PreprinttypesetusingLATEXstyleemulateapjv.5/2/11 A MULTI-WAVELENGTH STUDY OF STAR FORMATION ACTIVITY IN THE S235 COMPLEX L. K. Dewangan1,3, D. K. Ojha2, A. Luna1, B. G. Anandarao3, J. P. Ninan2, K. K. Mallick2, and Y. D. Mayya1 ABSTRACT We have carried out an extensive multi-wavelength study to investigate the star formation process in the S235 complex. The S235 complex has a sphere-like shell appearance at wavelengths longer than2µmandharborsanO9.5Vtypestarapproximatelyatitscenter. Near-infraredextinctionmap traces eight subregions (having A > 8 mag), and five of them appear to be distributed in an almost V 6 regularly spaced manner along the sphere-like shell surrounding the ionized emission. This picture is 1 also supported by the integrated 12CO and 13CO intensity maps and by Bolocam 1.1 mm continuum 0 emission. The position-velocity analysis of CO reveals an almost semi-ring like structure, suggesting 2 an expanding Hii region. We find that the Bolocam clump masses increase as we move away from the location of the ionizing star. This correlation is seen only for those clumps which are distributed n near the edges of the shell. Photometric analysis reveals 435 young stellar objects (YSOs), 59% of a J which are found in clusters. Six subregions (including five located near the edges of the shell) are very well correlated with the dust clumps, CO gas, and YSOs. The average values of Mach numbers 8 derived using NH data for three (East 1, East 2, and Central E) out of these six subregions are 2.9, 1 3 2.3, and 2.9, indicating these subregions are supersonic. The molecular outflows are detected in these three subregions, further confirming the on-going star formation activity. Together, all these results ] A are interpreted as observational evidence of positive feedback of a massive star. G Subject headings: dust, extinction – Hii regions – ISM: clouds – ISM: individual objects (S235) – stars: formation – stars: pre-main sequence . h p - 1. INTRODUCTION spanning near-infrared (NIR) to radio wavelengths. The o S235 complex is known as an active site of star forma- The energetics of massive stars can strongly influence r tion, harboring young stellar clusters (e.g., Kirsanova et t the surroundings (Zinnecker & Yorke 2007). Massive s al. 2008; Dewangan & Anandarao 2011; Camargo et al. starscandestroystar-formingclouds(i.e.,negativefeed- a 2011; Chavarr´ıa et al. 2014) associated with known star- [ back) and can also trigger star formation (i.e., positive feedback), leading to the formation of a new generation forming subregions, namely East 1, East 2, and Central 1 of stars including young massive star(s) (Deharveng et (e.g. Kirsanova et al. 2008). In our previous work on the v S235 complex using the Spitzer-IRAC data (Dewangan al. 2010). However, the feedback processes of massive 8 & Anandarao 2011, hereafter Paper I), we detected sev- stars are still poorly understood. 8 eral young stellar objects (YSOs) including a High Mass The extended star-forming region S235 is known to 4 Protostellar Object (HMPO) candidate as well as signa- be a part of the giant molecular cloud G174+2.5 in the 4 Perseus Spiral Arm (e.g. Heyer et al. 1996) and contains tures of outflow activities. Using 13CO (1−0) line data, 0 two known sites “S235 complex” and “S235AB region” Kirsanova et al. (2008) found three molecular gas com- . 1 (seeFigure1inDewangan&Anandarao(2011)andalso ponents (i.e., −18 km s−1 < Vlsr < −15 km s−1 (red), 0 in Kirsanova et al. (2014)). The present work is focused −21 km s−1 < V < −18 km s−1 (central), and −25 lsr 6 on “S235 complex” and does not include “S235AB re- km s−1 <V <−21 km s−1 (blue)) in the direction of lsr 1 gion”. Different values of the distance (1.36 kpc, 1.59 the S235 complex. However, the complex is well traced : v kpc, 1.8 kpc, 2.1 kpc, and 2.5 kpc) to the extended in mainly two molecular gas components (central and i star-forming region S235 are reported in the literature blue). More recently, Kirsanova et al. (2014) derived X (e.g. Georgelin et al. 1973; Israel and Felli 1978; Evans physical parameters of dense gas (i.e., gas density and r & Blair 1981; Brand & Blitz 1993; Burns et al. 2015; temperature) in subregions of the complex using ammo- a Foster&Brunt2015). Inthepresentwork,wehavecho- nia (NH ) line observations. However, the properties of 3 senadistanceof1.8kpcfollowingEvans&Blair(1981), dense gas are not explored with respect to the ionizing which is an intermediate value of the published distance star location. Previous studies indicated that the S235 range. The Hii region associated with the S235 com- Hii region is interacting with its surrounding molecular plex is predominantly ionized by a single massive star cloud and the S235 complex has been cited as a possible BD+35o1201ofO9.5Vtype(Georgelinetal.1973). The site of triggered star formation (Kirsanova et al. 2008, S235 complex has been studied using multiple datasets 2014; Camargo et al. 2011). Spitzerimagesrevealedthatthecomplexhasasphere- [email protected] like shell morphology and the ionizing star is approxi- 1InstitutoNacionaldeAstrof´ısica,O´pticayElectr´onica,Luis mately located at its center (see Figure 1 given in Pa- EnriqueErro#1,Tonantzintla,Puebla,M´exicoC.P.72840 2DepartmentofAstronomyandAstrophysics,TataInstitute per I). In addition to the observed interesting morphol- ofFundamentalResearch,HomiBhabhaRoad,Mumbai400005, ogy, the complex is a relatively nearby star-forming site, India making it a promising site to study the feedback of a 3Physical Research Laboratory, Navrangpura, Ahmedabad - massive star. In spite of the numerous existing observa- 380009,India 2 L. K. Dewangan et al. tionsandinterpretations, thefeedbackofanO9.5Vtype We obtained deep NIR photometric JHK images and star is not systematically explored in the S235 complex. the magnitudes of point sources in the region (∼17.(cid:48)6 Theaimofthepresentworkistostudythephysicalpro- × 15.(cid:48)5) from the UKIDSS 6th archival data release cesses governing the interaction and feedback effect of a (UKIDSSDR6plus) of the Galactic Plane Survey (GPS; massive star on its surroundings. Lawrenceetal.2007). ThesurveyusestheUKIRTWide In order to address above aim, we revisited the Field Camera (WFCAM; Casali et al. 2007). Following S235 complex using high sensitivity United Kingdom the selection conditions given in Lucas et al. (2008), we Infra-Red Telescope (UKIRT) Infrared Deep Sky Survey retrieved only reliable NIR sources in the region. One (UKIDSS)NIRdata,GiantMetre-waveRadioTelescope can also find more details about the selection procedure (GMRT) 610 MHz radio continuum map, and dust con- of the GPS photometry in the work of Dewangan et al. tinuum1.1mmdata, inconjunctionwithpublishednar- (2015). Two Micron All Sky Survey (2MASS; Skrut- rowbandH2map,Spitzermid-infrared(MIR)data,NH3 skie et al. 2006) data were obtained for bright sources linedata,NRAOVLASkySurvey(NVSS)1.4GHzcon- that were saturated in the GPS catalog. We found 2444 tinuummap,andCOlinedata. Weperformedadetailed sources detected in all the three NIR (JHK) bands. Ad- studyofthedistributionandkinematicsofmoleculargas ditionally, 326 sources were selected having detections in the complex. In order to systematically explore the only in the H and K bands. feedback of a massive star, we estimated various pres- sure components (such as pressure of an Hii region, ra- 2.3. Spitzer data diation pressure, stellar wind pressure, pressure exerted Thephotometricimages(3.6–24µm)oftheS235com- bytheself-gravitatingmolecularcloud,andratioofther- plex were obtained from the Spitzer Space Telescope mal to non-thermal gas pressure). For a detailed study Infrared Array Camera (IRAC; Fazio et al. 2004) and of the embedded young population in the complex, we the Multiband Imaging Photometer (MIPS; Rieke et al. employed different color-color and color-magnitude dia- 2004). The final processed IRAC images (3.6–8.0 µm) grams obtained using NIR and MIR data, as well as ex- and photometry of point sources were taken from Pa- tinction map generated from NIR data, and the surface per I. MIPS 24 µm observations were retrieved from the density analysis. Additionally, the physical properties Spitzer public archive1, which were carried out in MIPS of gas derived using the line and continuum data were scanmodeon15April2008(Programid40005; PI:Gio- investigated. vanniFazio). MIPSmosaicat24µmwasgeneratedusing In Section 2, we provide the description of various the Basic Calibrated Data (BCD) images. datasets along with reduction procedures. In Section 3, wesummarizetheresultsrelatedtothephysicalenviron- 2.4. Dust continuum 1.1 mm data mentandpoint-likesources. Thepossiblestarformation Bolocam1.1mmimage(Aguirreetal.2011)andBolo- scenario is discussed in Section 4. Our main conclusions cam source catalog at 1.1 mm (v2.1; Ginsburg et al. are summarized in Section 5. 