NGC 3503 and its molecular environment
N. U. Duronea, J. Vasquez, C. E. Cappa, M. Corti, E. M. Arnal
aa r X i v : . [ a s t r o - ph . GA ] N ov Astronomy&Astrophysicsmanuscript no. ngc3503˙2columns2 c (cid:13)
ESO 2018November 11, 2018
NGC 3503 and its molecular environment
N. U. Duronea , J. Vasquez , , C. E. Cappa , , M. Corti , , and E. M. Arnal , (A ffi liations can be found after the references) Received 2011 August 26; accepted 2011 November 14
ABSTRACT
Context.Aims.
We present a study of the molecular gas and interstellar dust distribution in the environs of the H ii region NGC 3503 associated with the opencluster Pis 17 with the aim of investigating the spatial distribution of the molecular gas linked to the nebula and achieving a better understandingof the interaction of the nebula and Pis 17 with their molecular environment. Methods.
We based our study in CO(1-0) observations of a region of ∼ . ◦ Results.
We found a molecular cloud (Component 1) having a mean velocity of –24.7 km s − , compatible with the velocity of the ionizedgas, which is associated with the nebula and its surroundings. Adopting a distance of 2.9 ± ± × M ⊙ and 400 ±
240 cm − , respectively.The radio continuum data confirm the existence of an electron density gradient in NGC 3503. The IR emission shows the presence of a PDRbordering the higher density regions of the nebula. The spatial distribution of the CO emission shows that the nebula coincides with a molecularclump, with the strongest CO emission peak located close to the higher electron density region. The more negative velocities of the molecular gas(about –27 km s − ), is coincident with NGC 3503. Candidate YSOs were detected towards the H ii region, suggesting that embedded star formationmay be occurring in the neighbourhood of the nebula. The presence of a clear electron density gradient, along with the spatial distribution of themolecular gas and PAHs in the region indicates that NGC 3503 is a blister-type H ii region that probably has undergone a champagne phase. Conclusions.
Key words.
ISM: molecules, Radio Continuum: ISM, Infrared: ISM, ISM: H ii regions, ISM:individual object: NGC 3503 , Stars: Pismis 17
1. Introduction
It is well established that massive stars have a significant impacton the dynamics and energetics of the interstellar medium (ISM)surrounding them. In the classical scenario, OB stars are borndeeply buried within dense molecular clumps and emit a copi-ous amount of far ultraviolet (FUV) radiation ( h ν > ii regionwhich expands into the molecular cloud due to the pressure dif-ference between the molecular and ionized gas. At the interfacebetween the ionized and molecular gas, where photons of lowerenergies (6 eV < h ν < ii regions has been studiedin detail by many authors (Keto & Ho 1989; Russeil & Castets2004). Many e ff orts have been devoted to observing the interac-tion between the H ii regions and their parental molecular cloudsand some evidence has been presented showing that moleculargas near H ii regions may be kinematically disturbed by severalkm s − (Elmegreen & Wang 1987). Understanding the complexinteraction between massive young OB stars, H ii regions, andmolecular gas is crucial to the study of massive star formationand the impact of massive stars on their environment.NGC 3503 ( = Hf 44 = BBW 335) ( l,b = . ◦ + . ◦
12) is abright and small ( ∼ ′ in diameter) optical emission nebula(Dreyer & Sinnott 1988) ionized by early B-type stars belong-ing to the open cluster Pis 17 (Herbst 1975; Pinheiro et al.2010) and it is seen projected onto an extended region ofmedium brightness in H α known as RCW 54. This last re- gion is centered at ( l, b ) = (289 . ◦
4, -0 . ◦ ∆ l × ∆ b ) = (3 . ◦ × . ◦
0) (Rodgers et al. 1960). The point sourceIRAS 10591-5934, which appears projected onto NGC 3503 , isthe infrared counterpart of the optical nebula. For the sake ofclarity, we show the main components and their relative locationin Fig. 1 superimposed on a DSSR image of the region of inter-est. The dashed white square delimits the region studied in thispaper.In their Southern Hemisphere catalogue of bright-rimmed clouds (BRCs) associated with IRAS point sources,Sugitani & Ogura (1994) quote SFO 62 as related to NGC 3503.BRCs are defined as isolated molecular clouds located atthe edges of evolved H ii regions, and are suspected to bepotential sites of star formation through the radiation-drivenimplosion (RDI) process (Sugitani et al. 1991; Urquhart et al.2009). Making use of the Australian Telescope Compact Array(ATCA), Thompson et al. (2004) carried out radio continuumobservations towards all BRCs catalogued by Sugitani & Ogura(1994). These authors classified SFO 62 as a broken-rimmedcloud associated with an evolved stellar cluster that is aboutto disrupt its natal molecular cloud. Latter on, Urquhart et al.(2009) claimed that SFO62 was incorrectly classified as a BRC,and that the RDI process is not working in this region.Using the [S ii ] λ / λ α survey of the Milky Way towards l = ◦ ,
1. U. Duronea et al.: NGC 3503 and its molecular environment
Fig. 1.
DSSR image of the brightest part of RCW 54. The image is ∼ . ◦ × . ◦ l,b ) = (289 . ◦
5, 0 . ◦ of –21 km s − towards NGC 3503 , and proposed that NGC 3503 is linked to acomplex of H ii regions placed at a distance of about 2.7 kpc,with radial velocities of about −
25 km s − . Published distancedeterminations of NGC 3503 vary between a minimum of 2.6kpc (Herbst 1975) and a maximum of 4.2 kpc (Mo ff at & Vogt1975). In this paper we shall adopt a distance of 2.9 ± CO (J = →
0) (HPBW = ′ ) line observationswere carried out by Yamaguchi et al. (1999) towards 23 H ii re-gions associated with 43 BRCs catalogued by Sugitani & Ogura(1994), with a grid spacing of ∼ ′ × ′ . The aim of that workwas to investigate statistically the dynamical e ff ects of H ii re-gions on their associated molecular clouds and star formation.A square region of ∼ . ◦ ′ diameter molec-ular cloud having a peak radial velocity of –24.9 km s − wasfound. Adopting a distance of 2.9 kpc, this molecular structurehas a linear radious of 2 pc and a total mass of ∼
500 M ⊙ .The authors rejected this cloud from the analysis sample, since Radial velocities are referred to the Local Standard of Rest (LSR) it was detected at only one position. CO, CO, and CO(J = →
0) MOPRA position-switched observations were carriedout by Urquhart et al. (2009). The ON position was centred onthe position of the IRAS source associated with NGC 3503. Theleast abundant of the isotopes was not detected, whilst the emis-sion lines of the other two isotopes show a double-peaked pro-file. From a Gaussian fitting, their peak radial velocities are ∼ +
20 km s − and –25.6 km s − .Molecular line observations provide an invaluable sup-port in pursuing a better understanding of the interaction ofH ii regions with their surroundings. Previous observations ofYamaguchi et al. (1999) and Urquhart et al. (2009), althoughproviding important information about the molecular gas asso-ciated with NGC 3503 , do not o ff er a complete picture of themolecular environment of the nebula. Bearing this in mind, thegoals of this paper are twofold, namely: a) to map the spatial dis-tribution of the molecular gas associated with NGC 3503 and tostudy its physical characteristics, and b) to achieve a better un-derstanding of the interaction of NGC 3503 and Pis 17 with theirmolecular environment. These aims were addressed by analyz-ing new CO (J = →
0) data gathered by using the NANTENtelescope to observe a square region ∼ . ◦
2. U. Duronea et al.: NGC 3503 and its molecular environment ( l,b ) = (289 . ◦ + . ◦
2. Observations and data reductions
The databases used in this work are:1. Intermediate angular resolution, medium sensitivity, andhigh-velocity resolution CO (J = →
0) data obtainedwith the 4-m NANTEN millimeter-wave telescope ofNagoya University. At the time the authors carried outthe observations, April 2001, this telescope was installedat Las Campanas Observatory, Chile. The half-powerbeamwidth and the system temperature, including theatmospheric contribution towards the zenith, were 2 . ′ ∼ ∼
