ATLASGAL-selected massive clumps in the inner Galaxy: VIII. Chemistry of photodissociation regions
W.-J. Kim, F. Wyrowski, J. S. Urquhart, J. P. Pérez-Beaupuits, T. Pillai, M. Tiwari, K. M. Menten
AAstronomy & Astrophysics manuscript no. mspdr_wjkim c (cid:13)
ESO 2020October 1, 2020
ATLASGAL-selected massive clumps in the inner Galaxy:
VIII. Chemistry of photodissociation regions
W.-J. Kim , , F. Wyrowski , J. S. Urquhart , J. P. Pérez-Beaupuits , T. Pillai , , M. Tiwari , , and K. M. Menten Max-Planck-Institut für Radioastronomie, Auf dem Hügel 69, 53121 Bonn, Germany Present address: Instituto de Radioastronomía Milimétrica, Avenida Divina Pastora 7, 18012 Granada, Spaine-mail: [email protected] School of Physical Sciences, University of Kent, Ingram Building, Canterbury, Kent CT2 7NH, UK European Southern Observatory, Alonso de Córdova 3107, Vitacura Casilla 7630355, Santiago, Chile Institute for Astrophysical Research, 725 Commonwealth Ave, Boston University Boston, MA 02215, USA Department of Astronomy, University of Maryland, College Park, MD 20742, USAReceived 29 July 2020 / Accepted 28 September 2020
ABSTRACT
Aims.
We study ten molecular transitions obtained from an unbiased 3 mm molecular line survey using the IRAM 30 m telescopetoward 409 compact dust clumps identified by the APEX Telescope Large Area Survey of the Galaxy (ATLASGAL) to understandphotodissociation regions (PDRs) associated with the clumps. The main goal of this study is to investigate whether the abundances ofthe selected molecules show any variations resulting from the PDR chemistry in di ff erent clump environments. Methods.
We selected HCO, HOC + , C H, c-C H , CN, H CN, HC N, and HN C as PDR tracers, and H CO + and C O as densegas tracers. By using estimated optical depths of C H and H CN and assuming optically thin emission for other molecular transitions,we derived column densities of those molecules and their abundances. To assess the influence of the presence and strength of ultra-violet radiation, we compare abundances of three groups of the clumps: H ii regions, infrared bright non-H ii regions, and infrared darknon-H ii regions. Results.
We detected C O, H CO + , C H, c-C H , CN and HN C toward most of the observed dust clumps (detection rate > CN is also detected with a detection rate of 75%. On the other hand, HCO and HC N show detection rates of 32%and 39%, respectively, toward the clumps, which are mostly associated with H ii region sources: detection rates of HCO and HC Ntoward the H ii regions are 66% and 79%. We find that the abundances of HCO, CN, C H, and c-C H decrease as the H columndensity increase, indicating high visual extinction, while those of high density tracers (i.e., H CO + and HC N) are constant. Inaddition, N (HCO) / N (H CO + ) ratios significantly decrease as H column density increase, and in particular, 82 clumps have X (HCO) (cid:38) − and N (HCO) / N (H CO + ) (cid:38)
1, which are the indication of far-ultraviolet (FUV) chemistry. This suggests the observed HCOabundances are likely associated with FUV radiation illuminating the PDRs. We also find that high N (c-C H ) / N (C H) ratios foundfor H ii regions having high HCO abundances ( (cid:38) − ) are associated with more evolved clumps with high L bol / M clump . This trendmight be associated with gain-surface processes, which determine initial abundances of these molecules, and time-dependent e ff ects inthe clumps corresponding to the envelopes around dense PDRs and H ii regions. In addition, some fraction of the measured abundancesof the small hydrocarbons of the H ii sources can be the result of the photodissociation of PAH molecules. Key words. astrochemistry – surveys – ISM: molecules – (ISM):H ii regions – (ISM): photo-dominated regions (PDR)
1. Introduction
Newly born high-mass stars ( > (cid:12) ) dramatically a ff ect theirenvironment in various ways, e.g. by powerful outflows, cre-ation of ionized gas regions via ultraviolet (UV) radiation, andstellar winds, etc. (Zinnecker & Yorke 2007). Their intense UVradiation impinging on the surrounding molecular gas, createsphotodissociation regions (PDRs, e.g., Tielens & Hollenbach1985; Tielens 2013). The UV inducing photo-chemistry andmany other factors, such as the geometric structure of moleculargas, volume density of H , and turbulence, control the forma-tion and destruction of molecules leading to a complex chem-istry in PDRs. So far, most detailed studies of PDRs have con-centrated on a few nearby regions (e.g., the Horsehead nebulaand the Orion Bar; Rodriguez-Franco et al. 1998; Teyssier et al.2004; Pety et al. 2005; Gerin et al. 2009; Cuadrado et al. 2015).According to the detailed molecular excitation studies ofHogerheijde et al. (1995) and Jansen et al. (1995) of the Orion Bar, the PDR layer consists of at least two components: a low-density inter-clump medium ( n (H ) ≈ × cm − ) and high-density clumps ( n (H ) ≈ × cm − ). More recent observa-tional studies have indicated the presence of dense PDRs be-tween the ionized and neutral gas in the Orion Bar and Mono-ceros R2 (Mon R2) (Rodriguez-Franco et al. 1998; Rizzo et al.2005). Especially toward Mon R2, high H densities ( > × cm − ) were found in its PDRs (Rizzo et al. 2005), which areilluminated by intense UV radiation ( G = × ; Rizzo et al.2003) unlike low- and moderate-UV illuminated PDRs (e.g.,the Horsehead nebula) ( G = − (cid:46) .
001 pc). In the other ex-treme, PDRs surrounding compact H ii or ultracompact (UC H ii )regions in molecular clumps have di ff erent physical conditions The UV radiation field is in Habing untits, which are a measure ofthe average far-UV interstellar radiation field; G = . × − erg cm − s − . Article number, page 1 of 19 a r X i v : . [ a s t r o - ph . GA ] S e p & A proofs: manuscript no. mspdr_wjkim than the aforementioned “weak” PDRs, such as that associatedwith the Horsehead nebula. Their high densities suggest thatUC H ii regions could be pressure-confined by the surroundinghigh-density neutral gas (Rizzo et al. 2005). The observations to-ward Mon R2 provided a good understanding of a particular typeof dense PDR illuminated by high-UV radiation ( h ν > ff erent lay-ers of the PDRs.In our previous studies (Kim et al. 2017, 2018), we iden-tified H ii regions toward compact dust clumps selected fromthe ATLASGAL compact source catalogues (Contreras et al.2013; Urquhart et al. 2014) using hydrogen radio recombinationlines (RRLs) at (sub)millimeter wavelengths and radio contin-uum surveys. The clumps with RRL detections are associatedwith more energetic UV radiation fields from mostly O-stars,whereas the clumps with only radio continuum associations mayhave weaker UV fields provided by early B-stars. Many of theremaining clumps without those signposts have embedded mas-sive young stellar objects (MYSOs) that provide even weakerradiation fields. Finally, some of the clumps without embeddedobjects, which are classified as infrared (IR) dark clumps, mayonly be a ff ected by cosmic-ray in their interior or by externalradiation fields from nearby star-forming complexes. The AT-LASGAL compact source catalogues, therefore, allow us to in-vestigate PDRs produced in a wide range of UV field strengthsand physical conditions.In general, dense and compact PDRs occur in complex star-forming regions that are embedded in their parental molecularclouds. It is critical to find molecular tracers whose abundanceis primarily driven by PDR chemistry. Infrared diagnostics pro-vide direct information about PDRs, but they are not easily ob-servable from the ground and accessible in the presence of highextinction. Alternatively, several molecular species are linked toUV photochemistry without su ff ering high extinction and alsoo ff er additional information on the velocity fields that allow us toinvestigate the kinematics and turbulence in PDRs. Rizzo et al.(2005) identified two groups of molecules as PDR tracers: thefirst group is related to the surface layers of the PDR (visual ex-tinction 2 mag < A V < + and HOC + ) and small hydrocarbons(e.g., C H and c-C H ; Sternberg & Dalgarno 1995). The secondgroup is found in deeper parts of PDRs (5 mag < A V <
10 mag)where they are less exposed to the UV field (i.e., low UV, 6 eV < h ν < H and CN) are considered to belong to this group.In this paper, we analyze 3 mm line survey data taken withthe Institut de Radioastronomie Millimétrique (IRAM) 30 mtelescope toward 409 ATLASGAL dust clumps. These observa-tions contain a large number of rotational transitions of molecu-lar species. Based on previous observational studies (e.g., Rizzoet al. 2005; Boger & Sternberg 2005; Gerin et al. 2009) we se-lected eight molecules that are regarded as typical PDR tracers:HCO, HOC + , C H, c-C H , H CN, HC N, HN C and CN. Inaddition, C O and H CO + were chosen as general probes ofcolumn density and dense gas in the dust clumps. The selectedmolecular transitions are given in Table 1.The structure of the paper is as follows. The selectedmolecules are described in Section 2. The observations, source types, and data reduction are explained in Section 3. Detectedmolecular lines and detection rates are presented in Section 4, in-cluding a description of several sources with CN self-absorptionline profiles. The estimated column densities and abundances ofthe selected molecules are given in Section 5 along with a com-parison with H column densities. Column density ratios of someof the selected molecular lines (i.e., HCO, H CO + , C H, and c-C H ) and their correlations are discussed in Section 6. Finally,we summarize our main results in Section 7.
