Forum Springer Astron. Astrophys.
Forum Whats New Search Orders

Astron. Astrophys. 327, 689-698 (1997)

Previous Section Next Section Title Page Table of Contents

5. Conclusions

We have mapped the protoplanetary nebula OH 231.8+4.2 (OH 231.8) in mm-wave lines of 12 CO (1-0 and 2-1), 13 CO (1-0 and 2-1), SiO (5-4), CS (5-4), SO2 (10 [FORMULA] -9 [FORMULA]), [FORMULA] (1-0), HCN(1-0), and HNC(1-0). See central position spectra and the total CO extent in Figs. 1 and 2. The molecular envelope of OH 231.8 is comparable in extent to the optical image. All the observed molecules share the strong velocity gradient found in CO (Alcolea et al. 1996), see Figs. 3 and 4. CO shows the largest total spatial and velocity extent, as expected in view of its high intensity. The southern lobe is clearly more extended and shows the highest flow velocity that, after deprojection (Sect. 3), reaches a value relative to the central core as high as 330 km s-1.

The structure of the spatial and kinetic extent of the observed molecules, except for [FORMULA], is similar to that of CO (Figs. 4 and 5). A central clump at the central velocity (around 33 km s-1 LSR) dominates the emission, and the line wings extend along the nebula axis. The redshifted emission comes from the extended southern lobe and the blueshifted one originates in the northern more compact lobe, which is known to point toward us from optical data (Sect. 1). This high-velocity axial flow (that is also observed in other PPNe, the best example being M1-92, e.g. Bujarrabal et al. 1997) seems to be due to the interaction between fast, bipolar post-AGB jets and the dense, mostly isotropic shell ejected during the past AGB mass-loss process. The fast jet is expected to impinge on polar regions of the dense (slow) shell communicating a significant amount of momentum to a good deal of this shell, probably by means of a bow-like shock. The central condensation of OH 231.8 shows a velocity dispersion of about 45 km s-1 and seems to be a (probably toroidal) clump that is weakly affected by the wind interaction.

It is also remarkable the clumped structure of the molecular envelope of OH 231.8. The nebula seems to be formed by a number of different components, although they show some continuity in the velocity-position diagram and do not separate from the velocity gradient mentioned above. (For simplicity in our calculations, we have divided OH 231.8 in six clumps, associated to six velocity ranges, I1 to I6, see Sect. 3, Fig. 2.)

This strong velocity gradient observed in OH 231.8 is expected from quite general (and robust) dynamical considerations (see Shu et al. 1991). The shells ejected by AGB stars are thought to show a strong density variation with the radius, proportionally to [FORMULA]. If (a part of) such a shell is accelerated by interaction with a bipolar jet, the outer layers produce on it a ram pressure that increases with the velocity of the shocked region, up to compensate the ram pressure of the wind interaction. From this point, both forces are expected to vary proportionally to [FORMULA], and then the velocity must be kept constant. Even if the balance is not perfect, the decrease of the forces with the distance to the star implies that the velocity must reach soon a final value. This means that during most of the path travelled by the accelerated clumps the velocity is approximately the same for each one (but probably not the same from one clump to another). If the interaction started at about the same time for all the clumps, the constant-gradient law must hold.

[FORMULA] shows a peculiar velocity-position distribution of flux. There is an intensity relative dip at the central core; the most intense emission comes from clumps displaced towards both lobes, showing the velocity shifts that correspond to the well established velocity gradient. This result cannot be explained invoking selfabsorption, since in expanding envelopes selfabsorption just appears at relatively negative velocities (Sect. 3; this property of selfabsorption is for instance observed in the [FORMULA] line from the PPN CRL618, Cernicharo et al. 1989). Moreover, the weak [FORMULA] intensity strongly suggests optically thin emission. Excitation effects cannot easily explain neither this peculiar [FORMULA] distribution, since the J =1-0 line is difficult to excite and one would expect the most intense emission to take place in the dense central part of the nebula. Therefore, we conclude that this peculiar flux distribution is probably due to an actual increase of the [FORMULA] abundance towards the lobe regions that have suffered a strong (shock) acceleration. As discussed in Sect. 4.3, this result suggests that [FORMULA] is efficiently formed in OH 231.8 by shock chemistry.

