5.4.2. Abundance inhomogeneities

Many studies have suggested that structures seen in planetary nebulae (extended haloes, condensations) have different composition from the main nebular body, indicating that they are formed of material arising in distinct mass loss episodes characterized by different chemical compositions of the stellar winds. However, these differences in chemical composition may be spurious, due inadequacies of the adopted abundance determination scheme. For example, the knots and other small scale structures seen in PNe are possibly the result of instabilities or magnetic field shaping, and their spectroscopic signature could be due to a difference in the excitation conditions and not in the chemical composition. In the following, some examples of such studies are presented, adopting the view of their authors.

a) Extended haloes

NGC 6720, the "ring nebula" is surrounded by two haloes: an inner one, with petal-like morphology, and an outer one, perfectly circular, as seen in the pictures of Balick et al. (1992). Guerrero et al. (1997) have studied the chemical composition of these haloes, and found that the inner and outer halo seem to have same composition, suggesting a common origin: the red giant wind. On the other hand, the N/O ratio is larger in the main nebula by a factor of 2, indicating that the main nebula consists of superwind and the haloes of remnants of red giant wind.

NGC 6543, the "cat eye nebula" also shows two halo structures: an inner one, consisting of perfectly circular rings, and an outer one with flocculi attributed to instabilities (Balick et al. 1992). Unlike what is advocated for NGC 6720, the rings and the core in NGC 6543 seem to have same chemical composition (Balick et al. 2001). It must be noted however, that the abundances may not be reliable, since a photoionization model for the core of NGC 6543 predicts a far too high [O III] 4363/5007 (Hyung et al. 2000). Another puzzle is the information provided by Chandra. Chu et al. (2001) estimated that the abundances in the X-ray emitting gas are similar to those of the fast stellar wind and larger than the nebular ones. On the other hand, the temperature of the X-ray gas (~ 1.7 × 10⁶ K) is lower by two orders of magnitudes than the expected post shock temperature of the fast stellar wind. This would suggest that the X-ray emitting gas is dominated by nebular material. These findings are however based on a crude analysis and more detailed model fitting is necessary.

b) FLIERs and other microstructures

A large number of studies have been devoted to microstructures in PNe, and their nature is still debated. Fast Low Ionization Emission Regions (FLIERs) have first been considered to show an enhancement of N and were interpreted as being recently expelled from the star (Balick et al. 1994). However, Alexander & Balick (1997) realized that the use of traditional ionization correction factors may lead to specious abundances. Dopita (1997) made the point that enhancement of [N II] 6584 / H can be produced by shock compression and does not necessarily involve an increase of the nitrogen abundance. Gonçalves et al. (2001) have summarized data on the 50 PNe known to have low ionization structures (which they call LIS) and presented a detailed comparison of model predictions with the observational properties. They conclude that not all cases can be satisfactorily explained by existing models.

c) Cometary knots

The famous cometary knots of the Helix nebula NGC 7293 have been recently studied by O'Dell et al. (2000) using spectra and images obtained with the HST. The [N II] 6584 / H and [O III] 5007 / H ratios were shown to decrease with distance to the star. Two possible interpretations were offered. Either this could be the consequence of a larger electron temperature close to the star due to harder radiation field. Or the knots close to the star would be more metal-rich, in which case they could be relics of blobs ejected during the AGB stage rather than formed during PN evolution. Obviously, a more thorough discussion is needed, including a detailed modelling to reproduce the observations before any conclusion can be drawn.

