,S.Paron5,6,M.E.Ortega5,SergioMolinari7,Maohai Huang1,AnnieZavagno8,ManashR.Samal8,Ya-FangHuang1,Si-JuZhang1,2 Inpreparation.RevisedVersion:February2,2016 ABSTRACT Thephysicalmechanismsthatinducethetransformationofacertainmassofgasinnewstarsarefarfrom beingwellunderstood. InfraredbubblesassociatedwithHiiregionshavebeenconsideredtobegoodsamples 6 forinvestigatingtriggeredstarformation. Inthispaperwereportontheinvestigationofthedustpropertiesof 1 theinfraredbubbleN4aroundtheHiiregionG11.898+0.747, analyzingitsinteractionwithitssurroundings 0 and star formation histories therein, with the aim of determining the possibility of star formation triggered 2 bytheexpansionofthebubble. UsingHerschelPACSandSPIREimageswithawidewavelengthcoverage, we reveal the dust properties over the entire bubble. Meanwhile, we are able to identify six dust clumps n surroundingthebubble,withameansizeof0.50pc,temperatureofabout22K,meancolumndensityof1.7 a ×1022 cm−2, mean volume density of about 4.4 ×104 cm−3, and a mean mass of 320 M . In addition, from J (cid:12) PAHemissionseenat8µm,free-freeemissiondetectedat20cmandaprobabilitydensityfunctioninspecial 1 regions, we could identify clear signatures of the influence of the Hii region on the surroundings. There are 3 hintsofstarformation,thoughfurtherinvestigationisrequiredtodemonstratethatN4isthetriggeringsource. ] Subjectheadings:ISM:bubbles-ISM:Hiiregion-stars: formation-stars: massive-ISM:individualobjects: N4 A G 1. INTRODUCTION implosion” (RDI, Bertoldi 1989; Lefloch & Lazareff 1994). h. Massive stars (of OB spectral type with masses greater IntheC&Cprocess, theultravioletradiationfromionizing p than ∼ 10M ) are thought to form in clusters within molec- sourcesproducesanionizationfront(IF),creatinganexpand- - ular cloud co(cid:12)mplexes. Intense radiative and mechanical out- ing Hii region. The supersonic expansion of the Hii region o putsfrommassivestarsstronglyaffecttheirparentmolecular drives the shock front (SF) in front of the IF. Finally, a shell r of circumstellar gas can be collected between the IF and the t cloudsintwooppositeways: bydisruptingmolecularclouds as and by triggering star formation. In effect, on the one hand, SF. In due time, the shell may become denser and collapse toformstars. Severalnumericalsimulations(e.g.,Hosokawa [ the natal molecular clouds can be accelerated beyond their 2 epsacnadpiengveHloiciirteiegsiobnys,feoeudtbflaocwksf,rostmellmaraswsiivnedss,taarnsdthsruopuegrhnoevxa- &expInauntdsiunkgaH20ii0r5e,g2io0n0s6;caDnalterigetgearl.s2ta0r07fo)rhmaavteiosnugthgreosutegdhththaet v C & C process only if ambient molecular material is mas- explosions,leadingtothedisruptionofthecloudswhichcan 4 siveenough. Thisprocesshasbeententativelydetectedtobe stoptheeventualstar-formingprocesses. Ontheotherhand, 1 at work in several well-known Hii regions such as Sh 104 thisfeedbackiscapableofsweepingupthesurroundingsand 1 (Deharvengetal.2003),RCW79(Zavagnoetal.2006),and accumulatingthemintocondensationsgravitationallybound, 1 RCW 120 (Deharveng et al. 2009). By contrast, in the RDI inducingtheformationofnewgenerationsofstars. 0 model,theIFdrivestheSFintomolecularcloudssurrounding Over the past several decades, numerous studies have fo- 1. cused on star formation triggered in the environs of Hii re- the Hii region, stimulating the collapse of pre-existing sub- 0 gions(e.g.,Deharvengetal.2005,2006;Zavagnoetal.2007; critical clumps. Recent numerical simulations (e.g., Miao et 6 Deharveng et al. 2008, 2009; Ogura 2010; Elmegreen 2011; al.2006,2009;Bisbasetal.2011)havedemonstratedthatthe 1 RDImodelcansuccessfullyinterpretstarformationinbright- Kendrewetal.2012;Daleetal.2013;Samaletal.2014;Liu v: etal.2015;Daleetal.2015). Twodifferentmechanismshave rimmed clouds (BRCs). This prediction seems to have been confirmed by observations of BRCs (e.g., Reach et al. 2009; i been proposed as models of star formation triggered in the X surroundings of Hii regions: the “collect and collapse” pro- Liu et al. 2012) indicating that star formation concentrated alongthecentralaxisofthecloudsmightbetheconsequence r cess(C&C,Elmegreen&Lada1977)and“radiation-driven a oftheRDIprocess. 1National Astronomical Observatories, Chinese Academy of Sciences, Thompson et al. (2012) and Kendrew et al. (2012) specu- 20ADatunRoad,ChaoyangDistrict,Beijing100012,China lated that around 14%-30% of the massive young stellar ob- 2UniversityofChineseAcademyofSciences,100049Beijing,China jects (MYSOs) in the Milky Way might have been induced 3DepartmentofAstronomy,PekingUniversity,100871Beijing,China byexpandingbubbles/Hiiregionsonthebasisoftheassocia- 4Korea Astronomy and Space Science Institute 776, Daedeokdae-ro, tionofalargesampleofIRbubbles(Churchwelletal.2006, Yuseong-gu,Daejeon,RepublicofKorea305-348 2007) with Red MSX MYSOs (Urquhart et al. 2008). How- 51InstitutodeAstronom´ıayF´ısicadelEspacio(IAFE,CONICET-UBA), ever, the exact processes of triggered star formation are far CC67,Suc.28,1428BuenosAires,Argentina 62FADU-UniversidaddeBuenosAires, CiudadUniversitaria, Buenos frombeingunderstood. Itishopedthatdetailedstudiesofthe Aires,Argentina physicalinteractionofbubbles/Hiiregionstogetherwiththeir 7IstitutodiAstrofisicaePlanetologiaSpaziali–IAPS,IstitutoNazionale environswillleadtoabetterunderstandingofstarformation diAstrofisica–INAF,viaFossodelCavaliere100,00133Roma,Italy associatedwithbubbles/Hiiregions(Liuetal.2015). 8AixMarseilleUniversit,CNRS,LAM(Laboratoired’Astrophysiquede Far-infraredobservationscarriedoutbytheHerschelSpace Marseille)UMR7326,13388,Marseille,France Observatory enable us to get a better insight into the physi- †[email protected] 2 calpropertiesofthebubble/Hiiregion,withunprecedented resolutionsandsensitivitiesinthatspectralrange. Thanksto TABLE1 SummaryofKVNsingle-dishobservations thewidespreadwavelengthcoveragefrom70to500µm,the dust temperature and column density distributions of all of the entire bubbles/Hii regions can be readily revealed by fit- Line GνHz θaHrcPsBeWc η TKsys T(Krm)s kmδVs−1 tingspectralenergydistributions(SEDs)pixelbypixel.Addi- tionally,Herschelobservationscanrevealstarformationsur- H2O61,6−52,3 22.235077 120 0.46 80 0.05 0.21 rounding the system. The widespread coverage is adequate CHHC3OO+H(17-00,)7−61,6 8494..108689542766 3610..55 00..3488 215700 00..0087 00..1015 for better constraining the SED profiles of YSOs at earlier o-H2CO21,2−11,1 140.83952 23.4 0.33 360 0.09 0.04 phasesandmoreaccuratelyestimatingtheirphysicalparame- terssuchasstellarmassandluminosity. Moreover,Herschel tor package (CuTeX) catalog. They are encompassed within observations coupled with suitable molecular lines offering the 2.5 times radii of the bubble, of detection in the PACS kinetic and dynamical information are very useful tools for 70 µm band. Details about the catalog and the photometric understandingtheinteractionsofthebubbles/Hiiregionswith proceduresconsideredinCuTeXcanbefoundinMolinariet theirsurroundings. Thepresentpaperpresentsacomprehen- al.