A study on subarcsecond scales of the ammonia and continuum emission toward the G16.59-0.05 high-mass star-forming region
L. Moscadelli, R. Cesaroni, Á. Sánchez-Monge, C. Goddi, R.S. Furuya, A. Sanna, M. Pestalozzi
aa r X i v : . [ a s t r o - ph . S R ] N ov Astronomy&Astrophysicsmanuscript no. main˙revised˙3 c (cid:13)
ESO 2018August 27, 2018
A study on subarcsecond scales of the ammonia and continuumemission toward the G16.59 − L. Moscadelli , R. Cesaroni , ´A. S´anchez-Monge , C. Goddi , R.S. Furuya , A. Sanna , and M. Pestalozzi INAF-Osservatorio Astrofisico di Arcetri, Largo E. Fermi 5, 50125 Firenze, Italye-mail: [email protected] Joint Institute for VLBI in Europe, Postbus 2, NL-7990 AA Dwingeloo, the Netherlands The University of Tokushima Minami Jousanjima-machi 1-1, Tokushima, Tokushima 770-8502, Japan Max-Planck-Institut f¨ur Radioastronomie, Auf dem H¨ugel 69, 53121 Bonn, Germany INAF-Istituto Fisica Spazio Interplanetario, Via Fosso del Cavaliere 100, I-00133 Roma, Italy
ABSTRACT
Aims.
We wish to investigate the structure, velocity field, and stellar content of the G16.59 − Methods.
We performed Very Large Array observations of the radio continuum and ammonia line emission, complemented byCOMICS / Subaru and Hi-GAL / Herschel images in the mid- and far-infrared (IR).
Results.
Our centimeter continuum maps reveal a collimated radio jet that is oriented E–W and centered on the methanol maser disk,placed at the SE border of a compact molecular core. The spectral index of the jet is negative, indicating non-thermal emission overmost of the jet, except the peak close to the maser disk, where thermal free-free emission is observed. We find that the ammoniaemission presents a bipolar structure consistent (on a smaller scale) in direction and velocity with that of the NE–SW bipolar outflowdetected in previous CO observations. After analyzing our previous N H + (1–0) observations again, we conclude that two scenariosare possible. In one case both the radio jet and the ammonia emission would trace the root of the large-scale CO bipolar outflow.The di ff erent orientation of the jet and the ammonia flow could be explained by precession and / or a non-isotropic density distributionaround the star. In the other case, the N H + (1–0) and ammonia bipolarity is interpreted as two overlapping clumps moving with di ff er-ent velocities along the line of sight. The ammonia gas also seems to undergo rotation consistent with the maser disk. Our IR imagescomplemented by archival data allow us to derive a bolometric luminosity of ∼ L ⊙ and to conclude that most of the luminosity isdue to the young stellar object in the hot molecular core. Conclusions.
The new data suggest a scenario where the luminosity and the outflow activity of the whole region could be dominatedby two massive young stellar objects: 1) a B-type star of ∼ M ⊙ at the center of the maser / ammonia disk; 2) a massive youngstellar object (so far undetected), very likely in an earlier stage of evolution than the B-type star, which might be embedded inside thecompact molecular core and power the massive, SE–NW outflow. Key words.
Techniques: interferometric – ISM: jets and outflows – ISM: molecules – Radio continuum: ISM – Infrared: ISM
1. Introduction
The study of high-mass ( M > M ⊙ ) star formation still facesfundamental questions. Among the most important issues still tobe clarified are 1) the role played by accretion disks to conveymass onto the (proto)star; 2) the properties of the (proto)stellaroutflows (neutral vs ionized; wide-angle vs collimated); 3) theway an H ii region develops and expands. From a theoreticalpoint of view, models of massive star formation are compli-cated by the need to consider the e ff ects of the intense stellarradiation of a massive young stellar object (MYSO), which, byheating, ionizing and exerting pressure on the circumstellar gas,strongly influences the process of mass accretion and ejection(e.g., Peters et al. 2010; Cunningham et al. 2011). Models pre-dict that an H ii region is quenched, remains trapped or expandshydrodynamically, depending on the balance between the stellarradiation and thermal pressure, on one hand, and the gravita-tional pull of the massive (proto)star and the ram pressure of theinfalling material, on the other (Keto 2002, 2003).From an observational point of view, the large (typically sev-eral kpc) distances of massive stars and their origin in clus-ters make it di ffi cult to disentangle the physical and kinemat- ical properties of a single massive (proto)star from those ofother cluster members. Prior to the advent of the Atacama LargeMillimeter Array (ALMA), the angular resolution ( ≥ a few 0 . ′′ ∼ ≤ ′′ ), thermal continuum sources nearby MYSOs, whichhas been recognized as an important tool for investigating ac-cretion and ejection phenomena. The origin of such emis-sion has been interpreted in terms of photo-ionized spherical(Panagia & Felli 1975) and collimated (Reynolds 1986) stellarwinds, shock-induced radiation (Ghavamian & Hartigan 1998),and, more recently, trapped H ii regions (Keto 2003). See alsoRodr´ıguez et al. (2012) and S´anchez-Monge et al. (2013b) and − references therein, for a list of di ff erent mechanisms proposed toexplain the observation of thermal centimeter continuum emis-sion in star-forming regions.With this in mind, in the past years we have focused onthe high-mass star-forming region G16.59 − ± ∼ ∼ M ⊙ (Beuther et al. 2002a; Hill et al. 2005), millime-ter continuum and line observations reveal a number of youngstellar objects (YSOs) and a complex flow pattern, dominated bytwo almost perpendicular (NE–SW and SE–NW), massive out-flows (see Fig. 8 of B2006). 18.1 µ m emission from the molec-ular clump is marginally detected by De Buizer et al. (2005)with the 3 m NASA Infrared Telescope Facility. The VLA C-array observations at 3.6, 1.3 and 0.7 cm of Zapata et al. (2006)identify three compact continuum sources, named “a”, “b” and“c”, within a distance of ≈ ′′ . Sources “a” and “b” belong tothe molecular clump and are separated by ∼ ′′ along the SE–NW direction: source “a”, the one to NW, is detected only at0.7 cm, whereas source “b”, to SE, is visible both at 3.6 and1.3 cm. Interferometric measurements at 3 mm (F2008) and1 mm (B2006) show that the continuum emission, engulfingboth “a” and “b” sources, peaks at the position of source “a”,indicating that this source is the most embedded, and plausi-bly youngest object in the clump. The SMA detection of high-density molecular tracers (B2006) with rotational temperaturesas high as 130–150 K (inferred from the CH CN lines by F2008)close to the position of “b”, indicate that this centimeter con-tinuum source is likely associated with a hot molecular core(HMC). For the sake of simplicity, hereafter we will refer tosource “a” as mm-core and to source “b” as