2. Line selection O As a rare isotopologue of carbon monoxide, C O is an excel-lent tracer of column density in star-forming regions. Chemi-cally, it is relatively stable compared to other species. C O, likeall carbon monoxide isotopologues, can be a ff ected by depletionon dust grains surfaces, but this only occurs in the innermost,densest and lowest temperature cloud core regions (e.g., Caselliet al. 1999; Bacmann et al. 2002; Giannetti et al. 2014). C O is,therefore, used as a reference molecule to measure the relativeabundances of the other molecules in this paper. CO + , HOC + and HCO H CO + is a rare isotopologue of HCO + (formylium) and mostlyoptically thin in molecular clouds. HCO + is a high-density tracer(critical density ∼ cm − for HCO + ( J = −
0) with the col-lisional rate at 20 K taken from the Leiden Atomic and Molec-ular Database (Schöier et al. 2005)) that shows enhanced abun-dances in regions of higher fractional ionization and toward out-flows where shock-generated radiation fields are present (Rawl-ings et al. 2000). Hydroxy methylidyne (HOC + ) is a reactive ionthat is almost exclusively related to regions with a high ioniz-ing flux (either PDRs or X-ray-dominated regions; Fuente et al.2003; Rizzo et al. 2003, 2005). However, there are few reporteddetections of this rare molecule. HOC + has been detected to-ward the UC H ii of Mon R2 together with the other reactive ionCO + (Rizzo et al. 2003). The Formyl radical (HCO) has beenmostly studied in PDRs toward the Orion Bar (Schilke et al.2001), the Horsehead nebula (Gerin et al. 2009), and even thestarburst galaxy M82 (García-Burillo et al. 2002). In the Horse-head nebula, its emission strongly correlates with polycyclic aro-matic hydrocarbon (PAH) and C H emission, and its measuredabundance reaches X (HCO) (cid:39) − × − (Gerin et al. 2009). Inaddition, the HCO / H CO + column density ratio in PDRs withH ii regions (e.g., Schenewerk et al. 1988; Schilke et al. 2001;Gerin et al. 2009) shows higher values than in regions withoutH ii regions or any other signpost of star formation. H and c-C H The ethynyl radical (C H) and cyclopropynylidyne (c-C H ) aresmall hydrocarbon species and well known for their associationwith PAH molecules (Rizzo et al. 2005). Enhanced abundancesof these small hydrocarbons were found in PDRs with intenseUV fields (Fuente et al. 2003; Teyssier et al. 2004; Fuente et al.2005; Pety et al. 2005; Rizzo et al. 2005; Ginard et al. 2012).Their spatial distribution follows PAH emission in general, but itis slightly di ff erent in extent (Pilleri et al. 2013; Cuadrado et al.2015). In the PDRs of Mon R2 and the Orion Bar, the abundanceof C H is constant for a broad range of incident UV radiationstrength, but the abundance of c-C H was appeared in usage Article number, page 2 of 19im et al.: PDRs in dust clumps
Table 1: The observed molecular transitions.
Species Transition Frequency g u E up µ S † ij Relative I b A ij [MHz] [K] [Debye] [s − ]C O J = − . × − HCO N K a , K c = , − , , J = / − / , F = − . × − HCO N K a , K c = , − , , J = / − / , F = − . × − HCO N K a , K c = , − , , J = / − / , F = − . × − HCO N K a , K c = , − , , J = / − / , F = − . × − H CO + J = − . × − HOC + J = − . × − C H N = − , J = / − / , F = − . × − C H N = − , J = / − / , F = − . × − C H N = − , J = / − / , F = − . × − C H N = − , J = / − / , F = − . × − C H N = − , J = / − / , F = − . × − C H N = − , J = / − / , F = − . × − c-C H J K a , K c = , − , . × − H CN J = − , F = − . × − H CN J = − , F = − . × − H CN J = − , F = − . × − HC N J = − . × − HN C J = − . × − CN N = − , J = / − / , F = / − / . × − CN N = − , J = / − / , F = / − / . × − CN N = − , J = / − / , F = / − / . × − CN N = − , J = / − / , F = / − / . × − Notes. ( † ) S ij is the line strength taken from the JPL and CDMS catalogs. (b) Expected relative intensities ( S ij / (cid:80) S ij ), assuming that lines areoptically thin ( τ ν < g u is the statistical weight of the upper state level,and E u is the energy of the upper level of the selected transition. µ is the permanent dipole moment of the species. A ul is the Einstein coe ffi cientfor spontaneous emission. Table 2: List of observed sources.
ATLASGAL RA. Dec. Dist. T dust Type Classification † Commentsname α (J2000) δ (J2000) (kpc) (K)AGAL006.216 − − ii
24 darkAGAL008.049 − − ii IR bright or H ii AGAL008.671 − − ii IR bright or H ii no C O, CN, H CN, HC N dataAGAL008.684 − − ii
24 dark no C H dataAGAL008.706 − − ii IR bright or H ii Notes. ( † ) The temperature of dust, distance, and classification are taken from Urquhart et al. (2018). This table is a fraction of the list of all theobserved sources (409). The full table is available at CDS via anonymous ftp. of high in low UV PDRs (Cuadrado et al. 2015). In addition,C H shows higher column densities in high-UV irradiated PDRs,whereas column densities of c-C H are predicted to decreasefor such PDRs (Cuadrado et al. 2015). C, H CN, HC N, and CN
Hydrogen cyanide (HCN) and hydrogen isocyanide (HNC) areused as tracers of dense gas within molecular clouds (e.g.,Vasyunina et al. 2011). The column density ratio HCN / HNC sig-nificantly depends on the temperature of the cloud as has beenfound in the Orion molecular cloud and several high-mass star-forming regions (Goldsmith et al. 1986; Schilke et al. 1992; Jinet al. 2015). The cyanide radical (CN) at millimeter wavelengthsis often used as a probe of dense gas and PDRs in the Galactic in-terstellar medium (Rodriguez-Franco et al. 1998; Boger & Stern-berg 2005). Previous observational and theoretical studies haveshown that CN abundances are enhanced in PDRs (Fuente et al. 1993; Rodriguez-Franco et al. 1998). In particular, the observa-tions toward the Orion Bar PDRs showed that the CN emissionis located between the molecular ridge and the ionization fronts(Jansen et al. 1995; Simon et al. 1997; Rodriguez-Franco et al.1998).
3. Observations, source type, and data reduction
The molecular line data were taken from unbiased spectral linesurveys covering a frequency range of ∼ −
115 GHz (see Csen-geri et al. 2016 for details). They contain a number of molec-ular lines with di ff erent rotational energy levels. The line sur-vey is observed with the Eight MIxer Receiver (EMIR) of theIRAM 30 m telescope (Project IR: 181-10 and 037-12) toward409 clumps selected from the ATLASGAL compact source cat- http: // / IRAMES / mainWiki / EmirforAstronomersArticle number, page 3 of 19 & A proofs: manuscript no. mspdr_wjkim
Fig. 1: Spectra of C H hyperfine lines toward two examplesources (AGAL015.013 − + ∼ − . The largest beam size in the 3 mm atmospheric win-dow is ∼ (cid:48)(cid:48) . We applied the forward e ffi ciency ( η l = ffi ciency ( η MB = T ∗ a , to the main beam brightness temperatures, T MB .Here we divide the ATLASGAL sample into two groups, whichare H ii and non-H ii regions, according to the presence of H ii re-gions based on detection of mm-RRL and radio continuum emis-sion (see Kim et al. 2017 for details). Table 2 displays infor-mation about the observed sources and the clump classificationfrom Urquhart et al. (2018). This classification is used to furtherdivide the group of non-H ii region sources into IR bright non-H ii and IR dark non-H ii regions. We utilized the CLASS software of the GILDAS package (Pety2005) for data reduction of the molecular line data and theWEEDS package within CLASS for line identification. The hy-perfine structures of the C H ( N = −
0) and H CN ( J = − of the CLASS software. The HFSfit method assumes that all hyperfine components share a sin-gle excitation temperature and local thermodynamic equilibrium(LTE). The HFS fit yields the radial velocity ( (cid:51) LSR ), the linewidth ( ∆ (cid:51) ) at the full-width of half maximum, and the total op-tical thickness ( τ tot ). The line parameters of the other molecularlines were fitted with multi-Gaussian components if one Gaus-sian fit was insu ffi cient.
4. Results
Since the local standard of rest (LSR) velocities of the sourceswere unknown when the observations were carried out, theywere observed with a velocity of 0 km s − . Consequently, dueto cases with a significantly di ff erent LSR velocity, the numbersof observed and detected lines in the sources vary slightly, and https: // / IRAMFR / GILDAS / doc / html / class-html / class.html https: // / IRAMFR / GILDAS / doc / html / weeds-html / node10.html http: // / IRAMES / otherDocuments / postscripts / classHFS.ps Fig. 2: From top to bottom, spectra of HCO, CN, c-C H ,HC N, H CN, HN C, H CO + , and C O toward the same ex-ample sources as in Fig. 1. Hyperfine structure fitting was usedfor H CN, and Gaussian profiles for the other molecular lines.The green lines represent Gaussian and HFS fitting results, andthe red lines indicate the systemic velocity. HCO (1 −
0) was notdetected toward AGAL045.474 + ∼ −
8% of the detected sources),we only selected a component with a common velocity in all de-tected lines. Figures 1 and 2 show the spectral lines and fitted lineprofiles of AGAL015.013 − + ff ect is stronger for the brightest three componentsthan for the weakest component, which is considered to be op-tically thin. To reduce optical depth e ff ects and to avoid non-LTE excitation, we used the weakest component of CN ( N JF = / / − / /
2) for all analysis presented in this work. Inaddition to these hyperfine anomalies, some sources show self-absorption in the three brightest components, which will be dis-cussed in Section 4.2. The four HCO hyperfine lines are con-siderably weaker than the other molecular lines, and thus only afew sources have all the four HCO hyperfine components clearly
Article number, page 4 of 19im et al.: PDRs in dust clumps
Table 3: The line parameters of hyperfine lines of C H ( N = −
0) and H CN ( J = − ATLASGAL name Line (cid:51)
LSR ∆ (cid:51) τ tot T MB rms(km s − ) (km s − ) (K) (K)AGAL008.671 − H 35.03 ± ± ± − H 9.55 ± ± ± CN 9.70 ± ± ± − H 14.30 ± ± ± CN 14.02 ± ± ± Notes.
