5.1. The model
The existence of a atomic stellar wind as the driving agent of the molecular outflows was very early suggested (Natta et al. 1988), but only the low-mass outflows HH7-11 and L1551 (Giovanaradi et al. 1992), T Tau (Ruiz et al. 1992) and the two high-mass bipolar outflows NGC 7021 (Bally & Stark 1983) and DR21 (Russell et al. 1992) have been detected in HI so far. Other well known outflows, like RNO 43, FU Ori, R Mon, NGC 2264, L1448, have not been detected (Giovanardi et al. 1992). The low-mass stellar outflows HH7-11, L1551, T Tau have been detected using the Arecibo telescope. In these sources the atomic gas reaches velocities up to 150 - 260 kms-1 and the momentum supply by the atomic gas can drive the molecular outflow in a momentum conserving interaction. Unfortunately, the low angular resolution of these observations prevent from a study of the spatial distribution of the atomic gas. It was interpreted as a neutral stellar wind that drives the molecular outflow. Toward the high-mass outflows, NGC 2071 and DR 21, the HI 21-cm line has been observed using the VLA. The high velocity atomic gas detected towards NGC 2071 is not able to drive the molecular outflow, and could be due to the photodissociation of the molecular gas by the UV radiation of the recently formed stars. However, towards DR21 Russel et al. (1992) detected a massive (24 ) atomic jet which is capable to drive the massive molecular outflow associated with this source.
The case of HD 200775 seems to be different from that of low-mass stars. The momentum of the atomic outflow is more than two orders of magnitude lower than the momentum of the molecular outflow. Though we cannot discard the possibility of a high velocity atomic wind flowing along the outflow axis, the morphology of the region suggests that this is not the case. Furthermore, we only detect high velocity HI gas in the eastern lobe. Thus the HI gas is not able to drive the molecular outflow at the present stage, and could not have been the driving agent of the outflow which excavated the cavity. We propose that the cavity was formed in an earlier evolutionary stage, and it is, in some sense, a fossil of the bipolar outflow. Most of the molecular gas was accelerated when the cavity was formed and the high velocity atomic gas we observe now drives only the highest velocity molecular gas of the eastern wall (V 6 kms-1 and V -1 kms-1). The study of the cavity gives important information about the history of the outflow. The four shells observed in 12CO and the symmetry between the two lobes, suggest that at least two episodes of energetic ejections took place along the outflow axis. But it seems that once the central channel was evacuated the high velocity gas started to outflow along the walls of the cavity instead of inside. The tunnels observed in 13CO were very likely formed at this stage. Surprisingly the symmetry of the region is not perfect. The center of symmetry of the multi-shell structure is not located towards the star, but about 40"- 50" (0.08 pc - 0.10 pc) to the west. This is also the position of the waist of the biconical cavity relative to the star. This lack of symmetry is more evident when we study the spatial distribution of gas and dust around the star at a smaller spatial scale. As commented in Sect. 4.3, the circumstellar material is located mainly to the west of the star. Then, the extinction from the star is higher towards the west than towards the east. We propose that this asymmetry is the responsible of the cometary shape of the HI outflow. First, because of the lower extinction towards the eastern lobe, the heating and photodissociation of the gas by the stellar UV radiation is producing mainly in this lobe. Secondly, the HI outflow could be driven by an isotropic stellar wind. In this case, the stellar wind would impinge to the west in a dense wall and turn back to the east producing a cometary outflow. This model is illustrated in Fig. 10.
5.2. The origin of a cometary nebula
In Sect. 5.1 we propose that the cometary shape of the HI emission is the consequence of an asymmetric distribution of matter around the star. But it is difficult to justify this asymmetric distribution if the disruption of the core has been produced by the bipolar outflow. At some point the bipolar symmetry disappeared. Two scenarios are proposed to account for this change: i) The existence of a binary could have introduced asymmetries in the disruption of the parent core. A close companion has been detected for this star by Pirzkal et al. (1997). ii) The star has moved away from its initial position. Once the bipolar outflow excavated the cavity, the star could have wandered to the east. Then the dispersal of material would be more efficient to the east than to the west producing an anisotropic distribution of the matter surrounding the star. The advantage of the latter explanation is that it can account for the "strange" position of the star relative to the biconical cavity . Both, the multi-shell structure observed in 12CO and the biconical cavity seem to be centered in a position located 40"- 50" (0.08 pc - 0.10 pc) to the west of the star. It seems as if the exciting star of the initial bipolar outflow were located at this position instead of at the position of HD 200775. One could think that HD 200775 is not responsible for the cavity. A possible low-mass companion located at the waist of the cavity could have produced it. This explanation is very unlikely. There is no evidence for the existence of a star around this position (see Sellgren et al. 1983). Furthermore, the amount of material swept up by the outflow suggests that the cavity has been formed by an intermediate-mass star (see Sect. 4.2). It is more plausible to think that the biconical cavity was excavated by HD 200775 and the star has changed its position. The proper motions of HD 200775 are (, ) = (+6.74 0.64 mas yr-1, -1.48 0.54 mas yr-1) (from Hipparcos Catalogue). The direction of these motions is in agreement with the observed displacement. In 3000 yrs the star could have moved from the center of symmetry of the bipolar outflow to its position nowadays. This time is negligible compared to the age of the outflow ( 105 yrs). We are aware that the proper motions measure the absolute motions of the star and do not give any information about the relative motion of the star with respect to the surrounding cloud. Both, the star and the parent molecular cloud could be moving at the same velocity. To have an estimate of this motion, we have compared the velocity of the star with that of the molecular gas in its surroundings. Finkezellet & Mundt (1984) obtained that the difference in velocity between the star and the molecular cloud is 3 kms-1. At the distance of HD 200775, a velocity of 3 kms-1 would correspond to 1 mas yr-1. This velocity is of the same order as the measured proper motions. Of course, there is a great uncertainty in the velocity of the star and consequently in this number. But the existence of a shift in radial velocity proves, at least, that the model is plausible. Cometary nebulae and cometary HII regions are very common in the interstellar medium. If intermediate-mass and massive star moves away from their initial position in the last stages of the mass-loss phase, the cometary shape of these regions has an easy and simple explanation. A similar observational work in other cometary nebulae is required to check this hypothesis.
© European Southern Observatory (ESO) 1998
Online publication: October 21, 1998