HMC .Intense ( ≥
10 Jy) Class II 6.7 GHz and Class I 44.1 GHzmethanol and 22 GHz water masers have been detected to-ward the G16.59 − OH masers are associated with the HMC and tracean elongated structure of ∼ of ∼
12 M ⊙ , assuming centrifugal equilibrium. Moscadelli et al.(2011) find that most of the 6.7 GHz maser features present aregular internal V LSR gradient, which can be also interpreted interms of Keplerian rotation around a star having a similar po-sition and mass as derived from the maser 3-D velocity distri-bution. Water masers, monitored with the Very Long BaselineArray (VLBA) by S2010, appear distributed close to the HMC,with measured proper motions tracing fast ( ≈
50 km s − ) andpoorly collimated expansion to the west (see Fig. 5b of S2010).A sketch of the main emission features in the G16.59 − The mass quoted here di ff ers from that computed by S2010(35 M ⊙ ), since we have taken into account the new distance estimateof d = ff erentassumptions concerning the systemic velocity adopted and the maserspots used for the calculation (2-3 spots might not be tracing the disk).Our new estimate of ∼
12 M ⊙ uses a rotation velocity of 5.1 km s − (fora systemic velocity of 60 km s − ) at a radius of 0 . ′′ This paper reports on new observations of G16.59 − / or outflow(s). In the follow-ing, after illustrating the observations in Sect. 2, our findings willbe presented and discussed in Sects. 3, 4 and 5 and a scenariowill be eventually proposed in Sect. 6 to explain in a consistentmanner all the features of this intriguing high-mass star-formingregion. Finally, conclusions are drawn in Sect. 7.
2. Observations emission and H O masers
We observed the NH (1,1) (at rest frequency of23694.5060 MHz) and NH (2,2) (at 23722.6336 MHz)main lines, and the water maser emission (at 22235.08 MHz) inG16.59 − in the B–Array configura-tion (project code 12A-054) in three di ff erent runs of two hourson July 5, July 30, and August 6 2012. We recorded dual polar-ization using two IFs (each comprising eight adjacent 4 MHzbandwidths), one centered at a rest frequency (23708.57 MHz)halfway between the NH (1,1) and (2,2) main lines, and theother at the water maser sky frequency. Considering the VLAcorrelator capabilities in Summer 2012, this frequency setupwas selected to observe both the NH and the water maser emis-sion, to maximize the number of observable NH (1,1) and (2,2)satellite lines and, at the same time, attain a high enoughvelocity resolution to resolve the NH lines. We could observethe NH (1,1) satellite lines with velocity separation (from themain line) of ± − and +
19 km s − , and the NH (2,2)satellite lines at +
16 km s − and +
26 km s − . Correlating each4 MHz bandwidth with 128 channels, we achieved a velocityresolution of 0.39 km s − .The primary and phase calibrators were the VLA calibra-tors 3C286 and J1832-1035, respectively. The phase calibratoris separated from G16.59 − . ◦
8. Each 2-hourrun included a single 5-min scan on the primary calibrator, eight3-min scans on the phase calibrator, and a total of 1.1 hoursof on-source time. First, we applied the amplitude and phasecorrections derived from the calibrators to the strongest (refer-ence) water maser channel, and then, we self-calibrated the ref-erence maser channel. Self-calibration was e ff ective in improv-ing the signal-to-noise ratio (SNR) of the image of the referencemaser channel by a factor ≥
10. For each of the three observingruns, the calibration of the NH line, water maser and contin-uum emission of G16.59 − The National Radio Astronomy Observatory is operated byAssociated Universities, Inc., under cooperative agreement with theNational Science Foundation.2oscadelli et al.: A subarcsecond study of G16.59 − Before imaging, the calibrated data of the three runs havebeen concatenated. We produced naturally-weighted and uv -tapered (Gaussian tapering with FWHM of 400 k λ ) imagesof the NH main and satellite lines, water maser and contin-uum emission. The continuum data was obtained by averagingthe 13 (out of the 16) 4 MHz bandwidths of the two IFs notcontaining NH (1,1) and (2,2), nor the water maser lines. Thenaturally-weighted beam has a FWHM size of 0 . ′′ × . ′′
31, atPA = ◦ , with small di ff erence between the NH line and con-tinuum maps. The continuum emission map, restored using around beam with FWHM size of 0 . ′′
4, has a rms noise level of0.06 mJy beam − . To increase the SNR on the NH lines, beforeimaging, the data have been smoothed in velocity, degrading thevelocity resolution to 0.8 km s − . The typical rms noise on thechannel maps of the NH main and satellite lines is of 1.3–1.5 mJy beam − . We used the VLA A–Array (code: 12B-044) to observe the con-tinuum emission of G16.59 − ff erentruns on October 2, October 21 and December 30 2012, andJanuary 17 2013. G16.59 − − and 13 km s − at 6 GHz and 22 GHz,respectively. Since the methanol and the water maser signals inG16.59 − ≈
20 Jy beam − ) and relatively wide( ≈
10 km s − ), we could detect them with high ( ≥ − ≥ / µ Jy beam − at 6 /
14 GHz and 23 µ Jy beam − at 22 GHz. The FWHM size of the naturally-weighted beams atthe di ff erent observing wavelenghts are listed in Table 1. Using the mid-infrared imaging spectrometer (COMICS;Kataza et al. 2000) at the Cassegrain focus of the 8.2 m SubaruTelescope, we carried out imaging observations of the 24.5 µ memission toward G16.59 − ∼ ′′ × ′′ with a pixel sizeof 0 . ′′
13. Flux calibration was performed toward four standardsources listed in Cohen et al. (1999): HD146051, HD186791,HD198542, and HD198542. We estimated the overall uncer-tainty in the flux calibrations to be less than 10%.
