4.1. Origin of the mm continuum emission
As mentioned above, the mm emission in I20126 is coincident with the mm emission peak measured by Cesaroni et al. (1997). The spectral index measured between mm and mm is . This is in good agreement with the recent measurement of Cesaroni et al. (1999) of the spectral index between mm and mm of 2.9. Hence we conclude that the mm emission arises from dust with an absorption coefficient proportional to .
4.2. An ultracompact H II region model
We now turn to the interpretation of the cm continuum emission. A possible explanation of the cm continuum emission shown in Fig. 2 is a cluster of three individual HII regions, each having a cm continuum flux density of about mJy. The two regions making up the northern component are unresolved, i.e. an upper limit on their size is , or pc. Assuming optically thin conditions with an electron temperature of K, neglecting the absorption of ionizing photons by dust and using the tables of Panagia (1973), we find that three ZAMS stars of spectral type B3 are required to explain the observed flux density. Three B3 stars only provide a luminosity of 3000 as compared with the FIR luminosity of measured by the IRAS satellite. This reflects the well known fact that the spectral types derived under these assumptions represent lower limits. Besides the internal absorption of ionizing photons by dust, and the presence of unrelated luminous sources in the IRAS beam, a large fraction of the luminosity in I20126 could derive from accretion: indeed, according to Cesaroni et al. (1999), the latter could be .
We tend to exclude that the cm continuum emission is optically thick. From the observed brightness temperature (6.5 K) we can derive the size of the emitting source of pc: this number is extremely small and even for a spectral type as late as B3 is of order of the initial Strömgren radius for densities of cm-3. This would imply that the age of the HII regions is of order a few hundred years, a somewhat unlikely scenario. Also, the continuum spectrum must become optically thin at frequencies larger than GHz, otherwise we should have detected excess emission over what is expected from dust alone at mm.
4.3. An ionized jet scenario
Both cm continuum components appear narrow and elongated in the direction of the large scale outflow in I20126. This morphology is strongly suggestive of ionized jets. In what follows, we will favor this scenario for the origin of the cm continuum emission. In order to estimate the physical parameters of the ionized jets, we have fitted single gaussians to both northern and southern sources. Assuming a two-sided symmetric jet, we can estimate its solid angle. In Table 1, we list the deconvolved sizes, solid angles, position angles, and total flux densities based on this model. At the assumed distance of kpc the physical dimensions of the jets are about AU along the major axis and AU along the minor axis, so that collimation must have occurred within a few hundred AU.
Table 1. Observed jet parameters
What is the origin of the continuum emission in the jet scenario? Since the data presented here are presently the only detection at cm-wavelengths toward I20126, we cannot exclude a non-thermal origin of the cm emission. However, since most radio jets detected toward highly embedded systems like I20126 are of thermal nature (e.g. Anglada 1996, but note Reid et al. 1995), we will restrict our discussion to the case of thermal radio jets. Detection of the radio emission from I20126 at other wavelengths would be extremely valuable to check this hypothesis.
Assuming then that the cm emission is of thermal nature, there are two possible sources of the ionized gas in I20126. First, we consider shock induced ionization, i.e. the gas could be ionized by UV photons produced when a neutral stellar wind shocks against the surrounding high density matter. We can test this by comparing the observed fluxes with those predicted from the measured momentum rate of the large scale molecular outflow (e.g. Anglada 1996). Using Eq. (8) of Curiel et al. (1989) with a terminal wind velocity of km s-1, we derive momentum rates of 0.01 and 0.004 yrkm s-1 for the northern and southern source respectively. These values are quite similar to what was obtained for the ionized jet from the massive object IRAS18162-2048 (HH 80-81) by Martí et al. 1995. Because the continuum sources are unresolved along their minor axis, these numbers represent lower limits on the momentum rate. The momentum rate as estimated from the HCO+(1-0) observations of Cesaroni et al. (1997) is 0.09 yr-1 km s-1, much larger and thus consistent with our lower limit. Clearly, higher resolution observations which resolve the ionized jet would be extremely useful to improve this comparison. Also, note that the momentum rate derived from the HCO+(1-0) data carries large uncertainties due to the unknown chemical abundance.
Second, the ionization could be caused by stellar UV photons. In this case we compute the total stellar Lyman continuum flux by assuming that only UV photons emitted into the solid angle of the jet contribute to the ionization while the UV photons emitted in the perpendicular plane are absorbed in the high density environment very close to the star which is possibly provided by an accretion disk. Because of dust absorption and again due to the fact that the jets are unresolved along the minor axis we obtain lower limits on the total Lyman continuum flux of and s-1 for the northern and southern jet respectively. These numbers correspond to ZAMS spectral types of B1 and B2: the expected FIR luminosity for such a system is 8000 , very close to the measured FIR luminosity of I20126.
In this scenario we can also obtain a limit on the momentum rate following the theory of Reynolds (1986). Using his Eq. (19) and under the assumption of an isothermal jet of temperature K with constant opening angle, a constant ionization fraction of and a terminal jet velocity of km s-1, we find upper limits on the momentum rate of about 0.001 yr-1 km s- 1. The momentum rate derived from the HCO+(1-0) observations is about two orders of magnitude larger which seems to contradict this theory. However we need to keep in mind that this comparison could be strongly affected by the uncertainty of the assumed values of wind velocity, ionization fraction and HCO+(1-0) abundance.
4.4. Water masers
As noted above (see Fig. 2) the water masers in I20126 are coincident with the northern 3.6 cm continuum source and are aligned at the same position angle. The observations of Tofani et al. (1995) show that the water maser spots at the NW and SE edges of the continuum have narrow velocity distributions with velocities blueshifted in the SE and redshifted in the NW with respect to the bulk velocity of the molecular gas in I20126 of km s-1. The maser spot close to the maximum of the continuum emission shows a wider range of velocities, but still distributed symmetrically around the bulk velocity. Possibly the central maser spot occurs close to the collimation region, so that both blue and redshifted velocities can be observed as the molecular gas expands away from the central object, whereas the maser spots at the NW and SE edges may occur in a turbulent molecular layer surrounding the ionized jet (e.g. Raga et al. 1993, Torrelles et al. 1997) and since the jet is almost in the plane of the sky the observed radial velocities here are relatively small. Thus the distribution and velocities of the water masers lend support for the ionized jet scenario.
© European Southern Observatory (ESO) 1999
Online publication: April 28, 1999