In this section we present the results of the line and continuum observations towards IRAS 20126+4104. The former consist only of the PdBI data, whereas the latter contain also the data obtained with the JCMT and UKIRT.
3.1. Molecular line data from the PdBI
The SiO(2-1), H13CO+(1-0), and CH3CN(12-11) lines mapped with the PdBI had been also observed with the 30-m telescope by various authors. The availability of single dish spectra makes possible a comparison with those obtained by integrating the line over the whole emitting region imaged with the interferometer. This is done in Fig. 1. The absorption features affecting the H13CO+ profile are due to extended emission which is lost in the high resolution map; on the contrary, all the SiO and CH3CN emission is recovered by the interferometer, which proves that it arises from a relatively compact structure. In Fig. 2 the same comparison is shown for all the K components of CH3CN(12-11), reinforcing the previous conclusion.
A detailed description of the structure of the emitting regions can be obtained from the study of the different lines in the PdBI maps. These are presented in the following sections.
3.1.1. Maps in the SiO(2-1) line
One of the goals of this work was to use the SiO emission to trace the spatial and velocity structure of the outflow seen by CFTWO in the HCO+(1-0) line. It is thus of great interest to compare maps in these two transitions, as done in Fig. 3. In the left panels we show the HCO+ maps (contours) as in Fig. 7 of CFTWO, whereas in the right panels similar maps for the SiO line are presented. Note that for both transitions the maps in the top panels have been obtained by integrating the emission in the outer wings of the line (i.e. from -28 to -9 km s-1 and from 2 to 21 km s-1), whereas the bottom panels correspond to integrals in the inner wings (i.e. from -9 to -5 km s-1 and from -2 to 2 km s-1). For a definition of the inner and outer wings see CFTWO. For the sake of comparison, also the 2.122 µm image in the H2 line is shown (grey scale).
A few interesting remarks can be made on the basis of Fig. 3. First of all, it is clear that the SiO emission traces the H2 emission better than HCO+: this confirms (see e.g. Bachiller & Gutiérrez 1997) that the SiO molecule is closely associated with shocked gas, whereas HCO+, although tracing the flow, seems to arise from a larger region. We stress that such a conclusion holds not only for the low velocity (bottom panels of Fig. 3) but also for the high velocity gas (top panels), where HCO+ still presents a larger structure than SiO.
The different distribution of these two outflow tracers is even more striking in Fig. 4, where the bulk emission (i.e. from -5 to -2 km s-1) in the two lines is compared. There is little doubt that HCO+ peaks at the core seen by CFTWO at the centre of the flow, whereas almost no SiO emission is detected at that position. This confirms that the SiO molecule is strongly enhanced in the shocked gas.
Finally, by comparing the left with the right panels of Fig. 3, one can see that the velocity structure of the outflowing gas is quite similar in the two tracers. However, an interesting difference can be pointed out: while at high velocities (top panels) the blue and red lobes of the flow are well separated in both lines, at low velocities (bottom panels) such a separation is sharper in the HCO+ map. In fact, in the SiO map both blue- (full contours) and red-shifted (dashed) emission is seen in each lobe of the flow: this makes the velocity reversal noted by CFTWO for HCO+ (i.e. the exchange between blue and red lobes going from low to high velocities) less striking for SiO, and suggests that the SiO emitting gas is confined into a smaller solid angle. We shall devote Sect. 4.1 to a better discussion of this point.
3.1.2. Maps in the CH3CN(12-11) line
As already found by CFTWO for the CH3CN(5-4) line, also the CH3CN(12-11) emission originates from a compact molecular core located at the nominal position of the IRAS source, namely at the centre of the outflow. Incidentally, we note that the CH3CN(12-11) lines are optically thick, as derived from the ratio between these and the corresponding transitions of the CCN isotopomer: for a relative abundance [CH3CN/CCN]=60 (see e.g. Wilson & Rood 1994), one obtains CH3CN optical depths ranging from 6 to 26 depending on the K component. The angular resolution achieved with the new observations (0:007) is 5 times better than that obtained by CFTWO and comparable to the estimated core diameter: this should allow to barely resolve the core and confirm the velocity gradient seen in the CH3CN(5-4) line.