(2011). sivestudyofthebubbleN4(Churchwelletal.2006)usinga combinationofHerschelandmolecularlineobservations. 2.2. MolecularLineObservations The bubble N4 appears as an almost complete ring in the IR regime (see Figure 1), enclosing the Hii region Single-dishobservationsofmolecularlineswereperformed G11.898+0.747 (Lockman 1989). N4 is centered at α = withtheKoreanVLBINetwork(KVN)21-mradiotelescope 2000 18h08m52s.80,δ = −18◦16(cid:48)22(cid:48).(cid:48)8, with a radius of ∼ 2(cid:48), atYonseisite,Seoulon2014December28. Thefrontendof 2000 corresponding to 1.9±0.5 pc at a distance of 3.2±0.9 kpc the single-dish telescope is equipped with the four receivers (Churchwelletal.2006).Byanalyzingthespatialdistribution working at 22, 43, 86, and 129 GHz simultaneously. A dig- ofcandidateYSOsaroundN4,Watsonetal.(2010)foundno ital filter bank (DFB) with a 64 MHz bandwidth split into evidencefortriggeredstarformation. However,theyclaimed 4096 channels was adopted, resulting in velocity resolutions thatthepresenceoftriggeredstarformationcannotberuled of 0.21 km s−1, 0.11 km s−1, 0.05 km s−1, and 0.04 km s−1, out due to the lack of a complete sample of the YSO popu- respectively.Thepointingandtrackingaccuracieswerebetter lation as well as maps of the molecular gas and ionized gas than3(cid:48)(cid:48)(Kimetal.2011). intheenvirons. Incontrast,theexpandingmotionofN4un- Sincemappingobservationsarequitetime-consuming,we coveredbytheobservationsof J = 1−0of12CO,13CO,and onlychosesixdenseclumps,locatedontheborderofN4,to C18O carried out by Li et al. (2013) suggests a signature of make single-point observations. To investigate the dynami- possibletriggeredstarformation. Therefore,theissueofstar cal and kinetic properties (see Section 3.2 and 4.1) and star formationinN4meritsamorecomprehensiveinvestigation. formationactivityintheclumps,H2O(22GHz),CH3OH(44 The purpose of this study is to analyze interactions of the GHz), HCO+ (89 GHz), and o-H2CO (141 GHz) were ob- bubble N4 with its surroundings and star formation histories served simultaneously. The data were calibrated to antenna therein,andtoexplorethepossibilityoftriggeredstarforma- temperature (T∗) using the chopper wheel method. A sum- a tion. This paper is organized as follows: the Herschel and mary of observations is given in Table 1, including the half- molecularlineobservationstogetherwitharchivalIRandra- powerbeamwidth(θHPBW), thebeamefficiency(η), thetyp- diodataarepresentedinSection2,theresultsaredescribedin ical systematic temperature (Tsys), the rms noise level (Trms) Section3,thediscussionisarrangedinSection4,andfinally, andthevelocityresolution(δV). Thedatawereanalyzedand wegiveasummaryinSection5. visualized with the software GILDAS (Guilloteau & Lucas 2000). 2. OBSERVATIONSANDDATAREDUCTION 2.3. Archivaldata 2.1. HerschelObservations To study the physical characteristics of the bubble and its ThebubbleN4wasobservedaspartoftheHi-GAL10 sur- associatedYSOcandidates,auxiliarydataofimagesandpoint vey. Inthissurvey,thePhotodetectorArrayCamera&Spec- sources were complemented by the GLIMPSE (Benjamin et trometer(PACS,Poglitschetal.2010)at70and160µmand al. 2003), MIPSGAL (Carey et al. 2009) and WISE (Wright theSpectralandPhotometricImagingReceiver(SPIRE,Grif- etal.2010)surveys. TheimagesoftheSpitzerInfraredArray fin et al. 2010) at 250, 350, and 500 µm worked simulta- Camera(IRAC)at3.6,4.5,5.8,and8.0µm,togetherwiththe neously in the parallel photometric mode. The observations Multiband ImagingPhotometer for Spitzer (MIPS) at24 µm wererunintwoorthogonalscanningdirectionsatascanspeed were retrieved from the Spitzer Archive.11. The angular res- of 60 arcsec s−1. The measured angular resolutions of these olutions of the images in the IRAC bands are < 2(cid:48)(cid:48) and it is bandsare10(cid:48).(cid:48)7,13(cid:48).(cid:48)9,23(cid:48).(cid:48)9,31(cid:48).(cid:48)3,and43(cid:48).(cid:48).8(Traficanteetal. ∼6(cid:48)(cid:48).0intheMIPS24µmband.Inaddition,theWISEsurvey 2011),respectively,correspondingto0.12±0.05to0.49±0.21 has mapped the entire sky in four infrared bands centered at pcatthedistancetoN4. Thedetaileddescriptionsofthepre- 3.4,4.6,12,and22µm,witharesolutionof6(cid:48).(cid:48)1,6(cid:48).(cid:48)4,6(cid:48).(cid:48)5,and processingofthedatauptousablehigh-qualityimagescanbe 12(cid:48).(cid:48)0, respectively (Wright et al. 2010). Point sources were foundinTraficanteetal.(2011). obtainedfromtheAllWISESourceandMIPSGALCatalogs.12 To search for early stages of star formation in N4, we Bothcatalogscontainthephotometriesofpointsourcescross took 56 point sources from the Curvature Threshold Extrac- matchedwiththeTwoMicronAllSky(2MASS,Skrutskieet al.2006)survey. The20cmradiocontinuumimagewasused 10Hi-GAL,theHerschelinfraredGalacticPlaneSurvey,isanOpenTime KeyprojectontheHershcelSpaceObservatory(HSO)aimingtomapthe entireGalacticPlaneinfiveinfraredbands.Thissurveycoversa|b|<1◦wide 11http://sha.ipac.caltech.edu/applications/Spitzer/SHA/ stripoftheMilkyWayGalacticplaneinthelongituderange−60◦≤l≤60◦ 12http://irsa.ipac.caltech.edu/frontpage/ Liuetal. 3 the bubble. The corresponding backgrounds were estimated −18◦12′ 4.5µm/8.0µm/24µm to be 634 MJy sr−1 for 160 µm, 387 MJy sr−1 for 250 µm, 205 MJy sr−1 for 350 µm and 79 MJy sr−1 for 500 µm. By treatingT andN asfreeparameters,theSEDfittingwas dust H2 performedusingtheIDLprogramMPFITFUN14(Markwardt 14 ′ 2009). Inprinciple,the870µmimageshouldhavebeencon- sidered in the SED fitting. Although the 870 µm data con- 0) strain the dust temperature, as they are sensitive to colder 00 16′ components than Herschel data, they miss the bulk of ex- 2 tended emission, which is the major interest in our current J ( work. Thus, the image at 870 µm has been excluded in the c e D SEDfitting. 18′ The resulting Tdust and NH2 maps are shown in Figure 2, wheretheblackcontoursstandfortheemissionat24µmand thepurplecontourdelineatesalevelofN =9×1021 cm−2, H2 sculptingashellstructure. Emissiontracedby24µmmainly 20 comesfromhotdustwhichcanreachveryhightemperatures ′ after absorbing high-energy photons (e.g., Deharveng et al. 2010, references therein). The hot dust distribution shown 10s 09m00s 50s 18h08m40s inFigure2(a)indicatesadustheatingfromtheinteriortothe RA(J2000) edgeofN4(alsoseeFigure1),therefore,possibletemperature gradients along this direction would be expected. However, Fig.1.