220 K (SSB) at 115 GHz,respectively. The data were gathered using the positionswitching mode. Observations of points devoid of CO emis-sion were interspersed among the program positions. Thecoordinates of these points were retrieved from a databasethat was kindly made available to us by the NANTENsta ff . The spectrometer used was an acusto-optical with2048 channels providing a velocity resolution of ∼ − . For intensity calibrations, a room-temperaturechopper wheel was employed (Penzias & Burrus 1973).An absolute intensity calibration (Ulich & Haas 1976;Kutner & Ulich 1981) was achieved by observing Orion KL(RA(1950.0) = h m . s = − ◦ ′ ′′ ),and ρ Oph East (RA(1950.0) = h m . s = − ◦ ′ ′′ ). The absolute radiationtemperature, T ∗ R , of Orion KL and ρ Oph East, as observedby the NANTEN radiotelescope were assumed to be 65 Kand 15 K, respectively (Moriguchi et al. 2001). The COobservations covered a region ( △ l × △ b ) of 35 . ′ × . ′ l, b ) = (289 . ◦ + . ◦
02 ) and the observed gridconsists of points located every one beam apart. A total of169 positions were observed. Typically, the integration timeper point was 16s resulting in an rms noise of ∼ ff ects.The spectra were reduced using CLASS software (GILDASworking group).2. Unpublished radio continuum data at 4800 and 8640MHz which were kindly provided by J. S. Urquhart andM. A. Thompson. The images were obtained in March 2005with the Australia Telescope Compact Array (ATCA) withsynthesized beams and rms noises of 23 . ′′ × . ′′
62 and0.82 mJy beam − at 4800 MHz and 14 . ′′ × . ′′
74 and0.56 mJy beam − at 8640 MHz. Radio continuum archivalimages were also obtained from the Sydney UniversityMolonglo Sky Survey (SUMSS) (Bock et al. 1999) at 843MHz, with angular resolution of 45 ′′ × ′′ cosec( δ ). http: // / cgi-bin / postage.pl
3. Infrared data retrieved from the
Midcourse SpaceExperiment (MSX) (Price et al. 2001), high reso-lution IRAS images (HIRES) at 60 and 100 µ m(Fowler & Aumann 1994), Spitzer images at 8.0 and4.5 µ m from the Galactic Legacy Infrared Mid-Plane SurveyExtraordinaire (GLIMPSE) (Benjamin et al. 2003), andMultiband Imaging Photometer for S pitzer (MIPS) imagesat 24 and 70 µ m from the MIPS Inner Galactic Plane Survey(MIPSGAL) (Carey et al. 2005).4. Optical data retrieved from the 2nd Digitized Sky Survey(red plate) (McLean et al. 2000).
3. Results and analysis
In order to illustrate in broad terms the molecular structures de-tected towards the region under study, a series of CO profiles aredisplayed in Fig. 2.Profiles a) and c) show the CO emission along the line ofsight to NGC 3503 and the bright edge of SFO 62 , respectively.In both spectra the bulk of the molecular emission is detectedbetween –30 and –20 km s − , and +
15 to +
25 km s − , in goodagreement with previous results by Urquhart et al. (2009).The medium brightness optical region northeast ofNGC 3503 (from here onwards MBO) though also depicts adouble peak structure (see profile b ), displays some small scalestructure in the feature peaking at ∼ +
20 km s − , while the mostnegative CO feature is detected at ∼ –18 km s − .The molecular gas along regions of high optical absorption isshown in profiles d) and e) . The former resembles the CO spec-trum observed towards NGC 3503 , and is characteristic of themolecular emission in the region of high optical absorption seennorthwest of NGC 3503 . The roundish patch of high absorptionseen at (l,b) = (289 . ◦ + . ◦
07) only shows the CO peak at posi-tive velocities.The spatial distribution of the molecular gas observed in thethree velocity intervals mentioned above is shown in the leftside panels of Fig. 3. In order of increasing radial velocity, theCO components will be referred to as Component 1 (peakingat ∼ –25 km s − ), Component 2 (peaking at ∼ -16 km s − ), andComponent 3 (peaking at ∼ +
20 km s − ). To facilitate the com-parison between the spatial distribution of molecular and ionisedgas, the right panels of Fig. 3 show the same molecular contoursas the left panels superimposed to the DSSR image in grayscale.The di ff erence in spatial distribution among the three compo-nents is readily appreciated. The molecular gas in the veloc-ity interval –29 to –20 km s − (Component 1) shows two welldeveloped concentrations, whose emission peaks are located at (l,b) = (289 . ◦ + . ◦
12) (clump A) and (l,b) = (289 . ◦ ′ , + . ◦ (l,b) = (289 . ◦ + . ◦
17) near bothNGC 3503 and MBO, and is projected onto a region withoutapreciable optical absorption. Finally, Component 3, peakingat (l,b) = (289 . ◦ − . ◦ http: // irsa.ipac.caltech.edu / Missions / msx.html http: // irsa.ipac.caltech.edu / applications / IRAS / IGA / http: // sha.ipac.caltech.edu / applications / Spitzer / SHA // http: // sha.ipac.caltech.edu / applications / Spitzer / SHA // http: // skyview.gsfc.nasa.gov / cgi-bin / query.pl
3. U. Duronea et al.: NGC 3503 and its molecular environment
Kilo VELOCITY-30 -20 -10 0 10 20 30 Kilo VELOCITY-30 -20 -10 0 10 20 30 Kilo VELOCITY-30 -20 -10 0 10 20 30Kilo VELOCITY-30 -20 -10 0 10 20 30Kilo VELOCITY-30 -20 -10 0 10 20 30 G A L A C T I C L A T . GALACTIC LONG.289 35 30 25 20 1500 15100500-00 05
NGC 3503SFO 62 a b cd −1 e MBO R E N
GALACTIC LONG. G A L AC T I C L A T . T ( K ) VELOCITY (KM S ) * Fig. 2.
CO emission profiles towards five positions around NGC 3503. The CO profiles are averaged over a square area ∼ ′ in size,centered on the black / white dots drawn on the DSSR image (center). The intensities are given as absolute radiation temperature T ∗ R .correspondence either with ionized regions or areas displayingstrong absorption.For the three molecular components, a mean radial velocity( ¯ V ) weighted by line temperature, was derived by means of¯ V = P i T Peak i × V Peak i P i T Peak i (1)where T Peak i and V Peak i are the peak T ∗ R temperature and the peakradial velocity of the i -spectrum observed within the 3 rms con-tour line defining the outer border. Mean radial velocities of–24.7 km s − , –16.6 km s − , and + − were obtainedfor Component 1, Component 2, and Component 3, respectively.Based on ¯ V , we tried to estimate the kinematical distanceof Component 1. The analytical fit to the rotation curve byBrand & Blitz (1993) along l = . ◦ − are forbidden for thisgalactic longitude. However, the ionised gas along this galac-tic longitude exhibits velocity departures more negative than − − relative to the circular rotation model, as the cases of theH ii regions Gum 37 (l,b) = (290 . ◦ + . ◦
26) ( d = (l,b) = (291 . ◦ + . ◦
71) ( d = −
25 to −
28 km s − and −
25 to − − , respectively (Georgelin et al. 2000). Bearing in mindthat close to the tangential point along l = . ◦ ii regions located atdistances between 2.7 and 2.8 kpc exhibit forbidden radial veloc-ities similar to those found for Component 1, we adopt for thelatter a distance of 2.9 ± α line towards the nebula( −
21 km s − ; Georgelin et al. 2000). It is also worth not-ing that the velocity interval of Component 1 is in ex-cellent agreement with the velocity of SFO 62 found bySugitani & Ogura (1994) (–28 ≤ v LS R ≤ –20 km s − , at CO),by Yamaguchi et al. (1999) (–24.9 km s − at CO), and byUrquhart et al. (2009) ( − − − , at COand CO, respectively). Based on above, we conclude thatComponent 1 is associated with NGC 3503 and its environs. Thebright rim of SFO 62 clearly follows the southernmost border ofclump A, which very likely indicates that the molecular gas inthe southern border of this clump is being ionized, originatingthe bright optical rim.On the other hand, considering that the mean radial velocityof Component 2 (–16.5 km s − ) is close to the radial velocity ofthe tangential point at l = . ◦ − ), we suggest thatComponent 2 may be located in the neighborhood of the tangen-tial point, close to Component 1. From the mean radial velocityof Component 3, a kinematic distance of ∼ ii re-gions at a velocity of ∼ +
20 km s − reported by Georgelin et al.(2000).
