Annu. Rev. Astron. Astrophys. 1999. 37: 487-531 Copyright © 1999 by Annual Reviews. All rights reserved

6. MORE INSIGHTS FROM GALACTIC CHEMICAL EVOLUTION

If we assume that QSO environments were indeed enriched by normal stellar populations, then we can use the results from galactic abundance and chemical-evolution studies to interpret the QSO data. Here we describe some relevant galactic results (see Wheeler et al. 1989 for a general review).

6.1. The Galactic Mass-Metallicity Relation

One important result from galaxy studies is the well-known mass-metallicity relationship among ellipticals and spiral bulges (Faber 1973, Faber et al. 1989, Bender et al. 1993, Zaritsky et al. 1994, Jablonka et al. 1996, Cozial et al. 1997). This relationship is attributed to the action of galactic winds; massive galaxies reach higher metallicities because they have deeper gravitational potentials and are better able to retain their gas against the building thermal pressures from supernovae (Larson 1974, Arimoto & Yoshii 1987, Franx & Illingworth 1990). Low-mass systems eject their gas before high Z's are attained. Quasar metallicities should be similarly tied to the gravitational binding energy of the local star-forming regions and, perhaps, to the total masses of their host galaxies (Section 7.1 below).

6.2. Specific Abundance Predictions

Another key result is the abundance behaviors of N and Fe relative to the elements such as O, Mg, and Si. Hamann & Ferland (1993b) constructed one-zone infall models of galactic chemical evolution to illustrate these behaviors in different environments. Figure 13 plots the results for two scenarios at opposite extremes. Both use the same nucleosynthetic yields, but the "Giant Elliptical" model has much faster evolution rates and a flatter IMF (more favorable to high-mass stars) compared with the "Solar Neighborhood" (or spiral disk) case. The Giant Elliptical evolves passively (without further star formation) after ~ 1 Gyr, because the gas is essentially exhausted. The parameters used in these calculations were based on standard galactic infall models (e.g. Arimoto & Yoshii 1987, Matteucci & Tornambé 1987, Matteucci & Francois 1989, Matteucci & Brocato 1990, Köppen & Arimoto 1990). However, the results are only illustrative and not intended to match entire galaxies. For example, evolution like the Giant Elliptical model might occur in just the central cores of extreme high-mass galaxies (cf Friaca & Terlevich 1998).

Figure 13. Logarithmic gas-phase abundance ratios normalized to solar for the two evolution models discussed in Section 6.2 of the text (adapted from Hamann & Ferland 1993b). Two scenarios for the N enrichment are shown (thin solid lines), one with secondary only and the other with secondary + primary (causing a plateau in N/O at low Z at early times). The thick solid curves represent O/H in both panels.

6.3. Fe/ as a Clock

At early times the abundance evolution is controlled by short-lived massive stars, mainly via Type-II supernovae (SN IIs). The elements, such as O and Mg, come almost exclusively from these objects, but Fe has a large delayed contribution from Type-Ia supernovae (SN Ias) - whose precursors are believed to be intermediate-mass stars in close binaries (Branch 1998). The predicted time delay is ~ 1 Gyr based on the IMF-weighted stellar lifetimes (Figure 13; Greggio & Renzini 1983, Matteucci & Greggio 1986). The actual delay is uncertain, but recent estimates are in the range of ~ 0.3 to 3 Gyr (Matteucci 1994, Yoshii et al. 1996, Yoshii et al. 1998). Because this delay does not depend on any of the global evolution time scales (e.g. the star formation rate, etc), Fe/ can serve as an absolute "clock" for constraining the ages of star-forming environments (Tinsley 1979, Thomas et al. 1998).

Observations of metal-poor galactic stars suggest that the baseline value of [Fe/] from SN IIs alone is nominally from -0.7 to -0.4 (Israelian et al. 1998, Nissen et al. 1994, King 1993, Gratton 1991, Magain 1989, Barbuy 1988; see also de Freitas Pacheco 1996), which is slightly larger than the prediction in Figure 13. The subsequent increase caused by SN Ias is a factor of a few or more. Note that the increase in Fe/ should be larger in rapidly evolving spheroidal systems because (a) by the time their SN Ia's turn on, there is relatively little gas left and each SN Ia has a greater effect; also (b) their rapid early star formation means that the SN Ia's occurring later are more nearly synchronized. The net result can be substantially super-solar Fe/ in the gas (even though Fe/ is sub-solar in most stars).

6.4. Nitrogen Abundances

Nitrogen also exhibits a delayed enhancement, although not on a fixed time scale like Fe/. Nitrogen's selective behavior is caused by secondary CNO nucleosynthesis, in which N forms out of pre-existing C and O. Studies of galactic HII regions indicate that secondary processing dominates at metallicities above ~ 0.2 Z, resulting in N/O scaling like O/H (or N Z²) in that regime. At lower metallicities, primary N can be more important based on an observed plateau in [N/O] at roughly -0.7 (see Tinsley 1980, Vila-Costas & Edmunds 1993, Thurston et al. 1996, Van Zee et al. 1998, Kobulnicky & Skillman 1998, Thuan et al. 1995, Izotov & Thuan 1999; but see also Garnett 1990, Lu et al. 1998). The models in Figure 13 show two N/O behaviors, for secondary only and secondary plus primary; the latter has a low-Z plateau forced to match the HII region data. Notice that the secondary growth in N/O can be shifted down considerably from the simple theoretical relation [N/O] = [O/H], e.g. in the Giant Elliptical case, because of the delays related to stellar lifetimes. We therefore have a strong prediction, based on both observations and these simulations, that measured values of [N/O] 0 imply Z Z - especially in quickly evolving spheroidal systems. This prediction was exploited above in the analysis of QSO BELs (Section 2.6; Shields 1976).