In the observed population of planetary nebula nuclei (PNNi) several spectral classes can be distinguished, e.g. O, Of, sdO, WR, WR-Of (e.g. Lutz 1978). From the point of view of the evolutionary status of these stars the most important classification is, however, that deduced from analyses of their surface abundances. In this case the PNN population can be divided into two well defined groups, i.e. hydrogen-rich and hydrogen-deficient (Méndez 1991). In the first group hydrogen is a dominant element in the PNN atmosphere and its abundance is close to the cosmic value. In the second case the star atmosphere is practically hydrogen-free while helium and carbon are the most abundant elements.
From spectral appearances one can distinguish two classes of the H-deficient PNNi. The spectra of many H-deficient PNNi are dominated by strong and wide emission lines characteristic for the Wolf-Rayet stars. At present we know about 50 objects of this kind and all of them are of [WC] type. On the other side of this scheme we have PNNi whose spectra show predominantly absorption lines. Many stars from these group have been found to pulsate like PG 1159-035. A few H-deficient PNNi do not fit this simple scheme. For example, the central stars of the planetary nebulae A 30 and A 78 show absorption lines and weak emissions (WR-Of(C) type according to Méndez (1991). Another case is the central star of the nebula Longmore 4 which has been observed to change its spectral appearance from PG 1159 to [WC 2-3] and back (Werner et al. 1992). The WR-type emission lines are interpreted as evidence of stellar winds. Detailed analyses of spectra of the [WC] PNNi give mass loss rates typically of . No other type of PNNi has been observed to lose matter at such a high rate. Recent reviews on the H-deficient PNNi can be found in Hamann (1996, 1997), Napiwotzki et al. (1996), Tylenda (1996), Werner et al. (1996), and Lundström & Stenholm (1996).
It is not clear at present which evolutionary path gives origin to the H-deficient PNNi. The principal problem are the observed surface abundances. It is evident that the [WC] PNNi burn helium in a shell (e.g. Tylenda & Górny 1993). Some [WC] PNNi are surrounded by rather young and dense nebulae which suggests that these objects have recently left the asymptotic giant branch (AGB) after having formed the planetary nebula (PN) presumably during a helium-shell flash (e.g. Méndez 1991). Then it seems natural to suppose that the PG 1159 stars descend from the [WC] PNNi. This idea is supported by the fact that the surface abundances are similar in both types of stars and that the PG 1159 stars, contrary to the [WC] PNNi, have old, faint or no detectable PN.
There is a number of theoretical evolutionary tracks which produce He-burning PNNi after having left the AGB during a He-shell flash (Iben 1984; Wood & Faulkner 1986; Vassiliadis & Wood 1994). However, these model PNNi keep a H-rich envelope throughout the PN phase so they cannot be interpreted as model [WC] PNNi. One can adopt that an intense mass loss continues after AGB until all the H-rich matter is lost. When it happens the surface layers are of almost pure helium and mass loss has to be continued in order to expose layers where carbon is as abundant as helium. When this is achieved the He-rich envelope is, however, too small to maintain He-burning (Schönberner & Blöcker 1992). Thus a high luminosity necessary to power the WR-type wind is no longer ensured.
The observed surface abundances of the H-deficient PNNi can be much more easily explained in an evolutionary scenario of a very late He-shell flash (also called born again AGB). In this case a PNN which has already finished its high luminosity phase as a H-burner suffers from a He-shell flash in the cooling phase. The star quickly expands, goes back to the AGB and repeats the PNN evolution, this time as a He-burner. The rests of hydrogen are lost in a wind and/or mixed and burnt. Strong convective mixing dredges carbon up to the surface. For the first time the born again AGB scenario has been described and analysed by Iben at al. (1983) in their study of the PNe A 30 and A 78. More recently, this scenario has been applied to explain the observed abundances in the PG 1159 stars (Schönberner & Blöcker 1992; Iben & MacDonald 1995).
The third possibility has been invoked and discussed in Tylenda & Górny (1993) and involves the AGB evolution in binaries. Evolutionary calculations of AGB stars (e.g. Vassiliadis & Wood 1994) show that it is during He-shell flashes when the stellar radius achieves local maxima (at least for initial star masses below 3.5 ). Therefore one may argue that if an AGB star happens to fill its Roche lobe it takes place most probably during a He-shell flash. If the common envelope phase ends up with a complete loss of the H-rich envelope of the AGB star than the outcome would be a close binary with a luminous H-poor He-burning star. Unfortunately not a single [WC] or PG 1159 central star is known to be a binary so this idea does not seem to be the major evolutionary canal for the H-poor PNNi.
In this paper we make an extensive analysis of the observational data for PNe with H-deficient central stars. We concentrate on a discussion and interpretation of diagrams and plots involving basic observational data such as PNN magnitude, nebular flux and PN dimensions. First, we investigate the hypothesis that the [WC] subclasses of the WR-type PNNi and other H-deficient (mostly PG 1159 type) central stars form an evolutionary sequence. Next, we compare our sample of H-deficient PNNi and their nebulae with the rest of the Galactic PN population. Then, the observed distributions on diagrams are confronted with predictions of theoretical evolutionary considerations within the born again AGB scenario. Our conclusion is that the very late He-shell flash cannot be the major evolutionary path leading to the formation of the H-deficient PNNi, at least of those of the [WC] type. We also discuss observational data for a few cases of PNe whose nuclei have very probably undergone a late He-shell flash. We show that it is very unlikely that these objects could give birth of [WC] PNNi. Thus the direct evolution from the AGB, presumably triggered by the last AGB He-shell flash, seems to be the only reasonable scenario for origin of the majority of the H-deficient PNNi.
© European Southern Observatory (ESO) 2000
Online publication: October 30, 2000