The main result from the present observations is that the far-infrared lines of water vapour are seen in absorption toward SgrB2, instead of in emission. Similar results are reported for other galactic molecular clouds by Cox et al. (1997). This suggests that in SgrB2, where the continuum emission in the far-infrared is optically thick (see below), the H2 O lines arise in a region where the excitation temperatures are smaller than the dust temperature ( 30 K). Nevertheless, the problem of H2 O line excitation is not simple because, apart from self-absorption and the possibility of collisional excitation, the excitation temperatures will also be affected by the continuum emission of the dust. The role of dust is essential in the excitation of molecules such as H2 O which have transitions at far- and near-infrared wavelengths, and must be taken into account in any realistic model of line excitation.
In the following, we present a simple model to define the conditions for which the far-infrared H2 O lines appear in absorption and to quantify the depth of the absorption lines. The model assumes a uniform and spherical cloud, with the following physical parameters appropriate for SgrB2 (see Baluteau et al. 1997 and references therein): an H2 column density of H2) cm-2, a dust to gas abundance of by mass, a grain radius 0.1 µm, an absorption efficiency , a dust temperature =30 K, a turbulent velocity km s-1. The H2 density H2), the kinetic temperature , and the water abundance are free parameters. The dust continuum observed in the LWS wavelength range is well reproduced with the above parameters (Fig. 3c). The continuum opacities are 3.5 and 1.9 at 100 and 180 µm respectively. The models compute the statistical equilibrium populations of the seven lowest levels of o-H2 O. The collisional coefficients are taken from Green et al. (1993) and the excitation by dust emission is treated by assuming that the dust grains and the water molecules are coexistent (González-Alfonso & Cernicharo, in preparation).
Figs. 3a and b display for the four lowest o-H2 O transitions with wavelengths in the LWS range the line over continuum flux ratio as a function of for two values of H2) and cm-3 and for an abundance of o-H2 O of 10-5. All four lines are very thick across the cloud, with maximum opacities always greater than . The lowest lying 212 -101 line at 179.5 µm has an opacity greater than 5 104 and its line over continuum flux ratio varies significantly with . Fig. 3 indicates that the H2 O lines are seen in absorption only if is below 35 and 30 K for H and cm-3, respectively. Hence, the water vapour lines detected in absorption originate in the low density regions of SgrB2, i.e. the extended and tenuous envelope of the molecular cloud, and not in the inner regions with K observed, e.g., in NH3 (Hüttemeister et al. 1993), where the water lines would be expected to be in emission. With the grating resolution, and even when collisions are neglected, the absorption depths in our models for the 108.1, 174.6 and 180.5 µm lines are always lower than 1 %. This is due to the fact that the dust emission increases the line excitation temperatures to values close to . This result is consistent with the Fabry-Perot observation of the 303 -212 line at 174.6 µm (Fig. 2), which shows an absorption of %. However, the absorption depth of the 212 -101 transition ( 15 % in the grating spectra) cannot be explained by this simple model. The reason is that the observed velocity coverage of this line is 200 km s-1 (Fig. 2), indicating that low excitation molecular clouds along the line of sight toward SgrB2 also contribute to the absorption. The excitation of the H2 O molecules in these intervening clouds must be low since the 303 -212 absorption line is not detected at velocities below 20 km s-1 (Fig. 2b). This is further confirmed by the lack of reemission of the 212 -101 line at those velocities. We have added to the previous model (with H2) cm-3 and K) an absorbing shell located between the observer and the SgrB2 model cloud. The parameters of this shell are the same as for the SgrB2 cloud except that H2) cm-2 and km s-1. Since the absorbing shell is optically thin in the continuum, the excitation temperatures are low and there is only appreciable absorption in the 179.5 µm line. Fig. 3c presents the resulting grating spectrum and the insert panel shows the line over continuum ratio around 180 µm. The present observations confirm the previous suggestion made by Cernicharo et al. (1994) that water vapour exits in large amounts in the direction of molecular clouds. However, the large opacity of the infrared H2 O lines detected in SgrB2 makes any estimate of the water abundance difficult. Our models indicate that the H O 212 -101 line must also be thick in order to explain the line absorption at 181.05 µm. Nevertheless, we note that the depth of the absorption is still sensitive to the optical depth even if the line is thick, because the velocity range where the line can absorb the continuum emission increases with the opacity. Hence, our simple models can derive the H O abundance from the observed absorption depth at 181.05 µm. The value depends on the assumed velocity dispersion in the cloud. Adopting km s-1, H2) cm-2, [H2 O]/[H O]=500 and neglecting collisional excitation, i.e. all the rotational levels are excited by infrared photons from the dust, we derive a lower limit for the water abundance in SgrB2 of 10-5.
Finally, the H O line profile in Fig. 2 shows an absorption of 10% between -50 and 50 kms-1. This absorption is produced in the low density gas associated with the molecular ring around the galactic center. Assuming that all the absorbing molecules are in the ground state, we derive N(H O) 5 1013 cm-2 and N(H2 O) 3 1016 cm-2. Assuming gas column densities of 1021 -1022 cm-2 the water vapour abundance in the clouds along the line of sight of SgrB2 is 0.1-1 10-5.
© European Southern Observatory (ESO) 1997
Online publication: June 5, 1998