The abundance distributions of Galactic bulge and disc planetary nebulae

Abstract

We present an abundance analysis of a sample of Galactic bulge planetary nebulae (GBPNe). The observational data set consists of spectra of 88 nebulae obtained with the FLAIR II multiobject spectrograph on the UK Schmidt Telescope, together with spectra of 42 nebulae obtained with the RGO Spectrograph on the Anglo-Australian Telescope. After selecting those for which reliable nebular electron temperatures could be derived, collisionally excited line abundances were derived for 45 GBPNe. These were then compared with similarly derived abundances for 54 Galactic disc PNe (GDPNe), taken from the work of Kingsburgh & Barlow. We find that within the errors the abundances have the same average values, essentially the same distributions – including that for the mass-sensitive N/O ratio – and show the same relationships. The width of the nitrogen distribution exceeds that arising from errors, which could be a consequence of the range of precursor masses. The ratio of Type-I to non-Type-I PNe in the bulge and disc samples is similar, 18 and 25 per cent respectively. For the GDPNe, we find larger N/H and N/O ratios for the small number of those with He/H > 0.14, compared with those with He/H < 0.14. For neither disc nor bulge sample is there any strong evidence for a depletion of oxygen for the higher-mass precursor stars (Type I PNe). We find no correlation between O/H and N/O or He/H. On the N/O to He/H plane, the bulge and disc PNe show a distribution whereby the low N/O values only occur for low He/H values, but at N/O > 0.25 the whole range of He/H values were sampled. The theoretical tracks to which we compare our data do not explain the PNe with low He/H abundance and high N/O ratio. Realistic uncertainties in collisionally excited lines (CEL) abundances for individual PNe are quite large, of the order of 40 per cent for oxygen. Large samples are therefore required to get good statistical accuracy. This is usually achieved by combining many studies, and so we have compared the results of a number of published studies with our own, to search for any systematic differences. The average abundances are found to be the same within the errors except for cases where the abundance derivation methods are dissimilar, where systematic differences can occur. The N/O ratio is especially sensitive to the details of the abundance derivations. Our bulge PN sample shows no evidence for either very-low-metallicity objects or for super-metal-rich objects – the implied mass and age distributions of the bulge PN precursor stars are indistinguishable within the observational errors from those in the local Galactic disc.