d) Planetary nebulae with Wolf-Rayet central stars

About 8% of PNe possess a central star having Wolf-Rayet charateristics, with H-poor and C-rich atmospheres. The evolutionary status of these objects is still in question. A late helium flash giving rise to a "born-again" planetary nebula, following a scenario proposed by Iben et al. (1983), can explain only a small fraction of them. The majority appear to form an evolutionary sequence from late to early Wolf-Rayet types, starting from the AGB (Górny & Tylenda 2000, Peñ at et al. 2001). This seemed in contradiction with theory which predicted that departure from the AGB during a late thermal pulse does not produce H-deficient stars. Recently however, it has been shown that convective overshooting can produce a very efficient dredge up, and models including this process are now able to produce H-deficient post-ABG stars following a thermal pulse on the AGB (Herwig 2000, 2001, see also Blöcker et al. 2001). It still remains to explain why late type Wolf-Rayet central stars seem to have atmospheres richer in carbon than early type ones (Leuenhagen & Hamann 1998, Koesterke 2001). Also, one would expect the chemical composition of PNe with Wolf-Rayet central stars to be different from that of the rest of PNe. This does not seem to be the case, as found by Górny & Stasinska (1995), basing on a compilation of published abundances: PNe with Wolf-Rayet central stars are indistinguishable from other PNe in all respects except for their larger expansion velocities. Peña et al. (2001) obtained a homogeneous set of high spectral resolution optical spectra of about 30 PNe with Wolf-Rayet central stars and reached a similar conclusion, as far as He and N abundances are concerned. Their data did not allow to draw any conclusion as regards the C abundances.

e) H-poor PNe

There are only a few PNe which show the presence of material processed in the stellar interior. They are referred to as H-poor PNe, although the H-poor material is actually embedded in an H-rich tenuous envelope. The two best known cases are A 30 and A 78, whose knots are bright in [O III] 5007 and He II 4686 but in which Jacoby (1979) could not detect the presence of H Balmer lines. With deeper spectra, Jacoby & Ford (1983) estimated the He/H ratio to be ~ 8 in these two objects. Harrington & Feibelman (1984) obtained IUE spectra of a knot in A 30, and found that the high C/He abundance implied by C II 4267 is not apparent in the UV spectra, suggesting that the knot contains a cool C-rich core. Guerrero & Manchado (1996) obtained spectra of the diffuse nebular body of A30, showing it to be H-rich. A similar conclusion was obtained by Manchado et al. (1988) and Medina & Peña (2000) for the outer shell of A 78. However, quantitatively, the results obtained by these two sets of authors are quite different and a deeper analysis is called for.

Three other objects belong to this group: A 58, IRAS1514-5258 and IRAS 18333-2357, the PN in the globular cluster M22, already mentioned in Sect. 3.7.5.

One common characteristic of this class of objects is their extremely high dust to gas ratio, and the fact that the photoelectric effect on the grains provides an important (and sometimes dominant) contribution to the heating of the nebular gas (see Harrington 1996). This may lead to large point-to-point temperature variations (see Sect. 3.7.5) and strongly affect abundance determinations.

Harrington (1996) concludes his review on H-poor PNe by noting that the H-poor ejecta cannot be explained by merely taking material with typical nebular abundances and converting all H to He. There is additional enrichment of C, N, perhaps O, and most interestingly, of Ne. However, more work on these objects is needed - and under way (e.g. Harrington et al. 1997) - before the abundances can be considered reliable. Stellar atmosphere analysis of H-deficient central stars (e.g. Werner 2001) is providing complementary clues to the nature and evolution of these objects.

In conclusion, we have shown how nebulae can provide powerful tools to investigate the evolution of stars and to probe the chemical evolution of galaxies. Nevertheless, is necessary to keep in mind the uncertainties and biases involved in the process of nebular abundance derivation. These are not always easy to make out, especially for the non specialist. One of the aims of this review was to help in maintaining a critical eye on the numerous and outstanding achievements of nebular Astronomy.

ACKNOWLEDGMENTS: It is a pleasure to thank the organizers of the XIII Canary Islands Winterschool, and especially César Esteban, for having given me the opportunity to share my experience on abundance determinations in nebulae. I also wish to thank the participants, for their attention and friendship. I am grateful to Miriam Peña, Luc Jamet and Yuri Izotov for a detailed reading of this manuscript, to Daniel Schaerer for having provided useful information on stellar atmospheres and to André Escudero for having kindly computed a few quantities related to planetary nebulae in the galactic bulge. I would like to thank my collaborators and friends, especially Rosa González Delgado, Slawomir Górny, Claus Leitherer, Miriam Peña, Michael Richer, Daniel Schaerer, Laerte Sodré, Ryszard Szczerba and Romuald Tylenda for numerous and lively discussions in various parts of the World.

Finally, I would like acknowledge the possibility of a systematic use of the NASA ADS Astronomy Abstract Service during the preparation of these lectures.