|Annu. Rev. Astron. Astrophys. 1999. 37:
Copyright © 1999 by . All rights reserved
3.1. Overview: Types of Absorption Lines
Quasar absorption lines can form in a variety of locations, from near the QSO engine (which we call "intrinsic," Section 1) to intervening gas at cosmologically significant distances. We exclude from our discussion the damped-Ly absorbers and the "forest" of many narrow Ly systems with weak or absent metal lines, because they form in cosmologically intervening gas (Rauch 1998). The remaining metal line systems can be divided into two classes by their broad or narrow profiles. This division is a gross simplification, but still useful because it distinguishes the clearly intrinsic broad lines from the many others of uncertain origin. Here we briefly characterize the two (broad and narrow) line types.
3.1.1 Broad Absorption Lines (BALs)
Broad absorption lines are blueshifted relative to the emission lines and have velocity widths of at least a few thousand kilometers per second (for example, Figure 8). They appear in 10%-15% of optically selected QSOs and clearly identify high-velocity winds from the central engines. The precise location of the absorbing gas is unknown, but there is little doubt that it is intrinsic - originating within at least a few tens of parsecs from the QSOs. See recent work by Weymann et al. (1991), Barlow et al. (1992), Korista et al. (1993), Hamann et al. (1993), Voit et al. (1993), Murray et al. (1995), Arav (1996), Turnshek et al. (1997), Brotherton et al. (1998b) and the reviews by Turnshek (1988, 1994), Weymann et al. (1985), Weymann (1994, 1997).
Figure 8. Spectrum of the BALQSO PG 1254+047 (emission redshift ze = 1.01) with emission lines labeled across the top and possible BALs marked below at redshifts corresponding to the three deepest minima in the CIV trough. Not all of the labeled lines are detected. The smooth dotted curve is a power-law continuum fit extrapolated to short wavelengths (from Hamann 1998).
3.1.2. Narrow Absorption Lines (NALs)
A practical definition of NALs would limit their full widths at half minimum (FWHMs) to less than the velocity separation of important doublets (e.g. < 500 km s-1 for CIV, < 1930 km s-1 for SiIV, or < 960 km s-1 for NV), because it is our ability to resolve these doublets that makes their analysis fundamentally different from the BALs (Section 3.2.2 below).
NALs can form in a variety of locations, ranging from very near QSOs, as in ejecta like the BALs, to unrelated gas or galaxies at cosmological distances (Weymann et al. 1979). It is not yet known what fraction of NALs at any velocity shift meet our definition of intrinsic (Section 1). Several studies have noted a statistical excess of NALs within a few thousand kilometers per second of the emission redshifts. These are the so-called "associated" or za ze absorbers (with redshifts close to the emission redshift; Weymann et al. 1979, 1981;, Young et al. 1982;, Foltz et al. 1986, 1988;, Anderson et al. 1987). Their strengths and frequency of occurrence appear to correlate with the QSO luminosities or radio properties, suggesting some physical relationship (see also Möller et al. 1994, Aldcroft et al. 1994, Wills et al. 1995, Barthel et al. 1997). These correlations may extend to NALs at blueshifts of 30,000 km s-1 (Richards et al. 1999). Nonetheless, we might expect a larger fraction of intrinsic NALs nearer the emission redshift and, if they are ejected from QSOs, they should appear at za < ze rather than za > ze.
Several tests have been developed to help identify intrinsic NALs, including (a) time-variable line strengths, (b) multiplet ratios that imply partial line-of-sight coverage of the background light source(s), (c) high gas densities inferred from excited-state absorption lines, and (d) well-resolved line profiles that are smooth and broad compared with both thermal line widths and the velocity dispersions expected in intervening clouds (e.g. Bahcall et al. 1967, Williams et al. 1975, Young et al. 1982, Barlow & Sargent 1997, Hamann et al. 1997b, Hamann et al. 1997c, Petitjean & Srianand 1999, Ganguly et al. 1999, and references therein). These criteria might not be definitive individually, but they sometimes appear in combination.
Figure 9 shows a za ze NAL system that is clearly intrinsic based on time-variable line strengths, partial line-of-sight coverage, and relatively broad profiles. High metallicities might be another indicator of intrinsic absorption (Section 3.4 below), but that criterion would bias abundance studies; we would like to determine the intrinsic versus intervening nature independently of the abundances. The other (nonabundance) tests indicate that bona fide intrinsic NALs can have velocity shifts out to 24,000 km s-1 and a wide range of FWHMs down to 30 km s-1. See the references above and the reviews of za ze systems by Weymann et al. (1981), Foltz et al. (1988).
Figure 9. Spectra of the za ze absorber in UM675 (ze = 2.15) showing its time-variability in two epochs (top panel) and broad, smooth profiles at higher spectral resolution (9 km s-1, bottom panels). From Hamann et al. (1995, 1997b).