— Compositethree-colorimageofN4. TheIRAC4.5µm,8.0µm, the dust distributions in the interior and on the edge of N4 and24µmarecoloredinblue,green,andred,respectively. Theimageis centeredatα2000=18h08m52s.80,δ2000=−18◦16(cid:48)22(cid:48).(cid:48)8. appeartobesmooth. Giventheresolutionof6(cid:48)(cid:48)at24µm,the smoothnesscouldbeinterpretedasthetemperaturegradients fromtheMulti-ArrayGalacticPlaneImagingSurvey(MAG- onsmallscalesnotresolvedbytheangularresolutionof43(cid:48).(cid:48)8 PIS, Helfand et al. 2006) archive13 was used to analyze the ofthedusttemperaturemap. Additionally,wenotethatthere properties of the Hii region associated with N4, . Its angu- is a cold region of ∼ 22 K within the bubble, which may be larresolutionislessthan6(cid:48).(cid:48)0andthesensitivityisbetterthan notrealduetothelowfluxfoundthere(Andersonetal.2012). 0.15mJybeam−1. On the large scale, the temperature distribution nearly matches with 24 µm emission (see Figure 2(a)). That is, the 3. RESULTS temperatures of cold dust in the bubble direction are hotter 3.1. DustEmission thanthoseawaythebubble,suggestingthattheHiiregionen- closedbyN4hasbeenheatingitssurroundings. Onthesmall 3.1.1. DustTemperatureandColumnDensityDistributions scale, therearethreeregionsthatareespeciallywarmerthan Herschel observations with the widespread wavelength otherplacesintheT map.Thefirstregion,locatednearthe dust coverage can reveal the dust properties of the entire bubble. bubblecenter,itcanbeattributedtotheheatingofthecentral Todothis,aSEDpixel-by-pixelfittingwasperformedtoob- ionizing stars (see Section 3.4). The second warm region, tain the distributions of dust temperature (Tdust) and H2 col- with which some YSO candidates are associated (see Sec- umndensity(NH2). Assumingopticallythindustemission,a tion3.3),istothenorthofN4.Hence,thiswarmregioncould graybodyfunctionforasingletemperaturecanbeexpressed be a result of the heating from embedded protostars therein. asfollows: However,thethirdwarmregioninthenorthernborderofN4 I =κ (ν/ν )β B (T )µm N(H ), (1) hasnobright24µmcounterpartorpotentialassociatedYSO ν ν0 0 ν d H 2 candidate. Figure2(b)showsthatthecolumndensityofthis where I is the surface brightness, B (T ) is the blackbody regionislowerthanthatofitsneighbors. Thisdifferencesug- ν ν d function for the dust temperature of T , the mean molecular gests that this warm region may be heated premarily by the d weight µ is assumed to be 2.8 (e.g., Kauffmann et al. 2008; ionizingphotonsleakingfromtheHiiregion. Sadavoyetal.2013);andm isthemassofahydrogenatom. For the H column density distribution in Figure 2(b), a H 2 Thedustopacityperunitmassofbothdustandgasisdefined shellmorphologyisinscribedasseeninIRbands.Intheshell, asκ = κ (ν/ν )β,whereκ isassumedtobe0.1cm2 g−1 at thereisananti-correlationbetween N andT . Thismay ν ν0 0 ν0 H2 dust 1 THz (Beckwith et al. 1990) under a gas-to-dust mass ratio be attributed to a lower penetration of the external heating of100,andβisfixedto2,whichisastatisticalvaluefoundin fromtheHiiregionintodenseregions. Iftheshellisbounded alargesampleofHiiregions(Andersonetal.2012). by a level of 9×1021 cm−2 as shown in Figure 2(b), which The Herschel data of the four bands at 160, 250, 350, covers a region of ∼ 1.5 radius of the bubble, the mean N H2 and 500 µm were used for the SED fitting. Since emission columndensityandtotalmassoftheshellcanbeestimatedto at 70 µm may trace hotter components such as very small be1.1×1022 cm−2 and5.2×103 M ,respectively. Thetotal (cid:12) grains (VSGs) and warmer material heated by protostars, it massoftheneutralmaterialoftheentirebubbleisestimated was not considered in the SED fitting for a single tempera- tobe∼ 5.5×103 M . (cid:12) ture. Thefourimageswereconvolvedtothesameresolution 43(cid:48).(cid:48)8, and rebinned to the same pixel size 11(cid:48).(cid:48)5. To mini- 3.1.2. DenseClumps mizethecontributionfromthelineofsight,wesubtractedthe Sixdenseclumpswereidentifiedwithintheshellstructure backgroundsforeachimageusingareferenceareaawayfrom intheH columndensityimage. Thesesourcesweredecom- 2 13http://third.ucllnl.org/gps/ 14http://www.physics.wisc.edu/˜craigm/idl/fitting.html 4 DustTemperature(K) ColumnDensity(1022cm−2) 15.0 16.5 18.0 19.5 21.0 22.5 24.0 25.5 27.0 28.5 30.0 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 2.25 2.50 1022 3 × (a) (b) 2 1 −18◦12′ D D C C 0) 15′ E E 0 20 B B J ( ec F F D 18′ A A 21′ 1pc 1pc 12s 09m00s 48s 18h08m36s RA(J2000) Fig.2.—Mapsofthedusttemperature(a)andtheH2columndensity(b)ofN4. ThetwomapswereallbuiltontheSEDfittingpixelbypixel,centeredat α2000 =18h08m52s.80,δ2000 =−18◦16(cid:48)22(cid:48).(cid:48)8. ThecrossessymbolizesixdenseclumpsextractedfromtheH2columndensitymap. Thepurplecontourshowsa levelof9×1021cm−2,whichcoversaregionof∼1.5radiusregionofthebubble.ColumndensityPDFswithinthethreemagentadashedcirclesaretakeninto account(seeSection4.2). . posed visually from the image by ellipses bounded by the levelof9×1021 cm−2 enclosingthemajorityoffluxes. The sixclumpshavedifferentphysicalsizesandpeakcolumnden- sities(N ). Theiraveragetemperatures(< T >)were H2,peak dust obtainedfromthedusttemperatureimage. Weestimatedthe totalmass(gas+dust)andthevolumedensityforeachclump. Table 2 gives a summary of the aforementioned parameters, where the main source of error is the uncertainty of the distance to the bubble. The average size, dust temperature, column density, number density, and mass of the six clumps are∼ 0.5pc, ∼ 22K,∼ 1.7×1022 cm−2, ∼ 4.4×104 cm−3, and∼ 3.2×102 M ,respectively. Theaveragedensity> 104 (cid:12) cm−3isconsistentwiththedetectionsofdensemolecularlines withintheseclumps(seeSection3.2). 3.2. MolecularLineEmission Figure3presentsHCO+(1-0)ando-H2CO21,2−11,1spec- Fig.3.