We only provide here the first three of the whole table which is available at the CDS via anonymous ftp. T MB is the peak intensity of thebrightest component of a given molecular hyperfine lines. The listed C H peak intensity indicates the transition
NJF = / − / CN is the transition JF = − Table 4: Gaussian line parameters of C O, HCO, c-C H , CN, HC N, and HN C. ATLASGAL Line Area (cid:51)
LSR ∆ (cid:51) T MB rmsName (K km s − ) (km s − ) (km s − ) (K) (K)AGAL008.671 − ± ± ± CO + ± ± ± H ± ± ± C 10.12 ± ± ± − O 17.65 ± ± ± ± ± ± CO + ± ± ± H ± ± ± ± ± ± N 0.35 ± ± ± C 3.13 ± ± ± Notes.
For multiple velocity components, we only tabulate line parameters of the component detected in all molecular species. The full table isavailable at the CDS via anonymous ftp.
Fig. 3: The self-absorption of CN lines ofAGAL011.936 − + N J = / − / CN( N = − , J = / − / , F = − , F = −
2) transitions.The red vertical lines indicate the systemic velocities measuredby H CO + lines.detected. The brightest component ( N JF = / − / / N), the detection rates andGaussian fits of HCO are based on stacked spectra. The stackedHCO lines are scaled to the brightest component transition. For the stacking, each HFS component is weighted and scaled byits relative intensity. The stacking increased the S / N by a fac-tor of 2. As shown in Fig. 2, the three hyperfine components ofH CN observed toward the clumps hosting H ii regions are of-ten blended. As a result, the measured line parameter of H CNtoward some sources are associated with high uncertainties, andwe will discuss this in later sections.
Self-absorption features are found in the three strongest CN lines( F = / − / , F = / − / F = / − /
2) to-ward 25 H ii and 15 non-H ii region clumps. The features in-dicate the clumps have inhomogeneous structures in tempera-ture and density, and, because of such structures, di ff use andlow-temperature gas in the foreground absorbs the emitted ra-diation from CN molecules along the line of sight. Figure 3displays two example sources showing self-absorption features.These line profiles represent di ff erent kinds of gas motionswithin the clumps. All CN self-absorption spectral lines plotsare available at CDS via anonymous ftp. In the left panel of Fig.3, AGAL011.936 − + Article number, page 5 of 19 & A proofs: manuscript no. mspdr_wjkim
Table 5: Detection rates of the observed molecules.
Molecule C O HCO H CO + Source type All H ii non-H ii All H ii non-H ii All H ii non-H ii rms (K) 0.050 − − − − − − Number of Observed source 385 102 283 405 103 302 403 103 300Number of Detected source 385 102 283 131 69 62 400 100 300Detection Rate (%) 100 100 100 32 66 20 99 97 100Molecule C H c-C H CNSource type All H ii non-H ii All H ii non-H ii All H ii non-H ii rms (K) 0.039 − − − − − − Number of Observed source 403 102 301 404 102 302 386 102 284Number of Detected source 399 102 297 403 102 301 363 102 261Detection Rate (%) 99 100 98 99 100 99 94 100 91Molecule HN C H CN HC NSource type All H ii non-H ii All H ii non-H ii All H ii non-H ii rms (K) 0.040 − − − − − − Number of Observed source 405 103 302 403 101 302 405 103 302Numnber of Detected source 384 94 290 305 94 211 161 82 79Detection Rate (%) 94 91 96 75 93 69 39 79 26
Notes.
For all observed sources, the rms of the line survey varies with observing frequency and is given for each molecule. C O H C O + C H c - C H C N H N C H C N H C N H C O P e r c e n t a g e o f D e t e c t i o n ( % ) HIInonHII
Fig. 4: Detection rates of typical PDR tracer molecules andC O as a reference molecule. Binomial statistics was used forestimating uncertainties marked with the red and blue error bars.
Table 5 shows detection rates for the whole sample, includingH ii and non-H ii regions. The detection of molecules was basedon applying a 3 σ rms threshold. We detect C O lines towardall the observed clumps (100%). This result shows that the AT-LASGAL dust clumps trace the high column density parts ofmolecular clouds. In addition, we also find high detection ratesfor H CO + (99%), C H (99%), c-C H (99%), CN (94%),and HN C (94%). We find a reasonably high detection ratefor H CN (75%). Even the rare isotopologue HC N is de-tected with a rate of 39%. However, the other cation, HOC + was detected in only five clumps (below 3 σ with an rms ∼ + frac-tional abundances toward typical PDRs, including the Orion Barand Mon R2, are in a range of ∼ × − − × − (Savage& Ziurys 2004). In addition, its large dipole moment and A ij im- M e a n C O H C O H C O + C H c - C H C N H C N H C N H N C M e d i a n C O H C O H C O + C H c - C H C N H C N H C N H N C L i n e w i d t h [ k m s ] Fig. 5: Mean, median, a standard deviation ( σ ) of line widths ofthe selected molecular lines. The σ shows the source to sourcevariation of the widths. The error bars for the mean and medianvalues are based on typical Gaussian fit errors with 2 σ . The errorbars of the standard deviation of line widths is the standard devi-ation error. Di ff erent colors show three subgroups of our sample,which are H ii in orange, IR bright in green, and IR dark non-H ii regions in black .ply the critical density of this molecule is probably higher thanothers with small dipole moment if collisional rates are similar.These tentative detections of HOC + are excluded from the fol-lowing statistical analyses.Figure 4 shows the detection rates of the nine molecules to-ward H ii (orange bars) and non-H ii (cyan bars) region sources.Di ff erences in detection rates for some molecular transitions(i.e., H CN, HC N, and HCO) toward the two groups aredistinct, with the H ii regions showing higher detection ratesthat can be related to higher column densities. According toKolmogorov-Smirnov tests and median values of H columndensities ( N (H )) for the ATLASGAL clumps, the median valueof the N (H ) determined for the H ii region sources is not signifi-cantly di ff erent from the value found for non-H ii region sources. Article number, page 6 of 19im et al.: PDRs in dust clumps
HCO
HIInon-HII 0 2 4 6 8 10 12 14 H CO + HIInon-HII0 2 4 6 8 10 12 140.02.55.07.510.012.515.0 C H HIInon-HII 0 2 4 6 8 10 12 14 c-C H HIInon-HII0.02.55.07.510.012.515.0 CN HIInon-HII 0 2 4 6 8 10 12 14 HC N HIInon-HII0 2 4 6 8 10 12 140.02.55.07.510.012.515.017.5 H CN HIInon-HII 0 2 4 6 8 10 12 14 HN C HIInon-HII L i n e w i d t h [ k m s ] Line width (C O) [km s ] Fig. 6: The orange circles indicate ATLASGAL clumps containing H ii regions and the cyan circles represent clumps without thepresence of H ii regions, which consist of IR bright non-H ii and dark non-H ii regions. The black dashed lines indicate equality. Theerror bars are obtained from Gaussian or HFS fits. The red and blue ellipses are covariance error ellipses with 1 σ (inner ellipse) and2 σ (outer ellipse) for the two groups.The N (H ) are taken from König et al. (2017), and Urquhartet al. (2018) and were derived by analysing dust continuum emis-sion with adopting dust opacity of 1.85 cm g − and µ H of 2.8that is mean molecular weight of the interstellar gas for a hy-drogen molecule. The N (H ) values of both H ii and non-H ii re-gions are mostly in a range of ∼ − × cm − (for H ii re-gions, 1 . × cm − ≤ N (H ) ≤ . × cm − comparedto 1 . × cm − ≤ N (H ) ≤ . × cm − for non-H ii re-gions). Another possibility is a distance e ff ect leading to weakermolecular lines for more distant sources. However, this can beexcluded because of the detection rates for all the molecules fora distance limited sample (3 − To investigate whether di ff erent molecular lines are originatingfrom the same gas or show di ff erent amounts of turbulence, we compared the mean, median, and standard deviations of the mea-sured line widths. Figure 5 shows these values with their uncer-tainties. The molecular line widths increase from less evolvedclumps to more evolved clumps. Such a trend, broader molec-ular lines detected toward clumps in evolved stages, was alsofound in ammonia (Wienen et al. 2012; Urquhart et al. 2013)and H CO + (Kim et al. 2017) data. The HCO line widths haveconsiderable higher fit uncertainties due to the low line intensi-ties (or lower S / N). Nevertheless, comparing the IR bright non-H ii and H ii regions, which have small fitting uncertainty, we findthat broad HCO lines are often associated with H ii regions. TheHCO, H CN, and HC N lines observed toward IR dark non-H ii regions seem to be broader than the other molecular lines,but their errors are also obviously larger than those of the othersource groups. The third panel from the top of Fig. 5 exhibitsthe standard deviations ( σ ) of line width and its associated un-certainty. Except for C H, CN, and HC N, the other molecular
Article number, page 7 of 19 & A proofs: manuscript no. mspdr_wjkim I (C H) / I (c-C H ) 0.00.20.40.60.81.0 C u m u l a t i v e D i s t r i b u t i o n HIIIR bright non-HIIIR dark non-HII 10 I (CN) / I (c-C H ) 0.00.20.40.60.81.0 C u m u l a t i v e D i s t r i b u t i o n HIIIR bright non-HIIIR dark non-HII 10 I (CN) / I (HN C) 0.00.20.40.60.81.0 C u m u l a t i v e D i s t r i b u t i o n HIIIR bright non-HIIIR dark non-HII
Fig. 7: Cumulative distribution plots of integrated intensity ratios of C H to c-C H (left panel) and CN to c-C H (middle panel)and to HN C (right panel). The orange curves are H ii regions. The green and black curves indicate IR bright non-H ii and IR darknon-H ii regions. Table 6: Column density of molecules in units of cm − . ATLASGAL C O HCO H CO + C H c-C H CN HC N H CN HN CName × × × × × × × × × AGAL006.216 − · · · − · · · · · · − · · · · · · · · · · · · − · · · · · · − · · · · · · Notes.
