Astron. Astrophys. 362, 921-936 (2000)

Astron. Astrophys. 362, 921-936 (2000)

5. Summary

Based on the successful chemical evolution model for the Milky Way disk developed by BP99, we calculated the corresponding abundance profile of elements up to the Fe-peak. For that purpose we used the metallicity dependent yields of WW95 for massive stars evolving at constant mass, along with those of Iwamoto et al. (1999) for SNIa. We also explored the yields of M92 for massive mass losing stars, concerning the elements He, C, N and O. We deliberately ignored yields from intermediate mass stars (IMS), our purpose being to check to what extent such stars are indeed required to explain observed abundance patterns. However, we did take into account ejecta from IMS, assuming that their net yield is zero for all elements; as explained in Goswami & Prantzos (2000), this procedure is crucial to ensure that the absolute metal abundances, expressed as X/H, are correctly evaluated.

We compiled a large sample of observational data, concerning abundance profiles of He, C, N, O, Ne, Mg, Al, Si, S and Ar in the Milky Way disk (Table 1). Most of the tracers are young objects (B-stars, HII regions, PNI), while PNII and PNIII abundances trace earlier stages of the Galaxy evolution (but with considerable uncertainties in the corresponding ages). Contrary to other authors, we did not consider data on Fe from open clusters, since the situation is rather uncertain at present (see Sect. 4.1). We simply note that our prescription for the rate of SNIa (major Fe producers), leads to a Fe abundance profile slightly steeper than the one of O in the disk (Fig. 4 and Fig. 5).

The main results of the model may be summarized as follows:

(i) We obtain abundance gradients for all elements from He to Zn. For primary elements, like O, the theoretical value of the abundance gradient at the present epoch (T = 13.5 Gyr) is: dlog(X/H)/dR [FORMULA] 0.06 dex kpc^-1. This value is compatible with most observational data concerning young objects in the disk (see Figs. 8 and Table 1). We note, however, the "puzzling" conclusion of Deharveng et al. (2000), pointing to an oxygen gradient smaller by about 40% than the commonly accepted value; if their conclusion is confirmed, our model parameters (the ratio of star formation to infall timescale as a function of galactocentric radius) should be appropriately modified. For secondary or "odd-Z" elements (like N and Al, respectively) we obtain slightly larger gradients; this is also the case for C when the M92 yields are used. Such values are marginally compatible with available observations. We find that the M92 yields can account for the totality of carbon production across the disk with no need for a contribution by IMS, in agreement with other recent studies (Prantzos et al. 1994, Carigi 1994, Gustafsson et al. 1999, Henry et al. 2000a, 2000b). On the contrary, the majority of N is produced by another source, most probably IMS.

(ii) Current observations show no trend of the abundance ratio X/O with galactocentric distance (or metallicity), for any element X. This is unexpected in the case of N and C, which do show such a trend in extragalactic HII regions where their abundance ratio to oxygen increases with metallicity (e.g. Henry & Worthey 1999). Our model shows a flat profile of S/O, Si/O and Ar/O (in agreement with observations) and a slowly decreasing ratio of Ne/O, Mg/O and Al/O with galactocentric distance (due to a "problematic" or overestimated metallicity dependence of the WW95 yields for those elements). They also show that C/O should decrease with galactocentric distance in the case of M92 yields (but this should not be the case if C is mainly produced as a primary by IMS). Clearly, precise observations of the C/O ratio across the disk are required in order to decide about its main production site. Similar conclusions hold for N: the M92 yields cannot account for its abundance profile, but is not clear whether primary or secondary production in IMS is the dominant mechanism. Observations of extragalactic HII regions suggest that the latter mechanism dominates in high metallicity regions (Henry et al. 2000b); if this is true, then an important N/O gradient should be observed in the Milky Way disk.

(iii) The evolution of abundance gradients provides a strong constraint for Galactic chemical evolution models. Despite a considerable amount of observational and theoretical work (e.g. Henry & Worthey 1999; Tosi 2000 and references therein), the question has not yet been settled. Observations of planetary nebulae (PN) of different ages could, in principle, help in that respect (Pasquali & Perinoto 1993, Maciel & Quireza 1999), but the ages and distances attributed to PN are not sufficiently well known at present. Our model predicts a steady flattening of the gradients with time, due to the adopted "inside-out" formation scheme for the disk. Such an evolution is also found in other works (e.g. Mollá et al. 1997, Portinari & Chiosi 1999) using similar assumptions. Our model makes also a testable prediction: the abundance scatter must be smaller in the inner disk than in the outer regions, if the observed dispersion between PN abundances at a given galactocentric distance is due to their intrinsic age differences.

In summary, we have shown that current massive star yields, combined with the BP99 chemical evolution model, reproduce fairly well most of the observed abundance gradients in the Milky Way disk. The assumptions of the model affect mostly points (i) and (iii) above, while the adopted yields affect point (ii). Some problems remain with C and N, and (to a smaller extent) Al. However, the major question of such studies, namely the evolution of abundance gradients, is far from being settled yet; a much more precise determination of ages, distances and abundances of planetary nebulae will be required for that.

Online publication: October 30, 2000
helpdesk@link.springer.de