Fig. 1.
Overlay of the radio continuum emission towardthe HMC in G16.59 − − , from 0.03 to 0.09in steps of 0.03 mJy beam − , from 0.174 to 0.348 in stepsof 0.058 mJy beam − , and from 0.075 to 0.25 in steps of0.025 mJy beam − . The solid circles and empty triangles rep-resent, respectively, the CH OH and H O masers from S2010,while the arrows are the proper motions of the H O masers. Thescale for the proper motion amplitude is given on the bottom leftof the panel. The size of the VLA beams for the di ff erent arrayconfigurations and wavelengths are reported in Table 1.To perform the astrometric calibration of the image, we havecompared the Subaru image with the MIPSGAL image at 24 µ m(Carey et al. 2009), after smoothing the former to the same an-gular resolution (6 ′′ ) as the latter. We have then aligned thetwo images by visually overlaying the non-saturated part of theMIPSGAL image on the corresponding part of the Subaru im-age. The original Subaru image was corrected by about 3 ′′ . Thisprocess, albeit not very accurate, should result in an absolute po-sitional uncertainty < ∼ ′′ , comparable to the astrometric accuracyof MIPSGAL.
3. Continuum emission: A bipolar jet
From previous studies we know that radio continuum emissionhas been detected in several locations in the G16.59 − ff erentwavelengths and with di ff erent resolutions.For the sake of comparison, we also show the H O andCH OH maser spots. The striking result is that, while the1.3 cm emission appears compact and almost coincident with theCH OH masers, consistent with Zapata et al. (2006) and S2010,at longer wavelengths (2 and 6 cm) the emission appears elon-gated E–W on both sides of the CH OH maser spots, extendingover ∼ ′′ (0.07 pc). This morphology is strongly suggestive of abipolar jet, whose powering source could lie close to the locationof the CH OH masers, near the geometrical center of the jet. InTable 1, we report the peak and integrated fluxes, derived fromnaturally-weighted maps, for the VLA A and B array observa-tions. The thermal and non-thermal fluxes refer to the compactand extended continuum components, respectively, and are ob-tained by integrating the continuum emission inside and outside,respectively, the area where the 1.3 cm emission is > σ (de-limited by the white contour of Fig. 2c). The outer limit of thearea where the non-thermal component has been computed, isdetermined by the 5 σ level of the 6 cm emission (see Fig. 2b). − Table 1.
Centimeter continuum fluxes for the VLA A and B-Array observations. θ beam P.A. beam rms I peak S thermal ν S non − thermal ν Wavelength ( ′′ × ′′ ) ( ◦ ) (mJy beam − ) (mJy beam − ) (mJy) (mJy)6.0 cm—A 0 . × . − . ± .
002 0 . ± . . × . −
14 0.009 0.117 0 . ± .
005 0 . ± . . × . + . ± . ≤ . . × . −
173 0.023 0.265 0 . ± . ≤ . Besides the morphology, a typical signature of thermal jetsis their spectral index, ranging from − + − uv range for all bands, as well as the same clean beam (0 . ′′ × . ′′ ff erentwavelengths minimizing the e ff ects of di ff erent samplings of theuv plane and thus obtaining a more reliable estimate of the spec-tral index. The new maps are shown in the top panel of Fig. 2,while in the other two panels we show the maps of the spec-tral index α (assuming S ν ∝ ν α ) between 6 and 2 cm (Fig. 2b)and between 2 and 1.3 cm (Fig. 2c). For these two estimatesthe uncertainty is, respectively, ± ± α hasbeen calculated only where the 6 cm (for Fig. 2b) and 2 cm (forFig. 2c) fluxes are > σ . Moreover, the white contour encom-passes the region where the 2 cm (for Fig. 2b) and 1.3 cm (forFig. 2c) fluxes are > σ . This means that the values of the spec-tral indices lying outside the white contour are upper limits.The striking result is that the spectral index between 6 and2 cm is < − O) (Reid et al. 1995) and, more re-cently, in HH 80–81 (Carrasco-Gonz´alez et al. 2010), indicat-ing that magnetic fields may play an important role in shapingsuch jets and, more in general, the accretion / ejection process inmassive (proto)stars. In the case of G16.59 − α in Fig. 2b is negative all over the region, butbecomes > − . OH masers; (ii) the spectral index is ∼ OH masers(outside the white contour of Fig. 2c), the value of spectral in-dex between 2 and 1.3 cm is only an upper limit and could beas low as measured between 6 and 2 cm. Closer to the CH OHmasers, given the low brightness temperature measured at 1.3 cm(69 ± α as large as ∼ ff erence between α (6–2 cm)and α (2–1.3 cm) is real. A possible explanation is that the radioemission is made of two contributions, with synchrotron dom-inating at long wavelengths and at larger separations from thepowering MYSO, and bremsstrahlung preponderant close to theMYSO. Note infact that, near the CH OH masers, α is consistentwith free-free at all frequencies. Interestingly, this is also wherethe most compact 1.3 cm emission is detected with the highestangular resolution (see gray scale in Fig. 1), which suggests that Fig. 2. a.