This table shows a portion of the full table that is available at the CDS via anonymous ftp. Not all lines were observed in all sources, andthus some column densities are missing in Table 6 (see Table 2 for missing observational data). lines have broader distributions toward the H ii region sourcescompared with those toward non-H ii region sources.Figure 6 shows scatter plots comparing line widths of thelines from selected molecules with C O line widths. Covari-ance error ellipses are overlapped over the data points to visu-alize two-dimensional Gaussian distributed data . The ellipsesand data points of H CO + and C O display pronounced cor-relations and well align on the equality line. On the other hand,the other molecular lines show some deviations from the equal-ity lines. Some sources have remarkably broader line widths inH CN with small uncertainty. Such broadening can be a resultof high optical depth, leading to a blending of the HFS lines.If any internal turbulence within clumps causes these broad linewidths, we expect to find such a trend also in HC N. But nosources with notably broader line widths of HC N are found.To identify similarities in the line widths between the clumpgroups, we performed KS tests for the molecular line widths.The null hypotheses are rejected with small p − values less than3 σ ( p-value (cid:28) . > ii regions than for the non-H ii regions, consistent withthe results in Fig. 5. One possible cause for this may be distancebias. To test this, we compared molecular line widths and dis-tances of the clumps (see Fig. B.1), but we do not find signif-icant correlations. The distance e ff ect cannot, therefore, be themain reason for the broad line widths of the H ii region group.For the H ii region group, we compared the line widths of molec-ular lines with RRL line widths. If the molecular lines are mainlyemitted in PDRs, we might expect to find some correlations be-tween them due to the dynamical interaction between ionized In the two-dimensional case, 1 σ and 2 σ show 39.4 % and 86.5 %confidences, respectively. and photodissociated gas. However, we could not find any sig-nificant correlations.Integrated intensity ratios of molecular lines have been usedas a chemical clock tracing di ff erent evolutionary sequences ofhigh-mass stars in molecular clumps (e.g., Rathborne et al. 2016;Urquhart et al. 2019). With the assumption that observed linesare optically thin, integrated intensity ratios gives us approxi-mate abundance ratios of molecular lines compared with eachother. Figure 7 shows cumulative distributions of the integratedintensity ratios of C H, c-C H , and CN (i.e., C H / c-C H in theleft panel, CN / c-C H in the middle panel, and CN / HN C inthe right panel). We performed KS tests for the ratio di ff erencesamong the three groups and found that we can reject the null hy-potheses for the similarities of their ratios with small p − values (cid:28) σ . The H ii regions show brighter ratios of those molecularlines than the other two groups. The well separated cumulativedistributions of C H / c-C H show that the relative intensities ofthe small hydrocarbons increase with the evolution of high-massclumps. This trend is also found for ATLASGAL sources in thesouthern hemisphere (Urquhart et al. 2019).
5. Analysis
To compare abundances of the selected molecules with respectto H for di ff erent environments in dust clumps, as the first step,we estimated column densities of the molecules. Since the ob-served data only contain a single transition per molecule, sev-eral assumptions are required to derive column densities ( N ): 1)the molecular lines are emitted under LTE condition because thecritical densities of the observed molecular transitions are lower Article number, page 8 of 19im et al.: PDRs in dust clumps r = 0.3 (HII) & 0.2 (non-HII) 11.011.512.012.513.013.5H CO + r = 0.8 (HII) & 0.8 (non-HII)14.014.515.015.5 C H r = 0.7 (HII) & 0.6 (non-HII) 12.012.513.013.514.0c-C H r = 0.7 (HII) & 0.7 (non-HII)13.514.014.515.015.5 CN r = 0.6 (HII) & 0.7 (non-HII) 11.011.512.012.513.013.5HC N r = 0.9 (HII) & 0.8 (non-HII)22.0 22.5 23.0 23.5 24.0 24.511.512.012.513.013.514.0 H CN r = 0.7 (HII) & 0.8 (non-HII) 22.0 22.5 23.0 23.5 24.0 24.5 11.512.012.513.013.5HN C r = 0.7 (HII) & 0.8 (non-HII) L o g N ( m o l ) [ c m ] Log N (H ) [cm ] Fig. 8: Column density of a given molecule as a function of H column density toward H ii and non-H ii regions with IR dark non-H ii regions superposed with black triangles. The ellipses show similar confidence contours as in Fig. 6. The r parameter indicatesPearson correlation coe ffi cients for H ii and non-H ii regions.than the H densities ( > cm − ) in high-mass star-forming re-gions; 2) at these high densities, the gas and dust temperaturesare in equilibrium; 3) the observed line emission is consideredto be optically thin. For optically thin emission, the total columndensity is given by, N thintot = (cid:32) πν c A ul (cid:33) (cid:32) Q ( T ) g u (cid:33) exp (cid:16) E u k B T ex (cid:17) exp (cid:16) h ν k B T ex (cid:17) − × J ν ( T ex ) − J ν ( T bg )] (cid:90) T MB f d (cid:51) , (1)where ν is the frequency of a selected molecular transition, A ul is the Einstein coe ffi cient for spontaneous emission, and g u isthe statistical weight of the upper state level. Q ( T ) and k B arethe partition function and Boltzmann constant, respectively. E u is the energy of the upper level of the selected transition. T MB is a main beam temperature of an observed source, and f is thebeam filling factor, which is the fraction of the beam filled by the source. Here we consider calculated column densities are beam-averaged values, and the medium is spatially homogeneous andlarger than the size of the beam. These assumptions return thebeam filling factor as 1. J ν ( T ) is the Rayleigh-Jeans temperature, J ν ( T ) ≡ h ν/ k B exp( h ν/ k B T ) − . T ex is the excitation temperature whichwe approximate with the dust temperature, and T bg is the back-ground emission temperature assumed to be 2.7 K that is the cos-mic microwave background radiation.Since the optical depths of the C H and H CN lines wereobtained by HFS fitting, correct column densities of these lineswere estimated by multiplying Eq. (1) by a factor of τ/ (1 − e − τ )for the non-optically thin case. For the other molecular lines,we utilized the optically thin case of Eq. (1), which gives us alower limit of column density in case the lines become opticallythick. We note that in some cases we failed to measure a τ tot with a fitting uncertainty below 50%, either because some ofC H and H CN hyperfine components are not separated su ffi -ciently (e.g., the H CN spectral line of AGAL045.474 + Article number, page 9 of 19 & A proofs: manuscript no. mspdr_wjkim f ( M o l ) H C O H C O + C H c - C H C N H C N H C N H N C X ( M o l ) Fig. 9: The abundance of the selected molecular species.
U pper :Relative abundances ( f ) normalized by C O column density.
