3 The N/O vs. O/H relation

Previous studies of galactic PN suggested the existence of an anti-correlation between N/O and O/H for type I objects, but the situation is still controversial when non-type I planetaries are considered. Here we revisit this question, using a homogeneous sample constituted of 74 PN (Costa & de Freitas Pacheco 1996) classified in different groups, according to chemical, morphological, kinematical and spatial criteria (Maciel & Dutra 1992). This classification, based on the original scheme proposed by Peimbert & Torres-Peimbert (1983), aggregates planetaries with progenitors of similar mass in the main sequence and, as we shall see, will play a major role in our understanding of the negative correlation between the N/O ratio and the oxygen abundance. For the sake of clarity, we recall in Table 5 the main properties of each class, valid only for galactic planetaries. Halo PN, classified as type IV objects, are not included in this table.

   

Table 5: Planetary Nebula classes

	Type I	Type IIa	Type IIb	Type III
population	young disk	disk	old disk	thick disk (?)
initial masses	2.5 - 8.0	1.8 - 2.5	1.2 - 1.8	0.95 - 1.2
$<Z_{{\rm kpc}}>$	0.15	0.28	0.42	0.66
mean age (Gyr)	0.3	1.4	4.3	10.2

Figure 1: N/O vs. O/H plot for galactic planetaries where only type I and non-type I objects are discriminated. Data are from Costa & de Freitas Pacheco (1996)

To analyse the data, we have first segregated the planetaries of our galactic sample in two main groups, type I and non-type I. Figure 1 shows these objects in the plane N/O vs. O/H. We note that type I objects (filled circles) show a net anti-correlation, but the remaining PN (open circles) display a scatter diagram. In Fig. 2 we have performed the same plot, but now taking into account the class membership. In order to avoid confusion and to strengthen our point, we plotted only type I (filled circles) and type IIb objects. In this plot we see that objects of the same class have a tendency to display a negative correlation, with planetaries having older progenitors being displaced towards lower oxygen abundances for a given N/O ratio. Type III PN follow a strip on the left of type IIb, whereas type IIa objects lie on an intermediate position between types I and IIa. A possible interpretation of these plots is that each class is characterized by progenitors within a given mass range (or within a given age interval) and the oxygen variation in each group reflects abundance gradients present in the galactic disk (Maciel & Köppen 1994). Under these conditions, the observed N/O - O/H anti-correlation indicates that the surface enrichment of the progenitor, as a consequence of a dredge-up episode, depends on the metallicity.

Figure 2: Same plot as Fig. 1, but for types I and IIb objects only

Figure 3: Same plot for LMC planetaries. Two groups are identified. Filled circles are probably young planetaries, while open circles correspond to objects 5 Gyr old

Guided by these results and using both data by de Freitas Pacheco et al. (1993a,b) and those compiled by Richer (1993), we were able to identify two main PN groups in the LMC. Figure 3 shows the N/O - O/H plot for LMC planetaries, where the different symbols distinguish objects belonging to these different groups. The anti-correlation is clearly detected, in agreement with the conclusions by Henry et al. (1989). The older population has a mean oxygen abundance of $<\varepsilon(\rm O)>$ = 7.87 and, using the evolutionary chemical model for the LMC by de Freitas Pacheco (1998), we found that the mean age of these planetaries is about 5 Gyr. This means that their progenitors have been formed just before the past enhanced star formation episode 3-4 Gyr ago. The second group is younger and should be a consequence of the star formation event extending over most of the past 3-4 Gyr, detected on color-magnitude diagrams (Westerlund et al. 1995; Vallenari et al. 1996; Gallagher et al. 1996; Ardeberg et al. 1997).

In Fig. 4 we plotted SMC data. Filled circles correspond to present data, while open circles correspond to literature data compiled by Richer (1993), excluding common objects. Again, the N/O - O/H anti-correlation is seen but we were not able to discriminate different age groups, in spite of the hint suggested by the observed dispersion of data. The bulk of the objects have abundances in the range 8.0 $< \varepsilon(\rm O) <$8.5 indicating ages less than 3 Gyr, according to our chemical models for the SMC, to be reported elsewhere. This is consistent with the fact that SMC planetaries have global mean abundances compatible with HII region values, as we have mentioned before. The only exception is SMP25. The rather high electron temperature, in agreement with the value found by Leisy & Dennefeld (1996), is a signature of a metal-poor nebula, consistent with our abundance determination. Our chemical evolutionary models suggest an age of about 12 Gyr for this object.

Figure 4: Same plot for SMC planetary nebulae. Present data are identified as filled circles

In the galactic case, the oxygen spread within a given class may be explained by an age spread as well as by abundance gradients existent in the disk. Probably this is also true for the LMC, since HII region data are consistent with a small gradient (Kolbulnick 1998). It is worth mentioning that such a gradient is not seen in F-supergiant stars (Hill et al. 1995). In the SMC case, no gradients were detected (Kolbulnick 1998) and abundance variations must be related to an age effect and/or to incomplete mixing.