Astron. Astrophys. 364, 613-624 (2000)

4. Discussion

4.1. General properties

IRAS 12326-6245 is associated with maser emission and with at least one of the two UC H II regions indicated in the Fig. 2 -Fig. 5.

Both phenomena are usually associated with the presence of very young massive stars. However, the UC H II regions remained unresolved with arcsec resolution (Walsh et al. 1998). This indicates that either the regions have a large distance or that they are very young. The UC H II regions have sizes smaller than 6 10³ AU which is rather small compared with other well-known UC H II regions, like G5.89-0.39.

Based on the free-free continuum emission measured by Walsh et al. (1998), we derived the brightness temperature , the optical depth , the emission measure EM, the electron density , the excitation parameter U, and the Lyman continuum photon flux (see Wood & Churchwell 1989). Using the calibration by Vacca et al. (1996) and assuming a single star as the excitation source for each of the UC H II regions, we obtain a spectral type B0 or earlier if dust extinction in the UC H II regions is important. This gives a luminosity of 8 10⁴  for the individual objects. The total luminosity derived from the IRAS and MSX measurements amounts to 3 10⁵  in good agreement with the estimates of the radio data.

Fig. 7 displays the spectral energy distribution (SED) of IRAS 12326-6245 taking into account the IRAS data, our infrared, and 1.3 mm continuum data. The data can be well fitted by a modified Planck function (1-exp( with The parameters of the fit are a temperature of 607 K and an index for the wavelength-dependent absorptivity of 1.20.5. The luminosity amounts to = 3.41.0 10⁵ . We note that the region has a very complex structure. Therefore, the use of a single modified black-body function is certainly oversimplified. In principle, we expect a distribution of temperatures inside the considered area. However, a radiative transfer calculation would require spatially resolved data to constrain the adopted model. The SED is similar to the SEDs of other UC H II regions (e.g. W 3(OH); Chini et al. 1986), but with less flux at near-infrared wavelengths pointing again to deeply embedded sources. This is clearly different to the BN-type objects (e.g. AFGL 490) which show even more near-infrared emission (Henning et al. 1990, 2000b; Gürtler et al. 1991).

Fig. 7. Spectral energy distribution of IRAS 12326-6245 using our infrared data together with the information from the MSX catalog (, Egan et al. 1999), the IRAS fluxes (), our 1.3 mm continuum fluxes, and the free-free emission measured at radio wavelengths by Walsh et al. (1998). In addition, we show a modified black-body curve (solid line). The fit parameters are a temperature of 607 K and an index for the wavelength-dependent absorptivity of 1.20.5.

4.2. Physical parameters of the core and envelope

Using the 1.3mm continuum measurements as well as the molecular line maps, we can estimate physical parameters as the column densities, the volume-averaged hydrogen number densities, and the masses of central cloud core as well as the total mass of the molecular cloud associated with IRAS 12326-6245.

The H₂ column density and the mass can be derived from the 1.3 mm continuum observations under the assumptions that the emission is completely dominated by thermal dust radiation (see, e.g., Watson et al. 1997) and that it is optically thin. The first assumption is fully justified when the low free-free emission at millimetre wavelengths, extrapolated from the radio data, is considered. The total gas mass is 11900 and the core has a mass of 2400 . Here, we used a gas-to-dust mass ratio of 150 based on solar metallicity and a correction factor of 1.36 to account for He and metals. For the mass absorption coefficient of the dust , we chose values of (1.3mm) = 0.9 cm² g^-1 for the dense core and (1.3mm) = 0.5 cm² g^-1 for the envelope (Ossenkopf & Henning 1994) which refer to coagulated dust particles with thin ice mantles in high and intermediate dense protostellar cores. Following Henning et al. (2000a), we applied a `beam-averaged' dust temperature of = 50 K for the core since the hot environment of O and early B type stars fills only a small percentage of the beam (see also the black-body fit to the SED in Sect. 4.1). For the more extended gas component, we adopted an average temperature of = 30 K.

In addition, the source-averaged molecular hydrogen column density for the core was determined from the 1.3mm continuum measurements (Henning et al. 2000a). We estimated a value of = 6.25 10²³ cm^-2 for IRAS 12326-6245 which is one order of magnitude larger then the estimates found by Henning et al. (2000a) for comparable massive star-forming regions.

