1998A&A...332.1044P

The problem of the accuracy that can be attained in the derivation of chemical abundances in Planetary Nebulae (PNe) with the best techniques of the ``constant T<sub>e</sub>, n<sub>e</sub>'' type currently in use, has been considered. This has been done by constructing two sequences (A, B) of radiation-hydrodynamical PNe models under the framework of the interacting stellar-winds theory which are intended to represent real planetaries in different evolutionary stages. Both sequences evolve along the 0.605 M<sub>☉</sub> stellar evolutionary track of Bloecker (1995b), but start from different initial conditions. Corresponding equilibrium models were also computed for a number of specific models along the stellar track to obtain an estimate of the errors that would result from stationary PNe models. The line intensities calculated from these models are interpreted under the scheme of the ``constant T_e, n_e'' method and the derived elemental abundances are compared with the original input values. The obtained deviations vary for the different elements. In the optically thin cases the discrepancies between the abundances derived using the ``constant T<sub>e</sub>, n<sub>e</sub>'' method and the corresponding input values amount to less than 10% for helium and to maximum factors of 1.5 to 3 for oxygen, nitrogen, neon and argon. The discrepancy is higher for sulphur, reaching an order of magnitude in the most excited models. Particular attention has to be given to optically thick models. It is found that the ``constant T_e, n_e'' method can yield rather erroneous results in these cases. Only minor deviations are found between abundances calculated from the hydrodynamical models and those in equilibrium.