4. Concluding remarks
Using a self-consistent method based on a simple model of PN evolution, we have derived central star masses for a sample of 125 objects in which we could use the following observables: nebular H flux, angular radius and expansion velocity, and stellar visual magnitude. Note that our approach does not require a knowledge of the distances to the objects. Our method can easily be generalized in the future to more realistic morpho-dynamical models for the evolution of PN.
We have shown how it is possible to empirically convert a distribution of central star masses of observed PN into a distribution of masses of zero-age post-AGB stars.
However, we also emphasized that the derived central star mass distribution depends on the - unknown - total nebular mass. Note that, even knowing the PN distances by independent means, such as expansion parallaxes (Hajian et al. 1995), would not suppress the uncertainty in the central star mass , which is linked to the fact that we do not know by how much a given PN is density bounded. Actually, this uncertainty is present in all the previously published determinations of central star masses that rely on nebular properties, although not explicitely stated. It is only by using more realistic models for the PN evolution, together with additional observables that one could perhaps improve the situation.
Actually, the only way to obtain accurate central star masses would be by model fitting of high resolution spectra of these stars. This procedure, however, has its own uncertainties (see the different results obtained by Méndez et al. 1988 and 1992 for the same objects). Moreover, it is restricted to a small number of objects, as it is so demanding on the observational material as well as on the modelling procedure.
We find that in our sample more than 80% of the central stars have masses between 0.55 and 0.65M , the true distribution being probably narrower. The median value of is ill defined, due to our absence of knowledge of the total nebular mass, which enters in the determination of with our method. Of course, the values derived for the central star masses strongly depend on the assumed stellar evolution. Our determinations were made using the H-burning evolutionary tracks of Blöcker (1995) and Schönberner (1983).
Interpretation of the difference between the PNN mass distribution and that of related objects like white dwarfs requires a detailed galactic evolution modelling combined with a simulation of selection effects, and was not attempted here.
© European Southern Observatory (ESO) 1997
Online publication: April 6, 1998