2013) were obtained from Bolocam Galactic Plane Sur- 2. DATAANDANALYSIS vey (BGPS). The effective full width at half maximum (FWHM) of the 1.1 mm map is ∼33(cid:48)(cid:48). Multi-wavelengthdataareemployedtostudythephys- ical conditions in the S235 complex. The size of the 2.5. Radio continuum data selected region is ∼17.(cid:48)6 × 15.(cid:48)5, centered at α = 2000 05h40m57.8s, δ = +35◦51(cid:48)13(cid:48)(cid:48), corresponding to a Weusedthearchivalradiocontinuumdataat610MHz 2000 physical scale of about 9.2 pc × 8.1 pc at a distance (50 cm) and 1.4 GHz (21 cm). The 1.4 GHz map was of 1.8 kpc. retrieved from the NVSS archive. The beam size of the NVSS image is ∼45(cid:48)(cid:48) (Condon et al. 1998). The 610 2.1. Narrow-band H image MHz continuum data were observed on 18–19 June 2005 2 (Project Code: 08SKG01) and were retrieved from the Narrow-band H (ν = 1 − 0 S(1) at λ = 2.122 µm 2 GMRTarchive. GMRTradiodatareductionwascarried (∆λ = 0.032µm)) imaging data of the S235 complex outusingAIPSsoftware, inamannersimilartothatde- were obtained from the survey of extended H emission 2 scribed in Mallick et al. (2013). Since the S235 complex (Navareteetal.2015), conductedattheCanada-France- was away from the center of the observed field, the pri- Hawaii Telescope (CFHT), Mauna Kea, Hawaii, using mary beam correction for 610 MHz was also carried out, Wide-field InfraRed Camera (WIRCam). The observa- using the AIPS PBCOR task and parameters from the tions were performed with the average seeing of ∼0(cid:48).(cid:48)7. GMRT manual2. The synthesized beam size of the final The survey also provided K-band continuum (λ = 2.218 610 MHz map is ∼48(cid:48)(cid:48) × 44(cid:48).(cid:48)2. µm; ∆λ = 0.033µm) images, which were used to get the final continuum-subtracted H2 map. Note that the 2.6. Molecular CO line data H image published by Navarete et al. (2015) contains 2 The J=1−0 lines of 12CO and 13CO data were ob- onlytheEast2andCentralstar-formingsubregions(see servedfromtheFiveCollegeRadioAstronomyObserva- source IDs 164 and 165 in Table B2 given in Navarete et tory (FCRAO) 14 meter telescope in New Salem, Mas- al. 2015), while the H image in this work is presented for a larger area ∼10.(cid:48)72× 9.(cid:48)9 toward the complex. One sachusetts. The FCRAO beam sizes were 45(cid:48)(cid:48) and 46(cid:48)(cid:48) for12COand13CO,respectively. TheS235complexwas can find more details about the survey in the work of observed as part of the Extended Outer Galaxy Survey Navarete et al. (2015). 1 seehttp://sha.ipac.caltech.edu/applications/Spitzer/SHA/ 2.2. Near-infrared data 2http://gmrt.ncra.tifr.res.in/gmrt hpage/Users/doc/obs manual.pdf Study of star formation activity in S235 3 (E-OGS, Brunt 2004), that extends the coverage of the The Central E subregion contains two bright sources, FCRAO Outer Galaxy Survey (OGS, Heyer et al. 1998) namelyEBIRS1/G173.6328+02.8064(hereafter,IRS1) to Galactic longitude l = 193◦, over a latitude range of and EB IRS 2/G173.6339+02.8218 (hereafter, IRS 2) −3◦.5 ≤ b ≤ +5◦.5. However, the data cubes of S235 (Evans & Blair 1981) with previously estimated lumi- complex were further re-processed and a document de- nosities of ∼2.5 × 103 L and ∼1.5 × 103 L (Evans (cid:12) (cid:12) scribingthere-processingdatamethodsisgiveninBrunt et al. 1981), respectively. Furthermore, radio continuum (2004) (also Brunt C.M. et al.; in preparation). These emissions are absent toward these two sources (IRS 1 CO data cubes were obtained from M. Heyer and C. andIRS2),aspointedoutbyNordhetal.(1984). Com- Brunt (through private communication). bining these information with the radio continuum data presentedinthiswork,onecannotethatnoradiocontin- 2.7. Other Data uumemissionpeaksareseentowardthesubregions. Fig- We utilized the publicly available archival WISE3 ure1bshowstheenhanced4.5µmandK-bandextended (Wright et al. 2010) image at 12 µm (spatial resolution emissionfeaturesintheEast1andEast2subregions. In ∼6(cid:48)(cid:48)). We also used previously published NH line data Paper I, the signatures of outflow activities were found 3 from Kirsanova et al. (2014). towardEast1andEast2subregionsduetothepresence of shock-excited H emissions as revealed by the IRAC 2 3. RESULTS ratio maps. Note that the embedded sources in East 1 3.1. Mid-infrared and radio continuum images subregionappeartobeseparatedbytheeasternemission wall from the sphere-like shell (see Figure 1a). The most prominent feature of the S235 complex ob- servableintheinfraredregime,thesphere-likeshellmor- phology,isdetectedatwavelengthslongerthan2µm(see 3.2. Near-infrared extinction map Figure 1). Figure 1a is a three-color composite image Thevisualextinction(A )mapoftheS235complexis V madeusingMIPS24µminred,WISE12.0µmingreen, generated using the publicly available GPS NIR (JHK) andIRAC8.0µminblue. Figure1bshowsathree-color photometric data. The extinction value of individual compositeimageusingMIPS24µm(red), IRAC4.5µm stars was computed using a color-excess along the line (green), and UKIDSS 2.2 µm (blue). Radio continuum of sight (E(H −K) = (H −K) − < (H −K) >; see obs 0 contours at 1.4 GHz and 610 MHz are also overlaid on equation given in Lada et al. 1994) using the reddening thecolor-compositeimagesinFigures1aand1b,respec- lawsofIndebetouwetal.(2005). The<(H-K) >isthe 0 tively. Theseradiocontinuumcontourstracetheionized meanintrinsiccolorofthestars,whichwasdeterminedto emission in the S235 complex. The distribution of ion- be∼0.25fromthenearbycontrolfield(similarsizeastar- izedemissionisalmostsphericalinboththeradiomaps. get region; central coordinates: α = 05h43m35s.8, J2000 The location of a previously characterized O9.5V star δ =+35◦20(cid:48)20(cid:48)(cid:48).4). Theextinctionmapwascreated J2000 (BD +35o1201) appears at the peak of radio continuum by mean value of 40 nearest neighbours. The resultant emission. The bulk of the ionized emission, as traced at A map of S235 complex is shown in Figure 2a, which V the sensitivity levels shown in the Figures 1a and 1b, is allows to trace several embedded subregions in the com- well located inside the sphere-like shell, as highlighted plex. All these subregions are indicated in Figure 2a, by a circle in Figure 1a. The radio emission at 610 MHz which are already highlighted in Figure 1b based on the appears elongated in north direction, as highlighted by visual appearance of stellar contents. The A values V an arrow (see Figure 1b). This particular feature does vary between 2 and 12.5 mag with an average of about not appear to be seen in the NVSS 1.4 GHz map. The 5.8 mag. The A map also resembles the sphere-like V WISE 12 µm band is dominated by the 11.3 µm poly- morphology as highlighted in Figure 1a, except in the cyclic aromatic hydrocarbon (PAH) emission as well as north direction, where it seems to be broken (shown by the warm dust continuum emission. The 24 µm image an arrow in Figure 2a). shows the warm dust emission, whereas the 8 µm band Fifteen Bolocam dust continuum clumps at 1.1 mm contains PAH emission features at 7.7 and 8.6 µm (in- were retrieved from Bolocam source catalog (v2.1) and cluding the continuum). In Figure 1, we notice that the areshowninFigure2a. Figure2aillustratesacorrelation warmdustemissionissurroundedbythe8µmemission. betweenextinctionanddustclumps. Mostoftheclumps Furthermore, the correlation of the warm dust and ion- are located in the region having A greater than 8 mag. V ized emission is evident, which has generally been found inHiiregions(e.g.Deharvengetal.2010;Paladinietal. 