Lower : Fractional abundances ( X ) with respect to H . The redcircles, cyan squares, and black triangles indicate H ii , IR brightnon-H ii , and IR dark non-H ii regions, respectively. The error barsshow full ranges of fractional abundances for each molecule.in Fig. 2) or their relative intensities have substantial deviationsfrom the predicted LTE values. Consequently, we only estimatedcolumn densities of these molecules using their optical thicknesswhen the uncertainty of the fit was smaller than 50 %. For the re-maining sources with high τ tot uncertainties, we calculated theircolumn densities as lower limits using the optically thin approx-imation (Eq. (1)) and their values in Table 6, but do not use themfor following statistical analysis. The dust temperatures used inthe analysis were determined by Urquhart et al. (2018) usingthe method for the dust temperature determination from Königet al. (2017). We estimated the partition function for an indi-vidual source and molecule by interpolating the data providedby the Cologne Database for Molecular Spectroscopy (CDMS,Müller et al. 2001) and the Jet Propulsion Laboratory (JPL, Pick-ett et al. 1998) line databases for a given temperature. The par-tition functions obtained from CDMS and JPL take into accountthe hyperfine splitting and ortho- and para-transitions. Overall,our column density measurements agree with those reported inprevious studies (e.g., Sanhueza et al. 2012; Gerner et al. 2014).However, we find no clear di ff erences between the sources rep-resentative of di ff erent evolutionary stages. We used simplifyingassumptions such as the optically thin case, equating the dusttemperature and gas temperature, and neglecting some factorslike size of the emitting region. The measured column densitiesof a portion of all the observed clumps are tabulated in Table 6.Figure 8 shows scatter plots of column density of H versusthat of selected molecules. H CO + and HC N exhibit an ex-cellent correlation (Pearson correlation coe ffi cient, r , ≥ . p − value (cid:28) ii and non-H ii region groups.In addition, C H, c-C H , CN, H CN and HN C also show agood correlation ( r ∼ p − values (cid:28) N (HCO) has poor correlations for both groups( r = ii and r = ii regions) with N (H ).HCO has been found to be enhanced in PDRs rather than in cold,dense molecular regions toward both Galactic objects (Schilkeet al. 2001; Gerin et al. 2009), and extragalactic sources (García-Burillo et al. 2002). The weak correlation supports the hypoth-esis that, where detected, the HCO emission is associated withPDRs on surfaces of the clumps and not their colder material.The column densities of H CO + , c-C H , CN, and HN C for the two groups are significantly di ff erent from each other (KStests gives p − values (cid:28) CN and HC N, we cannot find significant dif-ferences between these groups according to KS tests that yield p − values (cid:38) . N (C H) with evolutionary stages. In our re-sults, N (HN C) is higher toward H ii regions, but this trend wasnot apparent in the results of Sanhueza et al. (2012). For compar-ison, Miettinen (2020) found some evidence that the C2H abun-dance decreases with clump evolution. Overall, the column den-sities of all molecules are higher in H ii regions than in the othergroups. Figure 9 shows abundances relative to C O, f (mol), andfractional abundances relative to H , X (mol), of the selectedmolecules. The symbols and error bars represent median val-ues, and full ranges of the molecular abundances, respectively.The median abundances are listed in Table 7. In Fig. 9, thereare no significant abundance di ff erences among the three sourcegroups. When considering the uncertainties in the calculations,these results are not significantly di ff erent from the results ofGerner et al. (2014), who assumed higher temperatures forevolved sources such as H ii regions.In Fig. 10, we plot the relationships of X (mol) and N (H )of the observed clumps for the HCO, CN, C H and c-C H molecules; these molecules have been found to show good spa-tial correlations with some PDRs such as the Horsehead nebula,the Orion Bar and Mon R2 (Teyssier et al. 2004; Rizzo et al.2005; Gerin et al. 2009; Ginard et al. 2012; Cuadrado et al.2015). In the bottom panels of Fig. 10, the abundances of densegas tracers, X (H CO + ) and X (HC N), are found to be indepen-dent of N (H ). On the contrary, X (HCO) significantly drops forthe clumps with high N (H ) corresponding to a large A V . The X (CN), X (C H) and X (c-C H ) display similar trends as seen inthe X (HCO) plot, but are less pronounced.The X (HCO) measured in this work may also contain acontribution from HCO molecules originating in cold gas re-gions of clumps, not only from PDRs. To analyze this quanti-tatively, we use a simple toy model; it assumes a high abundance(2 . × − ) of HCO ( A V < × − ) in the dense molecular region. The resultingaverage abundance changes as a smooth progression between aPDR and a cold gas region. The gray dashed line on the HCOplot in the most upper left panel of Fig. 10 shows the result ofthe model. It fits our data points (orange, green and black sym-bols) and might indicate that the trends of decreasing abundanceswith N (H ) are caused by abundance jumps from the PDRs intothe cold molecular clouds. Besides, at a given N (H ), X (HCO)values determined for the H ii regions seem to be slightly higherthan toward non-H ii regions. The HCO molecule is referred toas a good tracer of FUV; its emission in the Horsehead nebulawas found in a range 1 . (cid:46) A V (cid:46) . (cid:46) A V (cid:46)
10 (Schilke et al. 2001). Some of the non-H ii region sources (all IR bright non-H ii s and potentially some of IRdark non-H ii s) host early stages of star formation (i.e., YSOs)but they do not yet show centimeter free-free emission or mm-RRLs. Thus, their embedded YSOs cannot provide FUV radia-tion, although shock associated with outflow from them poten-tially could. However, some of the IR dark non-H ii regions are Article number, page 10 of 19im et al.: PDRs in dust clumps
Table 7: Column densities ( N ) and fractional abundances ( X and f ) relative to H and C O for H ii , IR bright and dark non-H ii regions in a ( x ) = a × x . These values are median values for a given molecule.Molecule H ii IR bright non-H ii IR dark non-H ii N [cm − ] X f N [cm − ] X f N [cm − ] X f C O 9.5(15) 1.2( − − − − − − HCO 1.4(13) 1.9( −
10) 1.3( −
03) 8.9(12) 1.6( −
10) 1.1( −
03) 9.0(12) 2.2( −
10) 1.0( − CO + −
11) 2.6( −
04) 1.5(12) 3.8( −
11) 2.6( −
04) 1.2(12) 3.0( −
11) 2.1( − H 5.8(14) 6.4( −
09) 5.3( −
02) 3.5(14) 6.4( −
09) 5.3( −
02) 3.4(14) 6.0( −
09) 5.4( − H −
10) 1.1( −
03) 7.0(12) 1.6( −
10) 1.1( −
03) 6.4(12) 1.6( −
10) 1.1( − −
09) 5.4( −
02) 3.2(14) 7.2( −
09) 5.0( −
02) 2.7(14) 6.4( −
09) 4.8( − N 9.8(11) 1.3( −
11) 1.1( −
04) 8.7(11) 1.1( −
11) 1.0( −
04) 5.6(11) 1.2( −
11) 8.2( − CN 1.5(13) 1.1( −
10) 1.1( −
03) 1.6(13) 9.6( −
11) 1.7( −
03) 4.3(12) 6.0( −
11) 9.3( − C 3.0(12) 3.9( −
11) 2.8( −
04) 2.0(12) 4.6( −
11) 3.3( −
04) 1.8(12) 4.7( −
11) 3.5( − HCO
HIIIR bright non-HIIIR dark non-HIIcold dark cloudHCO peak (Horsehead)DCO + peak (Horsehead) CN10 C H 10 c-C H H CO + HC N X ( M o l ) N (H ) [cm ] Fig. 10: The fractional abundance of HCO, H CO + , C H, c-C H , H CO + and HC N relative to H as a function of N (H ). Theorange indicates H ii regions while green and black symbols are IR bright and dark non-H ii regions. In the most upper left panel, thebrown squares indicate 9 cold dark clouds from Agúndez et al. (2015). The purple and blue star symbols are HCO abundances froma PDR and cold gas region, which are resolved by the observation, in the Horsehead nebula (Gerin et al. 2009). In the top left panel,the gray dashed line represents a simple abundance jump model.possibly influenced by external radiation fields from nearby star-forming complex regions (e.g., IR dark non-H ii regions in M17SW). Therefore, we presume that bright PDRs are only found inthe H ii region sources because of abundant UV radiation fromtheir massive stars. This might be the reason that the H ii regionsshow slightly higher X (HCO) than the others. We also found thatall the ATLASGAL clumps have higher X (HCO) compared tonine local cold dark clouds (brown squares) taken from Agún-dez et al. (2015) with distances of around 140 −
500 pc, at a given N (H ) (the most upper left panel of Fig. 10). On the contrary, X (HCO) (blue star) in the cold gas region in the Horseheadnebula is similar to the abundances in the dust clumps, while X (HCO) (purple star) estimated toward the PDR of the Horse-head is higher than any of the sources in the plot. While Gerinet al. (2009) measured the HCO abundances from separated coldgas and PDRs, the estimated abundances in this study are likelyaverages of cold gas and PDRs in the clumps, similar to the resultof the simple abundance jump model (gray dashed line). Also,we cannot exclude the possibility that the low detection rates ofHCO toward non-H ii regions (especially toward IR dark non-H ii s) are due to a lack of UV radiation and / or a distance e ff ect.C H and c-C H are known as tracers of the surface layersof PDRs exposed to a strong UV field (their highest abundanceshave been found at visual extinctions of 2 mag < A V < Article number, page 11 of 19 & A proofs: manuscript no. mspdr_wjkim typical PDRs; Rizzo et al. 2005; Pety et al. 2005; Ginard et al.2012). These small hydrocarbons have also shown good spatialcorrelations with H ∗ ( ν = −
0) and PAH 8 µ m emission ob-served toward the Horsehead nebula, the Orion Bar, and MonR2 (Teyssier et al. 2004; Pety et al. 2005; Pilleri et al. 2013;Cuadrado et al. 2015). They were, however, also found in UVshielded regions (Beuther et al. 2008). In Fig. 10, their abun-dances as well as X (CN) decrease with a more moderate slopecompared to the X (HCO) plot. The abundance decrease of thefour molecules might be associated with the transitions of thePDRs and the cold envelopes in the clumps.
