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

1. INTRODUCTION

Quasistellar objects (QSOs or quasars) are valuable probes of the high-redshift universe (Schneider 1998). Their most distant representatives are now measurable out to redshifts of z ~ 5 (Schneider et al. 1991, Fan et al. 1999). In Big Bang cosmologies, these redshifts correspond to times when the universe itself was just ~ 1 gigayear (Gyr) old (see Figure 1).

Figure 1. Redshift versus age of the universe in Big Bang cosmologies. The three solid curves correspond to Ho = 65 km s-1 Mpc-1, = 0, and M = 0, 0.3, or 1. The dotted curve uses the same Ho but with = 0.7 and M = 0.3. The "error" bars show the range of ages possible for Ho between 50 and 80 km s-1 Mpc-1 (see Carroll et al. 1992).

Understanding the elemental abundances in these distant, early-epoch environments is a major goal of quasar research. Some of the first spectroscopic studies noted simply that quasars contain the usual array of "metals" (elements C, N, O and heavier) produced by stellar nucleosynthesis (Shklovskii 1965, Burbidge & Burbidge 1967). More quantitative estimates of the abundances came later from theoretical work on the broad emission lines, culminating in the important review by Davidson & Netzer (1979; also Baldwin & Netzer 1978, Shields 1976). Those studies inferred solar or slightly higher metal abundances, with large uncertainties. The past two decades have seen considerable progress. Today we have a better theoretical understanding of quasar environments, and greater abilities to both observe and model a range of abundance diagnostics.

We also have renewed motivation from the growing evidence that links quasars to galaxies. See, for example, Kormendy et al. (1998), Magorrian et al. (1998), Laor (1998) for black hole-host galaxy mass correlations; Chatzichristou et al. (1999), Hines et al. (1999), McLeod et al. (1999), Boyce et al. (1999), McLure et al. (1999), Aretxaga et al. (1998), Carballo et al. (1998), Bahcall et al. (1997), Miller et al. (1996), McLeod & Rieke (1995) for direct observations of QSO hosts; Cavaliere & Vittorini (1998), Shaver et al. (1998), Terlevich & Boyle (1993), Boyle & Terlevich (1998), Osmer (1998) for arguments based on QSO number-density evolution; McCarthy (1993), Saikia & Kulkarni (1998), Haas et al. (1998), Brotherton et al. (1998a) for radio galaxy-radio quasar unification schemes; and Turner (1991), Haehnelt & Rees (1993), Loeb & Rasio (1994), Katz et al. (1994), Haehnelt et al. (1998), Haiman & Loeb (1998), Taniguchi et al. (1999) for theoretical links between QSOs and galaxy evolution. If quasars reside, as expected, in galactic nuclei or dense protogalactic clumps, their abundances could yield unique constraints on the evolution of those environments. In particular, quasar abundances can indirectly probe the star formation that came before QSOs, possibly the first stars forming in massive collapsed structures after the Big Bang. Other studies, involving for example the "Lyman-break" objects (Steidel et al. 1998, Connolly et al. 1997) or the damped-Ly or Ly "forest" absorbers in QSO spectra (Pettini et al. 1997, Lu et al. 1998, Rauch 1998), probe galaxies and metal enrichment on much larger scales. The quasar results (on galactic nuclei) should therefore provide an important piece to the overall puzzle of high-redshift star formation and galaxy evolution.

Here we review the status and implications of quasar abundance work. We regret that many interesting related topics must be excluded; in particular, we will consider the quasars themselves to be simply light sources surrounded by emitting and absorbing gas. We discuss three abundance diagnostics that are readily observable in QSOs at all redshifts: the broad emission lines (BELs), the broad absorption lines (BALs), and the intrinsic narrow absorption lines (NALs). We include just these intrinsic spectral features to measure the abundances near QSO engines. We thereby exclude measures of more distant environments, such as the halos of the host galaxies, nearby cluster galaxies or cosmologically intervening material. If we could choose a maximum radius for the location of intrinsic gas, it would correspond to the size of the putative star clusters or galactic spheroids surrounding QSOs - perhaps a few hundred pc to a few kpc. Any material ejected from this region would also qualify as intrinsic. Ultimately, our interpretation of the abundance data depends critically on the location of the emitting/absorbing gas and its relationship to the quasar/host galaxy environment.

We begin with separate discussions of each abundance probe (Sections 2-3), followed by a summary of the overall results (Section 4). We then consider the plausible enrichment schemes, making a case for normal chemical evolution by stars in galactic nuclei (Section 5). Within that scheme, we use results from galactic studies (Section 6) to derive further implications of the QSO abundances (Section 7). We close with a brief outline for future work (Section 8).

In several sections below we present results of photoionization calculations performed with the numerical code Cloudy (version 90.05, Ferland et al. 1998). This code is freely available on the World Wide Web (http://www.pa.uky.edu/~gary/cloudy/). Finally, we define solar abundances based on the meteoritic results in Grevesse & Anders (1989).