3.3. Clump properties 2012). In addition to the extended emission, the embed- Dust continuum emission is known to trace the dense ded stellar contents are visually seen around the sphere- and cold regions. In Figure 2b, we show Bolocam dust like shell. The different subregions (viz. East 1, East 2, continuum emission at 1.1 mm, which is seen toward North, North-West, Central W, Central E, South-West, all the marked subregions in Figure 1b. It is impor- and South) in the S235 complex are shown in Figure 1b. tant to note that previously published SCUBA/JCMT Note that three subregions (East 1, South-West, and 850 µm data (beam size ∼14(cid:48)(cid:48)) are not available toward South)arelocatedawayfromtheedgesofthesphere-like the “North” and “North-West” subregions of the S235 shell(seeFigure1). Thereisnoticeablenebularemission complex (e.g. Klein et al. 2005; Francesco et al. 2008). seen between East 1 subregion and the sphere-like shell, Therefore, the 1.1 mm data (beam size ∼33(cid:48)(cid:48)) provide a whichisdesignatedaseasternemissionwallinFigure1a. morecompleteinformationofthedenseandcoldregions 3 WideField InfraredSurvey Explorer, whichis ajoint project oftheS235complex. Theidentifiedfifteenclumpsinthe oftheUniversityofCaliforniaandtheJPL,Caltech,fundedbythe 1.1mmmaparealsopresentedinFigure2b. Theassoci- NASA ationofclump(s)witheachsubregionislistedinTable1. 4 L. K. Dewangan et al. √ Fivesubregions(i.e.,North-West,CentralW,CentralE, solidangleΩ (=π(FWHM/(206265× 4ln2))2 = beam East 1, and East 2) contain at least two clumps, while 2.9×10−8 forBolocambeamFWHM=33(cid:48)(cid:48))isinsteradi- other two subregions (i.e., South-West and North) are ans, and N(H ) (= 2.0×1022 S (Jy) at T = 20 K; see 2 ν D associated with a single clump. In order to derive gas Bally et al. 2010) is per square centimeter. The column mass Mg for each clump at 1.1 mm emission, the follow- densities thus computed are listed in Table 1. Bolocam ing formula has been utilized (Hildebrand 1983): clumps have column densities between 0.33 × 1022 and 13.5 × 1022 cm−2. D2S R M = ν t (1) g B (T )κ ν d ν 3.4. IRAC ratio and H maps 2 whereS istheintegrated1.1mmflux(Jy),Disthedis- ν In order to trace molecular H emission in the com- tance (kpc), R is the gas-to-dust mass ratio (assumed 2 t plex, we examined the Spitzer-IRAC ratio (4.5 µm/3.6 to be 100), B is the Planck function for a dust temper- ν µm) map and the narrow H (2.12 µm) map. The ature T , and κ is the dust absorption coefficient. In 2 d ν Spitzer-IRACratiomapshavebeenutilizedbymanyau- the calculation, we use κ = 1.14cm2g−1 (e.g. Enoch et ν thors to study the interaction between a massive star al. 2008; Bally et al. 2010), D = 1.8 kpc, and T = 20 K d and its surrounding interstellar medium (ISM) (Povich (an average temperature value estimated from NH line 3 et al. 2007; Watson et al. 2008; Kumar Dewangan & data; see Kirsanova et al. 2014, for more details). The Anandarao 2010; Kumar & Anandarao 2010; Dewangan massescomputedfromtheobserveddustcontinuumdata & Anandarao 2011; Dewangan et al. 2012). IRAC 3.6 arelistedinTable1. Thetotalgasmassofthe15clumps µm and 4.5 µm bands have almost identical point re- comes out to be ∼1179 M . The clump masses vary be- (cid:12) sponse function (PRF), therefore, one can directly take tween 7 M and 285 M . All the clumps have masses (cid:12) (cid:12) the ratio of 4.5 µm to 3.6 µm bands. In Figure 4, we below 100 M except three clumps. These three clumps (cid:12) display a ratio map of 4.5 µm/3.6 µm emission. The are associated with Central E (clump IDs 6 and 9) and spherical shell-like morphology is evident and the map East 1 (clump ID 15) subregions and have masses more depicts the edges of the shell. Additionally, several fea- than 200 M (see Table 1). In an earlier work of the (cid:12) tures are seen that are present toward the edges of the S235 complex (excluding North and North-West subre- shell and in the vicinity of a massive star. In general, gions),Kleinetal.(2005)selectedthirteendustemission the bright emission region in ratio 4.5 µm/3.6 µm map clumps in the JCMT SCUBA 850 µm survey (beam size indicates the excess 4.5 µm emission, while the remain- ∼14(cid:48)(cid:48); see Fig. 1 of IRAS 05377+3548 in Klein et al. ing black or dark gray regions suggest the domination of 2005). They estimated a mass range of clumps between 3.6 µm emission. IRAC 4.5 µm band contains a promi- 10and120M ,usinganaveragedusttemperatureof20 (cid:12) nent molecular hydrogen line emission (ν = 0–0 S(9); K, a gas-to-dust mass ratio of 150, and a distance of 1.8 4.693 µm), which can be excited by outflow shocks, and kpc. Kirsanovaetal.(2014)alsocomputedamassrange a hydrogen recombination line Brα (4.05 µm). IRAC of dense gas in the clumps between 12 and 250 M (see (cid:12) 3.6 µm band harbors PAH emission at 3.3 µm as well Table6inKirsanovaetal.2014). Consideringtheclump as a prominent molecular hydrogen feature at 3.234 µm masses estimatedin thepresent work, onecan finda no- (ν = 1–0 O(5)). In ratio map, we also notice a bright ticeable difference in the clump mass at the high end of region in the vicinity of the massive star that coincides theestimatedrangecomparedtothepreviousworks. At withwarmdustemissiontracedat24µmaswellaswith the low end of the range, the gas mass is almost same in the peak of the radio emission (see arrow in Figure 4). the present calculations compared to the earlier works. This bright emission region probably traces the Brα fea- More extensive high sensitivity and high resolution ob- ture originated by photoionized gas. In East 1, East 2, servations at mm/sub-mm wavelengths would allow to Central E, and North-West subregions, we found bright further explore the clump mass at the high end of the emissions that are marked in the ratio map. Note that estimated mass range. there are no radio continuum emission detected toward In Figure 3, we show the distribution of clump masses these subregions. Therefore, these emissions are proba- as a function of distance from the location of the O9.5V blyH featurestracingoutflowactivities(seeSection3.5 star. Interestingly, we notice that the clump masses in- 2 for molecular outflows). Figure 5 shows a continuum- creaseaswemoveawayfromthelocationoftheionizing subtracted 2.12 µm H (ν = 1−0 S(1)) image, revealing star. Thiscorrelationisseenonlyforthoseclumpswhich 2 thepresenceofH emissioninthecomplex. Thecomplex aredistributedneartheedgesofthesphere-likeshell(e.g. 2 containsdiffuseorfilamentary-likeH featuresaswellas Central E, East 2, North, and North-West subregions). 