4. U. Duronea et al.: NGC 3503 and its molecular environment G A L A C T I C L A T . GALACTIC LONG.289 35 30 25 20 1500 15100500-00 05 G A L A C T I C L A T . GALACTIC LONG.289 35 30 25 20 1500 15100500-00 05 G A L A C T I C L A T . GALACTIC LONG.289 35 30 25 20 1500 15100500-00 05 G A L A C T I C L A T . GALACTIC LONG.289 35 30 25 20 1500 15100500-00 05 G A L A C T I C L A T . GALACTIC LONG.289 35 30 25 20 1500 15100500-00 05 G A L A C T I C L A T . GALACTIC LONG.289 35 30 25 20 1500 15100500-00 05 G A L A C T I C L A T . GALACTIC LONG.289 35 30 25 20 1500 15100500-00 05 clump Bclump A
E N −1 GALACTIC LONG. G A L A C T I C L A T . G A L A C T I C L A T . G A L A C T I C L A T . GALACTIC LONG. −1−1 [−29.0,−20.0] km s [−17.5,−15.5] km s [+14.0,+20.0] km s
Fig. 3.
Upper panels (Left): Averaged T ∗ R in the velocity range ∼ –29 to – 20 km s − (Component 1). Contour levels start at 0.42K (23 rms) and the contour spacing is 0.6 K. The beam size of the CO observations is shown by a circle in the lower left corner.(Right): Overlay of the T ∗ R values in the same velocity interval (contours), superimposed on the DSSR image (greyscale). Middlepanels:
Averaged T ∗ R in the velocity range ∼ –17.5 to –15.5 km s − (Component 2). Contour levels start at 0.7 K (23 rms) and thecontour spacing is 1 K. Lower panels:
Averaged T ∗ R in the velocity range ∼ +
14 to +
24 km s − (Component 3). Contour levels startat 0.35 K (23 rms) and the contour spacing is 0.45 K. In all cases, the molecular emission grayscale goes from 0.35 K to 3.9 K. Theorientation of the equatorial system is indicated in the top left panel.A direct comparison of Component 1 with the molecularcloud detected by Yamaguchi et al. (1999) (see Fig. 2u from thatwork), shows that the angular size of Component 1 is about a fac-tor 3 - 4 greater than the latter. Clearly, only the densest part ofthe molecular cloud associated with NGC 3503 (clump A) wasdetected in the CO observations of Yamaguchi et al. (1999).From here onwards, the analysis of the molecular gas associatedwith NGC 3503 will focus on Component 1.
The kinematics of Component 1 was studied by using position-velocity maps across selected strips. The map obtained along b = + . ◦
12 (corresponding to clump A) is shown in the upperpanel of Fig. 4. A noticeable velocity gradient is observed atvelocities from about −
23 km s − to about −
28 km s − .
5. U. Duronea et al.: NGC 3503 and its molecular environment G A L A C T I C L O N G . Kilo VELOCITY-30 -28 -26 -24 -22289 101520253035 G A L A C T I C L A T . GALACTIC LONG.289 35 30 25 20 1500 15100500-00 05
Angular resolution −24 km s −1 −24 km s −1 −26 km s −1 −30 −28 −26 −24 −22 −1 Velocity (km s )GALACTIC LONG. G A L A C T I C L O N G . G A L A C T I C L A T . Fig. 4.
Upper panel : Velocity-Galactic Longitude map obtainedin a strip along b = + . ◦
11 (corresponding to clump A) showing T ∗ R . Contour levels start at 0.7 K and the contour spacing is 0.7K. The dotted lines indicate the location of NGC 3503 . Lowerpanel : Mean velocity map of Component 1. Contours levels gofrom −
26 to −
24 km s − , with interval of 0.2 km s − . The dottedellipse depicts approximately the region of clump A.Additional support to the existence of this gradient can beobtained through a moment analysis. Due to the large angular di-mensions of Component 1 and the spatial sampling of NANTENobservations, 28 independient CO profiles were observed to-wards the region enclosed by the 0.42 K-contour line in Fig. 3.Having these spectra a high signal-to-noise ratio (S / N > ∼ − is observed across the region of clump A, which at a distance of2.9 kpc translates to a gradient ω ≈ − pc − . This panel G A L A C T I C L A T . GALACTIC LONG.289 35 30 25 20 1500 15100500-00 05 G A L A C T I C L A T . GALACTIC LONG.
Fig. 5.
Spatial distribution of T exc for Component 1. Contoursleves are 5.7, 7.9, 10.2, 12.4, 14.5, and 16.6 K.also shows that the CO emission corresponding to clump B doesnot show a significant velocity gradient.In order to o ff er a complete picture of the kinematical prop-erties of clump A, we show in Fig. 6 the spatial distribution ofthe CO emission within the velocity range from –27.4 to –23.4km s − . Every image depicts mean T ∗ R -values (in contours) overa velocity interval of 1 km s − superimposed on the 8.13 µ mMSX emission (in greyscale). In the velocity range from –27.4to –26.4 km s − the molecular emission arising from clump Ais slightly displaced from the brightest MSX emission located at (l,b) ≈ (289 . ◦ + . ◦ T exc ) obtained from the CO data to probe the sur-face conditions of Component 1. The excitation temperature ofthe CO line can be obtained considering that CO is opticallythick ( τ ν >> T peak ( CO ) = J ν ( T exc ) − J ν ( T bg ) (2)(Dickman 1978) where J ν is the Planck function at a frequency ν . Assuming gaussian profiles for the CO line, combining theorder zero moment map (i.e., integrated area) with the ordertwo moment map (i.e., velocity dispersion), and using Eq. 2we obtain the T exc distribution map (Fig. 5). As expected, the T exc distribution is quite similar to the CO emission distribu-tion of Component 1, reaching a maximum of ∼ l,b ) ≈ (289 . ◦ + . ◦ T exc towards the center of clump A is lower than the ob-tained by Urquhart et al. (2009) towards (l,b) ≈ (289 . ◦ + . ◦ T exc = ff erence may be explained in terms ofa beam smearing of our NANTEN data, which implies that thevalues of T exc shown in Fig. 5 must be considered as lower lim-its. The mass of the molecular gas associated with NGC 3503 canbe derived making use of the empirical relationship between the
6. U. Duronea et al.: NGC 3503 and its molecular environment
00 15100500-00 05 G A L A C T I C L A T . GALACTIC LONG.289 35 30 25 20 1500 15100500-00 05 [−27.4, −26.4] km s [−26.4, −25.4] km s[−25.4, −24.4] km s [−24.4, −23.4] km s −1 −1−1−1
GALACTIC LONG. G A L A C T I C L A T . Fig. 6.