Overlay of the VLA maps of the radio continuumemission at 6, 2, and 1.3 cm obtained using the common uvrange and the same clean beam (shown in the bottom right) atall wavelengths. Contours and symbols have the same meaningas in Fig. 1. The contour levels range from 0.0264 to 0.1584 insteps of 0.0264 mJy beam − at 6 cm, from 0.03 to 0.12 in stepsof 0.03 mJy beam − at 2 cm, and from 0.174 to 0.348 in stepsof 0.058 mJy beam − at 1.3 cm. b. Map of the spectral index be-tween 6 and 2 cm, computed where the 6 cm emission is > σ .The white contour encompasses the region where also the 2 cmemission is > σ . The values of the spectral index lying outsidethe white contour are upper limits. c. Same as panel b, for thespectral index between 2 and 1.3 cm.this compact emission could be tracing the dense ionized gas inthe neighborhoods of the star powering the jet.What is the origin of the free-free emission? This couldcertainly arise from material ionized by shocks along the jet.However, one cannot exclude that ionization is due to the stellarphotons. In this case, assuming optically thin emission, from the1.3 cm continum flux measured with the B-array (0.69 mJy) oneobtains a stellar Lyman continuum of ∼ s − , correspondingto a zero-age main-sequence (ZAMS) luminosity of ∼ L ⊙ and a stellar mass of ∼ M ⊙ (see Davies et al. 2011). Accordingto Beuther et al. (2002b, see their Fig. 5), this implies a mass lossrate for the associated outflow of a few times 10 − M ⊙ yr − . It isquestionable whether this value can be compared to those of thetwo CO outflows imaged on a larger scale by B2006, F2008,and L´opez-Sepulcre et al. (2009), because the direction of the − Fig. 3.
Spectral energy distribution of G16.59 − ∼ L ⊙ ) has been obtained by integrating the SED under the solidline.radio jet di ff ers by ∼ ◦ from that of either outflow. It is hencepossible that three massive objects are powering three distinctjets / outflows in the G16.59 − OH masers as tracing a ro-tating disk (see S2010), because the radio jet appears to originatefrom a MYSO located close to the center of the maser spots andalmost perpendicular to the elongated maser cluster.
4. Bolometric luminosity and its origin
Given the presence of multiple objects in G16.59 − / Spitzer 24 µ m (Carey et al. 2009), MSX(Egan et al. 2003), IRAS (Neugebauer et al. 1984), the APEXTelescope Large Area Survey of the Galaxy (ATLASGAL ;Schuller et al. 2009; Contreras et al. 2013), Bolocam GalacticPlane Survey (BGPS; Drosback et al. 2008), data from the lit-erature (Beuther et al. 2002a; De Buizer et al. 2005), and re-cent PACS and SPIRE continuum images obtained in the con-text of the Herschel infrared Galactic Plane Survey (Hi-GAL;Molinari et al. 2010a,b).The flux density estimates of IRAS, MSX, GLIMPSE,BGPS, and ATLASGAL were taken from the correspondingsource catalogues, while those of Hi-GAL and MIPSGAL werecomputed from the images with aperture photometry, by inte-grating the emission inside suitable polygons encompassing oursource and subtracting the background. Note that the MSX andIRAS flux densities are considered upper limits, because theHPBWs encompass a region significantly larger than that of in- The ATLASGAL project is a collaboration between the Max-Planck-Gesellschaft, the European Southern Observatory (ESO) and theUniversidad de Chile.
Fig. 4. a.
Map of the continuum emission at 3.6 cm (blackcontours; from S2010) overlaid on an 8.0 µ m image (resolu-tion 2 ′′ ) of the G16.59 − / Spitzersurvey (Benjamin et al. 2003). b. Same as panel a, for ourSubaru 24.5 µ m image smoothed to the same angular resolu-tion as the 8.0 µ m image (2 ′′ ). c. Same as panel a, for the Hi-GAL / Herschel (Molinari et al. 2010b) 70 µ m image (resolution ∼ ′′ ). The white dots indicate the positions of the centimetersources “a” (alias mm-core), “b” (alias HMC) and “c” detectedby Zapata et al. (2006).terest for us. The MIPSGAL flux is a lower limit, because theimage is partly saturated. We stress that such a limit is consis-tent with our Subaru measurement.To estimate the bolometric luminosity, the resulting SED(shown in Fig. 3) has been integrated by interpolating linearlybetween the known fluxes (solid line in the figure). We ob-tain ∼ L ⊙ , corresponding to a single star of ∼ M ⊙ (seeDavies et al. 2011).In the case of G16.59 − ∼ ′′ (i.e., ∼ µ m GLIMPSE / Spitzer image, as well asour 24.5 µ m Subaru image, shown in Figs. 4a and 4b, respec-tively. The problem is that the angular resolution at λ > µ m –i.e., close to the peak of the SED (see Fig. 3) – is not su ffi cientto resolve the three sources and find out which is responsible formost of the bolometric luminosity. To investigate this issue, inFig. 5 we present a composite color image of the same regiondisplayed in Fig. 4, where the blue, green, and red components − Fig. 5.
Composite color image of the same field shown in Fig. 4,where blue is associated with the GLIMPSE / Spitzer 3.6 µ m im-age, green with the GLIMPSE / Spitzer 8.0 µ m image, and redwith our 24.5 µ m Subaru image. The reddest and hence most ex-tincted source is the one to the north, coinciding with the HMC.The VLA 3.6 cm continuum (from S2010) is shown with whitecontours. The white dots have the same meaning as in Fig. 4.are associated, respectively, with the 3.6, 8, and 24.5 µ m im-ages (the first two from GLIMPSE / Spitzer). There is little doubtthat the reddest and hence most embedded object is the northernsource, namely the one coinciding with the HMC. Visual com-parison of the three panels in Fig. 4 leads to the same conclu-sion, as the northern source becomes progressively more promi-nent with increasing wavelength, until it dominates the whole IRemission at 70 µ m.We conclude that in all likelihood the HMC is the sourceresponsible for most of the ∼ L ⊙ measured over theG16.59 −
5. Ammonia emission
Our ammonia observations provide us with information that ishelpful to shed light on the structure and velocity field in thisregion. Such information is summarized by Fig. 6, which showsthree overlays between the NH line emission and other relevanttracers.In particular, in Fig. 6a the integrated NH (1,1) line map iscompared to the 24.5 µ m image obtained with Subaru, whichconfirms that the IR emission is tightly associated with theHMC, within the positional uncertainty. It is also interesting tocompare the 6 cm emission tracing the radio jet with the ammo-nia emission, as done in Fig. 6b. The strongest 6 cm continuumpeak, which likely traces shocked gas of the radiojet (see Fig. 1and Sect. 3), appears to coincide with the NH peak, while bothare o ff set by ∼ . ′′ ∼ OH masers – which likely pinpoint the po-sition of the star powering the radio jet (see Fig. 1). This factstrongly suggests that the jet is impinging against dense molec-ular gas to the west, where the brightest radio emission is seen.