2 extendedH emission. TheextendedH emissionisseen Theimplicationofthiscorrelationforthestarformation 2 2 as polar structures (i.e., monopolar, bipolar, and multi- process is discussed in Section 4. polar; see East 1, East 2, and Central E subregions in The column densities of clumps can also be computed Figures 5 and 6). The bipolar emission structure is of- using the Bolocam data. The column density per beam ten interpreted as a bipolar outflow excited by a young is given by (see Deharveng et al. 2010): star. The multipolar structures could be originated due S R to presence of bipolar outflows associated with multiple N(H2)= κ B (T ) µν tm Ω , (2) young stars. The monopolar H2 structure could be ex- ν ν d H2 H beam plainedbythe presenceof abipolar outflowexcited by a In the equation above, S , R , κ , and B (T ) are de- young star, however its red-shifted component is hidden ν t ν ν d fined as in Equation 1, µ = 2.8 is the mean molecular fromviewduetoinclinationangle,extinction,andopac- H2 weight per hydrogen molecule (Kauffmann et al. 2008), ity of the ISM around the young star. In Figure 6, we the hydrogen atom mass m is in grams, and the beam presentH ,4.5µm,and24µmimagesofthreesubregions H 2 Study of star formation activity in S235 5 (i.e., East 1, East 2, and Central-E), where polar struc- locity better than 12CO. Considering this fact, we pre- turesaretracedintheH map. InEast1subregion,one ferred to show only 13CO channel maps of the complex. 2 multipolarstructureisclearlytraced. Additionally,some In the velocity range from −22.75 to −21.25 km s−1, H knotsareseen. TheH mapconfirmsintenseoutflow Central W, East 2, and North-West subregions are well 2 2 or jet activities in this subregion. The extended 4.5 µm detected. In the velocity range from −21 to −19.75 km emission is also detected similar to those seen in the H s−1, Central E, Central W, East 1, East 2, North, and 2 map. The prominent 4.5 µm emission was reported in North-West subregions are traced. In the velocity range Paper I using the IRAC ratio map 4.5 µm/8.0 µm (see from−19.50to−18.0kms−1,CentralEandEast1sub- Figure 11 in Paper I). This subregion contains a clus- regions are seen. The 13CO gas kinematics presented in ter of embedded YSOs which makes difficult to pinpoint this work is consistent with the previous work of Kir- the exact exciting source(s) of outflows. However, some sanova et al. (2008). They suggested the presence of probable driving sources (i.e. YSOs) of outflows, that three molecular gas components in the complex, which are taken from Paper I, are marked in the 4.5 µm and correspond to quiescent undisturbed primordial gas (i.e. 24µmimages. ThesehighlightedYSOsappearbrightin −18kms−1 <V <−15kms−1 (red)),gascompressed lsr the 24 µm image. In East 2 subregion, a bipolar struc- by the shock from expanding S235 Hii region (i.e. −21 ture is detected and is also seen in the 4.5 µm image km s−1 < V < −18 km s−1 (central)), and gas ex- lsr (also see Figure 12 in Paper I). In Paper I, the driving pulsion from the embedded young star clusters driven sourceofthisoutflowwascharacterizedasaHMPOcan- by the combined effect of the cluster stars (i.e. −25 km didate. The position of this YSO is also marked in the s−1 < V < −21 km s−1 (blue)). Furthermore, Kir- lsr 4.5 µm and 24 µm images. The YSO is very bright in sanovaetal.(2014)utilizedNH linedataandfoundthe 3 the 24 µm image. In Central-E subregion (see Figure 6), densestmolecularclumpsassociatedwithEast1,East2, wefindabipolarandtwomonopolarstructures(alsosee and Central subregions, mostly belong to the “central” source IDs 164 and 165 in Table B2 given in Navarete molecular component. et al. 2015). The probable exciting sources of one bipo- Figure 8 shows the 12CO and 13CO integrated gray larstructure(Outflow-ce1)andonemonopolarstructure scale maps, obtained by integrating over the velocity (Jet-ce) are marked in the 4.5 µm and 24 µm images rangebetween−23and−18kms−1. Bolocamdustcon- (see Figure 6). The 24 µm counterparts of these driv- tinuum emission contours are also overlaid on the 12CO ing sources are also seen. However, the driving source integrated map to illustrate the physical association of of one monopolar structure (Outflow-ce2; Figure 6) is dust emission with the complex. The integrated CO not detected. In general, one can refer H nebulosity 2 mapstracearegionemptyofmoleculargasbetweentwo as an outflow when it traces back to a source. In the subregions, namely North and North-West. It is to be present case, considering the morphology of H nebulos- 2 noted that the region empty of CO gas is exactly coinci- ity (e.g. Takami et al. 2010), we refer this monopolar dent with the broken part in the extinction map, where structure as the outflow signature. It is generally ac- the radio continuum emission at 610 MHz is elongated cepted that the molecular outflows directly indicate the (see Figures 1b and 2a). This region also appears to presence of star formation processes. A comparison of be associated with Brα emission which is surrounded by the 2.12 µm H emission with the features detected in 2 H and/or PAH features (see Figure 4). It suggests a 2 the ratio map suggest that the black region in the ra- close interaction between ionized and molecular gas in tio map probably traces the H features (see Figures 4 2 the complex. It seems that the ionizing photons have and5). However,thepresenceof3.3µmPAHfeaturecan managed to escape in this direction. We propose that notbeignored. Theprominentdiffuseorfilamentary-like these features are best explained by the existence of a H features traced in the H map could be explained by 2 2 cavity between North and North-West subregions (also ultra-violet (UV) fluorescence. In summary, the IRAC see Section 4.1). ratio and H maps trace photodissociation regions (or 2 In Figure 9, we show position-velocity (p-v) analysis photon-dominated regions, or PDRs) around the Hii re- of 12CO and 13CO gas. The p-v plots (right ascension- gion and star formation activities in the complex. velocity and declination-velocity) of 12CO and 13CO gas have revealed noticeable velocity gradients and show es- 3.5. CO molecular gas associated with the S235 complex sentially the same structure. However, 13CO data pro- In the following, we present molecular line data anal- vide more information related to the velocity structure ysis, which enables us to examine the ongoing physical of the denser regions compared to 12CO emission (see processesandvelocitystructurespresentinthecomplex. declination-velocity plots in Figure 9). The p-v plots of 12COand13COgasrevealanalmostsemi-ring-likeorC- 3.5.1. 12CO and 13CO Distributions like structure at larger scale (about 8 parsec extended). In Figure 7, we present the velocity channel maps of The detection of such morphology in the p-v plot is of- the J = 1−0 line of 13CO (at intervals of 0.25 km s−1). ten interpreted as signature of an expanding shell (e.g. The maps reveal the distribution of molecular gas com- Wilson et al. 2005; Arce et al. 2011; Dewangan et al. ponents along the line of sight. The maps are shown for 2015). We find an expansion velocity of the gas to be thevelocityrangeof−22.75kms−1 to−15kms−1. One ∼3 km s−1. Detailed discussion about this morphology maynoticefromFigure7thatthedistributionofmolecu- can be found in Section 4.1. Note that the S235 com- largasisclumpy,asalsopreviouslypointedoutbyHeyer plexhoststheO9.5Vtypestar, thereforethepresenceof et al. (1996). In general, the 13CO line is more optically theC-likestructureinthecomplexcanbeinterpretedas thin compared to 12CO line. Therefore, the 13CO line caused by the expanding Hii region. Similar result was data can trace dense condensation and its associated ve- also observed in W42 star-forming region (see Figure 8 6 L. K. Dewangan et al. in Dewangan et al. 2015). more details about P in the work of Lada et al. TNT (2003). The sound speed a (= (kT /µm )1/2) can be s kin H 3.5.2. Molecular outflows estimatedwiththeknowledgeofgaskinetictemperature In addition to the C-like structure in the p-v plots, we (Tkin) and µ=2.37 (approximately 70% H and 28% He have also found noticeable velocity gradients toward the by mass). The non-thermal velocity dispersion