Overlay of mean T ∗ R -values (contours) in the velocity range from − − − and the MSX band A emission(greyscale). Every image represents the CO emission distribution averaged in a velocity interval of 1 km s − . The velocity intervalis indicated in the bottom right corner of each image. The lowest temperature contour is 0.84 K ( ∼
12 rms). The contour spacingtemperature is 1.4 K.molecular hydrogen column density, N (H ), and the integratedmolecular emission, I CO ( ≡ R T ∗ R d v). The conversion between I CO and N (H ) is given by the equation N (H ) = (1 . ± . × I CO (cm − ) (3)(Digel et al. 1996; Strong & Mattox 1996). The total molecularmass M tot , was calculated through M tot = ( m sun ) − µ m H X Ω N (H ) d (M ⊙ ) (4)where m sun is the solar mass ( ∼ × g), µ is the mean molec-ular weight, assumed to be equal to 2.8 after allowance of a rela-tive helium abundance of 25% by mass (Yamaguchi et al. 1999;Yamamoto et al. 2006), m H is the hydrogen atom mass ( ∼ × − g), Ω is the solid angle subtended by the CO featurein ster, d is the distance, expressed in cm, and M tot is given insolar masses. For Component 1, we obtain a mean column den-sity of N (H ) = (1.9 ± × cm − , and a total molecularmass of M tot = (7.6 ± × M ⊙ . The uncertainty in M tot ( ∼
28 %) stems from the distance uncertainty and from the errorquoted for the coe ffi cient in Eq. 3. Areas having T ∗ R ≥ ff erence with the value derived by Yamaguchi et al. (1999), (500 M ⊙ ), is due to both the sub-sampled observations of these authors and the fact that COline is a better tracer of high density regions.The mean volume density ( n H ) of Component 1 can be de-rived from the ratio of its molecular mass and its volume con-sidering that the volume of Component 1 is the result of theaddition of an ellipsoid with mayor and minor axis of 7 . ′ . ′
5, respectively, centered at (l,b) ≈ (289 . ◦ + . ◦
15) (which in-cludes clump A), and a sphere of 4 . ′ (l,b) ≈ (289 . ◦ + . ◦
04) (which includes clump B). We derived n H ≈
400 cm − . Taking into account the mass and volume un-certainties, we derive a conservative density error of about 60 %( ∼
240 cm − ). Figure 7 shows the radio continuum images at 843, 4800, and8640 MHz (in contours) superimposed on the DSSR image (ingrayscale) in a region of ∼ ′ × ′ centred on NGC 3503 .The continuum images show an extended source coinci-dent with NGC 3503 . The maxima at the three frequencies co-incide with the brightest optical emission region. The source,which has a good morphological correspondence with the op-
7. U. Duronea et al.: NGC 3503 and its molecular environment G A L A C T I C L A T . GALACTIC LONG.289 34 33 32 31 30 29 28 2700 121110090807060504 -0.005 0 0.005 0.01 0.015 0.02 0.025 300 320 340 360 380 400 420 440 460 480 G A L A C T I C L A T . GALACTIC LONG.289 34 33 32 31 30 29 28 2700 121110090807060504 G A L A C T I C L A T . GALACTIC LONG.289 34 33 32 31 30 29 28 2700 121110090807060504
012 0
G289.49+0.15
ANGULAR DISTANCE ( deg )
GALACTIC LONG. − E M ( p c c m ) G A L A C T I C L A T . G A L A C T I C L A T . GALACTIC LONG.
Fig. 7.
Upper left panel:
Radio continuum image at 843 MHz (contours) superimposed to the DSSR image (grayscale). Contours lev-els go from 4 mJy beam − ( ∼ − in steps of 4 mJy beam − , and from 20 mJy beam − in steps of 10 mJy beam − . Upper right panel:
Radio continuum image at 4800 MHz (contours) superimposed to the DSSR image (grayscale). Contours levelsgo from 2.4 mJy beam − ( ∼ − in steps of 2 mJy beam − , and from 10.4 mJy beam − in steps of 4 mJy beam − .The symmetry axis of the nebula is depicted by the dotted line. Lower left panel:
Radio continuum image at 8640 MHz (contours)superimposed to the DSSR image (grayscale). Contours levels go from 15 mJy beam − ( ∼ − . Lowerright panel:
Emission measure profile obtained from the 4800 MHz radio continuum image along the symmetry axis of NGC 3503 .tical emission, exhibits a cometary shape with its symmetryaxis indicated by the dotted line in the image at 4800 MHz(Fig. 7, top right panel). The fainter emission area detectedat 843 and 4800 MHz to the northeast of NGC 3503 coincideswith MBO and appears to be linked to the emission of thenebula at ( l,b ) ≈ (289 . ◦ + . ◦ (l,b) ≈ (289 . ◦ + . ◦
08) at 843 MHz depicts some morpho-logical correspondence with the bright rim of SFO 62. Thepoint-like source placed at ( l,b ) ≈ (289 . ◦ + . ◦
15) without op-tical emission counterpart is labelled G289.49 + F =
100 mJy, F = F = α = –0.92 ± ν ∝ ν α ), which suggests that G289.49 + α = − ± ii region. Table 1.
Radio continuum flux densities measurements ofNGC 3503
Frequency 843 MHz 4800 MHz 8640 MHzFlux density (mJy) 320 ±
30 270 ±
50 240 ± We estimate the emission measure EM (cid:16) = R n e dl (cid:17) along thesymmetry axis of the nebula using the image at 4800 MHz. Thebrightness temperature ( T b ) is related to the optical depth ( τ ) by T b ( ν ) = T e × (1 − e − τ ) (5)where T e = ±
700 K (Quireza et al. 2006) is the electrontemperature and the optical depth τ is given by τ = . × − T − . e ν − . EM (6)
8. U. Duronea et al.: NGC 3503 and its molecular environment
Fig. 8.
Left panel
Composite image of NGC 3503 and its environs. Red and green show emission at 8.13 µ m (MSX) and 24 µ m(MIPSGAL). The color scale goes from 1 × − to 4 × − W m − sr − and from 27 to 100 MJy ster − , respectively. The whiterectangle encloses the IR counterpart of NGC 3503 (IRK). Right panel : Composite image of NGC 3503 . Red, green and blue showemission at 8 µ m (IRAC-GLIMPSE), 24 and 70 µ m (MIPSGAL), respectively. Colors scales range from 40 to 500 MJy ster − (24 µ m), from 300 to 2000 MJy ster − (70 µ m), and from 30 to 400 MJy ster − (8 µ m). The location of the members of Pis 17 isindicated with black crosses. MSX and 2MASS candidate YSOs are indicated with red and green crosses (see text).In this last expresion, ν is given in GHz, and EM in pc cm − .The EM profile is shown in the lower right panel of Fig. 7. Asexpected, two maxima are observed, one at the brightest sec-tion of NGC 3503 and the other coincident with MBO, indicat-ing that the electron density is higher towards these regions thanin the sorroundings. The EM profile of NGC 3503 is consistentwith the cometary shape morphology of the nebula seen in theradio continuum and optical images. The EM profile shows asharp border towards the west of the nebula, while towards theeast it decreases smoothly, indicating the existence of an elec-tron density gradient. The observed EM profile is in agreementwith the results of Copetti et al. (2000), who reported the ex-istence of an electron density gradient in the east-west direc-tion. Using the peak values EM = − (obtained at (l,b) ≈ (289 . ◦ + . ◦
11) for NGC 3503 ) and EM = − (obtained at (l,b) ≈ (289 . ◦ + . ◦
18) for MBO), and consideringa pure hydrogen plasma with a depth along the line of sightequal to the size in the plane of the sky, electron density esti-mates n e = ±
14 cm − and n e = ± − are obtained forNGC 3503 and MBO, respectively.As a di ff erent approach to NGC 3503 , the rms electrondensity and ionized mass can be obtained using the spheri-cal model of Mezger & Henderson (1967) and the flux den-sity at 4800 MHz. Assuming a constant electron density, anda radious R HII = n e = ±
13 cm − and M HII = ± ⊙ . Thus, the electron density of NGC 3503 isabout a factor of 5 lower than the density of Component 1, whichclearly indicates that the nebula has expanded. An estimate ofthe filling factor ( f = p n e / n ′ e ) can be obtained by taking into ac- count the maximum electron density derived from optical lines( n ′ e = + − cm − ; Copetti et al. 2000). We obtain f = f = ⊙ .Regarding MBO, rms electron densities and masses es-timated from the image at 843 MHz are n e ≃
33 cm − and M HII ≃ ⊙ , respectively. The emission distribution at 8.13 µ m (MSX-A band) superposedto the image at 24 µ m (MIPSGAL) is shown in the left panel ofFig. 8. The MSX Band A includes the strong emission features at7.7 and 8.6 µ m attributed to PAH molecules, which are consid-ered tracers of UV-irradiated Photodissociated Regions (PDR)(Hollenbach & Tielens 1997). This image displays a small andstrong feature, from hereonwards dubbed the
IR Knot , or IRKfor short (indicated with a white rectangle), which is coinci-dent with the location of NGC 3503 , and a weaker and moreextended emission region detected in the north-western area ofthe image. The last feature will be referred to as the
ExtendedIR Emission , or EIE for short. IR emission at 8.3 and 24 µ m ap-pears mixed along the whole feature. A region of low IR emis-sion is seen between the IRK and the EIE. The bright rim ofSFO 62 is detected in the IR as the bright filament observedfrom ( l , b ) ≃ (289 . ◦ + . ◦
03) to ( l , b ) ≃ (289 . ◦ + . ◦
9. U. Duronea et al.: NGC 3503 and its molecular environment G A L A C T I C L A T . GALACTIC LONG.289 35 30 25 20 1500 15100500-00 05 30354045 G A L A C T I C L A T . GALACTIC LONG.289 35 30 25 20 1500 15100500-00 05 5.05.56.06.57.07.5
