|Annu. Rev. Astron. Astrophys. 1998. 36:
Copyright © 1998 by . All rights reserved
5.3. Quasistellar Object Absorption Lines
Our knowledge of QSO absorption line systems has increased tremendously in the last decade (see the review by Rauch 1998, in this volume). Structure formation simulations have played an important role in elucidating the nature of the Ly forest of narrow hydrogen absorption lines, which are seen in spectra of QSOs at redshift z > 2. Early models of the forest were based on isolated spherical or sheet-like clouds. About a decade ago, hierarchical clustering models like CDM ones were realized to have small-scale structure that might readily produce clouds with abundance (Bond et al 1988) and clustering (Salmon & Hogan 1986) comparable to the Ly forest lines. Another key was the realization by McGill (1990) that because of peculiar velocities, optically thin line profiles do not necessarily reflect the density profile of neutral hydrogen in space. Lines can form from velocity caustics (e.g. a structure that has just started to reverse the Hubble expansion) even if the gas is not particularly overdense.
Cosmological simulations of dark matter and gas with photoionization from the ultraviolet background (Cen et al 1994b, Zhang et al 1995, Hernquist et al 1996, Mücket et al 1996) have shown that the Ly forest arises naturally from the filamentary web (Bond et al 1996) of structure that occurs in hierarchical clustering models with an appropriate amount of small-scale power. When the ratio of ionizing flux to baryon density is set near the observationally favored value, hierarchical models almost automatically predict the correct density distribution, redshift evolution, and clustering of the Ly lines. Some of the lines form in well-defined clouds (particularly the damped lines; Katz et al 1996b), while others form in transient filamentary or sheet-like structures (Cen & Simcoe 1997), and still others are velocity caustics that may even be underdense in real space (Zhang et al 1995). Figure 4 shows the impressive match of simulated and measured column density distributions for the hydrogen Ly line at z = 3, along with a prediction for ionized helium.
Figure 4. Simulated column density distribution of Ly absorption lines (filled circles for H I, diamonds for He II) at redshift z = 3 in a standard CDM universe. Observational data are shown for H I (open symbols). From Zhang et al (1997).
The success of numerical simulations has inspired analytical models that can account very well for the column density distribution and provide an understanding of how it arises (Bi et al 1995, Gnedin & Hui 1996, Bi & Davidsen 1997, Hui et al 1997). The agreement of analytical and numerical models with each other and with observational data is a remarkable success story that supports the hierarchical clustering models of structure formation. These results also show that the concept of a uniform medium producing continuous absorption, used by Gunn & Peterson (1965) to sharply limit the neutral hydrogen density in the intergalactic medium, must be replaced by absorption arising in a fluctuating medium (Reisenegger & Miralda-Escudé 1995, Rauch et al 1997b). There is, however, predicted to be a Gunn-Peterson trough for He II absorption (Miralda-Escudé et al 1996, Croft et al 1997, Zhang et al 1997). Requiring the Ly forest to produce the observed overall opacity constrains the baryon abundance to b > 0.017 h-2 (Rauch et al 1997b, Weinberg et al 1997b), which is consistent with the high value obtained by Tytler et al (1996).
The computational modeling has been extended to include the physics of metal lines with a sophisticated treatment of photoionization equilibrium (Haehnelt et al 1996a, b, Hellsten et al 1997, Rauch et al 1997a, b), Voigt-profile fitting to the simulated absorption-line spectra for rigorous comparison with observations (Davé et al 1997c), and the examination of correlations between close lines of sight (Charlton et al 1997).
The high column density-damped Ly lines are thought to originate in dense gas associated with galaxies. Their abundance therefore can be used to constrain models of structure formation (e.g. Gardner et al 1997 and references therein).