is given East 1, East 2, Central E, and North-West subregions, by: suggesting the presence of outflow activities in each of (cid:115) (cid:114) subregions. Following Paper I, we know that the East 1, σ = ∆V2 − kTkin = ∆V2 −σ2, (3) East 2, and Central E subregions contain a cluster of NT 8ln2 17m 8ln2 T H YSOs. We have searched for outflows toward these sub- regions using the doppler shifted velocity components where ∆V is the measured FWHM linewidth of the ob- (red, green, and blue). Due to the coarse beam of CO servedNH3spectra,σT(=(kTkin/17mH)1/2)isthether- data (beam size ∼46(cid:48)(cid:48)), we cannot pinpoint the exact mal broadening for NH3 at Tkin (e.g. Dunham et al. exciting source of the outflows in each subregion. Con- 2011), and mH is the mass of hydrogen atom. The com- sidering this limitation, we have not shown the doppler puted values of as, σNT, and PTNT are given in Table 2. shifted components (red, green, and blue) toward the The value of PTNT is computed for each of the observed East 2, Central E, and North-West subregions. In Fig- positions. In each subregion, the variation of PTNT as a ure 10, we present the noticeable receding, approaching, functionofdistancefromthelocationofamassivestaris and rest gas components only in East 1 subregion. We shown in Figure 11b. We find an average PTNT value of have also marked the position of at least one probable 0.13,0.21,and0.22forEast1,East2,andCentralEsub- driving source of the CO outflow (also see Section 3.4). regions, respectively. It suggests that non-thermal pres- Highresolutionmolecularlineobservationsarenecessary sure is dominant in these densest subregions. The lower to examine better insight into the molecular outflows in values of PTNT might be due to contributions from the the complex. bipolaroutflows,asinvestigatedinprevioussection. The average values of Mach numbers (σ /a ) estimated for NT s 3.6. NH radial velocity and non-thermal velocity East1,East2,andCentralEsubregionsare2.9,2.3,and 3 dispersion 2.9,respectivelywhichindicatethatthesesubregionsare supersonic. As mentioned in the introduction, in general, it is knownthatNH3 datatracethedensestregionsofmolec- 3.7. Feedback of a massive star ularcloudinagivenstar-formingregion. Theproperties In this section, we derive the various pressure com- of densest gas traced by NH line observations have not 3 ponents (pressure of an Hii region (P ), radiation yet been studied with respect to the ionizing star in the HII pressure (P ), and stellar wind ram pressure (P )) S235 complex. The locations of NH line observations rad wind 3 driven by a massive star to study its feedback in the are marked in Figure 2b, which were mainly observed in vicinity. East 1, East 2, and Central E subregions (see Kirsanova It is found that the massive star is approximately lo- etal.2014,formoredetails). InFigure11a,weshowthe cated at the center of the sphere-like shell. The radio variationofNH (1,1)radialvelocityineachsubregionas 3 continuum images presented in this work can be used afunctionofdistancefromthelocationoftheO9.5Vstar. to estimate the number of Lyman continuum photons The NH (1,1) radial velocities vary between −22 and 3 (N ). The expression of N is given by (Matsakis et −18.5kms−1 inthesethreesubregions. Ingeneral,there uv uv al. 1976): are noticeable velocity variations within East 1 (−21.5 to −19.0 km s−1), East 2 (−21.5 to −20.5 km s−1), and (cid:18)S (cid:19)(cid:18) D (cid:19)2(cid:18) T (cid:19)−0.45 (cid:16) ν (cid:17)0.1 Central E (−22 to −18.5 km s−1) subregions. These ob- Nuv(s−1)=7.5×1046 Jνy kpc 104eK × GHz servational characteristics are also evident in CO data (4) (see Figure 9). Balser et al. (2011) reported the veloc- where S is the measured total flux density in Jy, D is ity of ionized gas to be about −25.61(±0.12) km s−1 ν the distance in kpc, T is the electron temperature, and in the S235 Hii region using a hydrogen radio recombi- e ν is the frequency in GHz. The calculation of N is uv nation line (H87-93α). Therefore, the molecular gas is performed for both the radio frequencies (0.61 GHz and red-shifted with respect to the ionized gas in the S235 1.4 GHz) separately. Substituting D = 1.8 kpc, T = Hii region. Interestingly, we investigate that the radial e 10000 K, and S = 1.92 Jy, we compute N = 4.9 × 1.4 uv velocity observed in Central E subregion shows a linear 1047 s−1. Similarly, we find N = 8.2 × 1047 s−1 for D, uv trend as we move away from the location of a massive T ,andS =3.49Jy. Theestimatesofionizingphoton e 0.61 star, which is not seen in other two subregions. In each fluxvaluesatdifferentfrequenciescorrespondtoasingle subregion, wehavelineofsightvelocitycomponentsand ionizingstarofO9.5Vspectraltype(Martinsetal.2005), itseemsthattheradialvelocityofgasinCentralEsubre- which is also in agreement with the previously reported gion might be almost equal to the real velocity, however spectral type of the ionizing source of the complex (e.g. it is not the case for other two subregions. The observed Georgelin et al. 1973). trend could be interpreted as the direct influence of the The different pressure components (P , P , and expandingHiiregionexcitedbytheO9.5Vstar. Further HII rad P ; as mentioned above) are defined as below (e.g. wind discussion on this result is presented in Section 4. Bressert et al. 2012): We computed thermal sound speed (a ), non-thermal s (cid:32)(cid:115) (cid:33) vneolno-ctihtyermdiaspleprrseisosnur(eσN(PTT)N, Tan=d ath2se/σrN2aTti)o. oOfntehecramnafilntdo PHII =µmHc2s 4π3αNuvD3 ; (5) B s Study of star formation activity in S235 7 P =L /4πcD2; (6) atedwithatypicalcoolmolecularcloud,weinferthatan rad bol s additional physical process has been acted to compress Pwind =M˙wVw/4πDs2; (7) thesurroundingmoleculargastoenhancethepressurein the complex. Consequently, this process may stimulate In the equations above, N , µ, and m are defined uv H the initial collapse and fragmentation of the extended earlier, the radiative recombination coefficient is “α ” B molecular cloud. The pressure calculations indicate that (= 2.6 × 10−13 × (104 K/T )0.7 cm3 s−1; see Kwan e the photoionized gas can be considered as the important 1997), c is the sound speed in the photo-ionized region s contributorforthefeedbackmechanismintheS235com- (=10kms−1), M˙ isthemass-lossrate, V isthewind w w plex. velocity of the ionizing source, L is the bolometric bol luminosity of the complex, and D is the projected s 3.8. Young stellar objects in S235 distance from the location of the O9.5V type star to the subregions where the pressure components are 3.8.1. Identification of young stellar objects estimated. Onecaninferfromequations5, 6, and7that A population of YSOs can be identified using NIR the pressures, P and P , scale as D−2 while P rad wind s HII and MIR color schemes as utilized by many authors scales as D−3/2. (e.g. Allen et al. 2004; Lada et al. 2006; Gutermuth et s al. 2009). In Paper I, we identified YSOs using only Note that the highlighted subregions are not located IRAC data. In the present work, we also employed at the same projected distance from the location of a the UKIDSS-GPS NIR data, in combination with massive star (see Figures 2 and 3). Therefore, we com- IRAC data for finding more deeply embedded and faint pute the pressure components driven by a massive star sources. NotethattheUKIDSS-GPSNIRdataarethree (i.e., P , P , and P ) at D = 3.4 pc (East 1), magnitudes deeper than 2MASS. Here, we describe the HII rad wind s 3.0 pc (East 2), 1.5 pc (Central E), 2.6 pc (North), 2.7 procedure of YSOs identification and classification using pc(North-West),1.9pc(CentralW),and3.0pc(South- photometric IRAC and GPS data. West) (see Figures 2 and 3). Evans&Blair(1981)estimatedatotalgasmassofthe 1. We selected sources having detections in all four S235 