30 K35 K40 K45 K G A L A C T I C L A T . NE GALACTIC LONG. G A L A C T I C L A T . G A L A C T I C L A T . Fig. 9.
Upper panel: T d distribution estimated from HIRESIRAS images at 60 and 100 µ m. Contour levels go from 25 to31 K in steps of 1 K, and from 33 K in steps of 2 K. Lowerpanel:
Overlay of the T d distribution (contours) and the DSSRimage.The right panel of Fig. 8 shows a detailed image of the IRK,at 8.0 µ m ( Spitzer -IRAC band 4, in red), 24 µ m (MIPSGAL,in green), and 70 µ m (MIPSGAL, in blue). The emissions at 8and 70 µ m show a bright half shell-like feature, which enclosesthe position of the stars of Pis 17, indicated as black crosses.Arc-shaped faint filaments are also detected in the 8 µ m emis-sion in the northeastern section of IRK. The morphology ofthe IR emission is compatible with the cometary-shape of theH ii region and very likely indicates the presence of a PDR be-tween NGC 3503 and clump A. The MIPSGAL emission at 24 µ m, which is projected at the center of the structure, indicatesthe existence of warm dust inside the H ii region (as the cases ofN10 and N21, Watson et al. 2008).IRAS 60 and 100 µ m data show dust with color tempera-tures between about 20 to 100 K, which corresponds to the “cooldust component” (see Sreenilayam & Fich 2011, and referencestherein). A dust temperature ( T d ) map was produced using theequation T d = . / ln( B n ) (K) (7)(Draine 1990; Whittet 1992; Cichowolski et al. 2001), where B n = (3 + n ) ( F / F ) is the modified Planck function, with F and F being the 100 µ m and 60 µ m fluxes, respectively.The parameter n = ffi ciencyof the dust ( k ν ∝ ν n , normalized to 40 cm g − at 100 µ m). The obtained dust temperature map is shown in the upper panelof Fig. 9. Dust temperature goes from 25 K to 46 K. Fig. 9(lower panel) also shows that the region with the highest dusttemperatures ( ∼
46 K) coincides with the brightest section ofNGC 3503 . This temperature is in good agreement with thoseobtained in RCW 121 and RCW 122 (Arnal et al. 2008) and inNGC 6357 (Cappa et al. 2011) although they are slightly higherthan typical values for H ii regions ( ∼
30 K) (see Cappa et al.2008; Cichowolski et al. 2009; Vasquez et al. 2010). The stellarUV radiation field of the stars in NGC 3503 is responsible forthe heating of the dust immersed in the H ii region. A noticeablefeature is a fringe of cold ( ∼
25 K) dust surrounding the po-sition of NGC 3503 , which coincides with the region showinglow emission at 8.3 µ m placed between IRK and EIE describedabove. An extended region of warmer ( ∼
36 K) dust is detectedto the northeast of NGC 3503 , coincident with the position ofMBO, indicating that this optical feature has a radiatively heateddust component. In Table 2 we summarized the IR and dust pa-rameters for IRK and EIE. The averaged dust temperatures ( ¯ T d )are calculated using Eq. 7 and flux densities at 60 and 100 µ mobtained integrating several polygons around the sources. Dustmasses are derived from M d = m n F d ( B . n −
1) (M ⊙ ) (8)where d the distance in kpc, F is given Jy, and m . = × − . Based on the molecular gas mass derived forComponent 1, the ionized gas of NGC 3503 and MBO, and thedust masses obtained for the IRK and the EIE ( ∼
10 M ⊙ ), wederive a mean weighted gas-to-dust ratio of about ∼ ± ∼ ±
50; Tokunaga 2000). Note how-ever that dust with temperatures T d .
20 K (“cold dust compo-nent”), which is detected at wavelengths not surveyed by IRAS( λ > µ m), dominates the dust mass by a factor larger than ∼
70 with respect to the cool dust component (Sreenilayam & Fich2011). This notoriously decreases the gas-to-dust ratio. The dusttemperature estimated for EIE may suggest the existence of ion-izing sources inside Component 1. However, a search for OBstars embedded in Component 1 performed using the availableVIZIER catalogues failed to detect such sources.To investigate the presence of protostellar candidates possi-bly related to NGC 3503, we used data from the MSX, 2 MASS,and IRAS point source catalogs. We searched for point sourcesin a region of about 10 ′ in size centered at the position of the H ii region. Taking into account sources with flux quality q >
2, a to-tal of 69 MSX point sources were found projected onto the area.Based on F / F and F / F ratios, where F , F , F , and F are the fluxes at 8.3, 12, 14, and 21 µ m, respectively, and af-ter applying the criteria summarized by Lumsden et al. (2002),we were left with three sources with F / F > F / F <
1, which are classified as compact H ii regions (CH ii ). This isindicative of stellar formation. The location of these sources isindicated with red crosses in Fig. 8.Using the 2 MASS catalog (Cutri et al. 2003), which pro-vides detections in J , H and K s bands, we searched forpoint sources with infrared excess. Taking into account sourceswith signal-to-noise ratio (S / N) >
10 (corresponding to quality“AAA”), we found 5528 sources projected onto a circular regionof 5 ′ in radius. Following Comer´on et al. (2005), we determinedthe parameter q = ( J - H ) - 1.83 × ( H - K s ). Sources with q < –0.15 are classified as objects with infrared excess , i.e. can-didate YSOs. By applying the above criteria, we found only 9sources projected onto NGC 3503 . The location of these sourcesis also indicated in Fig. 8 with green crosses. The location of the
10. U. Duronea et al.: NGC 3503 and its molecular environment
Table 3.
Candidate YSOs obtained from the MSX and 2 MASS catalogues. ( l , b ) MSX source F (Jy) F (Jy) F (Jy) F (Jy)1 289.48, + + + + + + l , b ) 2 MASS source J (mag) H (mag) K (mag)4 289.49, + + + + + + + + + Table 2.
Main infrared parameters inferred from IRAS fluxes at60 µ m and 100 µ m Parameter
IRK EIES (Jy) 1470 ±
20 700 ± (Jy) 2170 ±
25 1970 ± T d (K) ∼ ∼ M d (M ⊙ ) ∼ ∼ + l , b ) = (289 . ◦ + . ◦
14 ) (see Fig. 8, right panel), and ontoa small enhancement in the radio continuum emission detectedat 4800 MHz (see Fig. 10 below). These characteristics makethis object an excellent candidate for investigating star forma-tion with high angular resolution observations. The names of theYSO candidates, their position, fluxes and magnitudes at di ff er-ent IR wavelengths are listed in Table 3.