In Fig. 5a we show the map of the CH3CN(12-11) line emission integrated under the K=0 and 1 components. Although most of the emission is coming from the central unresolved portion of the core, a minor fraction of it arises from a more extended region 3" in size, which looks marginally elongated in the NE-SW direction. Such an elongation is consistent with that found in the unresolved core (see Fig. 5b) by applying the same method used by CFTWO to investigate the structure and velocity gradient of the CH3CN emission on a scale smaller than the synthesised HPBW. This consists in making channel maps in the lines and fitting a 2-D gaussian to the map in each channel: in this way one can determine the peak position of the CH3CN emission at each velocity, namely at the associated with each channel. The peak positions thus computed are plotted in Fig. 5b: their distribution is elongated in the NE-SW direction. The position uncertainty, represented by the error bars in the figure, has been set equal to 0.45 HPBW/(S/N) (see e.g. Zhang et al. 1998a) with S/N signal-to-noise ratio. One can then plot the velocity associated with each of these peak positions against the corresponding offset measured along the line of symmetry of the distribution in Fig. 5b: this is done in Fig. 5c, which shows a nice correlation and demonstrates the existence of a steady velocity gradient (4300 km s-1 pc- 1) through the core. As in CFTWO, we stress that the trend shown in Fig. 5c is not a real rotation curve, because the HPBW (0:007) is larger than the angular range shown in this figure (0:004). This also demonstrates that to really see the structure of the gas around the star one needs still higher angular resolution than that attained by us.
The previous results fully confirm the findings of CFTWO, thus supporting the rotating disk interpretation proposed by them. In Sect. 4.2 we shall discuss this topic in some better detail. Note that in Fig. 10 of CFTWO the size of the CH3CN distribution looks larger (1") than in our Fig. 5b (0:004): this is due to the fact that the 2-D gaussian fit tends to privilege the strong, compact (and hence unresolved) core with respect to the fainter, extended emission around it (represented by the lowest contour in Fig. 5a). However, the velocity gradient seen in the core and shown in Fig. 5c is also present in the faint extended CH3CN emission: for example, the velocities of the spectra at positions (+1",+1") and (-1",-0:005) in Fig. 5a are respectively -5.6 and -3.1 km s-1.
3.1.3. Maps in the H13CO+(1-0) line
The u-v coverage obtained in our observations turned out to be lacking of sufficiently short baselines to properly map the extended emission in the H13CO+(1-0) transition, as witnessed also by the line profile shown in Fig. 1: this is seriously affected by absorption features due to the existence of extended structures that cannot be imaged by the interferometer in the configuration used. However, our purpose was to map the regions on scales not larger than the size of the HCO+ outflow seen by CFTWO, namely 20 ": this is comparable to the maximum imageable size and hence the data are still worth for studying the structure of the flow and disk.
For the sake of comparison with HCO+ and SiO, one would like to produce maps of the H13CO+ emission by integrating under the line in the same velocity intervals used in Fig. 3: however, the H13CO+(1-0) line is more narrow than the same line in the main species, so that no emission is detected in the outer wings, namely below -9 km s-1 and above 2 km s-1. Therefore, in the bottom panel of Fig. 6 we can show only the outflow maps obtained integrating the H13CO+ line emission in the inner wings, namely the equivalent of the maps in the bottom panel of Fig. 3. The agreement between the H13CO+ and HCO+ low velocity outflows is remarkable.
The grey scale in the same figure represents the bulk emission (i.e. from -5 to -2 km s-1) in the H13CO+(1-0) line: this peaks at the hot core position. Thanks to the high spectral resolution used to observe this line (see Table 1), one can separate the bulk emission red-shifted with respect to the line centre (-3.5 km s-1), from that blue-shifted. This is done in the top panel of Fig. 6, where we show maps of the H13CO+ emission integrated in the velocity intervals from -5 to -3.5 km s-1 (full contours) and from -3.5 to -2 km s-1 (dashed), overlaied with the H2 line emission from CFTWO. In this map the H13CO+ emission concentrates in two regions, one associated with the core at the H2O maser position, the other (represented by the red-shifted emission to the top right of the figure) belonging to a more extended structure which is not of interest for the present work. The latter could be related to a large scale outflow revealed in recent CO(1-0) observations (Shepherd priv. comm.): this could indicate the existence of multiple outflow episodes originating from the same region, possibly from distinct YSOs. However, although important, a study of these phenomena requires more extended maps than those presented here and hence goes beyond our purposes. Confining thus our attention only to the blue- and red-shifted H13CO+ emission at the centre of the map, we note that, unlike the outflow axis, oriented from SE to NW, in this case the blue-shifted gas lies to the NE and the red-shifted to the SW: such a structure resembles - although on a much larger scale - that seen in CH3CN.