6. Column density ratios
Molecular abundances are used to quantify the number ofmolecules in dense clumps. However, column density ratios ofchemically related molecules are a more direct way to diagnosechemical enhancement of a specific molecule within a clump.Furthermore, these ratios are insensitive to the unknown beamfilling factor, the H column density and to uncertainties in theassumed physical conditions for molecular species with similarexcitation (Ginard et al. 2012). We here focus on column den-sity ratios of HCO / H CO + and c-C H / C H. Before discussingthese column density ratios for each pair, we will briefly explainthe proposed formation and destruction of HCO and small hy-drocarbons in order to assist in interpreting the column densityratios of these molecules. CO + column density ratio In prior studies (Gerin et al. 2009; Goicoechea et al. 2009), theHCO and H CO + abundances in cold gas were found to beconstant, whereas in PDRs the abundances of these moleculeschanged due to di ff erent chemical reactions. The primary de-struction process of H CO + in PDRs is dissociative recom-bination with electrons, and this happens quickly: H CO + + e − → CO + H. On the other hand, the formation of HCO isclosely related to FUV radiation in PDRs, and several formationroutes have been proposed to describe the high HCO abundancein PDRs via gas-phase reaction or photodissociation (Gerin et al.2009). The gas-phase formation route has two possible chemicalreactions. The first one is a vital formation route of HCO in FUVshielded regions (Schenewerk et al. 1988; Gerin et al. 2009):Metals (Mg or Fe) + HCO + → HCO + metals + (Mg + or Fe + ) . (2)The most plausible pure gas-phase route is a reaction of atomicoxygen with carbon radicals in PDRs (Watt 1983; Leung et al.1984; Schenewerk et al. 1988):O + CH → HCO + H . (3)Another suggested formation route is the gas-grain reactionthrough FUV radiation: photodissociation of formaldehyde(H CO) or grain photodesorption. The first route by FUV hasbeen proposed by Schilke et al. (2001),H CO + photon → HCO + H . (4)The second route is by grain photodesorption: HCO forms ongrain mantles, and subsequently it is desorpted from the grainsinto the gas-phase via thermal or / and photo-desorption processes(Gerin et al. 2009). In cold gas regions below ∼
30 K, thermaldesorption does not play a primary role (Gerin et al. 2009), N (H ) [cm ]1.00.50.00.51.01.52.0 L o g N ( H C O ) / N ( H C O + ) HIIIR bright non-HIIIR dark non-HIIHCO peak (Horsehead)DCO + peak (Horsehead)
10 15 20 25 30 35 40 45Dust Temperature [K]1.00.50.00.51.01.52.0 L o g N ( H C O ) / N ( H C O + ) HIIIR bright non-HIIIR dark non-HII
Fig. 11: Column density ratio of HCO and H CO + moleculesas a function of N (H ) (upper panel) and T dust (lower panel). Thesymbols are the same as explained in Fig. 10. The bigger dots in-dicate sources with their X(HCO) higher than 10 − (Gerin et al.2009). The purple and blue stars (upper panel only) represent N (HCO) / N (H CO + ) ratios from the PDR and cold gas regionin the Horsehead nebula (Gerin et al. 2009). The red and bluedashed line ellipses are confidence ellipses showing distributionsof these two populations (H ii and non-H ii regions) with 1 and 2 σ levels. In the upper plot, the red and blue lines indicate the best-linear regression fits data points of H ii and non-H ii regions (IRbright and IR dark non-H ii s). In the upper panel, the black hor-izontal dashed line presents that N (HCO) / N (H CO + ) ratios is1.whereas, in warm regions like H ii regions, it possibly contributesto HCO desorption. If thermal desorption is not essential, ice-mantle photo-desorption induced by FUV radiation is an alterna-tive process (Willacy & Williams 1993; Bergin et al. 1995). Thefact that high X (HCO) is always found in PDRs also supportsthis photo-desorption process (Schenewerk et al. 1988; Schilkeet al. 2001; Gerin et al. 2009).Since HCO and HCO + result from di ff erent chemical re-actions in PDRs and cold gas regions, we compare columndensities of these molecules to constrain the origin of the de-tected HCO and H CO + toward ATLASGAL clumps. H CO + Article number, page 12 of 19im et al.: PDRs in dust clumps N (H ) [cm ]2.502.252.001.751.501.251.00 L o g N ( c - C H ) / N ( C H ) HIIIR bright non-HIIIR dark non-HII
10 15 20 25 30 35 40Dust Temperature [K]2.502.252.001.751.501.251.00 L o g N ( c - C H ) / N ( C H ) HIIIR bright non-HIIIR dark non-HII
Fig. 12:
U pper : Column density ratio of c-C H and C H as afunction of N (H ). Lower : N (c-C H ) / N (C H) ratios as a func-tion of T clump . Same symbols and lines as Fig. 11.is used to avoid high optical thicknesses. In the upper panel ofFig. 11, N (HCO) / N (H CO + ) ratios decrease with an increase of N (H ), which is closely related to a decrease of X (HCO) because X (H CO + ) is mostly independent of N (H ) (Fig. 10). Besides,results of linear fits for di ff erent source types show that H ii re-gion sources (red line, slope = − N (HCO) / N (H CO + ) ratios than non-H ii region sources(blue line, slope = − p -value (0.01) from a2-dimensional KS test on the two group distributions indicatesthat these groups are significantly di ff erent samples.We added X (HCO), and N (HCO) / N (H CO + ) values ob-served toward a PDR and a cold gas region in the Horseheadnebula (Gerin et al. 2009) in the upper panel of Fig 11, as areference for the origin of HCO toward ATLASGAL clumps. N (HCO) / N (H CO + ) ratios toward H ii (orange color) and non-H ii regions (green and black colors) are distributed between thevalues in the PDR (purple star, HCO peak in the Horsehead neb-ula) and the cold gas region (blue star, DCO + peak in the Horsenebula). The HCO in the cold gas area is considered to be orig-inated from the low column density surface of the cold gas re-gions, not from the UV shielded cold gas region. The clumps with bigger symbol size indicate X (HCO)greater than 10 − , which Gerin et al. (2009) proposed asan indicator of the presence of FUV radiation. All of themare located above a black dashed line corresponding to N (HCO) / N (H CO + ) of 1. According to modelling results forthe Horsehead nebula (Gerin et al. 2009), X (HCO) (cid:38) − and N (HCO) / N (H CO + ) (cid:38) G < in unitsof the Habing field), but not with HUV (like the center of MonR2, 5 × ). A high HCO / HCO + ratio ( > X (HCO) (cid:38) − and N (HCO) / N (H CO + ) (cid:38) X (HCO) and N (HCO) / N (H CO + ) ratios are not always foundfor H ii region sources but also for non-H ii regions (especiallyIR dark non-H ii regions marked with black symbols). In pre-vious HCO observations toward starless cores (e.g., Frau et al.2012; Agúndez et al. 2015), HCO lines were often detected to-ward sources with H column densities lower than 10 cm − ( ∼ some 10 cm − ). High X (HCO) ( (cid:38) − ) in our sources isfound mostly from the clumps with N (H ) < cm − . HCOemission in the clumps with H column density of ≥ cm − isdetectable because bright, massive star-forming regions provideenough radiation to irradiate such high H column density re-gions, or their surfaces are illuminated by nearby massive stars.Also, it is possible that the UV radiation from the deeply em-bedded high-mass stars can still penetrate the dense clumps andcreate the embedded PDRs. The line widths of HCO emissionfound toward the H ii region clumps are significantly broader thanthose of H CO + , which is abundant in cold, dense gas regionsas seen in the Horsehead nebula (Gerin et al. 2009), with a smallfitting uncertainty (see Fig. 5). This also supports that the HCOemission lines detected toward the clumps with high H columndensities are not emitted from the dense regions.In the lower panel of Fig. 11, we compare the column densityratio with dust temperature because one possible HCO formationroute is related to thermal grain desorption. The plot shows thatthermal-desorption from grains does not seem to play an impor-tant role since no correlation is found between the ratios anddust temperatures. The lack of correlation with the dust temper-ature might indirectly explain the reason that we also found high X (HCO) for non-H ii regions, not only H ii regions. Even if grainthermal-desorption processes contributes to the observed abun-dances, it is di ffi cult to probe such reaction with our current lowangular resolution observational data that average HCO emis-sion and dust temperatures over large clumps within single-dishtelescope beams. Therefore, the possibility of thermal desorp-tion is not fully excluded from the formation of HCO. If HCO ismainly formed on grain mantles and is desorpted by photons, weneed to compare HCO with H CO, CH O and CH OH. Thesemolecules are also formed on grain mantles by hydrogenationreactions of CO-ice (Tielens & Whittet 1997; Charnley et al.1997). Studying both HCO and the other molecules allows usto test whether grain photo-desorption processes cause the mainformation reaction of HCO in massive clumps. In addition, wealso need to have better angular resolution to separate relativecontributions of HCO emission from PDRs and cold gas regions.Our observed abundances are lower limit because their emissionis averaged over the clumps. Observing with interferometers to-ward the HCO clumps will allow us more accurate HCO abun-dances and to have a better understanding of its formation and