cloud to be about 3000 M in a radius of 3 pc. Spitzer-IRACbands,whichweretakenfromPaperI.We (cid:12) Nordh et al. (1984) computed a total luminosity of S235 found131YSOs(52ClassIand79ClassII),4ClassIII, of ∼8 × 104 L . We use M ≈ 3000 M , R ≈ 3.0 133 photospheres, and 38 contaminants in our selected (cid:12) cloud (cid:12) c pc, L ≈ 8 × 104 L , M˙ ≈ 1.58 × 10−9 M yr−1 region shown in Figure 1a. The details of YSOs clas- bol (cid:12) w (cid:12) (for O9.5V star; Marcolino et al. 2009), V ≈ 1500 km sification can be found in Paper I. Figure 12a displays w s−1 (for O9.5V star; Marcolino et al. 2009) in the above the different selected sources. In Figure 12a, we show Class I YSOs (red circles), Class II YSOs (open blue equations to compute the pressure contributions driven triangles), Class III YSOs (black squares), photospheric by a massive star on different subregions (see the list in emissions (gray dots), and PAH-emission-contaminated Table 3). The error associated with each pressure com- apertures (magenta multiplication symbols). ponentrelatedtoeachsubregionisalsogiveninTable3. In the S235 complex, we find that the pressure of the Hii region exceeds the radiation pressure and the stellar 2. Then, we considered sources lacking detections in two longer wavelengths of IRAC bands (5.8 and windpressure(seeTable3). Thetotalpressure(P = total 8.0 µm). For such sources, GPS-IRAC (H, K, 3.6, P + P + P ) driven by a massive star on each HII rad wind and 4.5 µm) classification scheme was utilized, as subregion is also tabulated in Table 3, which is found to be ∼10−10 dynes cm−2 for each subregion. described in details by Gutermuth et al. (2009). In this method, the dereddened color-color space ([K−[3.6]] Additionally,pressureexertedbytheself-gravityofthe 0 and [[3.6]−[4.5]] ) was used to identify infrared-excess surrounding molecular gas is also estimated using the 0 sources. These dereddened colors were computed relation (e.g. Harper-Clark & Murray 2009): using the color excess ratios given in Flaherty et al. P ≈πGΣ2 (8) (2007). This scheme also offered to identify possible dim cloud extragalactic contaminants from YSOs with additional where Σ (=Mcloud/πRc2) is the mean mass surface den- conditions (i.e., [3.6]0 < 15 mag for Class I and [3.6]0 sityofthecloud,Mcloud isthemassofthemoleculargas, < 14.5 mag for Class II). We used the observed color andRcistheradiusofthemolecularregion. Thevalueof and the reddening law (from Flaherty et al. 2007) to Pcloud for the entire cloud is estimated to be ∼(1.0±0.5) compute the dereddened 3.6 µm magnitudes. In the × 10−10 dynes cm−2. The values of Pcloud associated end, 16 Class I and 133 Class II YSOs are obtained with different subregions are also found to be ∼0.1–3.0 using GPS-IRAC data (see Figure 12b). ×10−10 dynes cm−2 (see Table 1 and Figure 2a). In subregions, we find Ptotal (cid:38) Pcloud (see Table 3). This 3. NIR color-color diagram (H−K vs J−H) is a very argument is still valid if we put all the subregions at useful tool to select infrared-excess sources. We applied the same projected distance (i.e. Ds = 3.0 pc; see pres- this scheme for sources, having detections in all the sure calculations related to East 2 in Table 3). Addi- three JHK bands. Figure 12c shows a NIR color-color tionally, the Pcloud is relatively higher than the pressure diagram of such sources. The reddening lines are drawn associated with a typical cool molecular cloud (∼10−11 using the extinction law of Indebetouw et al. (2005). – 10−12 dynes cm−2 for a temperature ∼20 K and the The color-color diagram is divided into three different particledensity∼103–104 cm−3)(seeTable7.3ofDyson regions, namely “I”, “II”, and “III”. One can find more &Williams1980). Consideringthepressurevalueassoci- details about NIR YSOs classification in Sugitani et 8 L. K. Dewangan et al. al. (2002) and Dewangan et al. (2012). The sources ysis, such as empirical cumulative distribution (ECD) of distributed in the “I” region represent the likely Class I YSOs as a function of NN distance, is often studied to objects (protostellar objects). T Tauri-like sources infer the clustered YSO populations in star-forming re- (Class II objects) are identified within the “II” region gions (see Chavarr´ıa et al. 2008; Gutermuth et al. 2009; along the T Tauri locus with large NIR excess. The Dewangan & Anandarao 2011, for more details). In the sources that fall between the reddening bands of the ECD analysis, a cutoff length (also referred as the dis- main-sequence and giant stars are located in the “III” tance of inflection d ) is chosen, to delineate the low- c region. With this method, we obtain 5 Class I and 96 density/distributed populations. For the present case, Class II sources (see Figure 12c). we selected a cutoff distance of d ∼52(cid:48)(cid:48) (0.45 pc at a c distance of 1.8 kpc), which separates the cluster mem- 4. Finally, we considered sources (significant in num- bers within the contour level of 10 YSOs/pc2. Such ber) that have detections only in the H and K bands. cutoff distance has been determined for different star- In order to further identify infrared excess sources from formingregions, suchastheDiamondRingregioninthe such sample, we utilized a color-magnitude (H−K/K) Cygnus-X star-forming complex (d ∼61(cid:48)(cid:48) or 0.43 pc at c diagram (see Figure 12d). The diagram allows to select 1450 pc; Beerer et al. 2010), the Cygnus-OB2 region in red sources having H−K > 1.04. The color-magnitude theCygnus-Xcomplex(d ∼72(cid:48)(cid:48) or0.51pc;Guarcelloet c analysis of the nearby control field allowed us to infer al. 2013), and the W5 star-forming region (d ∼33(cid:48)(cid:48) or c this color criterion. This condition provides us 54 0.32pcat2kpc;Chavarr´ıaetal.2014). Thecomparison additional deeply embedded infrared excess sources. of d values suggests that W5 is more densely populated c with YSOs compared to Cygnus-X. In general, the value Recently, Chavarr´ıa et al. (2014) also presented pho- of d offers to infer the most densely YSOs populated c tometry of the entire extended star-forming S235 region star-forming regions. The ECD analysis yielded a clus- (including S235AB region) using NIR and Spitzer data. teredfractionofabout59%YSOs(i.e. 258fromatotalof These authors obtained K-band observations using the 435 YSOs). The YSO clusters are mainly distributed in 2.1-m telescope located at Kitt Peak National Observa- East1,East2,CentralE,CentralW,North-West,South, tory. Additionally, they observed J and H bands using and South-West subregions. In Central E and East 1 the 6.5-m Multiple Mirror Telescope (MMT) telescope subregions, we found the highest level of surface density located at Fred Lawrence Whipple Observatory. They contours. North subregion is not associated with any identified 690 YSOs in the entire S235 complex. How- YSO cluster; however, it contains a few Class II YSOs. ever, they did not separate the contaminations in their TheclusteredandscatteredYSOlocationsareidentified analysis. Additionally,theydidnotprovidethepositions in Figure 13 by filled and open symbols, respectively. of their identified YSOs. Therefore, a comparison of our Wealsoconstructedseparatelythesurfacedensitycon- selected YSOs to the sources identified by Chavarr´ıa et tour maps of Class I and Class II YSOs, using the same al. (2014) cannot be thoroughly performed in this work algorithm as explained above. The surface density con- (see Figure 13 in this work and Figure B2 in Chavarr´ıa tour maps of Class I and Class II YSOs are shown in et al. (2014)). In our YSOs analysis, we have carefully Figure 14a. The surface density of Class I YSOs lies be- separatedthecontaminationsfromYSOpopulations. As tween 0.04 and 42.75 YSOs/pc2 with a dispersion (σ) a result of the four schemes explained above, a total of of 