In conclusion, it is tempting to speculate that the H13CO+ emission arises both from the outflowing gas (on a scale of 20" and in a velocity range of 11 km s-1) and from a flattened rotating structure (on a scale of 10" and in a velocity range of 3 km s-1) roughly perpendicular to the outflow axis. Such a structure, might correspond to the outer layers of the disk seen in CH3CN. We shall come back to this point in Sect. 4.2.
3.2. Continuum data
An updated version of the continuum spectrum of IRAS 20126+4104 is shown in Fig. 7. With respect to Fig. 11 of CFTWO, the new figure contains the flux measured at 7 mm (3.10.8 mJy) by Hofner (priv. comm.) and those obtained in the present work at 1.3 mm (with the PdBI), in the submillimeter (with the JCMT), and in the MIR (with the UKIRT). As in CFTWO, we stress that the data in the figure have been taken with different angular resolutions. The estimate of the bolometric luminosity, obtained by linear interpolation of the points of the spectrum, remains approximately the same as in CFTWO, namely . However, the slope of the spectrum in the millimeter looks less steep (), although still consistent with optically thin dust emission. Using 0.9 as the spectral index of the dust absorption coefficient, one can fit the spectrum of Fig. 7 with a grey-body of temperature 60 K (dotted line in the figure): this is much less than =200 K estimated by CFTWO for the core, thus demonstrating that a substantial fraction of the continuum emission towards IRAS 20126+4104 arises from cooler regions more extended than the hot core. One can reach the same conclusion just noting that the 3 mm and 1.3 mm fluxes measured with the PdBI (filled circles in Fig. 7) are 3-5 times less than the corresponding values obtained with single dish telescopes (empty circles and filled squares). A direct measure of the distribution of the dust continuum emission is represented by the maps in the sub-mm obtained with SCUBA and discussed in Sect. 3.2.2.
In the following we describe the results obtained with the PdBI, JCMT, and UKIRT.
3.2.1. PdBI continuum maps
With respect to CFTWO, we have obtained two major improvements: a better angular resolution at 3 mm and a map at 1.3 mm. The latter presents the advantage of a HPBW half as that at 3 mm and a signal strength ten times stronger: this allows to obtain a good picture of the core structure, which looks barely resolved. In Fig. 8 we compare the map at 3 mm with that at 1.3 mm. In order to increase the sensitivity to extended structures the 3 mm map was obtained by merging the new data with those of CFTWO, which differ in frequency by only 3 GHz. One can see that 60% of the flux at 3 mm comes from the compact core at the centre of the map, while the rest originates from a faint halo surrounding it. Such a halo very likely extends over a larger region than that covered by us and probably explains the discrepancy found between single-dish and interferometer measurements (see Fig. 7). The 1.3 mm map looks similar to the CH3CN map of Fig. 5a and presents an elongated shape along the NE-SW direction. This is consistent with the emission originating from a flattened structure, as expected in the disk model discussed by CFTWO, although much better angular resolution is needed to assess this beyond any doubt.
In Table 4 we give the main parameters of the continuum maps, namely the peak position, the integrated flux (), the measured full width at half power (FWHP), the deconvolved diameter (), the linear diameter (D), the brightness temperature measured in the synthesised beam (), and the value of this after correction for beam dilution (). For the sake of comparison, also the same values for the CH3CN map of Fig. 5a are listed. All tracers peak essentially at the same position. The diameter in the 3 mm continuum is greater than that in the methyl cyanide line and - even more - in the 1.3 mm continuum: this is probably due to the extended structure previously discussed, which could not be imaged at 1.3 mm. The increase of from 3 mm to 1.3 mm is consistent with the fact that also the dust optical depth increases with frequency. Finally, it is worth noting that the brightness temperature of the CH3CN(12-11) line is an order of magnitude larger than that quoted by CFTWO for the (5-4) transition (see their Table 4): this proves that we are beginning to resolve the optically thick core, thus approaching the value of the excitation temperature estimated by CFTWO (200 K).
Table 4. Main parameters of the continuum and methyl cyanide emission in hot core. A distance of 1.7 kpc is used to estimate the linear diameters D.