Article number, page 13 of 19 & A proofs: manuscript no. mspdr_wjkim destruction, and furthermore, complex organic molecules thatare formed after HCO. H /C H column density ratio
The formation of small hydrocarbons is still not fully under-stood, but several processes of their formation in PDRs havebeen proposed: gas-phase reactions and grain-surface reactionssuch as photodestruction of PAHs or very small grains (VSGs).In highly UV-illuminated PDRs with high gas temperatures (afew hundreds ∼ H is formed before c-C H becausec-C H needs an additional carbon atom. C H can be formed byrecombination of C H + , C H + , or C H + with electrons in PDRs(e.g., Mookerjea et al. 2012; Cuadrado et al. 2015) via barrierlesshydrogenation reactions, for example,C H + + e − → C H and 2H . (5)Also, C H can be formed by photodissociation of acetylene(C H ) in gas phase (Lee 1984);C H + h ν → C H + H . (6)For the gas-phase formation of c-C H , C H and C + formC H + that reacts with H and then produces the linear and cyclicC H + isomers (Maluendes et al. 1993; McEwan et al. 1999).Through dissociative recombination of these molecules (Fosséet al. 2001), linear and cyclic-C H are formed as c / l − C H + + e − → c / l − C H + H . (7)Both C H isomers are destroyed by photodissociation ina strong-UV radiation field, which in general should destroyssmall hydrocarbons. Thus, their observed abundances in suchPDRs cannot be explained with gas-phase reactions only. Be-sides, the c-C H abundances in Mon R2 were found to be higherclose to the UV-exposed PDR than the C H abundances (Pilleriet al. 2013). The formation sequences mentioned above cannoteasily result in the high c-C H abundances, and reproducing ob-served c-C H values with PDR models has been di ffi cult (e.g.,c-C H abundances in M8, Tiwari et al. 2019). Another possibleway to produce c-C H is grain-surface formation via photons.The photodissociation of small PAHs or VSGs in PDRs providesfresh small hydrocarbons in highly-illuminated PDRs (Fuenteet al. 2003; Pety et al. 2005) and laboratory experiments alsohave demonstrated the production of small hydrocarbons fromsmall PAHs (number of carbons ≤
24, Useli Bacchitta & Joblin2007). In particular, PAH-related photochemistry enhances theabundance of c-C H in PDRs via dissociative recombinationof C H + with electrons ejected from PAHs or fragmentation ofPAHs (Mookerjea et al. 2012; Pilleri et al. 2013).Although those small hydrocarbons are well-known PDRtracers, their abundances vary from one PDR to the other (e.g.,Pety et al. 2005; Mookerjea et al. 2012; Pilleri et al. 2013;Cuadrado et al. 2015; Tiwari et al. 2019). By investigating theirabundances for ATLASGAL clumps, we may gain insight intothe origin of the detected small hydrocarbons (i.e., C H and c-C H ). Unlike the other PDR tracer, HCO, N (c-C H ) / N (C H)in the upper panel of Fig. 12 does not show any correlationwith N (H ). The presence of UV radiation and C + is a nec-essary factor for the formation of C H and c-C H , and thus, L bol / M clump [L /M ]2.502.252.001.751.501.251.00 L o g N ( c - C H ) / N ( C H ) HIIIR bright non-HIIIR dark non-HII
Fig. 13: Column density ratio of c-C H and C H as a functionof clump bolometirc luminosity over their mass ( L bol / M clump ).Same symbols and lines as in Fig. 11.we mark with bigger dots sources with X (HCO) (cid:38) − and N (HCO) / N (H CO + ) (cid:38)
1, which are proposed as diagnostics ofthe presence of FUV radiation fields and probably C + (Gerinet al. 2009). However, these sources also do not show any corre-lation with any of the parameters.While HCO is found toward a few sources, these small hy-drocarbons are ubiquitous in molecular clouds although they areprobably more abundant in PDRs. To compare with our beam( ∼ (cid:48)(cid:48) corresponding to a size of 0.73 pc at a median distanceof 5 kpc for the observed sources), highly illuminated or densePDRs surrounding young H ii regions are spatially very small andthin ( ∼ ii regions is presumablydominating the observed emission, as in the case of the molec-ular envelope of Mon R2. The small hydrocarbons column den-sities we find in this study are close to averaged values of themolecular envelope and PDRs in Mon R2 or to those of the en-velope (Pilleri et al. 2013). We suggest that the origin of the mea-sured abundances of C H and c-C H might be from the enve-lope of clumps rather than the PDR regions.In low-density ( n H ∼ × cm − ) envelopes with low ki-netic temperature ( T kin ∼
35 K), many species are presumablylocked in the icy mantles of dust grains, and some fraction ofthem sublimate into the gas-phase during the collapse of prestel-lar cores (Pilleri et al. 2013). We, therefore, investigated the in-fluences of the kinetic temperature on the small hydrocarbonabundances in ATLASGAL clumps. The lower panel of Fig. 12shows the N (c-C H ) / N (C H) ratio versus the dust temperatureof the clumps. Without separating di ff erent source groups (H ii and non-H ii regions), the ratios seem to be constant over therange of dust temperatures (10 −
40 K) with a large scatter. Previ-ously, Viti et al. (2004) and Pilleri et al. (2013) showed that fi-nal abundances of these small hydrocarbons did not show di ff er-ences in the temperature range of 10 −
35 K because most of themolecular species are still on grain mantles below T dust =
35 K.However, when we focus on only H ii regions (orange dots), thereis a trend of N (c-C H ) / N (C H) ratios increasing with dust tem-perature, and in particular, the sources with high HCO abun-dances (bigger dots) show an even better correlation (red solid-
Article number, page 14 of 19im et al.: PDRs in dust clumps line confidence ellipses). The dust temperatures of clumps areassociated with the evolution of the clumps. In Fig. 13, we com-pare the small hydrocarbons with bolometric luminosity overclump mass ( L bol / M clump ) that is used as an evolutionary stageindicator of clumps, and thus higher L bol / M clump refer to moreevolved clumps. It is evident that high N (c-C H ) / N (C H) ratiosare found toward more evolved clumps that are associated withH ii regions and have high HCO abundances indirectly probingthe presence of FUV radiation fields. According to results of themodels for the molecular envelope in Mon R2 and DR21 (Mook-erjea et al. 2012; Pilleri et al. 2013), in the cold and low-densitymolecular envelopes of the regions, grain-surface processes (i.e.,CO freeze-out) and time-dependent e ff ects become an importantrole in the formation and destruction of these molecules. In par-ticular, while abundances of C H and c-C H do not vary until10 years, after about 10 years with the development of UC H ii regions, the N (c-C H ) / N (C H) ratios steeply go up. Depend-ing on early freeze-out fractions, which control the initial smallhydrocarbons abundances in the gas-phase, final abundances ofthese molecules predicted in the models vary. Since we do nothave any information about the initial physical conditions in ourclumps, it is hard to predict how the gain-surface processes andtime-dependent e ff ects a ff ect our observed results. The Strato-spheric Observatory for Infrared Astronomy (SOFIA) observa-tions toward M17SW (Pérez-Beaupuits et al. 2012) showed that[C ii ] emission was found interestingly throughout deeper partsof the molecular cloud, and a significant fraction of its emissionwas not associated with other PDR tracers (i.e., CO J = − i ]). This means that C + gas, probed by [C ii ] spectral lines,can exist in deeper molecular clouds. Thus, we cannot excludethat C + can contribute to the formation of small hydrocarbon viagas-phase reactions in such molecular envelop regions. In addi-tion, photodissociation of PAH molecules in di ff erent strengthsof UV radiation fields can add some fraction of small hydrocar-bons abundances in gas-phase. According to Murga et al. (2020),in high-UV illuminated regions (i.e., the Orion Bar), the PAHdissociation becomes more important at A V > . H , which is oneof precursors of C H and c-C H as mentioned above. On theother hand, in low-UV radiation, pure gas-phase reaction is amain route to increase the production rate of C H . Since themore evolved clumps shown in Fig. 13 emit stronger UV radi-ation, it is possible that the measured abundances of the smallhydrocarbons are still associated with the photodissociation. Wesuggest that the small hydrocarbons we observe toward clumpshosting H ii regions likely originate mostly in the molecular en-velope around dense PDRs and H ii regions. In particular, abun-dances of these molecules might be a combined result of gain-surface processes controlling initial molecular abundances in thegas-phase and subsequent time-dependent evolution in environ-ments with a significant C + abundance. Also, some fraction ofthese small hydrocarbon abundances in the H ii sources can beresulted of photodissociation of PAH molecules by the UV radi-ation from the newly born high-mass stars.
7. Summary and conclusion