—HCO+ (1-0)ando-H2CO21,2−11,1 spectraofthesixclumps. traofthesixclumps. SincethemoleculesH OandCH OH Theblackdashedlinesindicateasystematicvelocityof24.7kms−1forthe 2 3 bubble.Thesystematicvelocityisdeterminedfromtheaveragespectraofthe wsioerneonfoHtCdeOte+c(te1d-0,)thaenidr sop-Hec2tCraOar2e1,2no−t1b1e,1diisspsltaroynegd.enEomugish- HofC1O2C+Oa,n1d3HC2OC,OanldinCes18oOft(h1e-0si)xocblsuemrvpast,iownhsic(Lhiisetcaoli.n2c0id1e3n)t.wTimthbrtheperreesseunlttss to be detected in all clumps. Their detections indicate high themain-beamtemperature.Thenameofeachclumpisnotedinthetopleft densities for the six clumps, since the critical densities of ofeachpanel. both lines are 4.5×104 cm−3, and 1.3×105 cm−3 for a kinetic temperature of 20 K (Shirley 2015). Using the soft- vealed the same velocity difference between the SE and the wareGILDAS,weretrievedtheobservedparametersfromthe NW (see Figure 3 of Li et al. 2013). Due to the lack of spectraofthesixclumps,includingthepeakvelocity(VLSR), blueshifted profiles in the front of the bubble and redshifted line width (FWHM) and the main-be(cid:82)am temperature (Tmb), ones in the back cap of the bubble in the CO observations, andthevelocity-integratedintensity( T dV)ofthediffer- Lietal.(2013)suggestedthatthebubbleN4maybeexpand- mb entspectra(seeTable3). ingalongtheSE-NWdirectionwithaninclinationrelativeto It is interesting to analyze the centroid velocity variances the sky plane. This scenario is compatible with Beaumont of the six clumps. The spectra of HCO+ and o-H CO obvi- & Williams (2010) interpretations of molecular gas studies 2 ously show the velocities blueshifted to a systematic veloc- aroundanumberofIRbubbles,wheretheyconcludethatthe ity of 24.7 km s−1in the southeastern (SE) part of the bub- SFs driven by massive stars tend to create rings. Therefore, ble (i.e., Clumps A and B) and the redshifted velocities in the shell of dust and gas in the bubble N4 is presumably as- thenorthwestern(NW)part(i.e.,ClumpsDandE).Thetwo sembledbytheexpandingHiiregion. remaining clumps show no evident velocity shifts. Observa- Assuming that the observed o-H CO line is optically thin 2 tionsofJ=1-0of12CO,13CO,andC18Oalsohavealreadyre- and in local thermodynamic equilibrium (LTE), we estimate Liuetal. 5 TABLE2 Characteristicsofsixdustclumps Name (hh:mRm.A:s.s.ss) (dd:Dmemcl:.ss.s) (arrmcasjeocr) (arrmcisneorc) r(epfcf)a <T(dKu)st> (1N02H22c,pmea−k2) (104ncHm2−3) M(1(0ga2s+Mdu(cid:12)s)t) A 18:08:53.14 -18:18:25.5 24.3 44.6 0.4±0.1 22.3±0.2 1.0±0.1 4.0±1.4 0.8±0.5 B 18:09:00.66 -18:16:30.0 36.7 51.9 0.6±0.2 22.1±0.5 1.9±0.3 5.0±2.1 3.5±2.5 C 18:08:55.59 -18:14:44.5 34.9 49.9 0.6±0.2 23.1±0.9 2.8±0.5 7.8±3.6 4.5±3.4 D 18:08:48.57 -18:14:23.1 34.1 40.8 0.5±0.1 22.3±0.7 1.6±0.2 5.2±2.1 1.9±1.3 E 18:08:45.39 -18:15:34.6 35.8 41.0 0.5±0.2 21.7±0.7 1.5±0.1 4.7±1.8 1.9±1.3 F 18:08:45.18 -18:17:40.7 37.4 35.2 0.5±0.1 23.6±0.7 1.3±0.1 4.4±1.6 1.4±0.9 areffisthedeconvolvedeffectiveradius. TABLE3 Parametersofmolecularlinesofthesixdustclumps HCO+ o-H2CO (cid:82) (cid:82) Name VLSR FWHM Tmb TmbdV VLSR FWHM Tmb TmbdV (kms−1) (kms−1) (K) (Kkms−1) (kms−1) (kms−1) (K) (Kkms−1) A 23.61±0.06 2.62±0.11 1.30±0.12 3.62±0.16 23.53±0.25 2.30±0.52 0.40±0.11 0.99±0.19 B 23.58±0.06 2.03±0.08 1.84±0.12 4.01±0.15 23.88±0.08 1.91±0.19 0.97±0.11 1.98±0.17 C-I 23.56±0.21 2.54±0.21 0.99±0.11 3.84±0.31 24.51±0.38 4.32±0.70 0.53±0.12 2.42±0.41 C-II 26.68±0.06 1.79±0.21 2.91±0.11 4.74±0.26 26.87±0.04 1.16±0.11 1.78±0.12 2.19±0.29 D 25.00±0.06 3.06±0.10 1.47±0.11 4.77±0.17 25.11±0.12 2.00±0.35 0.67±0.11 1.43±0.19 E 25.28±0.06 2.46±0.06 2.33±0.11 6.12±0.17 25.47±0.09 2.23±0.28 0.96±0.12 2.26±0.21 F 24.86±0.09 4.02±0.24 0.85±0.12 3.64±0.23 25.36±0.20 2.91±0.51 0.49±0.11 1.53±0.22 the column density of this molecular species toward each TABLE4 dense clump presented in Section 3.1.2 using (Goldsmith & o-H2COLTEcolumndensitiesandabundances. Langer1999): Clump N(o-H2CO) X 8πkν2 Q E (cid:90) (×1012cm−2) (10−10) N = hc3A grot exp(T u ) Tmbdv, (2) A 6.26 6.2 ul u rot B 12.66 6.6 C 28.20 9.9 whereA =5.4×10−5s−1,E =21.9K,andg =15.Assum- D 9.05 5.6 ul u u ingthatthedustandthegasarecoupledatthesametemper- E 14.61 9.7 ature, in LTE conditions we can approximate T = T = F 9.13 7.0 dust kin T . Thus,forT weusedtheT valueobtainedforeach rot rot dust clump (see Section 3.1.2). Given a temperature of ∼ 22 K towardPDRs. averagedoverthesixclumps, theo-H COpartitionfunction 2 Q wasestimatedtobe∼ 50bytheextrapolationtothere- 3.3. YSOsAssociatedwithN4 rot lationbetween Q andT fromtheCDMSCatalog.15 The To understand more about star formation histories associ- rot rot obtainedcolumndensityvaluesforeachclumparepresented atedwiththebubble,candidateYSOsassociatedwithN4have inTable4,whichalsoincludestheo-H COabundances(X = beeninitiallyidentifiedusingtheonlineSEDfittingtool16 of 2 N(o-H CO)/N(H )). The abundances were estimated using Robitaille et al. (2006, 2007) and then further demonstrated 2 2 theH columndensitiespresentedinSection3.1.2. Wenote andclassifiedwithcolor-colordiagrams.TheSEDfittingtool 2 that o-H CO emission is generally optically thick, therefore, invokes a grid of20,000 two-dimensional Monte Carlo radi- 2 the assumption of optically thin emission of o-H CO results ation transfer models (e.g., Whitney et al. 2004). With 10 2 inunderestimatedcolumndensityandabundance. viewing angles for each model, there are actually 200,000 Theobtainedo-H COcolumndensitiesandabundancesare YSO SED models. This tool works as a linear regression 2 quitesimilartothoseobtainedtowardotherphoto-dominated methodtofitthesemodelstothemulti-wavelengthphotome- regions (PDRs), such as the Horsehead PDR (Guzmań et al. try measurements of a given source. The goodness/badness 2011), and the Orion Bar (van der Wiel et al. 2009, 2010). of each fit could be quantified by a specified value of the Considering that the dust grains are cold (T ∼ 22 K) in best χ2 (χ2 ), whereby YSO candidates could be robustly dust best the dense clumps around the N4 bubble and given the abun- distinguishedfromreddenedphotospheresofmain-sequence dancespresentedinTable4,wesuggestthatinthisregionthe and giant stars since YSOs require a thermal emission com- formaldehyde may be formed mainly in the gas phase with ponent to reproduce the shapes of their mid-IR excesses. In aprobablecontributionofphoto-desorptionofthegrainman- addition, this method can make use of any available data to tlesaswasfoundintheHorseheadPDR(Guzmańetal.2011). constrain the shape of SED profiles as well as possible (see Observations of other o-H CO and p-H CO lines toward AppendixA).However,wehavetoacknowledgethatthedisk 2 2 this region would be very useful to give an important obser- and stellar parameters inferred from the SED models can be vational probe of the gas phase and grain surface chemistry very uncertain (Offner et al. 2012, and T.P. Robitaille 2015, 15www.astro.uni-koeln.de/cdms/catalog 16http://caravan.astro.wisc.edu/protostars/ 6 privatecommunication). Therefore,iftheseuncertainparam- eters are used to classify the identified