6. Conclusions and discussion
We have made a thorough analysis of the available observational data for the PNe with H-deficient central stars. As we have shown in Sect. 3 an evolutionary sequence can be identified within this group of objects. The stars start their evolution as late-[WC], then become early-[WC] and end up as PG 1159 stars. This is however a general sequence and it can be that some individual objects evolve in a different way.
The analyses in Sects. 4 and 5 give arguments that the very late He-shell flash (born again AGB) is not the main evolutionary path giving origin to the H-deficient PNNi. We argue that the H-deficient central stars evolve directly from the AGB as do other PNNi in their majority. This conclusion is most evident for the [WC] PNNi and is consistent with the observed characteristics of the PG 1159 stars and their nebulae. Nevertheless it cannot be excluded that some H-deficient PNNi have been produced in the late He-shell flash. Indeed for some objects, i.e. A 30, A 58, A 78, FG Sge and V4334 Sgr, we have direct evidences that they have experienced a late He-shell flash. However, these objects are expected to evolve directly to the region of PG 1159 stars without passing through the [WC] region.
Our conclusion that the H-deficient PNNi mostly evolve directly from the AGB has serious implications. First, the mass loss from (at least some) stars leaving the AGB while burning helium in the shell has to remove all the mass of the H-rich envelope before the PN phase. This has not yet been predicted from any standard evolutionary modelling of the post-AGB phase. The point is that when the H-rich envelope drops below typically the stellar effective temperature starts increasing, accordingly mass loss drops and is not able to remove the rest of the H-rich envelope before the onset of the PN phase. The analysis in Sect. 4 shows that the duration of the post-AGB (or proto-PN) phase is similar for the H-deficient PNNi as for the others. Thus a proto-[WC] PNN during a few thousand years gets rid of the entire H-rich envelope which implies a mass loss rate typically of . Similar mass loss rates are also observed in the [WC] phase. This can be interpreted as an indication that the yet unidentified mass loss mechanism responsible for the appearance (and evolution) of the [WC] PNNi resides beneath the H-rich envelope. One may speculate that He- and C-rich layers pushed up during the helium shell flash cool down and form a high opacity layer well below the stellar photosphere. Increased radiation pressure (due to high opacity and high luminosity during the flash) might lift up upper stellar layers and form an intense wind. A detailed radiation-hydrodynamic modelling of flashing AGB stars with small H-rich envelopes are obviously necessary.
The second implication results from the observed abundances in the [WC] PNNi. Spectroscopic analyses of the latest [WC] (i.e. youngest according to our analysis) stars show that from the very beginning of the [WC] phase the observed C abundance is comparable to that of He (e.g. Leuenhagen et al. 1996). This implies a very intense mixing in the He-rich layers during the He-shell flash. Standard treatment of convection gives models in which upper He-rich layers are composed of almost pure He (see e.g. discussion in Schönberner & Blöcker 1992). Significant overshooting seems to be required in order to account for the observed abundances (Herwig et al. 1997).
© European Southern Observatory (ESO) 2000
Online publication: October 30, 2000