Results and discussion

Many of the models described in Section 1 adopt nucleosynthetic yields from Woosley & Weaver (1995) or one of its predecessors. However, as discussed by Gibson et al. (1997) in a different context (clusters of galaxies), while stellar nucleosynthesis models may have identical global metal (i.e. Z) yields, the relative distribution of elements and isotopes therein may be quite different (driven by differences in the treatment of reaction rates, mass loss, convection, etc.). Figure 45.1, adapted from Gibson (1995), provides a graphic demonstration for the important a-element pair of oxygen and magnesium - a factor of 5-10 difference in O/Mg exists, for example, between Woosley & Weaver (1995) and Arnett (1991), at Solar metallicity, in the mass range 15 < m/MQ < 25.

The shaded region highlights an interesting puzzle - as hinted at already in the seminal work of McWilliam & Rich (1994) and confirmed recently by Lecureur et al. (2007), O/Mg in the bulge appears to be a factor of two lower than that of the Sun, over the metallicity range -0.5 < [Fe/H] < + 0.5 (i.e. a range spanning the bulk of the stars in the bulge). As discussed by Gibson (1995), such subsolar [O/Mg] values are essentially impossible to recover with any chemical-evolution model employing the Woosley & Weaver (1995) yields, because not a single model in the grid lies within the shaded region (thus, no IMF that would lead to a model matching the data could be constructed a posteriori); the situation does not appear particularly tenable with the Thielemann et al. (1996) models either. What is interesting from Figure 45.1, though, is the location of the little-used Arnett (1991) yields in this particular plane - specifically, a natural byproduct of the models is the increased production of magnesium in the mass range 15 < m/MQ < 25, shifting those models into the abundance-pattern regime populated by the stars of the Galactic bulge; (this is a not entirely surprising consequence, in the light of Gibson (1997, Figure 3).

Using the Arnett (1991; henceforth A91), Woosley & Weaver (1995, henceforth WW95) and Thielemann et al. (1996; henceforth TNH96) Type-II supernova yields, we have constructed representative models of the Galactic bulge using the one-zone infall precursor analogue (Gibson & Matteucci 1997) to GEtool (Fenner & Gibson 2003). Our fiducial bulge model is patterned after the (x = 0.95, k = 20 Gyr-1, n = 0.1 Gyr) model of Matteucci et al. (1999), the primary differences being (i) the reduction of the upper mass limit of the IMF from 100M© to 35M© and (ii) the inclusion of three different Type-II supernova yield options. Note, in keeping with our philosophy, we do not perform any a posteriori normalisation of our models. As in Koppen & Arimoto (1989, 1990), we have stopped the simulations at 1 Gyr, although this does not alter the general thrust of our conclusions.

The predicted chemical evolution of this fiducial model, using the three different yield sources, is shown in Figure 45.2. It should come as little surprise (in the light of the discussion surrounding Figure 45.1) that only the model incorporating Arnett's (1991) yields successfully predicts the bulk of the bulge stars to have -0.4 < [O/Mg] < -0.2. Conversely, Matteucci etal. (1999, Figure 3) and Molla etal. (2000, Figure 3) predicted [O/Mg] « +0.05 ± 0.05 for the stars of the Galactic bulge. It should be stressed that this was a natural prediction of Arnett's yields, and required no a posteriori rescaling of the magnesium abundances, as is normally done when employing the method of Woosley & Weaver (1995).

Having said all this, it would be foolish to suggest that this is definitive proof in favour of the Arnett (1991) compilation; all of the caveats noted in Gibson et al. (1997) regarding their input physics remain valid today. More importantly, that which might ameliorate one important abundance-ratio problem in the bulge may also lead to irreparable consequences for other patterns, or which is more likely, problems for the Solar neighbourhood (although a cynic could turn the problem around and say that having to resort to a different IMF for the bulge, as opposed to the Solar neighbourhood, is not necessarily 'better'). Indeed, we suspect that a detailed accounting of all relevant observables will suggest that the A91 yields are

Figure 45.2. Predicted chemical-evolution trajectories in the [O/Mg]-[Fe/H] plane, for Type-Ia supernovae, under the assumption of a massive-star (mu = 35M0 )-biased initial mass function, x = 0.95, rapid infall (on a free-fall timescale) of primordial gas, and the three Type-II supernova yield compilations shown in Figure 45.1: WW95, Woosley & Weaver (1995); TNH96, Thielmann etal. (1996); and A91, Arnett (1991). As in Figure 45.1, the shaded region corresponds to the subsolar [O/Mg] range encountered in the Galactic bulge.

Figure 45.2. Predicted chemical-evolution trajectories in the [O/Mg]-[Fe/H] plane, for Type-Ia supernovae, under the assumption of a massive-star (mu = 35M0 )-biased initial mass function, x = 0.95, rapid infall (on a free-fall timescale) of primordial gas, and the three Type-II supernova yield compilations shown in Figure 45.1: WW95, Woosley & Weaver (1995); TNH96, Thielmann etal. (1996); and A91, Arnett (1991). As in Figure 45.1, the shaded region corresponds to the subsolar [O/Mg] range encountered in the Galactic bulge.

not a panacea for the chemical evolution of the bulge, but our goal here was not to prove (or disprove) that statement, but simply to remind the end-user of such yield tables that, while WW95 is an extraordinarily beautiful suite of models, one should be cautious in assuming that their use constitutes the elimination of nucleosynthesis as a significant systemtic uncertainty in models of galactic chemical evolution!

Was this article helpful?

0 0

Post a comment