4. Discussion
The number of ionizing Lyman continuum photons ( N Lyc )needed to sustain the ionization in NGC 3503 can be calculatedusing the equation given by Simpson & Rubin (1990) N Lyc = . × T − . e S (4800 MHz) d (9)where d is the distance in kpc, and S (4800 MHz) is the flux den-sity at 4800 MHz in Jy. We obtain N Lyc = (1.8 ± × s − .This number is a lower limit to the total number of Lymancontinuum photons required to maintain the gas ionized, sinceabout 25 - 50 % of the UV photons are absorbed by inter-stellar dust in the H ii region (Inoue 2001). Consequently, weneed N Lyc ≈ × s − . The number of Lyman continuumphotons emitted by a B0 V star is N ∗ Lyc = × s − (Sternberg et al. 2003), which is capable of ionizing NGC 3503 ,in agreement with previous results by Thompson et al. (2004)and Pinheiro et al. (2010). The last authors asserted one B0 Vstar and three B2 V stars, belonging to the open cluster Pis17,inside the nebula as ionizing sources. Since the ionizing photonsemitted by a B2 V star are significantly fewer than those of a B0V star, their contribution to the energetics of the nebula can beneglected.In Sec. 3.2 we pointed out on the discrepancy between theelectron and molecular densities. The density of Component 1 isabout a factor of 5 higher than the electron density of NGC 3503 ,which clearly indicates that NGC 3503 has been expanding asa result of unbalanced pressure between ionised and moleculargas. In order to estimate the dynamical age ( t dyn ) of the H ii regionwe used the model of Dyson & Williams (1997). The radious ofan H ii region ( R HII ) in a uniform medium is given by R HII R S = + s t dyn R S ! / (10)were R S is the radious of the Str¨omgren sphere (Str¨omgren 1939)before expanding, given by R S = (cid:16) N ∗ Lyc / π (2 n H ) α β (cid:17) / , andv s is the sound speed in the ionized gas ( ∼
10 km s − ). For R HII ≃ . ′ n H = ±
240 cm − (see Section 3.1.3), and N ∗ Lyc = × s − , we infer t dyn ≈ × yr. The expansion velocity ofthe H ii region can be estimated by means of ˙ R HII = v s (cid:16) R HII R s (cid:17) − / ,yielding an expansion velocity of about ∼ − . The ob-tained dynamical age and expansion velocity are in agreementwith those obtained in typical H ii regions (Gum 31, Cappa et al.2008; Sh2-173, Cichowolski et al. 2009).The total number of ionizing Lyman continuumphotons needed to sustain the ionization in MBO is N Lyc ≈ × seg − . To search for stars that can pro-vide the necessary UV photons, we used the availableVIZIER catalogues. CP-59 2951 is a B1 V star placed at (l,b) = (289 . ◦ + . ◦ d = ± N ∗ Lyc = × s − (Smith et al. 2002). Although
11. U. Duronea et al.: NGC 3503 and its molecular environment the distance of this star is in agreement (within errors) with thedistance of NGC 3503 , the value of N ∗ Lyc is almost one order ofmagnitude lower than the required to ionize MBO. The presenceof Pis17 at a projected distance of 3.4 pc ( ∼ ′ at a distance of2.9 kpc) indicates that the contribution of the star cluster to theionization of MBO can not be ruled out. The presence of 12 candidate YSOs projected onto theNGC 3503 -region suggests that they may have been triggeredby the expansion of the H ii region through the “collect and col-lapse” model, which indicates that expanding nebulae compressgas between the ionization and the shock fronts, leading to theformation of molecular cores where new stars can be embed-ded. Using the analytical model of Whitworth et al. (1994) forthe case of expanding H ii regions, we derived the time whenthe fragmentation may have occurred ( t frag ), and the size of theH ii region at t frag ( R frag ), which are given by t frag [10 yr ] = . a / n − / N − / (11) R frag [ pc ] = . a / n − / N / (12)where a is the sound velocity in units of 0.2 km s − , n ≡ n H / N ≡ N ∗ Lyc / . Adopting for this region 0.3 kms − for the sound velocity, which corresponds to temperaturesof 10-15 K in the surrounding molecular clouds (see Section3.1.2), we obtained t frag ∼ × yr, and R frag ∼ t frag and R frag are larger than t dyn and R HII (see Section 4.1), we can conclude that the fragmenta-tion in the edge of NGC 3503 is doubtful.
In section 3.2 we reported the existence of a velocity gradientacross clump A. Velocity gradients in molecular cores / clumpswere usually interpreted as gravitationally bound rotation mo-tions (particulary in dense and small molecular cores). Severaltheoretical models predict cloud flattening perpendicular to therotation axis in response to centrifugal stress, which at firstglance seems to be suitable for clump A. In order to investi-gate the dynamical stability of clump A, we use the parameter β defined by Goodman et al. (1993) to quantify the dynamicalrole of rotation by comparing the rotational kinetic energy to thegravitational energy. Thus, β can be written as β = (1 / I ω ′ q G M / R = (1 /
2) ( p / q ) ω ′ R G M (13)where I is the moment of inertia ( I = pMR ), qGM / R isthe gravitational potential energy, and ω ′ = ω/ sin ( i ), where i is the inclination of the cloud along the line of sight.Considering p / q = i ) = ω = − pc − (see Sect 3.1.2), R ∼ M ∼ × M ⊙ ), we obtain β ≈ β indicates that the e ff ect of rotation,if exists, is not significant in mantaining the dynamical stabilityof clump A. This might weaken the rotating cloud interpreta-tion. Furthermore, Figs. 4 and 6 show that only molecular gas atmore negative velocities ( ∼ −
27 km s − / −
26 km s − ) coincidewith NGC 3503 , which suggests that the velocity gradient mightbe a direct consequence of an interaction between clump A andNGC 3503 . Therefore, although rotation can not be entirely ruled outwith the present data, we consider di ff erent origins for the ve-locity gradient observed across clump A, namely, 1) the expan-sion of the nebula at ∼ − (see Sect 4.1) has been ac-cumulating molecular gas behind the shock front which orig-inated the expansion of clump A at approximately the samevelocity of the ionized gas, as expected according to the mod-els of Hosokawa & Inutsuka (2006), 2) the velocity gradient ofclump A is the consequence of a collision between Component1 and another molecular cloud, which in turn, might have in-duced the formation of Pis 17 and the candidate YSOs reportedin Sect. 3.3. Although they are rare, cloud-cloud collisions canlead to gravitational instabilities in the dense, shocked gas, re-sulting in triggered star formation (see Elmegreen 1998, and ref-erences therein), 3) clump A is actually composed by di ff erentsubclumps at di ff erent velocities, which are not resolved by theNANTEN observations. In this case, NGC 3503 might be relatedwith a subclump at more negative velocities.Further high-resolution studies with instruments like APEXmay help to clarify this question. As mentioned in Sect 3.1.2, clump A exhibits an excita-tion temperature T exc ≥ T ∼ G ) in unitsof the Habing (1968) FUV flux (1.6 × − ergs cm − s − ), with G ∝ N − / n / e L ⋆ χ (Tielens 2005), where χ is the fraction lu-minosity over 6 eV, and L ⋆ is the luminosity of the star. Adopting N Ly = × s − and L = × L ⊙ (Sternberg et al.2003), χ =
1, and considering n e ≃
75 – 154 cm − (see Sect.3.2), a range of G ≃ (0.5 - 1.5) × is obtained. The dis-tribution of dust temperature in the PDR ( T pdrd ) is governed bythe absorption of the stellar photons which are reemited by dustgrains as IR photons. Following Tielens (2005), we can relate T pdrd with the visual absorption along the PDR ( A v ) as( T pdrd ) ≃ ν G e ( − . A v ) + ln(3 . × − τ T ) τ T (14)where ν is the frequency at 0.1 µ m, τ = − is the e ff ec-tive optical depth at 100 µ m, and T is the temperature of theslab estimated as T = G / . Considering both, the posi-tion of NGC 3503 at the edge of clump A, and that the low valueof visual absorption derived for the members of Pis 17 ( ¯ A v = ii region, we can as-sume A v ∼ G and Eq. 14, we obtain T pdrd ≃
40 - 55 K, in ac-
12. U. Duronea et al.: NGC 3503 and its molecular environment cordance with the observed dust color temperature estimated forNGC 3503 ( ∼