In order to estimate the volume-averaged hydrogen number density n(H₂) from our 1.3 mm continuum measurements, we assume that the source has the same extent along the line of sight as in the plane of the sky. Based on the barely resolved Gaussian core profile in the NE-SW direction, a beam-deconvolved source size of 7.3" was determined from the FWHM of the Gaussian core, if we use a beam size (HPBW) of 23 Since, the Gaussian core profile in the NE-SW direction is unresolved, we assumed a mean beam-deconvolved source size of 6.5 The result is n(H₂) = 3.5 10⁷ cm^-3 which implies a very high central core density in a rather large volume (source size of 6.5" is equal to 0.49 pc in a distance of 4.4 kpc) compared to other well-known high-mass star-forming regions, like e.g. Orion IRc2 or W3 IRS5. If we use the beam-convolved source size of 23 n(H₂) amounts to a much lower value of n(H₂) = 8 10⁵ cm Based on the large distance and the low angular resolution of the 1.3 mm continuum measurements, a clear distinction of both the core component and the envelope is not possible. However, the strong intensity contrast between the extended emission and the central component implies that most of the core emission is dominated by radiation from dense and warm regions being the birth places of very young massive stars. In this respect, we note that all N- and Q-band objects (as well as both UC H II regions) shown in Fig. 4 are located in a region covered by the 23" beam.

The column density and the mass can be also determined by the presumably optically thin C¹⁸O emission. Simple LTE estimates give the relation

assuming an excitation temperature of  K, which is similar to the relation based on LVG calculations by Mauersberger et al. (1992). This value of is close to the CO peak brightness temperature. The estimate changes by less than a factor of 2 when varies between 10 and 100 K. The hydrogen column densities and masses are derived under the assumption of a C¹⁸O relative abundance of (Frerking et al. 1982). For the hydrogen column density, we obtain N(H₂) = 2.0 10²³ cm This value is quite similar to the estimate we obtained from the dust continuum measurement. The beam-averaged extinction can be estimated from the relation by Frerking et al. (1982) and ranges between 200 (from C¹⁸O) and 600 mag (from continuum measurements).

The masses were determined from

where m is the mass of the hydrogen molecule, µ is the ratio of total gas mass to hydrogen mass, (e.g. Hildebrand 1983), and is the pixel size (10"10" in our case). In this way we obtain   within the C¹⁸O(2-1) map which is covering only the central cloud core. This is remarkably close to the core mass derived from the dust emission. However, taking into account the difference in the peak gas column densities derived from C¹⁸O(2-1) and the dust emission this coincidence can be explained only by different source sizes in C¹⁸O(2-1) and in the 1.3 mm continuum. In fact the "average" C¹⁸O(2-1) FWHM is 34" which is much larger than that of the continuum emission. Using the virial theorem (density gradient of p = 2, line width of V = 8.7 km s and radius of r = 50" of an spherically symmetric cloud model), we obtain a virial mass for the whole cloud of 10⁴ which agrees very well with the total mass derived from the 1.3 mm continuum which includes core and envelope.

4.3. Outflow

The data clearly show the presence of a high-velocity molecular outflow. The terminal line of sight velocity reaches  km s^-1 relative to the bulk of the cloud (from the CO(2-1) and SO data). The CO(2-1), SO, and C¹⁸O(2-1) "wing" maps show that the direction of this outflow is approximately SE-NW and roughly perpendicular to the major axis of the dense core traced in CS, in the = 1.3 mm continuum emission, and in the bulk of the C¹⁸O(1-0) emission mapped by Lapinov et al. (1998).

The Fig. 4a and Fig. 5 clearly show that the centre of the outflow probably coincides with the southern UC H II region which is also the strongest infrared source in the N and Q bands. However, it is necessary to note that the positional uncertainties of the SEST maps can be substantial on this scale.

The outflow is most pronounced in the CO and SO lines (as well as in the SiO lines measured by Harju et al. 1998). However, the CO(2-1) and SO lines can be hardly used for a derivation of outflow mass and related parameters due to the apparently high optical depths and the uncertain SO abundance in shocked gas. This fact is illustrated by Fig. 8 where we plot ratios of the main beam brightness temperatures in CO(2-1) and C¹⁸O(2-1) as well as in ³⁴SO and C¹⁸O(2-1), respectively, as a function of the velocity.

Fig. 8a and b. Ratio of the brightness temperatures in for CO(2-1):C18O(2-1) a and for 34SO:C18O(2-1) b as a function of the LSR velocity.

Fig. 8a shows that the CO(2-1)/C¹⁸O(2-1) brightness ratio increases with increasing flow velocity and reaches a value of in the outer wings. This means that the optical depth in CO(2-1) decreases with increasing velocity but still remains very high even in the wings because the ratio is still much lower than the abundance ratio (the terrestrial value is ). A simple estimate shows that in the wings (assuming similar excitation for both isotops).

Both the C¹⁸O(2-1) and ³⁴SO lines are presumably optically thin so that strong variations of their intensity ratio most probably reflect variations in abundance ratio because changes in excitation lead only to moderate changes in intensities. Therefore, Fig. 8b shows that the SO/CO abundance ratio increases significantly with the flow velocity (by about an order of magnitude in the wings). This is in qualitative agreement with observations of other outflows (e.g. Chernin et al. 1994).