2.64 and is shown by contour levels of 5, 8, 10, and 435 YSOs are yielded in the complex (S235AB region is 20 YSOs/pc2. Similarly, the surface density of Class II not included). The positions of all YSOs are shown in YSOs is found to lie between 0.17 and 174 YSOs/pc2, Figure 13. with a dispersion of 6.62 and is drawn at contour levels of 10, 20, and 40 YSOs/pc2. All the Class I and Class II 3.8.2. Spatial distribution of YSOs YSO clusters are exclusively found in the regions with In recent years, the surface density of young stellar A > 8 mag. The Class I YSO clusters are traced in V populations in star-forming regions has been adopted East 1, East 2, Central E, South, and South-West sub- to identify and to study the young stellar clusters (e.g. regions (see Figure 14a). Gutermuth et al. 2009; Bressert et al. 2010). In Paper I, thesurfacedensitymapofYSOswascomputedusingthe 4. DISCUSSION nearest-neighbour (NN) technique (also see Gutermuth 4.1. Expanding Hii region et al. 2009; Bressert et al. 2010, for more details). Here, we also followed the same procedure and generated the Oneoftheimportantresultsofthepresentworkisthe surface density map of all the selected 435 YSOs. The identification of an almost semi-ring-like or C-like struc- map was created using a 5(cid:48)(cid:48) grid and 6 NN at a distance tureinthep-vplots,whichissimilartotheoneobserved of1.8kpc. Figure13showstheresultantsurfacedensity inOrionnebula(seeFigure10binWilsonetal.2005)and contours of YSOs overlaid on the extinction map. The to that in W42 (see Figure 8 in Dewangan et al. 2015). contourlevelsareshownat10,20,40,and70YSOs/pc2, Arce et al. (2011) also studied such observed structures increasingfromtheoutertotheinnerregions. Thefigure in Perseus molecular cloud along with the modeling of clearly illustrates the spatial correlation between YSOs expanding bubbles in a turbulent medium. These au- surface density and extinction. thorssuggestedthatthesemi-ring-likeorC-likestructure The surface density analysis allows to trace the indi- in the p-v plots is the signature of an expanding shell. vidualgroupsorclustersofYSOs. Now,wewanttoiden- They also mentioned that the ring-like morphology can tify the clustered populations from distributed sources. beseeninthep-vplotwhenthepowerfulionizingsource However, there is no unique method to find a cutoff dis- is situated at the center of the region. Additionally, the tancefortracingthesetwopopulations. Statisticalanal- NH radial velocity observed in Central E subregion ex- 3 Study of star formation activity in S235 9 hibitsalineartrendasonemovesawayfromthelocation traced toward these subregions in a velocity range −17 of a massive star. The analysis of pressure contributions km s−1 to −15 km s−1 (see Figure 7). In Figures 15a driven by a massive star (i.e., P , P , and P ) and 15b, we present the distribution of molecular gas HII rad wind indicates that the P value is higher than the other toward the “S235 complex” and the “S235AB region”. HII pressurecomponents. Therefore,itseemsthattheC-like The CO map is integrated in the [-25,-15] km s−1 veloc- structure in the p-v plots corresponds to the expanding ity range (see Figure 15a). In Figure 15b, the p-v plot HiiregionexcitedbytheO9.5Vstaratthecenterofthe of13COgasrevealsanalmostbroadbridgefeature. The sphere-like shell. Negative feedback of the O9.5V star p-v plot shows that two peaks (a red shifted and a blue is also evident by the presence of a cavity (i.e. empty shifted; as mentioned above) are separated by lower in- of molecular CO gas region) between North and North- tensity intermediated velocity emission. The presence of West subregions. Previously, Evans & Blair (1981) also such broad bridge feature in the p-v plot might indicate suggestedthattheS235molecularcloudisheatedbythe the signature of a collision between two clouds (e.g. Ha- excitingstar. Alltheseobservedfeaturessuggesttheim- worth et al. 2015a,b). Therefore, there is a possibility pact of ionizing photons on the low density regions in of the formation of YSO clusters associated with South the complex, through which the ionized gas might have and South-West subregions by the interaction between escaped. these two clouds. In the present work, our analysis is mainlyfocusedontheS235complex,thereforetheresults 4.2. Morphology and signposts of star formation related to the S235AB region are not presented in this work. Hence, a detailed investigation of the interaction TheNIRextinctionmapoftheS235complexprovides between these two clouds is beyond the scope of present a more complete picture of the spatial distribution of work. It is important to note that the broad bridge fea- embedded subregions. There are eight subregions which ture is not seen in the p-v maps of “S235 complex” (see arewellcorrelatedwiththelocationsofthedustclumps, Figure9),suggestingtheapplicabilityofcloud-cloudcol- CO gas, and YSOs. Five of these subregions (i.e. Cen- lision scenario is unlikely in the S235 complex. Hence, tral E, East 2, North, North-West, and Central W) ap- we rule out the cloud-cloud collision process within the pear to be nearly regularly spaced along the sphere-like S235 complex. shell surrounding the ionized emission. Deharveng et al. (2003) also reported similar configuration in Sh 104 re- 4.3. Triggered star formation gion that was classified as a site of triggered star for- InSection3.7,weshowedthatthetotalpressuredriven mation. Furthermore, three subregions (i.e. East 1, by a massive star is in equilibrium with the pressure ex- South, and South-West) in the S235 complex are spa- erted by the self-gravitating molecular cloud. In Sec- tially away from the sphere-like shell. The presence of tion 4.1, we discussed that the signature of the expand- H2 and PAH emissions at the boundaries of the sphere- ing Hii region is evident. Additionally, we inferred that like shell indicates the presence of PDR surrounding the the densest subregions (i.e. East 1, East 2, and Cen- ionized emission (see Figure 4). In the present work, we tral E) traced by NH data are supersonic (see Sec- 3 investigated more number of YSOs using GPS, in addi- tion 3.6). We also pointed out that the observed higher tiontoIRACdata, comparedtoPaperI.TheYSOclus- velocity dispersions could be due to the presence of out- ters are found in all the subregions except North, where flows in the subregions. Morgan et al. (2010) studied 44 only Class II YSOs are seen without any clustering. Ad- bright-rimmedclouds(BRCs)usingNH dataandfound 3 ditionally, the clusters of Class I YSOs are associated that the potentially triggered samples of BRCs are asso- with Central E, East 1, East 2, South, and South-West ciatedwithhighervelocitydispersionscomparedtonon- subregions. Themoleculardataalsorevealedoutflowsig- triggeredsources. Intheirsamples,non-triggeredsources naturesintheCentralE,East1,East2,andNorth-West were largely subsonic, whereas the triggered samples of subregions. We find the probable outflow driving candi- BRCs were supersonic. These results indicate that the date (i.e. Class I YSO) in each subregion. All together, S235complexcouldbeasiteoftriggeredstarformation. the early stages of star formation activity are evident in This interpretation is further supported by the observed all the eight subregions. configuration as traced in the extinction map, dust con- Concerning the molecular gas distribution toward the tinuum emission, molecular gas, ionized emission, and SouthandSouth-Westsubregions(seeFigure7),wesug- distribution of YSOs. Therefore, the on-going star for- gest that these two subregions seem to be located at the mation in the S235 complex may have been influenced interfacebetweentheS235molecularcloud(−23kms−1 by the expanding Hii region. <Vlsr <−18 km s−1) and the S235AB molecular cloud Theoretically, there