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

Next Contents Previous


Several scenarios have been proposed for the production of heavy elements near QSOs, including (a) the normal evolution of stellar populations in galactic nuclei (Hamann & Ferland 1992, 1993b), (b) central star clusters with enhanced supernova (and perhaps nova) rates caused by mass accreted onto stars as they plunge through QSO accretion disks (Artymowicz et al. 1993), (c) star formation inside QSO accretion disks (Silk & Rees 1998, Collin & Zahn 1999), and (d) nucleosynthesis without stars inside accretion disks (Jin et al. 1989, Kundt 1996).

5.1. Occam's Razor: the Case for Normal Galactic Chemical Evolution

The first scenario listed above, for normal galactic chemical evolution, is most compelling because (a) it is the only one of these processes known to occur and (b) it is sufficient to explain the QSO data. In particular, the stars in the centers of massive galaxies today are (mostly) old and metal rich (Bica et al. 1988, 1990; Gorgas et al. 1990; Bruzual et al. 1997; Vazdekis et al. 1997; Jablonka et al. 1992; Jablonka et al. 1996; Feltzing & Gilmore 1998; Worthey et al. 1992; Kuntschner & Davies 1997; Sansom & Proctor 1998; Ortolani et al. 1996; Sil'chenko et al. 1998;, Idiart et al. 1996; Fisher et al. 1995; Bressan et al. 1996). The exact ages are uncertain, but there is growing evidence for most of the star formation in massive spheroids (ellipticals and the bulges of large spiral galaxies) occurring at redshifts z gtapprox 2-3, especially (but not only) for galaxies in clusters (see also Renzini 1997, 1998; Bernardi et al. 1999; Bruzual & Magris 1997; Ellis et al. 1997; Tantalo et al. 1998; Ivison et al. 1998; Kodama & Arimoto 1997; Ziegler & Bender 1997; Kauffmann 1996; Van Dokkum et al. 1998;, Mushotzky & Loewenstein 1997; Spinard et al. 1997; Stanford et al. 1998;, Heap et al. 1998; Barger et al. 1998). The star-forming (Lyman break or Lyalpha emission) objects measured directly at z gtapprox 3 might be galactic or protogalactic nuclei in the throes of rapid evolution (Friaca & Terlevich 1999; Baugh et al. 1998; Steidel et al. 1998, 1999; Connolly et al. 1997; Lowenthal et al. 1997;, Trager et al. 1997; Hu et al. 1998; Franx et al. 1997; Madau et al. 1996; Giavalisco et al. 1996). These objects are more numerous than QSOs and some have been measured at z > 5 (Dey et al. 1998, Hu et al. 1998, Weymann et al. 1998), beyond the highest known QSO redshift of z approx 5.0. On the theoretical side, recent cosmic-structure simulations show that protogalactic condensations can form stars and reach solar or higher metallicities at z gtapprox 6 (Gnedin & Ostriker 1997).

These studies all suggest that there was considerable star formation at epochs preceding, or concurrent with, the QSOs. Quasars might form in the most massive and most dense of the early-epoch star-forming environments (Turner 1991, Loeb 1993, Haehnelt & Rees 1993, Miralda-Escude & Rees 1997, Haehnelt et al. 1998, Spaans & Corollo 1997). They might also form preferentially in globally dense cluster environments, based on the higher detection rates of star-forming galaxies near high-z QSOs (Djorgovski 1998).

The gas in these environments might have been long ago ejected via galactic winds, consumed by central black holes, or diluted by subsequent gaseous infall, but its signature remains in the old stars today. The mean stellar metallicities (4) in the cores of massive low-redshift galaxies are typically < Zstars> ~ 1-3 Zodot (see references to metallicities listed above). [It is worth noting here that, because of a significant time-delay in the iron enrichment, O/H and Mg/H are better measures of the overall metallicity than Fe/H (see Section 6 and Wheeler et al. 1989)]. Individual stars are distributed about the mean with metallicities reflecting the gas-phase abundance at the time of their formation. If the interstellar gas is well mixed and the abundances grow monotonically [as expected in simple enrichment schemes (Section 6)], the gas-phase metallicity, Zgas, will always exceed <Zstars>. Only the most recently formed stars will have metallicities as high as the gas. Therefore, the most metal-rich stars today should reveal the gas-phase abundances near the end of the last major star-forming epoch.

In the bulge of our own Galaxy, the nominal value of <Zstars> is 1 Zodot and the tail of the distribution reaches Zstars gtapprox 3 Zodot, with even higher values obtaining near the Galactic center (Rich 1988, 1990;, Geisler & Friel 1992;, McWilliam & Rich 1994;, Minniti et al. 1995;, Tiede et al. 1995;, Terndrup et al. 1995;, Idiart et al. 1996;, Castro et al. 1996;, Bruzual et al. 1997). The gas-phase metallicity should therefore have been Zgas gtapprox 3 Zodot after most of the bulge star formation occurred. Simple chemical evolution models indicate more generally that Zgas should be ~ 2 to three times <Zstars> in spheroidal systems like galactic nuclei (Searle & Zinn 1978, Tinsley 1980, Rich 1990, Edmunds 1992, de Fretas Pacheco 1996). Thus the observations of <Zstars> ~ 1-3 Zodot suggest that gas with Zgas ~ 2-9 Zodot once existed in these environments.

We might therefore expect to find 2 ltapprox Z ltapprox 9 Zodot in QSOs, as long as most of the local star formation occurred before the QSOs "turned on" or became observable. These expectations are consistent with the QSO abundance estimates reported above (Section 4). More exotic enrichment schemes are therefore not needed to explain the QSO data.

4 It is worth noting here that, because of a significant time-delay in the iron enrichment, O/H and Mg/H are better measures of the overall "metallicity" than Fe/H (see Section 6 and Wheeler et al. 1989). Back.

Next Contents Previous