Intrinsic absorption lines provide independent probes of the elemental abundances near QSOs that can test and extend the emission-line results. Intrinsic absorbers include the BALs, at least some of the ``associated'' NALs (with similar absorption and emission redshifts, i.e. za ze ), and any other systems that form in (or were ejected from) the vicinity of the QSO engine. Figures 4 and 5 show examples of QSO spectra containing BALs and NALs, respectively.
Figure 4. Spectrum of the BALQSO PG 1254+047 (ze = 1.01) with emission lines labeled across the top and possible BALs marked at redshifts corresponding to the 3 deepest minima in CIV. Not all of the labeled BALs are present. The smooth dotted curve is a continuum fit extrapolated to short wavelengths (from Hamann 1998).
BALs, with their broad troughs and maximum velocity extents often exceeding 10,000 km s-1, clearly form in high-velocity outflows from the QSOs. NALs might form in a variety of environments, ranging from QSO winds like the BALs to cosmologically intervening gas like galactic halos. Each narrow-line system must be examined individually. Several indicators of intrinsic absorption have been developed, including (1) time-variable line strengths, (2) line multiplet ratios that imply partial line-of-sight coverage of the background light source(s), (3) high space densities inferred from excited-state fine-structure lines, and (4) line profiles that are broad and smooth compared to thermal line widths (see Hamann et al. 1997b, Barlow & Sargent 1997 and references therein).
Figure 5. Normalized spectra of the QSO UM 675 (ze 2.15) show that its za ze absorption lines (at za = 2.134) varied between two observations. The time variability, together with the relatively broad line profiles and partial line-of-sight coverage revealed by other higher-resolution data, imply that this NAL system is intrinsic to the QSO.
One well-studied example of an intrinsic narrow-line system is in the QSO, UM 675 (Hamann et al. 1995b, 1997b). Figure 5 shows that this system varied between two observations. At higher spectral resolution (~ 9 km s-1 ) the line profiles appear much broader than the thermal speeds (with full widths at half minimum of ~ 470 km s-1 ) and the resolved CIV and NV line troughs appear too shallow for the optical depths required by their doublet ratios. The troughs are evidently filled-in by unabsorbed flux. This filling-in probably results from partial coverage of the background light source(s) (see HF99 for a sketch of possible partial-coverage geometries). The coverage fractions in UM 675 are ~ 50% for CIV and NV and > 85% for HI. The variability time scale implies that the absorber is not more than 1 kpc from the central continuum source, and very likely much nearer. The diverse absorption lines detected in UM 675 (from CIII 977 and NIII 990 to OVI 1034 and NeVIII 774) imply a range of ionization states, consistent with a factor of 100 range in density or 10 in distance from the ionizing continuum source (see Hamann et al. 1997b).