YSO candidates with TABLE5 ParametersofYSOcandidates the scheme of YSO classification of Robitaille et al. (2006), the classification results must be unreliable. To avoid this No. Name RA Decl. N χ2 Class uncertainty, the color selection schemes of Gutermuth et al. best (2009) and Koenig et al. (2012) have been adopted to group Y1 J180831.89-181639.1 272.133 -18.278 7 13.4 II Y2 J180832.49-181616.6 272.135 -18.271 7 35.7 TD the YSO candidates identified by the SED fitting into three Y3 J180834.88-181751.0 272.145 -18.298 7 18.3 TD “Classes”: ClassI,ClassII,andTransitionDisk(TD)YSOs. Y4 J180835.68-181814.8 272.149 -18.304 7 11.0 TD ClassIobjectsaredeeplyembeddedprotostarswithadomi- Y5 J180839.36-181813.8 272.164 -18.304 7 20.9 TD nant infalling envelope, Class II objects are surrounded by a Y6 J180839.71-181841.5 272.165 -18.312 7 47.7 TD Y7 J180840.11-181902.3 272.167 -18.317 7 21.6 TD substantial accreting disk (Lada 1987), and TD objects are Y8 J180841.62-181600.0 272.173 -18.267 7 12.9 TD more evolved protostars where the inner parts of the disks Y9 J180842.08-181457.5 272.175 -18.249 7 8.2 TD havebeenclearedbyphotoevaporationofthecentralstarsor Y10 J180842.19-181329.1 272.176 -18.225 7 34.6 TD byplanet-formingprocesses(Gutermuthetal.2009;Yuanet Y11 J180842.69-181930.1 272.178 -18.325 7 34.0 TD Y12 J180843.58-181441.8 272.182 -18.245 7 38.3 TD al. 2014; Paron et al. 2015). We believe that using the in- Y13 J180844.11-181456.4 272.184 -18.249 7 30.4 TD frared color schemes is a better way to categorize the YSO Y14 J180844.15-181401.9 272.184 -18.234 7 43.4 TD candidates because the infrared color schemes are a power- Y15 J180844.97-181457.3 272.187 -18.249 7 40.2 TD fulproxyformeasuringtheexcessemission(e.g.,Allenetal. Y16 J180845.97-181612.4 272.192 -18.270 8 32.2 II Y17 J180846.14-181844.0 272.192 -18.312 7 32.0 TD 2004;Gutermuthetal.2008,2009;Koenigetal.2012).Apart Y18 J180847.47-181416.2 272.198 -18.238 7 22.3 TD from this, the candidate YSOs, which have been singled out Y19 J180847.65-181450.0 272.199 -18.247 7 32.9 TD by the SED fitting, can be further confirmed by such color Y20 J180847.67-181441.7 272.199 -18.245 7 12.6 TD Y21 J180848.33-181357.2 272.201 -18.233 7 33.9 TD schemes. Y22 J180849.08-181441.8 272.205 -18.245 4 0.2 I SixtypotentialYSOswithin5(cid:48)fromthebubblecenterhave Y23 J180849.78-181340.5 272.207 -18.228 7 6.5 TD been identified and classified. Of these, there are 8 Class I, Y24 J180850.19-181816.2 272.209 -18.305 7 27.3 II 12ClassII,and40TDobjects. Theclassificationresultsare Y25 J180850.83-181836.6 272.212 -18.310 7 14.0 II Y26 J180851.55-181831.9 272.215 -18.309 7 7.6 TD listedinTable5andthedetailedprocessescanbefoundinthe Y27 J180852.42-181744.2 272.218 -18.296 7 22.5 II Appendix A. We note that the resulting 60 YSO candidates Y28 J180852.59-181356.6 272.219 -18.232 7 7.4 II areincompletebutrobust(seetheAppendixA).Despitethis Y29 J180852.97-181336.2 272.221 -18.227 7 43.8 TD incompleteness,theresultingpopulationwithrobustidentifi- Y30 J180853.04-181742.2 272.221 -18.295 8 5.5 I Y31 J180853.35-181130.9 272.222 -18.192 7 41.9 TD cations is enough for simply learning about the distributions Y32 J180853.65-181208.4 272.224 -18.202 7 29.8 TD oftheYSOsassociatedwithN4. Y33 J180853.94-181422.4 272.225 -18.240 5 0.3 I Figure4showsthespatialdistributionofthe60YSOcan- Y34 J180854.47-181406.5 272.227 -18.235 4 0.1 I didates. Onecanseethatthemajority(∼ 82%)ofYSOsare Y35 J180854.69-181230.5 272.228 -18.208 7 39.4 TD Y36 J180854.87-181409.5 272.229 -18.236 4 0.1 I spatially correlated with the PDRs as traced by 8 µm emis- Y37 J180855.80-181136.3 272.233 -18.193 7 16.7 TD sion(seeSection3.4),whichindicatesthegoodspatialasso- Y38 J180856.26-181505.7 272.234 -18.252 7 26.5 II ciation of these potential YSOs with the bubble N4 and the Y39 J180856.64-181420.4 272.236 -18.239 7 12.4 II strong impacts that the enclosed Hii region is having on the Y40 J180857.00-181354.7 272.238 -18.232 4 0.1 II Y41 J180857.48-181853.3 272.240 -18.315 7 25.0 TD surroundingstarformation. Inaddition, thereisanoverden- Y42 J180858.07-181506.0 272.242 -18.252 7 12.8 TD sityofthenumberofYSOcandidatesresidingonthevergeof Y43 J180858.12-181807.2 272.242 -18.302 7 26.4 TD the bubble. This is also demonstrated by the statistics of all Y44 J180858.75-181806.5 272.245 -18.302 7 22.4 TD Y45 J180858.77-181629.7 272.245 -18.275 11 25.1 I candidatesasafunctionofthenormalizedbubbleradius(see Y46 J180859.32-181316.1 272.247 -18.221 7 9.8 II Figure5). Thestatisticssuggestthatalargenumber(∼40%) Y47 J180859.44-181332.8 272.248 -18.226 7 0.9 I ofYSOcandidatesislocatedwithin1.0−1.5timestheradius Y48 J180859.47-181855.8 272.248 -18.316 7 14.1 TD ofthebubble. SuchspatialdistributionofYSOsinN4agrees Y49 J180900.26-181347.5 272.251 -18.230 7 11.3 TD Y50 J180900.28-181921.8 272.251 -18.323 7 31.4 TD wellwiththestatisticalresultsbasedonthestudyoftheasso- Y51 J180900.48-181903.2 272.252 -18.318 7 28.2 TD ciationofalargesampleofYSOswithIRbubbles(Kendrew Y52 J180902.92-181316.2 272.262 -18.221 7 41.9 TD etal.2012;Thompsonetal.2012). Y53 J180904.56-181638.3 272.269 -18.277 7 6.6 TD Y54 J180904.95-181806.2 272.271 -18.302 7 23.3 TD 3.4. PropertiesoftheIonizedRegion Y55 J180905.60-181621.4 272.273 -18.273 8 2.1 II Y56 J180906.19-181639.1 272.276 -18.278 7 0.9 TD Figure 4 (a) presents the 8 µm image overlaid with radia- Y57 J180906.60-181847.9 272.278 -18.313 7 19.7 TD tionat20cm. The8µmemissionpredominantlycomesfrom Y58 J180908.60-181706.1 272.286 -18.285 7 39.2 TD two prominent polycyclic aromatic hydrocarbons (PAHs) at Y59 MG011.8545+00.7327 272.216 -18.313 10 8.7 I Y60 MG011.9455+00.7481 272.248 -18.226 9 6.6 II 7.7 µm and 8.6 µm. They are good tracers of PDRs delin- eating the IF created by massive stars. The 20 cm radiation NotesCol.1isthenumberofthesources,Col.2isthesourcenameretrieveddirectlyfrom thearchivedcatalog,Cols.3-4arethecoordinates,Col.5givesthecountofdatapointsused representsfree-freecontinuumemissionfromionizedgas.As intheSEDfitting,Col.6providesthebestsquareoftheresultingSEDs,andCol.7the showninFigure4(a), mostoftheionizedgasissurrounded classification. bytheringofPDRs,showingthestrongeffectsoftheHiire- gion on its surroundings. Additionally, the ring of PDRs is cm free-free continuum emission. Churchwell et al. (1978); also spatially well correlated with the 250 µm emission (see Wink et al. (1983); Quireza et al. (2006) demonstrated the Figure4(b)), whichissensibletocolddustinthermalequi- existence of a relationship between the Hii region electron librium.Assuch,thecolddustwrappingionizedgassuggests temperature,T ,andtheGalactocentricdistance,R . Ade- e Gal a positive association of the bubble N4 with the Hii region, tailedanalysisofasampleof76Hiiregionswiththehighest enhancingtheirintensivemutualinteraction. qualitydatasuggestedthattheelectrontemperaturealongthe SeveralpropertiesoftheHiiregioncanbederivedfrom20 Galactic disk decreases approximately as a function of R Gal Liuetal. 