Fig. 6. a.
Map of the NH (1,1) emission averaged over the mainline (white contours) overlaid on an image of the 24.5 µ m con-tinuum emission obtained with Subaru. The circles in the bottomleft and right indicate, respectively, the angular resolution of theIR and radio data. b. Same as panel a, for the NH (2,2) line (con-tours) and the 6 cm radio continuum map (background image). c. Overlay between the NH (2,2) (white contours), CH CN(5–4) (black contours), and 3.3 mm continuum (image) maps. Theammonia map has been smoothed to the same angular resolution(ellipse in the bottom right) as the methyl cyanide data (fromF2008).This hypothesis is supported by the distribution and proper mo-tions of the H O masers (see Fig. 1), which are known to be as-sociated with shocks. The proposed scenario could also explainwhy the jet appears slightly more extended to the east, as in thisdirection the gas would be much less dense. − Figure 6c sheds further light on the structure of the region.This presents an overlay between our map of the NH (2,2) in-version line and the maps of the CH CN(5–4) line and 3.3 mmcontinuum emission from F2008. Clearly, ammonia and methylcyanide trace the same region (the HMC), as expected, whileboth are significantly o ff set (by ∼ . ′′ ∼ ff set between the HMC and themm-core. The latter must be quite massive, because it appearsto dominate the mm continuum emission despite its low tem-perature ( ∼
38 K; see Williams et al. (2005)) with respect to theHMC ( > ∼
130 K; see F2008). Using the same dust properties andcontinuum flux density (37 mJy) as in F2008, one can estimate amass of ∼ M ⊙ for the mm-core. Such a large value suggeststhat other MYSOs, possibly in a much earlier evolutionary phasethan the star in the HMC, might be forming inside the mm-core. emission In Fig. 7, we show the average spectra of the NH (1,1) and (2,2)inversion transitions over the HMC. The most intriguing fea-ture is the presence of a double peak, clearly detected in bothtransitions. The fact that such a double peak is seen also in theoptically thin satellites proves that this is not the e ff ect of self-absorption, but is due to two distinct velocity components. Thered profiles in Fig. 7 are fits to the NH lines taking into accountthe hyperfine structure of the transitions, and clearly confirm thatboth the main line and satellite emission can be fit with two ve-locity components.It is important to investigate the spatial distribution of thesetwo components, as done in Fig. 8, where we compare the mainline NH (1,1) and (2,2) emission integrated under the red andblue components, with the radio jet and the core traced, respec-tively, by the centimeter and millimeter continuum emission. Forthe sake of comparison, also the CH OH and H O masers areshown. We have used di ff erent colors for the blue- and the red-shifted CH OH spots, to outline the rotation of the disk about aNE–SW axis. From this figure one sees that ammonia is tracing acompact bipolar structure roughly perpendicular to the CH OHdisk and slightly o ff set (by ∼ . ′′
4) to the NW with respect to it.Such a small o ff set (less than the NH map angular resolution)might be justified by the fact that the gas density is higher to theNW of the disk than to the SE.The orientation and velocity of the ammonia structure beingsimilar to the (larger scale) NE–SW bipolar outflow detected byB2006 and L´opez-Sepulcre et al. (2009) suggests that both couldbe manifestations of the same ejection phenomenon. Indeed, thispossibility cannot be excluded, as ammonia has been observed inassociation with bipolar jets in other massive sources (see, e.g.,Zhang et al. 1999). In this scenario, the ammonia gas would bepreferentially entrained by the flow on the NW side – i.e., thatfacing the core – where the density is higher. This would explainwhy the NH emission is slightly skewed to the NW. To verify the outflow hypothesis, one may estimate the relevantparameters from the ammonia emission and compare them tothose obtained for the larger-scale NE–SW bipolar outflow. We
Fig. 7.
Spectra of the NH (1,1) (top) and (2,2) (bottom) in-version transitions obtained by averaging the emission over theHMC. The red solid curves are multiple Gaussian fits taking intoaccount the hyperfine structure of the lines and two distinct ve-locity components (at about 58.2 km s − and 61.8 km s − ). Thered dotted curves indicate the contribution of each velocity com-ponent to the total fit.have computed the mass loss rate, momentum rate, and mechan-ical luminosity by integrating the blue- and red-shifted emissionin the NH (1,1) main line over the respective lobes, and dividingby the dynamical time scale. In this calculation we have assumedthat the line is optically thin for reasons that will be explained inSect. 5.2. The dynamical time scale, obtained from the ratio be-tween the length of the lobes, and the maximum speed in theline wings, is ∼ × yr, which is only a lower limit of theMYSO age. We use 0 . ′′ emission (see Fig.8). For themaximum speed in the line wings, we take 6.3 km s − , whichwas evaluated taking the di ff erence in velocity between the mostblue- and red-shifted channels where ammonia emission is de-tected. We caution that the derived dynamical time scale is notcorrected for the (unknown) inclination, i o , of the outflow withrespect to the line of sight. Such a correction, corresponding tothe factor coth( i o ), becomes important if the outflow is closeto the plane of the sky or the line of sight. The dynamical timescale measured for the NE–SW outflow by L´opez-Sepulcre et al.(2009) using CO single-dish observations is ≈ yr, signifi-cantly larger than the value determined from the NH data andlikely close to the MYSO age. For our estimates we used a gastemperature of 140 K, equal to the maximum brightness tem-perature measured in the NH line. This is consistent with therotational temperature (130 + − K) computed by F2008 from theCH CN(5–4) lines. Note that reducing the temperature to, for − Fig. 8.