We have investigated 10 molecular transitions from HCO,H CO + , HOC + , C H, c-C H , CN, H CN, HC N, HN Cand C O covered in an unbiased molecular line survey of the3 mm band that was observed using the IRAM 30 m telescope to-ward 409 ATLASGAL clumps. The ATLASGAL clumps weredivided into three groups based on the presence of H ii regionsand infrared emission: H ii , IR bright non-H ii , and dark non-H ii region sources. We carried out an analysis of the column densityand abundance of the selected molecules, and our main resultsare as follows: − C O, H CO + , C H, c-C H , CN, and HN C. H CN showhigh detection rates (higher than 94 %), whereas H CN,HCO and HC N are detected with detection rates of 75%,32%, and 39%, respectively. The non-detections of the HCO,H CN and HC N transitions are mostly from the IR-brightand IR-dark non-H ii regions, while high detection rates of themolecular transitions toward H ii region sources were found,namely HCO in 66%, H CN in 93% and HC N in 79% ofthe regions. − While the abundances of high column density tracers (i.e.,H CO + and HC N) are almost constant over the range ofH column densities, the abundances of HCO, CN, C H andc-C H drop with an increase of H column density. In par-ticular, the HCO abundances are prominently reduced in highH column density, and they seem higher toward H ii regionsthan toward non-H ii regions for a given H column density. − We also find that N (HCO) / N (H CO + ) ratios decrease asH column density increase, and 61 clumps have X (HCO) (cid:38) − and N (HCO) / N (H CO + ) (cid:38)
1. This implies thatthe HCO detected toward ATLASGAL clumps is likely con-nected to PDRs, and the sources with high HCO abundancesare associated with on-going FUV chemistry in their PDRs.However, due to low angular resolution of our data, the mea-sured HCO abundances in the ATLASGAL clumps are av-erages over their PDRs and FUV shielded molecular gas re-gions. − The c-C H / C H ratios toward the dust clumps are con-stant with H column density and dust temperature, al-though with large scatter. However, toward only H ii re-gions having high HCO abundances, their c-C H / C H ra-tios evidently rise with dust temperatures. Especially, highc-C H / C H ratios are found toward more evolved clumpswith high L bol / M clump . These results show similar trends withthe model results of Pilleri et al. (2013) that predict increas-ing c-C H / C H ratios with the time after the age of 10 years. Therefore, the measured abundances of small hydro-carbons in this study are possibly results of gain-surface pro-cesses (CO freeze-out fractions) and time-dependent e ff ectsin the clumps rather than in the PDRs. In addition, somefraction of the measured abundances toward the H ii sourcescan be added as the result of the photodissociation of PAHmolecules. Acknowledgements
Article number, page 15 of 19 & A proofs: manuscript no. mspdr_wjkim
References
Agúndez, M., Cernicharo, J., & Guélin, M. 2015, A&A, 577, L5Bacmann, A., Lefloch, B., Ceccarelli, C., et al. 2002, A&A, 389, L6Bergin, E. A., Langer, W. D., & Goldsmith, P. F. 1995, ApJ, 441, 222Beuther, H., Semenov, D., Henning, T., & Linz, H. 2008, ApJ, 675, L33Boger, G. I. & Sternberg, A. 2005, ApJ, 632, 302Caselli, P., Walmsley, C. M., Tafalla, M., Dore, L., & Myers, P. C. 1999, ApJ,523, L165Charnley, S. B., Tielens, A. G. G. M., & Rodgers, S. D. 1997, ApJ, 482, L203Contreras, Y., Schuller, F., Urquhart, J. S., et al. 2013, A&A, 549, A45Csengeri, T., Leurini, S., Wyrowski, F., et al. 2016, A&A, 586, A149Cuadrado, S., Goicoechea, J. R., Pilleri, P., et al. 2015, A&A, 575, A82Fossé, D., Cernicharo, J., Gerin, M., & Cox, P. 2001, ApJ, 552, 168Frau, P., Girart, J. M., & Beltrán, M. T. 2012, A&A, 537, L9Fuente, A., García-Burillo, S., Gerin, M., et al. 2005, ApJ, 619, L155Fuente, A., Martin-Pintado, J., Cernicharo, J., & Bachiller, R. 1993, A&A, 276,473Fuente, A., Rodrıguez-Franco, A., Garcıa-Burillo, S., Martın-Pintado, J., &Black, J. H. 2003, A&A, 406, 899García-Burillo, S., Martín-Pintado, J., Fuente, A., Usero, A., & Neri, R. 2002,ApJ, 575, L55Gerin, M., Goicoechea, J. R., Pety, J., & Hily-Blant, P. 2009, A&A, 494, 977Gerner, T., Beuther, H., Semenov, D., et al. 2014, A&A, 563, A97Giannetti, A., Wyrowski, F., Brand, J., et al. 2014, A&A, 570, A65Ginard, D., González-García, M., Fuente, A., et al. 2012, A&A, 543, A27Goicoechea, J. R., Pety, J., Gerin, M., Hily-Blant, P., & Le Bourlot, J. 2009,A&A, 498, 771Goldsmith, P. F., Irvine, W. M., Hjalmarson, A., & Ellder, J. 1986, ApJ, 310, 383Hogerheijde, M. R., Jansen, D. J., & van Dishoeck, E. F. 1995, A&A, 294, 792Jansen, D. J., Spaans, M., Hogerheijde, M. R., & van Dishoeck, E. F. 1995, A&A,303, 541Jin, M., Lee, J.-E., & Kim, K.-T. 2015, ApJS, 219, 2Kim, W. J., Urquhart, J. S., Wyrowski, F., Menten, K. M., & Csengeri, T. 2018,A&A, 616, A107Kim, W.-J., Wyrowski, F., Urquhart, J. S., Menten, K. M., & Csengeri, T. 2017,A&A, 602, A37König, C., Urquhart, J. S., Csengeri, T., et al. 2017, A&A, 599, A139Lee, L. C. 1984, ApJ, 282, 172Leung, C. M., Herbst, E., & Huebner, W. F. 1984, ApJS, 56, 231Lucas, R. 1976, A&A, 46, 473Maluendes, S. A., McLean, A. D., Yamashita, K., & Herbst, E. 1993,J. Chem. Phys., 99, 2812McEwan, M. J., Scott, G. B. I., Adams, N. G., et al. 1999, ApJ, 513, 287Miettinen, O. 2020, A&A, 639, A65Mookerjea, B., Hassel, G. E., Gerin, M., et al. 2012, A&A, 546, A75Müller, H. S. P., Thorwirth, S., Roth, D. A., & Winnewisser, G. 2001, A&A, 370,L49Murga, M. S., Kirsanova, M. S., Vasyunin, A. I., & Pavlyuchenkov, Y. N. 2020,MNRAS[ arXiv:2007.06568 ]Myers, P. C., Mardones, D., Tafalla, M., Williams, J. P., & Wilner, D. J. 1996,ApJ, 465, L133 + Pérez-Beaupuits, J. P., Wiesemeyer, H., Ossenkopf, V., et al. 2012, A&A, 542,L13Pety, J. 2005, in SF2A-2005: Semaine de l’Astrophysique Francaise, ed. F. Ca-soli, T. Contini, J. M. Hameury, & L. Pagani, 721Pety, J., Teyssier, D., Fossé, D., et al. 2005, A&A, 435, 885Pickett, H. M., Poynter, R. L., Cohen, E. A., et al. 1998,J. Quant. Spectr. Rad. Transf., 60, 883Pilleri, P., Treviño-Morales, S., Fuente, A., et al. 2013, A&A, 554, A87Rathborne, J. M., Whitaker, J. S., Jackson, J. M., et al. 2016, PASA, 33, e030Rawlings, J. M. C., Taylor, S. D., & Williams, D. A. 2000, MNRAS, 313, 461Rizzo, J. R., Fuente, A., & García-Burillo, S. 2005, ApJ, 634, 1133Rizzo, J. R., Fuente, A., Rodríguez-Franco, A., & García-Burillo, S. 2003, ApJ,597, L153Rodriguez-Franco, A., Martin-Pintado, J., & Fuente, A. 1998, A&A, 329, 1097Sanhueza, P., Jackson, J. M., Foster, J. B., et al. 2012, ApJ, 756, 60Savage, C. & Ziurys, L. M. 2004, ApJ, 616, 966Schenewerk, M. S., Jewell, P. R., Snyder, L. E., Hollis, J. M., & Ziurys, L. M.1988, ApJ, 328, 785Schilke, P., Pineau des Forêts, G., Walmsley, C. M., & Martín-Pintado, J. 2001,A&A, 372, 291Schilke, P., Walmsley, C. M., Pineau Des Forets, G., et al. 1992, A&A, 256, 595Schöier, F. L., van der Tak, F. F. S., van Dishoeck, E. F., & Black, J. H. 2005,A&A, 432, 369Simon, R., Stutzki, J., Sternberg, A., & Winnewisser, G. 1997, A&A, 327, L9Smith, R. J., Shetty, R., Stutz, A. M., & Klessen, R. S. 2012, ApJ, 750, 64Sternberg, A. & Dalgarno, A. 1995, ApJS, 99, 565Teyssier, D., Fossé, D., Gerin, M., et al. 2004, A&A, 417, 135 Tielens, A. G. G. M. 2013, Interstellar PAHs and Dust, ed. T. D. Oswalt &G. Gilmore, 499Tielens, A. G. G. M. & Hollenbach, D. 1985, ApJ, 291, 722Tielens, A. G. G. M. & Whittet, D. C. B. 1997, in IAU Symposium, Vol. 178,IAU Symposium, ed. E. F. van Dishoeck, 45Tiwari, M., Menten, K. M., Wyrowski, F., et al. 2019, A&A, 626, A28Urquhart, J. S., Csengeri, T., Wyrowski, F., et al. 2014, A&A, 568, A41Urquhart, J. S., Figura, C., Wyrowski, F., et al. 2019, MNRAS, 484, 4444Urquhart, J. S., König, C., Giannetti, A., et al. 2018, MNRAS, 473, 1059Urquhart, J. S., Thompson, M. A., Moore, T. J. T., et al. 2013, MNRAS, 435,400Useli Bacchitta, F. & Joblin, C. 2007, in Molecules in Space and Laboratory, 89Vasyunina, T., Linz, H., Henning, T., et al. 2011, A&A, 527, A88Viti, S., Collings, M. P., Dever, J. W., McCoustra, M. R. S., & Williams, D. A.2004, MNRAS, 354, 1141Watt, G. D. 1983, MNRAS, 205, 321Wienen, M., Wyrowski, F., Schuller, F., et al. 2012, A&A, 544, A146Willacy, K. & Williams, D. A. 1993, MNRAS, 260, 635Wyrowski, F., Güsten, R., Menten, K. M., et al. 2016, A&A, 585, A149Zinnecker, H. & Yorke, H. W. 2007, ARA&A, 45, 481
Article number, page 16 of 19im et al.: PDRs in dust clumps
Appendix A: Detection rates of molecular lines with source distance CO + H 0.000.050.100.150.20c-C H N0 2 4 6 8 10 12 14 160.000.050.100.150.20 H CN 0 2 4 6 8 10 12 14 160.000.050.100.150.20HN C D e t e c t i o n R a t e s ( X % ) Distance [kpc]
Fig. A.1: Detection rate as a function of distance.
Article number, page 17 of 19 & A proofs: manuscript no. mspdr_wjkim
Appendix B: Linewidth of molecular lines with source distance CO + H 02468101214c-C H N0 5 10 15051015 H CN 0 5 10 15 051015HN C L i n e w i d t h [ k m s − ] Distance [kpc]
Fig. B.1: Line width as a function of distance.
Article number, page 18 of 19im et al.: PDRs in dust clumps
Appendix C: Integrated intensity ratios CO + H 1.51.00.50.00.5c-C H N0.0 0.5 1.0 1.5 2.02.01.51.00.50.00.5 H CN 0.0 0.5 1.0 1.5 2.0 2.01.51.00.50.00.5HN C L o g ( I n t e g r a t e d i n t e n s i t y r a t i o o f X / C O ) Log (Integrated intensity of C O) [K km s ] Fig. C.1: Integrated intensity ratio of X / C O as a function of integrated intensity of C O. The X refers noted molecular species.The orange and cyan symbols indicate H ii and non-H ii regions, respectively. The red (H ii regions) and blue (non-H ii regions) eclipsesrepresent covariance error ellipses showing the distributions of the data points.regions) eclipsesrepresent covariance error ellipses showing the distributions of the data points.