7 (a) -37 -31 8µm 20 -32 -35 -52 -46 15 -4--96407 --4309-3-3-634-3-2298 -23-2-118 -14-10 mber -22--2109 -1-51-132-9 Nu10 -42-38 -8 -55 -16 -2 5 -45 -56-53 -1 -58 -3-207 -3 00.5 1.0 1.5 2.0 2.5 -54 -4-443 -24 -5 -4 Distance to the bubble's center -57 -51-48-41 --52-9625 -17 -7-6 aFniggu.l5ar.—disPtalontcoefftrhoemntuhmebceerncteoruonftsthoefbthueb6b0leYNS4O.Tchaneddiidsatatensceasisanfuonrmctaiolinzeodf -50 bythebubbleradius.ErrorbarsaredeterminedviaPoissonstatistics. -11 matedfollowingKurtzetal.(1994): θ D ν T S 1pc n =2.878×104[( )−3( )−1( )0.1( e)0.35( ν)]0.5cm−3, e arcsec kpc GHz K Jy (4) (b) 4.5µm 4 M = πr3 n m , (5) ion 3 Hii e p T ν S D N =7.588×1048( e)−0.5( )0.1( ν)( )2phs−1, Ly K GHz Jy kpc (6) whereS istheintegratedfluxdensityatthespecialfrequency ν νovertheangularsizeθ; r istheradiusoftheHiiregion, Hii and m is the proton mass. If the θ and r are considered p Hii to be approximately equal to those of the bubble, a total of 32.8±0.01Jyat20cmmeasuredbyintegratingoverthe5σ contour(seeFigure4)yieldsanestimatedn , M ,andN e ion Lym of75±15cm−2,50±38M ,and(1.5±1.1)×1048phs−1,re- (cid:12) spectively. Theerrorsoftheseparametersresultmainlyfrom theuncertaintyinthedistanceofN4. AccordingtoMartinset al.(2005),theestimatedN (Log(N )(cid:39) 48.18)indicates Lym Lym thatamainO8.5V-O9Vstarwasresponsiblefortheionization ofN4. Thesearelowerlimitsgiventhatanyionizingphotons absorbed by dust or moving away from the bubble were not 1pc accountedfor. The central exciting stars of N4 may be located near the very bright 24 µm emission inside the bubble. Based on the Fig.4.— (a) Image at 8 µm (grayscale) overlaid with free-free contin- statisticalanalysisforthesevenwell-definedbubblesinclud- uumemissionat20cm(contours). Theyellowcontoursstartfrom5σ(i.e., ingN4,Deharvengetal.(2010)arguedthatthehotdustgrains 1σ=0.15mJyBeam−1)withastepof8σ. ClassI,ClassII,andtransition seenat24µminsidetheionizedregioncanbeheatedbyab- diskYSOcandidatesaresymbolizedbysquare,triangle,andcirclesymbols, sorptionoftheLymancontinuumphotonswhicharemorenu- respectively.Thedashedrectangleisaselectedregionforaclose-upviewof 4.5µmemission. (b)Imageat4.5µm(colorscale)overlaidwithdustemis- merous near exciting stars. This indicates that the ionizing sionat250µm(contours).Thewhitecontoursstartfrom18JyBeam−1with stars of N4 may reside near the very bright 24 µm emission astepof5.5JyBeam−1. Theellipsesdepictthedustclumpsextractedin inside the bubble (see Figures 1 and 2(a)). This explanation Section3.1.2,andthecrosssymbolsrepresenttheirpeakpositions. Ascale canbefurthersupportedbythepositiveassociationofintense barof1pcisshownonthebottomleft. free-freecontinuumradiationfromionizedgaswiththebright (Quirezaetal.2006): 24µmemission(seeFigures2(a)and4(a)). Observationsof spectroscopy of ionizing stars (Martins et al. 2010) are ex- Te =(5780±350)+(287±46)RGal. (3) pectedtorevealtheirnatures. WithanestimatedGalactocentricdistanceof5.5kpctoN4,it 4. DISCUSSION yieldsanelectrontemperatureof7400±600K.IftheHiire- 4.1. StarFormationinClumps gionreachestheequilibriumatthistemperature,theelectron density,n ,themassofionizedgas,M ,andtheLymanpho- The good agreement of the PDRs with cold dust emission e ion tons per second, N , from the exciting star(s), can be esti- andtheassociatedYSOcandidatesthereinaregoodindicators Lym 8 ofthestrongimpactoftheHiiregiononitssurroundingsand on star formation processes. In what follows, we investigate 103 (rMaMwc⊙ri)d=at8a70.0×(prc)1.33 the six identified dust clumps in a context of star formation C withinthem. B In the case of the Aquila Rift complex, a column density ttihinornNeso4hf,otplhdreeosmfteNellaHan2rcc>oo7rlue×sm1(nA02dn1edcnrmesí−ety2tawolf.a2s10.e41st1×i)m.1aF0to2e2rdctfhmoer−st2ihxseucfglougrmmespatss- ()MMcore⊙ A DEF 102 that they may be capable of forming stars. A mass-size re- lationshipisgenerallyadoptedtopredictwhethertheclumps will form low-mass stars or high-mass stars (Kauffmann & Pillai2010).Basedonthestatisticalanalysisofnearbyclouds without high-mass star formation (i.e., Pipe Nebula, Taurus, 100 Perseus,andOphiuchus)andknownsampleswithhigh-mass reff(pc) starformationsuchasthoseofBeutheretal.(2002);Mueller Fig.6.—Mass-sizerelationshipofthesixclumps. Thedashedlinerepre- etal.(2002);Hilletal.(2005),andMotteetal.(2007),alimit- sentsanempiricalrelationofm(r)=870M(cid:12)(r/pc)1.33,whichisexpectedto ingmass-sizerelationofm(r)≥870M (r/pc)1.33wasfound beasathresholdforforminghigh-massstars(Kauffmann&Pillai2010). (cid:12) to be an approximate threshold for high-mass star formation (Kauffmann & Pillai 2010). Figure 6 shows the mass versus extended4.5µmemissionarethoughttobemassiveYSOout- size relation for the six clumps in N4. Given uncertainties flowcandidates(Cyganowskietal.2008). Figure4(b)shows ofthetwoparameters,thetwoclumps(ClumpsBandC)are Spitzer 4.5 µm emission overlaid with cold dust radiation at abovethethreshold,indicatingthattheymaybeformingmas- 250 µm. One can see that six clumps except for Clumps C sivestars,whereastheotherfourclumpsbelowthethreshold and D peak at dark 4.5 µm emission, and no brighter ex- maybeinclinedtoformlow-massstars. Indeed,severalYSO tended 4.5 µm structures with respect to their backgrounds candidatesareclosetoallclumpsbutClumpFandsixofthese aredetectedforClumpsCandD,indicatingthattheremaybe (Y22, Y33, Y34, Y36, Y45andY59, seeFigure4)areclas- either no active ongoing star formation or deeply embedded sified as Class I YSOs, however, there is no YSO candidate protostarsatthecentersoftheseclumps. located at the peak positions of the six clumps. These facts DuetotheaforementionedlackoftheH O,ClassICH OH suggestthattheinteriorsofsixclumpsmaybeintheprocess 2 3 maser detections, and other signs of ongoing star formation, of forming stars but without active ongoing star-forming ac- it is difficult to confirm that the asymmetric red profiles are tivities. produced by outflow motions. The fact that Clump C is lo- Noclearevidenceforongoingstarformationhasalsobeen cated close to the strong ionization zone implies that it is demonstrated in the observations