Maps of the two velocity components (see Fig. 7) iden-tified in the NH (1,1) (top) and (2,2) (bottom) lines. Blue andred contours are the maps obtained by averaging the main lineemission, respectively, from 55.8 to 60.6 km s − and from 60.6to 63.0 km s − . The gray scale is the same map of the 6 cmcontinuum emission as in Fig. 2a, while the backgroud image(green tones) is the map of the 3.3 mm continuum emissionobtained by F2008. The red and blue dots give the position ofthe red- and blue-shifted, respectively, methanol masers fromS2010. The black triangles and arrows are the water maser spotsand corresponding proper motions from S2010. The amplitudescale for the proper motions is given on the bottom left of theupper panel. The continuous and dashed line in the lower panelindicate the axes of projection to produce the position-velocityplots of Figs. 10 and 11, respectively. The ellipses in the bot-tom left and right of the lower panel indicate, respectively, theHPBW of the 6 cm continuum and ammonia line maps.example, 50 K would decrease the outflow parameters only by afactor ∼ M out ≃ × − / X M ⊙ yr − , ˙ P out ≃ × − / X M ⊙ km s − yr − , ˙ E out ≃ × − / X L ⊙ , where X isthe relative abundance of NH with respect to H . For a fidu-cial X = × − (see e.g., Van Dishoek et al. 1993, S´anchez-Monge et al. 2013c and references therein), the value of ˙ P (5 × − M ⊙ km s − yr − ) is only a few times less than that es-timated for the NE–SW outflow by L´opez-Sepulcre et al. (2009)from the CO(2–1) line (0 . × − M ⊙ km s − yr − ). Sucha discrepancy could be explained by the uncertainty on the am- monia abundance and we thus conclude that ammonia might betracing the root of the larger scale CO outflow oriented NE–SW. We also consider another explanation for the two ammonia com-ponents, namely that they are tracing two distinct, nearby clumpswith slightly di ff erent velocities. To investigate this possibility,we have analyzed the N H + (1–0) data of F2008 again. N H + isfound in close association with NH and the OVRO maps ob-tained by F2008 are sensitive to more extended emission thanthe higher-resolution ammonia maps obtained by us. We havefit the N H + spectra taking into account the hyperfine splittingof the transition and created maps of the average emission overthe main line, peak velocity, and line FWHM. These are shown,respectively, in Figs. 9a, 9b, and 9c. Overlaid on these, we alsoshow contour maps of the CH CN(5–4) line emission, HCO + (1–0) blue- and red-shifted emission, and 3.3 mm continuum emis-sion.The most striking feature is the sharp change in velocity seenin Fig. 9b along a NE–SW line, which looks consistent with thebipolar structure seen in HCO + . This result weakens the outflowinterpretation significantly, because N H + is a well known tracerof dense gas, but is only exceptionally detected in outflows (seethe case of L1157; Tobin et al. 2011). Moreover, for a bipolaroutflow one would not expect such a sharp change in velocitybecause the line peak should shift from red to blue velocitiesgradually, going from one lobe to the other. We thus believe thatthe existence of two clumps is at least as likely as the outflowscenario. Indeed, looking at Fig. 9a, one sees that while mostof the emission outlines a big, complex clump to the NE (wherethe compact CH CN core is located), a minor peak of emission isalso seen to the SW, possibly tracing another less massive clump.Let us hence assume for the moment that we are dealingwith two nearby clumps moving with di ff erent velocities alongthe line of sight. The question is whether these two clumps arephysically interacting or only overlapping in projection on theplane of the sky. Comparison between Figs. 9c and 9b revealsthat the line FWHM attains a maximum just along the line wherethe sudden velocity change is seen. The fact that such a max-imum coincides with a tail of millimeter continuum emission(contours in Fig. 9c) seems to confirm that the FWHM increaseis not due to two velocity components overlapping along theline of sight, but to a real enhancement of the column density.Although hazardous at this stage, one might even speculate thatthe formation of the mm-core has been triggered by the colli-sion of the two clumps, as it has been observed in other regions(Duarte-Cabral et al. 2011; Henshaw et al. 2013).Clearly, further observational evidence is needed beforedrawing any conclusion on the nature of the velocity field in thisregion. However, if the two-clump hypothesis were correct, thiscould imply that the NE–SW bipolar outflow claimed by B2006and L´opez-Sepulcre et al. (2009) does not exist.The scenarios proposed here and in the Sect. 5.1.1 will befurther discussed in Sect. 6.Before that, however, we wish to analyze the velocity fieldof the ammonia gas in some better detail, searching for a con-nection with the CH OH maser disk proposed by S2010. OH maser disk
Our data make it possible to find out whether the maser diskproposed by S2010 is detected also in the ammonia emission. − The most direct approach consists in studying the gas velocityalong the disk plane by means of a position–velocity plot of theNH emission. Since the NH (1,1) main line is slightly blendedwith the inner satellites (see Fig. 7) and the (1,1) transition tracesmore extended (and colder) regions than the (2,2) line, we havechosen the latter for our purposes. The plane of the disk is as-sumed to lie along the direction joining the centers of the red-and blue-shifted maser clusters (see e.g., Fig. 8), i.e., along aP.A. of 140 ◦ .The position-velocity plot of the NH (2,2) main line isshown in Fig. 10, where we also report the points represent-ing the CH OH maser spots. One sees that ammonia emissionis indeed detected at the same positions and velocities as thetwo groups of masers, consistent with emission from a rotatingdisk / ring. However, the emitting region is not limited to the lo-cations of the masers, but extends up to an o ff set of ∼ . ′′ emission appears to attain its maximum,both at blue- and red-shifted velocities. Such an o ff set betweenthe methanol masers center and the location of the two ammo-nia velocity components was already pointed out in Sect. 5.1(see also Fig. 8). In fact, the position–velocity plot along the di-rection perpendicular to the disk plane and passing through theammonia peak (i.e., at an o ff set of 0 . ′′
