5.4. PNe probe the nucleosynthesis in their progenitor stars
5.4.1. Global abundance ratios
It is clear from the diagrams presented by Henry et al. (2000) that PNe show significantly higher values of He/H, N/O, C/O than H II regions of the same O/H. This indicates that He, N and C have been synthesized in PNe progenitors, as theory predicts. More quantitative comparison with theory is difficult because of the number of determining parameters (stellar mass, parametrization of the mixing processes) and of complex selection effects. In the following we draw a few examples of more detailed interpretations of abundance ratios that have been proposed.
The nature of Type I PNe is a good example of the difficulty in the interpretation. Peimbert (1978) had defined type I PNe as objects having He/H > 0.125 and N/O > 0.5. Kaler et al. (1978) interpreted the high N/O and He/H together with the (He/H , N/O) correlation observed in such objects as due to second dredge up, implying initial stellar masses larger than 3 M. Later, the He/H criterion to define Type I PNe was abandoned (it must be noted that old determinations of He/H did not include proper correction for collisional excitation of He lines). Henry (1990) found that Type I PNe showed an (N/O, O/H) anticorrelation and concluded that in these objects N is produced at the expense of O (due to ON cycling). Kingsburgh & Barlow (1994) contested the existence of such an anticorrelation and propose a new definition of Type I PNe, as being PNe that underwent envelope burning conversion to N of dredged up primary C. Thus they are objects in which the present N/H is larger than the initial (C+N)/H (equal to 0.8 in the solar vicinity). Costa et al. (2000) on the contrary define Type I PNe using only the criterion He/H > 0.11. They find a (N/O, O/H) anticorrelation when PNe are segregated by types. They interpret this by saying that the oxygen abundance is not modified by the PN progenitor but reflects the metallicity of the site where the progenitor was born, and that dredge-up is more efficient at low metallicity. It must be noted that, whatever the definition, there is actually no clear dichotomy between Type I and other PNe, the distribution of the N/O ratios is rather continuous (and this is also what is predicted at least at solar and half solar metallicities from the models of Marigo 2001).
Concerning carbon, (C+N+O)/H is found to increase with C/H and becomes dominated by C/H for the most carbon rich objects. This is seen both in Galactic samples (Kingsburgh & Barlow 1994) and in Magellanic Clouds samples (Leisy & Dennefeld 1996). This is in agreement with a scenario where carbon is produced by 3- from He and brought to the surface by third dredge up. From the number of PNe with observed C enhancement, one concludes that third dredge up is common in PNe progenitors. Among the PNe in which the carbon abundance could be determined, about 40% (in the Galactic sample) and 70% (in the Magellanic Clouds sample) have C/O > 1. This is well in line with theoretical predictions that third dredge up is more efficient at low metallicity. Note that PNe with C/O > 1, the so-called carbon-rich PNe, are likely to contain carbon rich dust, since their progenitors must have developed a carbon chemistry to form grains in their atmospheres.
More detailed comparisons of PNe results with the predictions of post AGB models have been attempted by Henry et al. (2000) and Marigo (2001). Interpretations are difficult, due to the number of parameters involved and to the difficulty to derive accurate central star masses and to relate them to initial masses.
Péquignot et al. (2000) discuss two PNe in the Sgr B2 galaxy, He 2-436 and Wray 16-423, whose nuclei are interpreted as belonging to the same evolutionary track. The authors perform a differential analysis of these two PNe, based on tailored photoionization modelling, and argue that while systematic errors may substantially shift the derived abundances, the conclusions based on differences between the two models should not be influenced. The main conclusion is that third dredge up O enrichment is observed in He 2-436, at the 10% level.