of the four molecular lines fullyexposedtointensefeedbackfromthemainO8.5V-O9V towardthepeakpositionsofthesixclumps.The22GHzH O 2 star such as the compression of the ionized region and stel- and 44 GHz Class I CH OH masers are known to occur in low-and/orhigh-masssta3rformationregions(e.g.,Forster& lar winds, forming the different two different velocity com- ponentsandleadingtotheasymmetriclineprofiles. Alterna- Caswell1999;Kurtzetal.2004). Thesetwotypesofmasers tively, this type of line profile may originate from the clump are generally thought to be associated with molecular out- rotation. Hence,itisworthwhiletodeterminethemechanism flows(e.g.,Codellaetal.2004;Kurtzetal.2004). Optically thicklinesofHCO+ andH COarewidelyusedtosearchfor responsiblefortheasymmetriclineprofileinClumpCusing 2 amapobservationwithhigherangularresolution. asymmetric line profiles indicative of infall or outflow motions (e.g., Wu et al. 2007; Chen et al. 2010, and references 4.2. FeedbackFromMassiveStars therein). Therefore, thesefourmolecularlinesareanimpor- tantsignpostofongoingstarformation. InthebubbleN4, it Massivestarsareexpectedtoaffecttheirsurroundings. In- turns out that no detection of the 22 GHz H O and 44 GHz side the bubble, the hot dust traced by 24 µm emission has 2 Class I CH OH masers has been obtained toward the peak been presumably heated by absorption of Lyman continuum 3 positionsofallclumps,andtheasymmetricredprofiles(i.e., photons whose free-free radiation is detected through radio adoublepeakspectralprofilewithredpeakhigherthanblue continuumemissionat20cm. Attheedgeofthebubble,the peak)havebeendetectedonlyinClumpCbytheHCO+ and shell of cold dust emission is associated with the PDR that H COlines(seeFigure3). delimits the Hii region, indicating the bright illumination of 2 As a result, no scenario of ongoing star formation is con- the far UV photons created by massive stars. In the north- firmedinthecenterofanyofthesixclumpsexceptforClump east of the bubble, brighter PAH emission is seen extending C. On the one hand, this result may be a consequence of beyondtheIF.ThiscanbeattributedtofarUVphotonsleak- the poor angular resolution of the present observations (23- ingfromtheHiiregionduetosmall-scaleinhomogeneitiesin 120(cid:48)(cid:48), corresponding to 0.26-1.35 pc). On the other hand, theIFandinthesurroundingmedium(Zavagnoetal.2007). it can result from the incompleteness of the current sample Likewise, the southwestern filament with bright PAH emis- of YSO candidates. Therefore, we have correlated the six sion can also be attributed to leaking radiation. All of the clumps with the YSOs in the RMS survey database (Lums- aboveobservedfactsdemonstratethestronginfluenceofthe denetal.2013),thesourcesintheMethanolMulti-Beamsur- expanding Hii region on the surroundings. Therefore, the vey (Green et al. 2010), and the compact Hii region catalog YSO candidates associated with them must suffer the phys- in the CORNISH survey (Purcell et al. 2013). It turns that icalactionoftheHiiregion. Moreover,8µmemissionshows thereisonlyonesourcefoundintheRMSsurvey,whichcor- concave-convex characteristics, indicating that the IF is lo- responds to the YSO candidate Y45 (see Figure 4). In addi- cally distorted. The outline of the cold dust emission, and tion,extendedgreenobjects(EGOs)identifiedbasedontheir likewise that of the IF, shows that this type of distortion oc- Liuetal. 9 Circle 1 TABLE6 10-1 ParametersofthePDFfitting Region p0 µ0 σ0 p1 µ1 σ1 1 0.035 -0.091 0.200 0.006 0.430 0.238 10-2 23 00..003450 --00..007881 00..128010 00..000053 00..348838 00..227211 bability Pro cursintheadjacentclouds. Inthiscasesuchdistortioncould beattributedtocompressionsoftheexpandingIF.Suchcom- 10-3 pressioncanbecharacterizedbytheprobabilitydensityfunc- tion(PDF)ofthecolumndensity. In turbulent simulations (Tremblin et al. 2012a,b), it is ar- gued that when the ionized gas pressure overweighs the ram 10-4 −0.5 0.0 0.5 1.0 1.5 pressure of the turbulence, the ionization compression could ln(NH2/NH2) producePDFformswithtwolognormaldistributions: Circle 2     10-1 p(η)= (cid:113)p0 exp−(η2−σ2µ0)2+(cid:113)p1 exp−(η2−σ2µ1)2, 2πσ2 0 2πσ2 1 0 1 (7) 10-2 wdihspereersηio=nolnf(eNa/chN¯)c,opmi,pµoni eanntd,rσeispareecttihveeliyn.teN¯grraelp,rmeseeannt,satnhde bability mean column density over a large region. The first lognor- Pro malformatlowdensitiesisgenerallythoughttobearesultof 10-3 isothermal supersonic turbulence (Vazquez-Semadeni 1994; Va´zquez-Semadeni et al. 2008; Federrath et al. 2008, 2010). The secondlognormal distributionin thePDF athigh densi- tiesisbelievedtobecausedbythecompressionfromtheion- 10-4 izedgaspressure(Tremblinetal.2012a,b,2014). Currently, −0.5 0.0 0.5 1.0 1.5 this sort of PDF form has been observed in several massive ln(NH2/NH2) star-forming regions, and interpreted as ionization compres- Circle 3 10-1 sions(e.g.,Schneideretal.2012;Tremblinetal.2014). Therefore, the exploration of the PDF form in the bubble N4 can help to improve our understanding of the influences of the Hii region on the surroundings. The column density P3,DsFeeisFiingvuerseti2g)atceodveorvienrgtthhreeeiornegizieodnsga(is.ea.,ndCitrhcelemsa1j,o2r saunrd- bability10-2 roundings. The three regions are concentric with an equal Pro separationof0.01degree. Figure7displaysthecolumnden- sity PDFs toward the three regions well fitted by the sum 10-3 of two lognormal distributions. The mean column density N¯ = 8.0 × 1021 cm−2 is averaged over the region of Cir- cle 3. The fitting parameters can be found in Table 6. As shown in Figure 7, all the column density PDFs of the three 10-4 −0.5 0.0 0.5 1.0 1.5 regionsshowthesecondlognormalforms. Giventheassoci- ln(NH2/NH2) abteiocnauosfetdhebHy iciormegpiroensswioitnhothfeiobnuibzbeldegNa4s,.thWeseenfootremtshmatigthhet reFpirge.s7e.n—tsmCoealunmthneddeennssiittyy.PTDhFesboevsetrlothgenotarmrgaeltPreDgFiofints((ssoeleidFciguurvree)2i)s.dNoHne2 regions smaller than the Circle 1 do not fit two lognormal bythesumoftwolognormaldistributions(dashedcurves,seeEq. 