We have calculated for the six clumps the total mass and the molecular abundances. The mass is calculated from the 13 CO data, after estimating the excitation state of this molecule, which is found to be almost constant across the nebula and equivalent to a rotational (and kinetic) temperature [FORMULA] 10 K (Sect. 4.2, Table 1). For the other molecules we also assume that the excitation of the low-J levels can be described by a single rotational temperature (Sect. 4). The resulting abundances (Table 2, Sect. 4.3) confirm, in particular, the high [FORMULA] abundance in the nebula lobes. We also note the different behavior of SiO and SO2: while SiO is clearly more abundant in the northern lobe, the SO2 abundance is larger in the south (we also argue that this result is probably not related to excitation effects, which are expected to be similar in both cases). The fact that SiO is observed in the axial flow at large distances from the star (contrarily to the case of AGB envelopes, in which SiO is only found very close to the star) also indicates that shock-induced chemistry is efficient in the OH 231.8 lobes, in this case extracting molecules from dust grains (see discussion on SiO abundance in molecular flows by Bachiller 1996). The molecular abundances given here are in general lower than those obtained by Morris et al. (1987). We argue (Sect. 4.3) that our abundance determination represents a significant improvement with respect to previous results, because of our better observational data, including mapping, and a more reliable model for the distribution of excitation and density across the nebula.

The calculated values of the total mass and momenta of the different clumps are given in Table 1; as discussed in Sect. 4, they can be somewhat underestimated but probably by no more than a factor 2. The total mass of the molecular envelope is very large, probably between 0.5 and 1 [FORMULA]. Note that at least 0.2 [FORMULA] have been accelerated in the axial direction by more than 40 km s-1. We find that the northern lobe is about 20% more massive than the southern one, but that their total momenta are almost the same (within less than 10%), due to the highest velocities in the southern lobe. This is in agreement with theoretical explanations of the peculiar optical appearance of OH 231.8 (the Calabash Nebula). Icke and Preston (1989) have proposed that the shape in the visible of the OH 231.8 can be explained if the bipolar post-AGB jet has communicated almost the same momentum to both lobes, but the northern one is slightly more massive. In such a case, the light lobe must attain a larger velocity and a wider structure; in order to explain the observations, their calculations require that the northern lobe is about 30% more massive.

It is believed that the mass loss in AGB stars is driven by the pressure exerted by the stellar radiation on the envelope material; in our case the situation is different because of the high measured momenta. As we have mentioned, the high-velocity molecular material was probably accelerated by a short-duration interaction with the bipolar post-AGB wind. Since the total post-AGB time is probably [FORMULA] 1000 yr, the wind interaction process would have lasted [FORMULA] 100 yr in order to explain the constant velocity gradient. We calculate from our results that during this time period the molecular gas won an energy [FORMULA] 2.2 1046 erg and a momentum [FORMULA] 3 1039 gr cm s-1. For a stellar luminosity [FORMULA] 104 [FORMULA], the above total energy is released by stellar radiation in about 17 yrs, therefore the conversion from radiative to kinetic energy does not need to be anomalously efficient to explain the observations. However, the momentum per year of the stellar radiation is equivalent to 4 1034 gr cm s-1 yr-1. Therefore, more than 70000 yr would have been necessary to power the observed outflow by radiation pressure: the radiation momentum is too low to explain the momentum of the molecular outflow by at least a factor 700. The situation is worse if we take into account the high collimation of the material outflow, which is not expected for radiation.

The conversion factor from radiative to kinetic momentum (due to photon pressure acting onto dust grains) can be somewhat larger than one (but not much larger than 10) due to multiple scattering (e.g. Netzer and Elitzur 1993); such an effect helps to understand the very copious mass loss in the last AGB phases but, as we see, it is still unable to explain the high axial momentum presently measured in OH 231.8. Even if we assume that the wind interaction is taking place during the whole post-AGB time (about 1000 yr), the disagreement persists. Errors in the mass values also seem unable to explain this problem. As we have mentioned, the assumptions we have made in the mass calculation (low excitation temperature, relatively high 13 CO abundance, optically thin emission) should always lead to an underestimation of the mass. If the distance we assumed is wrong, this would affect in the same way the calculation of the kinetic momentum and of the stellar luminosity, the radiation/mass momentum ratio being unchanged. Errors in the assumed inclination of the nebula axis can produce at most an overestimation of the velocity by a (moderate) factor [FORMULA] 1.5. We conclude that the high kinetic momentum in the axial direction of the molecular clumps has not been radiatively released by the star. We note that the situation in OH 231.8+4.2 is quite similar to that often present in bipolar outflows from young stars, except for that we cannot invoke here momentum transfer from interstellar material accretion. We speculate that the needed momentum was ejected by the central star of OH 231.8 by another mechanism, like the conversion of gravitational energy of the binary system (in the case the star is multiple, Sect. 1) or the ejection of material during a peculiar phase of stellar pulsation.

Previous Section Next Section Title Page Table of Contents

© European Southern Observatory (ESO) 1997

Online publication: April 6, 1998