46 K). This gives additional support to the exis-tence of a PDR between NGC 3503 and clump A.
Figure 10 displays a composite image of NGC3503 and its en-virons. White and light blue contours show the radio continuumemission at 4800 MHz corresponding to NGC3503 and MBO,and the CO emission, respectively. The color scale shows theemission at 8 µ m (red) and 4.5 µ m (green).As described in Sect. 3.3 and 4.3, emission attributedto PAHs encircles the southern and western borders of theH ii region, indicating the location of the PDR. The positionof NGC 3503 near the border of clump A and the existenceof an electron density gradient along the symmetry axis ofthe H ii region (with the higher electron densities close to thestrongest CO emission) suggest that the ionised gas is beingstreamed away from the densest part of the molecular cloud. Thisis indicative that NGC 3503 is a blister-type H ii region (Israel1978) that probably has undergone a champagne phase.The so-called Champagne flow model (Tenorio-Tagle 1979;Bedijn & Tenorio-Tagle 1981; Tenorio-Tagle & Bedijn 1982)proposes that an expanding H ii region placed at the edge of amolecular cloud eventually reaches the border of the cloud andexpands freely in the lower density surrounding gas, originat-ing an extended H ii region with a characteristic density distri-bution. The H ii region becomes density bounded towards thelower density region, while it is ionization bounded towardsthe molecular cloud. This scenario was formerly proposed byCopetti et al. (2000) to explain the electron density gradient ob-served across NGC 3503. The location of Pis 17 close to thebrightest radio continuum region is consistent with a projec-tion e ff ect (Yorke et al. 1983). In this scenario, Pis 17 originatedNGC 3503, which reached the northeastern border of clump Aafter 2 × yr. This time represents a lower limit to the ageof the H ii region, since the inferred main-sequence life time ofthe main ionizing star is about (3 -5) × yr (Massey 1998).The leakage of ionized gas and UV photons might have con-tributed to the formation of MBO, since its location (along thesymmetry axis of NGC 3503 ) and its low density (about half ofNGC 3503 ) suggest that this feature may consist of ionized gaswhich has scaped from NGC 3503 after a time of 2 × yr.The velocities of both the molecular and ionized gas arecompatible with the champagne scenario. The mean velocity ofComponent 1 (–24.7 km s − , see Sect. 3.1.1) corresponds to thenatal cloud where NGC 3503 originated. Molecular gas havingmore negative velocities is mainly linked to clump A and mayrepresent material moving away from the ionized gas placed infront of the H ii region (as a result of expanding motions or anexternal impact over Component 1). The ionized gas, having avelocity of –21 km s − (Georgelin et al. 2000), might be reced-ing the observer. We note, however, that the Fabry-Perot obser-vations by Georgelin et al. (2000) have a spectral resolution of 5km s − .NGC 3503 resembles the two well studied cases of S305 andS307. These H ii regions show a non spherical morphology in the1465 MHz map of Fich (1993), and they are placed close to COemission peaks, which suggests that they are located close to thehottest part of their parental molecular clouds (see Russeil et al.1995, and references therein). Both objects show significantelectron density dependence on position, which can also be in-terpreted as radial gradients (Copetti et al. 2000). Furthermore,both H ii regions show discrepancies of about ∼
12 km s − be- tween the velocities of the ionized and molecular gas, prob-ably due to a champagne e ff ect producing a flow toward theobserver (Russeil et al. 1995). In the case of S307, the cham-pagne scenario is reinforced by its half-shell shape, and the highdensity in the brightest part of the nebula (Felli & Harten 1981;Albert et al. 1986)According to the champagne scenario, a velocity gradi-ent along the symmetry axis of the nebula can be expected(Garay et al. 1994). In the case of NGC 3503 increasing veloci-ties with the distance to the molecular cloud are expected. Thus,high spectral resolution radio recombination are required to an-alyze the ionized gas velocities and may give additional supportto our interpretation.
5. Summary
NGC 3503 is a bright H ii region of ∼ ′ in size centered at (l,b) = (289 . ◦ + . ◦
12) located at a distance of 2.9 ± ii region and analyzingthe interaction of the nebula and Pis 17 with their molecularenvironment, we analyzed CO(1-0) data of a region of 0 . ◦ . ′
7, radio continuum data of NGC 3503 at 4800and 8640 MHz obtained with the ATCA telescope (with syn-thesized beams of 21 ′′ and 13 ′′ , respectively), and data at 843MHz retrieved from SUMSS, and available IRAS, MSX, IRAC-GLIMPSE, and MIPSGAL images.The analysis of the CO data allowed the molecular gas linkedto the nebula to be mapped. This molecular gas (Component 1)has a mean velocity of –24.7 km s − , a total mass and densityof (7.6 ± × M ⊙ and 400 ±
240 cm − , respectively, anddisplays two clumps centered at (l,b) = (289 . ◦ + . ◦
12) (clumpA) and (l,b) = (289 . ◦ + . ◦
03) (clump B). The morphologicalcorrespondence of the molecular emission with a large patchof high optical absorption adjacent to NGC 3503 is excellent.NGC 3503 is projected near the border of clump A, with thestrongest molecular emission adjacent to the highest electrondensity regions. The agreement of the molecular velocities withthe velocity of the ionizad gas (–21 km s − ) as well as the mor-phological correspondence with the nebula indicate that clumpA is associated with NGC 3503. The more negative velocities ofthe gas in clump A, coincident with the H ii region, are probablydue to the expansion of the H ii region or an external impact overComponent 1.The analysis of the radio continuum images confirms theelectron density gradient previously found by Copetti et al.(2000). The images show that NGC 3503 exhibits a cometarymorphology, with the higher density region near the maximumof clump A. The rms electron density and the ionized massamount to 54 ±
13 cm − and 9 ± ⊙ , respectively. A low den-sity ionized region (MBO) located close to the lower electrondensity area of NGC 3503 is also identified in these images.Strong emission at 8 µ m surrounds the bright radio contin-uum region of the nebula, indicating the presence of a photodis-sociated region at the interface between the ionized region andclump A. MIPSGAL emission at 24 µ m shows the existence ofwarm dust inside the H ii region. Based on high resolution IRASimages at 60 and 100 µ m, a mean dust color temperature and dustmass of 37 K and 3 M ⊙ were estimated for NGC 3503 . The pres-ence of candidate YSOs projected onto the HII region, detectedusing MSX and 2MASS point source catalogues, suggests the
13. U. Duronea et al.: NGC 3503 and its molecular environment
Fig. 10.
Composite image of NGC 3503 and its environs. Red and green show emission at 8 and 4.5 µ m (IRAC-GLIMPSE),respectively. White and light blue contours show the radiocontinuum 4800 MHz and CO line emission, respectively.existence of protostellar objects in the neighbourhood of NGC3503, although there is no clear evidence of triggered star for-mation.The location of NGC 3503 at the edge of a molecular clumpand the electron density gradient in the H ii region suggeststhat NGC 3503 is a blister-type H ii region that has undergonea champagne phase. In this scenario, the massive stars of Pis 17created NGC 3503, which reached the border of the molecularcloud after 2 × yr. The leakage of ionized gas and UV pho-tons have probably contributed in the formation of MBO. Thus,the spatial distribution of the molecular gas and PAHs give ad-ditional support to a scenario first proposed by Copetti et al.(2000). The proposed scenario for NGC 3503 may explain theslight di ff erence between the main velocity of its parental molec-ular cloud (–24.7 km s − ) and the velocity of its ionized gas(–21 km s − ). High resolution radio recombination line obser-vations may help to confirm the proposed scenario.. Acknowledgements.