7). Error barsaredeterminedviaPoissonstatistics. distributions well. This may be due to the fact that in this region there are not enough sample points. Furthermore, we observethattheintegralofthesecondlognormalcomponent (p )decreasesastheradiusofthetargetregionincreases. As 1 suggested by Tremblin et al. (2014), this trend could be due tothefactthatthelargertheregionis,thelessimportantthe amplitudeofthecompressedlognormalbecomessincemore thecontextoftheC&Cprocess,suchanimpactmaystimu- unperturbedgasisaddedtothedistribution,indicativeofless latetheformationofanewgenerationofstars. Totesttheoc- significance of compression of the ionized gas in larger re- currenceofthisprocess,thedynamicaltimeoftheHiiregion gions. canbecomparedwiththefragmentationtimeofthesurround- ingmolecularclouds(Deharvengetal.2003,2005;Zavagno 4.3. CollectandCollapseProcess etal.2006,2007,2010). The association of collected molecular gas with PDR re- Accordingtothewell-knownexpansionlaw(Spitzer1978; gions and the compression from ionized gas confirm the Dyson & Williams 1980), if an Hii region evolves in an ho- stronginfluenceoftheHiiregionontheadjacentmedium. In mogeneous molecular cloud, the dynamical age (t ) of the dyn 10 Hiiregioncanbeexpressedasfollows: out of the six may be massive enough to form high- mass stars, while the remaining could form low-mass RSt=(3NLy/4πn2H,0αB)1/3, stars. Atthepresentsensitivityandangularresolution, 4 R R theobservationsofthefourmolecularlinestowardthe tdyn=7 CSt [(RIF)7/4−1], (8) six clumps did not reveal a clear evidence of ongoing Hii St starformation. where R is the radius of the Stro¨mgren sphere, n is the St 0 initial particle density of ambient neutral gas, αB = 2.6 × 3. The Hii region associated with the bubble is likely to 10−13(104K/T)0.7 cm3 s−1 (Kwan1997)isthethecoefficient beexcitedbyanO8.5V-O9Vtypestarwithadynami- of the radiative recombination, CHii is the isothermal sound calageof∼ 1.0Myr. Thevelocitydifferencebetween speed of ionized gas assumed to be 10 km s−1, and RIF is the southeastern clumps and the northwestern ones as the radius of the IF. If we assume that RIF is approximately shown in the spectra of HCO+ and o-H2CO, suggests equaltotheradiusofN4, thetotalmassof∼ 555 M(cid:12) ofthe thatthebubbleisexpanding. neutralandionizedgascanyieldaninitialnumberdensityof nH,0 =∼1.2×104cm−3. GiventheestimatedNLy(seeSection 4. TheshellofcolddustemissionassociatedwithPDRs, 3.4),tdyn ofN4istherefore∼ 1.0Myr. Thistimescale,how- and the compressions of ionized gas characterized by ever,isuncertainsincetheactualevolutionoftheHiiregions the column density PDF of the target regions demon- is not in a strictly uniform medium. Bearing this in mind, stratethattheexpandingbubblehasastronginfluence the estimated dynamical age should be considered with this ontheitssurroundings. caveat. Ontheotherhand,followingWhitworthetal.(1994),grav- 5. In the context of the C & C mechanism, we find that itational fragmentation of the shell of collected material can the shell of collected matter has enough time to frag- beexpectedwhen ment during the lifetime of N4. From compression of t =1.56a 7/11N −1/11n −5/11, (9) ionizedgas,theoverdensityofthenumberofYSOcan- frag .2 49 3 didatesattheedgeofthebubble,andthetimescalesin- where a is the sound speed (a ) inside the shell in units of volved, wesuggestthattriggeredstarformationmight .2 s 0.2 km s−1, N is the ionizing photon flux (N ) in units of have taken place in the bubble N4 but its definitive 49 Ly 1049 ph s−1, and n is the initial gas atomic number density demonstration requires more detailed molecular lines 3 (n )inunitsof103 cm−3. Anestimateofa = 0.3kms−1at observations. H,0 s adusttemperatureof22Kgivest =∼0.3Myr. frag Incomparison,thefragmentationtimeismuchshorterthan the dynamical age. This indicates that the shell of collected We thank the anonymous referee for the comments that gas has had enough time to fragment during the lifetime of much improved the quality of this paper. This work is N4, which is consistent with the signature of six dust frag- supported by the National Natural Science Foundation of mentscondensedoutoftheshell. Therefore,theC&Cpro- China through grants of 11503035, 11573036, 11373009, cessispresumablyatworkinthebubbleN4. Inaddition,the 11433008, and 11403040; the International S&T Coopera- spatial distribution of some YSO candidates shows an over- tionProgramofChinathroughgrandof2010DFA02710;and densityofthenumberofYSOsattheedgeofthebubble(see the Beijing Natural Science Foundation through the grant of Figure4and5). Incombinationwithsuchanoverdensityand 1144015; andtheChineseAcademyofSciences. G.D., S.P., the existence of ionization compression, the aforementioned and M.O. acknowledge support from ANPCyT and CON- timescalesshowthatascenariooftriggeredstarformationin ICET (Argentina) grants. SPIRE has been developed by a thebubbleN4throughtheC&Cprocessispossible. consortiumofinstitutesledbyCardiffUniv. (UK)withUniv. 5. SUMMARY Lethbridge (Canada); NAOC (China); CEA, LAM (France); IFSI, Univ. Padua (Italy); IAC (Spain); Stockholm Ob- Taking advantage of observations of Herschel and of four molecularlines(i.e.,H O6 −5 ,CH OH7 −6 ,HCO+ servatory (Sweden); Imperial College London, RAL, UCL- 2 1,6 2,3 3 0,7 1,6 MSSL, UKATC, Univ. Sussex (UK); Caltech, JPL, NHSC, (1-0),ando-H CO2 −1 ),togetherwithauxiliaryarchival 2 1,2 1,1 Univ. Colorado (USA). This development has been sup- data involving four public surveys (i.e., GLIMPSE, MIPS- ported by national funding agencies: CSA (Canada); NAOC GAL, WISE, 2MASS, and MAGPIS), we have investigated (China); CEA,CNES,CNRS(France); ASI(Italy); MCINN the interactions of the bubble N4 with the adjacent medium (Spain); SNSB (Sweden); STFC (UK); and NASA (USA). andexplored thepossibilityoftriggered starformation. The PACS has been developed by a consortium of institutes mainresultsaresummarizedbelow. led by MPE (Germany) with UVIE (Austria); KU Leuven, 1. The distributions of the dust temperature and column CSL, IMEC (Belgium); CEA, LAM(France); MPIA (Ger- densitytowardN4showananti-correlation: thehigher many); INAFIFSI/ OAA/OAP/OAT, LENS, SISSA (Italy); the column density, the colder the dust temperature. IAC (Spain). This development has been supported by the This is attributed to the penetration degree of the ex- funding agencies BMVIT (Austria), ESA-PRODEX (Bel- ternalheatingfromtheassociatedHiiregion. gium), CEA/CNES (France), DLR (Germany), ASI/INAF (Italy), and CICYT/MCYT (Spain). We are grateful to the 2. A shell structure standing out of the column density KVN staff. The KVN is a facility operated by the Korea mapharborssixdensedustclumps. Theseclumpshave Astronomy and Space Science Institute. We have used the a mean size of ∼ 0.5 pc, temperature of ∼ 22 K, col- NASA/IPAC Infrared Science Archive to obtain data prod- umn density of ∼ 1.7×1022 cm−2, volume density of ucts from the Spitzer-GLIMPSE, Spitzer-MIPSGAL, WISE, ∼ 4.4×103 cm−3, and mass of ∼ 3.2×102 M . Two and2MASSsurveys. (cid:12)