Apparently better estimates of the outflow parameters can be obtained from C¹⁸O where the abundance is rather constant and the emission is optically thin. For this purpose, we constructed an integrated (over the source) C¹⁸O(2-1) spectrum plotted in Fig. 9. The quantity plotted here is . It is proportional to the mass per unit velocity interval.

Fig. 9. The integrated C18O(2-1) spectrum multiplied by the pixel size. The thin solid line represents a 2-dimensional Gaussian fit. The dashed line shows the broad Gaussian profile separately.

This spectrum can be well fitted by two Gaussians with different widths as shown in Fig. 9. It seems to be reasonable to identify the broad component (V  km s^-1) with the outflow. Another approach is that the emission from outflowing gas is identified with residual "wings" in the line profile after subtraction of a single Gaussian. In this case the velocity distribution of the outflowing gas is arbitrarily "zeroed" at lower velocities which cannot be justified by the data. Therefore this approach gives essentially a lower limit for the outflow mass. In the following we shall call it the "traditional" estimate. A detailed discussion of the amount of high-velocity emission within the ambient cloud line core and the uncertainties of different approaches to derive flow parameters can be found in the paper by Cabrit & Bertout (1990).

In the following, we give two sets of the estimates for physical parameters: (1) the "traditional" one which refers to the gas producing wings in the line profile above a single Gaussian and (2) that which refers to the broad Gaussian component in the Gaussian fit. We shall call it the "Gaussian" estimate. In the latter case we subtract from the source-integrated spectrum the narrow (V = 4.5 km s^-1) Gaussian component.

The main integral outflow parameters are mass (), mechanical momentum (), and kinetic energy (). From the residual C¹⁸O spectrum, we can derive these quantities per unit velocity interval under the same assumptions as above. In this way, we obtain from Eq. (2) applying it to an unit velocity interval, and . These quantities are plotted in Fig. 10. The shaded areas are used to derive the integral parameters in the "Gaussian" (G) approach. The heavily shaded areas correspond to the "traditional" (T) approach. From the 2-component Gaussian fit, we obtain:  ,   km s and   km² s^-2  erg. The outflow is neither located in the plane of the sky nor completely perpendicular to that plane. Here we assumed a value of 57.3^o relative to the plane which is the average value for a uniform distribution of random orientations.

Fig. 10a-c. The outflow mass in a , the mechanical momentum in b , and the kinetic energy in c per unit velocity interval are shown as function of the LSR velocity. The light-shaded areas are used to derive the integral parameters in the "Gaussian" (G) approach. The dark-shaded areas correspond to the "traditional" (T) approach (see Sect. 4.3).

The mass estimate for the "traditional" approach is much lower,  . For the momentum we get   km s^-1 and for the kinetic energy   km² s The mass of the outflow can also be estimated from the CO(2-1) map (Fig. 4a) (see, e.g., Henning et al. 2000a). Here, we obtain a value of 320 , well in the range of the above estimates. The momentum we get from the CO(2-1) map is 5000  km s^-1 comparable with the "Gaussian" estimate.

Other often determined outflow parameters are the dynamical timescale , the mass outflow rate , the force or momentum supply rate , and the mechanical luminosity . The dynamical flow age is not a very well-defined quantity. It would be better to replace this quantity by the time in which the "information" travels from the central source to the most distant points of the outflow. However, such an estimate is not yet possible because the information speed cannot be directly estimated from the observational data.

For the sake of comparison with estimates in the literature, we derived the dynamical timescale by the expression , where is assumed to be equal to half of the distance between the high-velocity emission peaks and (e.g. Cabrit & Bertout 1990; Shepherd & Churchwell 1996a,b). The separation between the peaks of the blue-shifted and red-shifted emission is in C¹⁸O but much lower in SO. An estimate of the average velocity is very different for the two approaches outlined above due to the difference in mass estimates. With the "traditional" approach we obtain  pc,  km s^-1 and years without correction for the inclination angle. With an inclination angle of 57.3 we obtain 7 10³ years. Then, the other parameters are the following:  /yr,  km s^-1/yr and  . All these values are among the highest known from the literature for objects of similar luminosity (see Shepherd & Churchwell 1996a,b; Acord et al. 1997; Henning et al. 2000a). The origin of the outflow must be very close to the mid-infrared sources. The near-infrared nebulosity points to the northern UC H II region as the origin of the outflow. However, without interferometric measurements of the outflow, the question from which source the outflow exactly comes cannot be finally answered.

© European Southern Observatory (ESO) 2000

Online publication: January 29, 2001
helpdesk.link@springer.de