are two main scenarios discussed (−17 km s−1 <Vlsr <−15 km s−1). As previously also in the literature which explain the triggered star for- reported by Evans & Blair (1981), the extended star- mation by the expansion of the Hii region (Elmegreen forming region S235 is comprised of two velocity compo- 1998; Deharveng et al. 2005): “collect and collapse” (see nents at −20 and −17 km s−1. The velocity information Elmegreen & Lada 1977; Whitworth et al. 1994; Dale suggests that the molecular gas associated with S235AB et al. 2007) and radiation-driven implosion (RDI; see region is red-shifted with respect to the S235 molecu- Bertoldi 1989; Lefloch & Lazareff 1994). In the “col- lar cloud. In the integrated 12CO map, some molecu- lect and collapse” scenario, massive and dense shell of lar gas is seen toward the South and South-West subre- cool neutral material can be accumulated around an ex- gions(seeFigure8a),howeverthemoleculargasisabsent pandingHiiregion,andstarformationoccurswhenthis there in the integrated 13CO map (see Figure 8b). Ad- material becomes gravitationally unstable. In the RDI ditionally, A value is found significant there (see Fig- model, the expanding Hii region causes the instability V ure 2a). Furthermore, the molecular gas is noticeably and aids in the collapse of a pre-existing dense clump in 10 L. K. Dewangan et al. themolecularcloud. Kirsanovaetal.(2014)utilizedNH etal.2009),respectively. Consideringtheseages,onecan 3 line data and suggested that the subregions (i.e. S235 notice that the dynamical age of the Hii region is com- East 2 and S235 Central) were formed via triggering by parable to the ages of YSOs. These calculations suggest a “collect-and-collapse” process. Furthermore, they sug- that the molecular materials have been fragmented into gested that S235 East 1 region was formed as a result of several condensations around the sphere-like shell (also aninteractionoftheshockfrontfromS235complexwith seeFigure13). Therefore,the“collectandcollapse”sce- a pre-existing dense clump. nario seems to be applicable in the S235 complex, which In the present work, Bolocam data analysis revealed could be responsible for the origin of Central E, East 2, that the clump masses grow as one goes away from the North, North-West, and Central W subregions. location of the ionizing star. This pattern is observed NotethatEast1subregionisfarawayfromthesphere- only for those clumps which are located near the edges like shell and is the most massive clump in the com- of the sphere-like shell (e.g. Central E, East 2, North, plex. Additionally, this subregion is separated by the and North-West subregions). It seems that the material eastern emission wall from the sphere-like shell and is in has been collected on the edges of the sphere-like shell pressure equilibrium and is associated with Class I and by the expanding Hii region (also see Figure 2a), sug- Class II YSO clusters. This subregion is considered as gesting the “collect and collapse” scenario applicable in the youngest star-forming site in the complex (see Kir- thecomplex. Inordertocheckthe“collectandcollapse” sanovaetal.2014)andissupersonic. Therefore,itseems process as triggering mechanism, we calculated the dy- that the YSO clusters in East 1 subregion could be orig- namical age (t ) of the Hii region and the fragmenta- inated by compression of the pre-existing dense material dyn tion time scale (t ) following the equations given in by the expanding Hii region through the RDI process. frag Dyson & Williams (1980) and Whitworth et al. (1994), respectively. Onecanfindmoredetailsofsimilaranalysis 5. SUMMARYANDCONCLUSIONS in Dewangan et al. (2012). The condition for the “col- Inthispaper,wehavemadeanextensiveinvestigation lect and collapse” process is t ≥ t . To find the of the S235 complex using the multi-wavelength data dyn frag values of ambient density (n =) for t ≥ t con- covering radio through NIR wavelengths. The results 0 dyn frag dition, the diagram of t and t as a function of obtained in this work are based on the study of ion- frag dyn “n ” is shown in Figure 15c. The t is estimated for ized emission, molecular emission, cold dust emission, 0 frag different“a ”valuesof0.2, 0.3, and0.4kms−1 (seeSec- embedded young populations, and various physical s tion 3.6 for “a ” estimation). Here, we use N = 4.9 × calculations. The important findings of this work are as s uv 1047 s−1 (see Section 3.7) for the entire radio emission follows: of the region (spatial extent ∼6.(cid:48)40 or 3.35 pc) and α • The most prominent structure in the S235 complex B = 2.6 × 10−13 cm3 s−1 at T = 10000 K. If the condi- is the sphere-like shell morphology, as traced at wave- e lengths longer than 2 µm. tion “t ≥ t ” is satisfied then the values of “n ” dyn frag 0 • The distribution of ionized emission observed in the should be greater than 5750, 7700, and 9240 cm−3 for GMRT 610 MHz and NVSS 1.4 GHz continuum maps different “a ” values of 0.2, 0.3, and 0.4 km s−1, respec- s is almost spherical. The ionizing photon flux values tively (see Figure 15c). On the other hand, if n < 5750 0 estimated at both the frequencies correspond to a single cm−3, then the condition “t ≥ t ” is not fulfilled. dyn frag ionizing star of O9.5V spectral type. The location of Consequently, the fragmentation of the molecular mate- the counterpart of this ionizing star (i.e. BD +35o1201) rials into clumps will not occur due to the “collect and appears at the peak of radio continuum emissions as collapse” process in the complex. well as approximately at the center of the sphere-like Using 13CO, NH , H CO, HCO+, and HCN line data, 3 2 shell. a significant density variation (i.e. 1 × 103 – 2 × 104 • The NIR extinction map depicts eight subregions cm−3) has been reported in the S235 complex (Evans & (East 1, East 2, North, North-West, Central W, Cen- Blair 1981; Kirsanova et al. 2014). Nordh et al. (1984) tral E, South-West, and South; having A > 8 mag), V noted that the density structure of the S235 cloud is and five of them seem to be located in an almost non-homogeneous with high density clumps engulfed in regularly spaced manner along the sphere-like shell a medium with lower density. Kirsanova et al. (2014) surrounding the ionized emission. adopted the value of volume density of 7000 cm−3 to • Bolocam dust continuum emission at 1.1 mm is compute the age of the S235 Hii region. In order to found toward all the eight subregions and provides compare the previous results, following Kirsanova et al. more insight of the dense and cold regions of the S235 (2014), we also adopt the same value of volume density complex,whichislackinginthepublishedsub-mmmaps (i.e. n0 = 7000 cm−3) in the present work. Substituting in the literature. Nuv,n0,andanaverageradiusoftheHiiregionR≈1.67 • Bolocam clump masses increase as one moves away pc, we find tdyn ∼1 Myr. Our estimated tdyn value is in fromthelocationoftheionizingstar. Thischaracteristic agreementwiththeworkofKirsanovaetal.(2014). The is found only for those clumps which are located near tfrag is computed to be about 0.85 Myr, 1.1 Myr, and the edges of the shell. 1.3 Myr, respectively, for a = 0.2, 0.3, and 0.4 km s−1. • The position-velocity analysis of 12CO and 13CO s WefindthatthedynamicalageoftheHiiregion(t ≈ emissions depicts an almost semi-ring like structure, dyn 1 Myr) is comparable with the fragmentation time scale indicating the signature of an expanding Hii region. (tfrag ≈ 0.85–1.3 Myr) for n0 = 7000 cm−3 and as = • The pressure calculations (PHII, Prad, and Pwind) 0.2–0.4 km s−1. In general, the average ages of Class I indicate that the photoionized gas associated with andClassIIYSOsare∼0.44Myrand∼1–3Myr(Evans the S235 Hii region can be considered as the major contributor for the feedback mechanism in the S235