We especially thank Dr. James S. Urquhart and Dr. Mark A.Thompson for making their unpublished radio continuum images at 4800 MHzand 8640 MHz available to us. We acknowledge the anonymous referees fortheir helpful comments that improved the presentation of this paper. This projectwas partially financed by the Consejo Nacional de Investigaciones Cient´ıficasy T´ecnicas (CONICET) of Argentina under projects PIP 112-200801-02488and PIP 112-200801-01299, Universidad Nacional de La Plata (UNLP) under project 11G / / ff members ofLas Campanas Observatory of the Carnegie Institute of Washington. We thank allmembers of the NANTEN sta ff , in particular Prof. Yasuo Fukui, Dr. ToshikazuOnishi, Dr. Akira Mizuno, and students Y. Moriguchi, H. Saito, and S Sakamoto.We also would like to thank Dr. D. Miniti (Pont´ıfica Universidad Cat´olica, Chile)and Mr. F Bareilles (IAR) for their involment in early stages of this project References
Albert, C. E., Schwartz, P. R., Bowers, P. F., & Rickard, L. J. 1986, AJ, 92, 75Arnal, E. M., Duronea, N. U., & Testori, J. C. 2008, A&A, 486, 807Bedijn, P. J. & Tenorio-Tagle, G. 1981, A&A, 98, 85Benjamin, R. A., Churchwell, E., Babler, B. L., et al. 2003, PASP, 115, 953Bock, D. C.-J., Large, M. I., & Sadler, E. M. 1999, AJ, 117, 1578Brand, J. & Blitz, L. 1993, A&A, 275, 67Buscombe, W. & Foster, B. E. 1999, MK spectral classifications. Fourteenthgeneral catalogue, epoch 2000, including UBV photometry., ed. Buscombe,W. & Foster, B. E.Cappa, C., Niemela, V. S., Amor´ın, R., & Vasquez, J. 2008, A&A, 477, 173Cappa, C. E., Barb´a, R., Duronea, N. U., et al. 2011, MNRAS, 1027Carey, S. J., Noriega-Crespo, A., Price, S. D., et al. 2005, in Bulletin of theAmerican Astronomical Society, Vol. 37, American Astronomical SocietyMeeting Abstracts, 1252– + Cichowolski, S., Pineault, S., Arnal, E. M., et al. 2001, AJ, 122, 1938Cichowolski, S., Romero, G. A., Ortega, M. E., Cappa, C. E., & Vasquez, J.2009, MNRAS, 394, 900Comer´on, F., Schneider, N., & Russeil, D. 2005, A&A, 433, 955
14. U. Duronea et al.: NGC 3503 and its molecular environment
Copetti, M. V. F., Mallmann, J. A. H., Schmidt, A. A., & Casta˜neda, H. O. 2000,A&A, 357, 621Cutri, R. M., Skrutskie, M. F., van Dyk, S., et al. 2003, VizieR Online DataCatalog, 2246, 0Dickman, R. L. 1978, ApJS, 37, 407Digel, S. W., Grenier, I. A., Heithausen, A., Hunter, S. D., & Thaddeus, P. 1996,ApJ, 463, 609Draine, B. T. 1990, in Astrophysics and Space Science Library, Vol. 161, TheInterstellar Medium in Galaxies, ed. H. A. Thronson Jr. & J. M. Shull, 483–492Dreyer, J. L. E. & Sinnott, R. W. 1988, NGC 2000.0, The Complete New GeneralCatalogue and Index Catalogue of Nebulae and Star Clusters by J.L.E. Dreyerand R. W. Sinnott , ed.Dyson, J. E. & Williams, D. A. 1997, The physics of the interstellar medium, ed.Dyson, J. E. & Williams, D. A.Elmegreen, B. G. 1998, in Astronomical Society of the Pacific ConferenceSeries, Vol. 148, Origins, ed. C. E. Woodward, J. M. Shull, & H. A. ThronsonJr., 150– + Elmegreen, B. G. & Wang, M. 1987, In Molecular Clouds in the Millky Way andexternal galaxies (eds. R. L. Dickman, R. L. Snell, & J. S. Young), (Berlin:Springer)Felli, M. & Harten, R. H. 1981, A&A, 100, 42Fich, M. 1993, ApJS, 86, 475Fowler, J. W. & Aumann, H. H. 1994, in Science with High Spatial ResolutionFar-Infrared Data, ed. S. Terebey & J. M. Mazzarella, 1– + Garay, G., Lizano, S., & Gomez, Y. 1994, ApJ, 429, 268Georgelin, Y. M., Russeil, D., Amram, P., et al. 2000, A&A, 357, 308Goodman, A. A., Benson, P. J., Fuller, G. A., & Myers, P. C. 1993, ApJ, 406,528Habing, H. J. 1968, Bull. Astron. Inst. Netherlands, 19, 421Herbst, W. 1975, AJ, 80, 212Hollenbach, D. J. & Tielens, A. G. G. M. 1997, ARA&A, 35, 179Hosokawa, T. & Inutsuka, S.-i. 2006, ApJ, 646, 240Inoue, A. K. 2001, AJ, 122, 1788Israel, F. P. 1978, A&A, 70, 769Keto, E. R. & Ho, P. T. P. 1989, ApJ, 347, 349Kutner, M. L. & Ulich, B. L. 1981, ApJ, 250, 341Lumsden, S. L., Hoare, M. G., Oudmaijer, R. D., & Richards, D. 2002, MNRAS,336, 621Massey, P. 1998, in Astronomical Society of the Pacific Conference Series, Vol.142, The Stellar Initial Mass Function (38th Herstmonceux Conference), ed.G. Gilmore & D. Howell, 17– + McLean, B. J., Greene, G. R., Lattanzi, M. G., & Pirenne, B. 2000,in Astronomical Society of the Pacific Conference Series, Vol. 216,Astronomical Data Analysis Software and Systems IX, ed. N. Manset,C. Veillet, & D. Crabtree, 145– + Mezger, P. G. & Henderson, A. P. 1967, ApJ, 147, 471Mo ff at, A. F. J. & Vogt, N. 1975, A&AS, 20, 125Moriguchi, Y., Yamaguchi, N., Onishi, T., Mizuno, A., & Fukui, Y. 2001, PASJ,53, 1025Penzias, A. A. & Burrus, C. A. 1973, ARA&A, 11, 51Pinheiro, M. C., Copetti, M. V. F., & Oliveira, V. A. 2010, A&A, 521, A26 + Price, S. D., Egan, M. P., Carey, S. J., Mizuno, D. R., & Kuchar, T. A. 2001, AJ,121, 2819Quireza, C., Rood, R. T., Bania, T. M., Balser, D. S., & Maciel, W. J. 2006, ApJ,653, 1226Rodgers, A. W., Campbell, C. T., & Whiteoak, J. B. 1960, MNRAS, 121, 103Russeil, D. & Castets, A. 2004, A&A, 417, 107Russeil, D., Georgelin, Y. M., Georgelin, Y. P., Le Coarer, E., & Marcelin, M.1995, A&AS, 114, 557Schmidt-Kaler, T. 1982, In Landolt-Bornstein New Series, Group VI, Vol. 2b.(eds. K. Schaifers & H. H. Voigt), (Springer-Verlag, Berlin)Simpson, J. P. & Rubin, R. H. 1990, ApJ, 354, 165Smith, L. J., Norris, R. P. F., & Crowther, P. A. 2002, MNRAS, 337, 1309Sreenilayam, G. & Fich, M. 2011, AJ, 142, 4Sternberg, A., Ho ff mann, T. L., & Pauldrach, A. W. A. 2003, ApJ, 599, 1333Str¨omgren, B. 1939, ApJ, 89, 526Strong, A. W. & Mattox, J. R. 1996, A&A, 308, L21Sugitani, K., Fukui, Y., & Ogura, K. 1991, ApJS, 77, 59Sugitani, K. & Ogura, K. 1994, ApJS, 92, 163Tenorio-Tagle, G. 1979, A&A, 71, 59Tenorio-Tagle, G. & Bedijn, P. J. 1982, A&A, 115, 207Thompson, M. A., Urquhart, J. S., & White, G. J. 2004, A&A, 415, 627Tielens, A. G. G. M. 2005, The Physics and Chemistry of the InterstellarMedium, ed. Tielens, A. G. G. M.Tokunaga, A. T. 2000, Infrared Astronomy, ed. Cox, A. N., 143– + Ulich, B. L. & Haas, R. W. 1976, ApJS, 30, 247Urquhart, J. S., Morgan, L. K., & Thompson, M. A. 2009, A&A, 497, 789 van der Tak, F. F. S. & van Dishoeck, E. F. 2000, A&A, 358, L79Vasquez, J., Cappa, C. E., Pineault, S., & Duronea, N. U. 2010, MNRAS, 405,1976Watson, C., Povich, M. S., Churchwell, E. B., et al. 2008, ApJ, 681, 1341Whittet, D. C. B. 1992, Dust in the galactic environment, ed. Whittet, D. C. B.Whitworth, A. P., Bhattal, A. S., Chapman, S. J., Disney, M. J., & Turner, J. A.1994, MNRAS, 268, 291Yamaguchi, R., Saito, H., Mizuno, N., et al. 1999, PASJ, 51, 791Yamamoto, H., Kawamura, A., Tachihara, K., et al. 2006, ApJ, 642, 307Yorke, H. W., Tenorio-Tagle, G., & Bodenheimer, P. 1983, A&A, 127, 313 Instituto Argentino de Radioastronomia, CONICET, CCT-La Plata, C.C.5., 1894, Villa Elisa, Argentina e-mail: [email protected] Facultad de Ciencias Astron´omicas y Geof´ısicas, UniversidadNacional de La Plata,Paseo del Bosque s //