6. A model for the G16.59 − Before proposing an interpretation for the G16.59 − – A compact ( ∼ CN and NH peakat a position o ff set by ∼ . ′′ ∼ – Three mid-IR sources are seen within ∼ ′′ from the HMC,but only that associated with the HMC appears to dominatethe bolometric luminosity of the whole region ( ∼ L ⊙ , cor-responding to a ∼ M ⊙ ZAMS star). – The detection of two outflows extending over ∼ ′′ (0.35 pc)has been claimed by various authors, one directed NE–SW(B2006; L´opez-Sepulcre et al. 2009) and the other SE–NW(B2006; F2008). – Our ammonia observations have detected two velocity com-ponents arising from a compact region ( ∼ . ′′ ff erent velocities along theline of sight. – A radio jet directed E–W is imaged from 6 to 1.3 cm. Whilethe 6 cm continuum is dominated by non-thermal (possiblysynchrotron) emission, at 1.3 cm compact free-free emis-sion appears to arise very close to the center of the groupof CH OH maser spots detected by S2010. A 10 M ⊙ ZAMSstar is su ffi cient to explain the observed flux in terms of ion-ization by the stellar Lyman continuum – although ionizationby shocks along the radio jet cannot be excluded. – The H O maser proper motions measured by S2010 seem totrace expansion to the west, along the western border of theradio jet. The CH OH maser spots cluster in two groups thatappear to rotate about the free-free continuum peak, with a rotation axis oriented NE–SW and an equilibrium mass of ∼ M ⊙ (from S2010, after correcting for the new distanceestimate). A similar velocity field is seen also in the ammo-nia lines.We have proposed two scenarios to interpret the NH andN H + observations and in the following we will refer to them as“case A”, for that depicted in Sect. 5.1.1 (ammonia emission as-sociated with the NE–SW bipolar outflow) and “case B”, for thatillustrated in Sect. 5.1.2 (ammonia emission from two overlap-ping clumps with di ff erent velocities). Figure 12 shows a sketchthat summarizes both scenarios.Neither case A nor case B can be explained with only onedominant MYSO in the G16.59 − OH maserdisk (see Sect. 3), plus the SE–NW outflow detected by B2006and F2008. On the one hand, methanol masers have never beenfound in association with low-mass YSOs; on the other, the mo-mentum rate of the SE–NW outflow is typical of MYSOs. Wethus conclude that both the radio jet and the outflow must bepowered by MYSOs. In this scenario the disk-jet system is as-sociated with the HMC, while the outflow might be originatingfrom a young massive protostar deeply embedded in the mm-core.The situation of case A is more complex. Here we have thedisk-jet system, the SE–NW outflow, and the NE–SW outflowthat we detect in ammonia on a small scale. Like in case B, the(so far undetected) MYSO powering the SE–NW outflow mightbe lying inside the mm-core. As for the radio jet and NE–SWoutflow, we propose that both could be associated with the sameMYSO. The problem with this hypothesis are the di ff erent di-rections of the E–W jet and NE–SW outflow. This discrepancycan be explained if the jet / outflow is undergoing precession, be-cause the direction of the jet / outflow axis would change from thesmall to the large scale, as in the case of the massive protostarIRAS 20126 + ff set observedbetween the CH OH masers and the geometrical center of theNH flow (see Fig. 8).Is the case-A scenario compatible with the physical parame-ters estimated by us? We note that the values of the stellar mass(10–13 M ⊙ ) derived from the bolometric luminosity, CH OHmaser rotation velocity, and free-free 1.3 cm continuum, areall in good agreement with each other. Using models of out-flow evolution recently presented by Duarte-Cabral et al. (2013),a star of final mass of 10–15 M ⊙ , evolved to the stage of aHMC, should emit outflows with typical momentum rates of afew 10 − M ⊙ km s − yr − . Considering the uncertainty on the in-clination angle of the outflow, such a value agrees well with themomentum rate, 5 × − M ⊙ km s − yr − , we have derived inSect. 5.1.1 for the NH flow.Our conclusion is that the observational evidence collectedso far is insu ffi cient to decide between case A and B, althoughwe tend to favor the latter because case A implies the existenceof a NE–SW outflow traced also by the N H + (1–0) line: to ourknowledge, this would be the second case ever of N H + tracingan outflow – and the very first case for outflows from MYSOs. −
7. Summary and conclusions
We performed ammonia line and radio and IR continuum obser-vations of the G16.59 − OH maser disk imagedby S2010, and establish the origin of the two bipolar outflowsobserved in the region. We discover a radio jet oriented E–Wand centered on the CH OH masers, whose western lobe ap-pears to expand to the west, impinging against the HMC, as sug-gested by the H O maser proper motions measured by S2010.The radio continuum from the jet has negative spectral indexbetween 6 and 2 cm, indicating non-thermal continuum (syn-chrotron) emission, with the sole exception of a compact free-free continuum source close to the CH OH masers.Our 24.5 µ m image complemented by Herschel continuumdata from the Hi-GAL survey indicate that the bolometric lumi-nosity of the region is dominated by the emission from the neigh-borhoods of the HMC and mm-core. The stellar mass impliedby the observed luminosity ( ∼ M ⊙ ) is in good agreement withboth the equilibrium mass of the CH OH maser disk ( ∼ M ⊙ ),and the stellar mass ( > ∼ M ⊙ ) needed to explain the free-freecontinuum emission – assuming the latter is due to ionization bystellar photons.We find that the ammonia emission presents a bipolar struc-ture consistent (on a smaller scale) in direction and velocitywith that of the NE–SW bipolar outflow reported by B2006and L´opez-Sepulcre et al. (2009). After analyzing the N H + (1–0) observations of F2008 again, we conclude that two scenar-ios are possible. Both imply the existence of two MYSOs, butdi ff er in the interpretation of the NE–SW bipolar structure. Inone case this is a bipolar outflow (in agreement with B2006 andL´opez-Sepulcre et al. (2009)), which represents the prosecutionon the large scale of the small-scale radio jet. In the other case,the bipolarity is interpreted as two overlapping clumps movingwith di ff erent velocities along the line of sight. We slightly fa-vor the second hypothesis, because the first would imply that theNE–SW outflow is detected also in the N H + (1–0) line emis-sion, which appears unlikely. Discriminating between the twoscenarios requires observations with higher sensitivity than ourVLA ammonia images and we believe that the Atacama LargeMillimeter Array will be the right instrument to establish thenature of the NE–SW bipolar structure and investigate the cir-cumstellar disk associated with the methanol masers. Acknowledgements.