3.2.2. JCMT data
Table 5 details the multi-wavelength flux measurements determined from the JCMT data. These data show that there is much more flux in extended structure than picked up by the photometry data with approximately 4 to 5 times as much flux present in the map data than seen in a single beam. Although the source size determined from the interferometric observations (1" at 1.3 mm) suggests that the core is approximately a point source at all SCUBA wavelengths, the photometry observations show that there is still extended structure being detected by the long-wavelength measurements. This can be seen in Fig. 9 where the emission at all wavelengths extends over a region much larger than the corresponding HPBW. Equivalently, one may note that the FWHPs listed in Table 5 are significantly larger than the corresponding HPBWs. It is also worth pointing out that, despite the similar angular resolution, the FWHP at 350 µm is slightly larger than that at 450 µm: this difference, if significant, suggests that the emission is getting optically thick at short wavelengths. Such an effect can be seen better in Fig. 10: here we show a comparison between the normalised radial profiles at 350 and 450 µm, obtained by averaging the flux density over circular annuli centred at the position of the 1.3 mm continuum peak. This comparison seems to indicate that the profile is broader at 350 µm than at 450 µm.
Table 5. Flux density and size measurements from SCUBA observations. Measurements were taken using photometry (single pixel) mode and mapping mode. Peak and integrated flux densities have been derived from mapping data. The quoted 3 errors are a combination of the standard error and estimated calibration error
Probably, the most striking feature of the maps in Fig. 9 is the almost perfect symmetry: the emission peaks at the core position and then falls down smoothly in all directions. This proves that the energy source heating the dusty cloud traced by the sub-mm emission is located inside the inner 10" of it. This is consistent with IRAS 20126+4104 being the heating agent of the surrounding molecular, dusty clump. It is worth noting that the 850 µm map presents a north-south elongation which cannot be seen at the other wavelengths: this is likely due to the contribution of the CO(3-2) line which happens to fall in the bandwidth used to measure the continuum emission. Such a contribution is estimated to amount to 5% of the continuum flux. Indeed, the existence of a bipolar molecular outflow in the N-S direction on a large scale has been recently revealed by interferometric maps in the CO(1-0) transition (Shepherd priv. comm.).
3.2.3. UKIRT images
The IRAS data available for IRAS 20126+4104 are affected by two major problems: the poor angular resolution and the non-detection of the region at 12 µm, where such a resolution would be best. Therefore, it is of interest to study the structure of the source at these wavelengths and determine the value of the flux in the MIR. To this purpose, we have used the MAX camera on the UKIRT to perform observations towards IRAS 20126+4104 at 10 and 20 µm. The fluxes measured at 10 and 20 µm are respectively equal to 0.32 and 30 Jy. The 20 µm image is shown in Fig. 11; the 10 µm image is essentially identical, although the S/N is poorer. In the figure, we have overlaied the MIR emission with the images obtained by CFTWO in the H2 line and in the broad K-band (2.2 µm continuum). As explained in Sect. 2.3, no astrometry was made to fix the absolute position of the MIR images and we assumed the 10 and 20 µm peaks to be coincident with the 1.3 mm continuum peak. Such an assumption seems sensible - although arbitrary - and allows a direct comparison between the structure of the MIR emission and that seen at other wavelengths.
The 20 µm emission extends mostly over the same region as that occupied by the core seen in the CH3CN lines and in the millimeter continuum. However, one sees also a faint tail of emission elongated to the NW: interestingly, this tail overlaps very well with a peak of the H2 line emission close to the core (see Fig. 11, top), whereas only some faint 2.2 µm continuum is detected at the same position (see Fig. 11, bottom). One may wonder whether the MIR emission from the tail is thermal in origin or just scattered radiation coming from the core. In fact, in the latter case the spectral index between 10 and 20 µm should be the same both for the core and the tail, whereas in the former the tail is likely to have a lower colour temperature than the core, which contains the embedded source of energy. The ratio between the 20 and 10 µm fluxes changes from 80 (corresponding to a colour temperature of 111 K) in the core to 110 (106 K) in the tail: such a difference is not large enough to unambiguously prove the thermal origin of the MIR emission in the tail, but nevertheless indicates that some difference does exist. In particular, using Eq. (31) of Goldreich & Kwan (1974) one can estimate the dust temperature of the tail: assuming a luminosity of for the embedded star and a distance of 2" from the star to the tail, one obtains 50 K, to be compared with a colour temperature of 106 K. Given the uncertainties and the rough correspondence between the colour temperature and the dust temperature, the discrepancy of a factor 2 does not seem dramatic and supports the hypothesis that in the MIR we are observing thermal emission from dust.
We conclude that the previous findings are consistent with the suggestion of CFTWO that the H2 line arises in shocked, heated gas along the jet feeding the outflow (see also Ayala et al. 1998): in this case, the MIR must be thermal in origin, being coincident with the H2 line emission. Instead, no correlation between mid- and near-infrared continuum emission is seen because the latter is just scattered radiation along the walls of the cavity created by the jet.
© European Southern Observatory (ESO) 1999
Online publication: April 28, 1999