We thank the anonymous referee for his constructive crit-cisms and the suggestion to analize the N H + data by F2008 again, which helpedus to shed new light on this complex region. R.S.F. acknowledges T. Usuda, T.Inagaki , S. S. Hayashi, and H. Shinnaga for their help with the Subaru ob-servations and data reduction. We are greatful to Beatriz S´anchez Monge andRafael Delgado Romero for preparing the sketch shown in Fig. 12. A. Sanna ac-knowledges the financial support by the European Research Council for the ERCAdvanced Grant GLOSTAR under contract no. 247078. References
Anglada, G. 1996, in Astronomical Society of the Pacific Conference Series,Vol. 93, Radio Emission from the Stars and the Sun, ed. A. R. Taylor & J. M.Paredes, 3–14Benjamin, R. A., Churchwell, E., Babler, B. L., et al. 2003, PASP, 115, 953Beuther, H., Schilke, P., Menten, K. M., et al. 2002a, ApJ, 566, 945Beuther, H., Schilke, P., Sridharan, T. K., et al. 2002b, A&A, 383, 892Beuther, H., Zhang, Q., Sridharan, T. K., Lee, C.-F., & Zapata, L. A. 2006, A&A,454, 221 (B2006)Carey, S. J., Noriega-Crespo, A., Mizuno, D. R., et al. 2009, PASP, 121, 76Carrasco-Gonz´alez, C., Rodr´ıguez, L. F., Anglada, G., et al. 2010, Science, 330,1209 Cesaroni, R., Galli, D., Lodato, G., Walmsley, C. M., & Zhang, Q. 2007,Protostars and Planets V, 197Cesaroni, R., Neri, R., Olmi, L., et al. 2005, A&A, 434, 1039Cohen, M., Walker, R. G., Carter, B., et al. 1999, AJ, 117, 1864Contreras, Y., Schuller, F., Urquhart, J. S., et al. 2013, A&A, 549, A45Cunningham, A. J., Klein, R. I., Krumholz, M. R., & McKee, C. F. 2011, ApJ,740, 107Davies, B., Hoare, M. G., Lumsden, S. L., et al. 2011, MNRAS, 416, 972De Buizer, J. M., Radomski, J. T., Telesco, C. M., & Pi˜na, R. K. 2005, ApJS,156, 179Drosback, M. M., Aguirre, J., Bally, J., et al. 2008, in Bulletin of the AmericanAstronomical Society, Vol. 40, American Astronomical Society MeetingAbstracts + (S2010)Schuller, F., Menten, K. M., Contreras, Y., et al. 2009, A&A, 504, 415Shepherd, D. S., Yu, K. C., Bally, J., & Testi, L. 2000, ApJ, 535, 833Tobin, J. J., Hartmann, L., Chiang, H.-F., et al. 2011, ApJ, 740, 45Williams, S. J., Fuller, G. A., & Sridharan, T. K. 2005, A&A, 434, 257Zapata, L. A., Rodr´ıguez, L. F., Ho, P. T. P., Beuther, H., & Zhang, Q. 2006, AJ,131, 939Zhang, Q., Hunter, T. R., Sridharan, T. K., & Cesaroni, R. 1999, ApJ, 527, L117 − Fig. 9.
Obtained from the data of F2008: a. Map of theCH CN(5–4) line emission (contours) tracing the HMC, over-laid on a map of the N H + (1–0) emission (from F2008) aver-aged over the main line (color scale). The black dots indicate thepositions of the centimeter sources “a” (alias mm-core) and “b”(alias HMC), detected by Zapata et al. (2006). b. Maps of theblue- (solid white contours) and red-shifted (dashed black) emis-sion of the HCO + (1–0) line overlaid on a map of the N H + (1–0)line velocity. The HCO + (1–0) maps are derived using the mostextreme velocity channels (1.68 km s − wide) at which emissionwas detected by F2008, centered at 55.8 km s − and 62.5 kms − for the blue- and red-shifted line, respectively. c. Map ofthe 3.3 mm continuum emission (contours) tracing the mm-core,overlaid on a map of the N H + (1–0) line FWHM. Fig. 10.
Position–velocity plot of the NH (2,2) line emissionalong a direction crossing the centers of the red- and blue-shiftedCH OH maser clusters (P.A. = ◦ ). Contour levels range from2 to 6 mJy / beam in steps of 1 mJy / beam. The red and blue circlesrepresent the methanol maser spots. The cross in the bottom leftindicates the angular and spectral resolutions. Fig. 11.
Position–velocity plot of the NH (2,2) line emissionalong a direction perpendicular to that of Fig. 10 (i.e., withP.A. = ◦ ) and passing at an o ff set of 0 . ′′ OH masers. Contour levels range from 2 to 7 mJy / beam insteps of 1 mJy / beam. The cross in the bottom left indicates theangular and spectral resolutions. − Fig. 12.
Sketch of the scenarios proposed to explain all the fea-tures observed in the G16.59 − ff erent structures (disk, jet, outflow, etc.) have beenarbitrarily chosen to make the figure more readable.erent structures (disk, jet, outflow, etc.) have beenarbitrarily chosen to make the figure more readable.