Annu. Rev. Astron. Astrophys. 1998. 36: 267-316 Copyright © 1998 by Annual Reviews. All rights reserved

1. INTRODUCTION

The Lyman forest is an absorption phenomenon in the spectra of background quasistellar objects (QSOs). It can be observed in the ultraviolet (UV) and optical wavelength range, from the local universe up to the highest redshifts where QSOs are found (currently z ~ 5). Neutral hydrogen intersected by the line of sight (LOS) to a QSO will cause absorption of the QSO continuum by the redshifted Ly (1215.67 Å) UV resonance line. In an expanding universe homogeneously filled with gas, the continuously redshifted Ly line will produce an absorption trough blueward of the QSO's Ly emission line (independent predictions by Gunn & Peterson 1965; Scheuer 1965; Shklovski 1965). Gunn & Peterson found such a spectral region of reduced flux, and used this measurement to put upper limits on the amount of intergalactic neutral hydrogen. The large cross-section for the Ly transition makes this technique by far the most sensitive method for detecting baryons at any redshift.

Bahcall & Salpeter (1965) suggested that there should also be a population of discrete absorption lines from a more clumpy gas distribution, specifically from intervening groups of galaxies. Discrete lines were observed shortly thereafter (Lynds & Stockton 1966; Burbidge et al 1966; Stockton & Lynds 1966; and Kinman 1966), but the quest for their precise origin has given rise to a long and, at times, controversial debate; only in recent years does the issue appear to have been resolved (see below). Soon thereafter, the simultaneous detection of higher order lines of the Lyman series (e.g., Baldwin et al. 1974) had confirmed the suggestion (Lynds 1970) that most of the absorption is indeed from HI Ly. At higher spectral resolution, the Ly forest can be resolved into hundreds (in z > 2 QSO spectra) of distinct absorption lines, the appearance of which gave rise to the label Ly forest (Weymann et al 1981); see Figure 1. A small fraction of the lines hidden in the forest are not caused by HI but belong to UV transitions from several common metal or heavy element ions (various ionization stages of C, O, Mg, Si, Fe and Al are most frequently seen). These metal lines are invariably associated with strong Ly lines. At column densities N(HI) exceeding 10¹⁷ cm^-2 the gas becomes optically thick to ionizing radiation and a discontinuity at the Lyman limit (912 Å) is detectable. In systems with N(HI) larger than ~ 10¹⁹ cm^-2, selfshielding renders the gas predominantly neutral. The damping wings of the Lorentzian component of the absorption profile beginn to be detected from about the same column density, reaching their maximum in the "damped Ly systems".

Figure 1. High resolution (FWHM 6.6 km s-1) spectrum of the zem = 3.62 QSO 1422+23 (V = 16.5), taken with the Keck HIRES (signal-to-noise ratio ~ 150 per resolution element, exposure time 25000 s). Data from Womble et al (1996).

The question of whether the majority of the absorption systems are truly intervening at cosmological distances from the quasar, or ejected by it, which had received considerable interest in the earlier days, is now settled in favor of the intervening hypothesis. The huge momentum requirements for ejection (Goldreich & Sargent 1976), the outcome of the Bahcall & Peebles test (1969; Young et al 1982a) for a random redshift distribution of absorbers to different QSOs, the discovery of galaxies at the same redshifts as metal absorption systems (Bergeron 1986), and the detection of high metallicity gas in systems close to the QSO and low metallicities more than 30000 km^-1 away from it (Petitjean et al 1994) leave no doubt that most of the systems are not physically related to the QSO against which they are observed.

The basic observational properties of the Ly forest were established in the late 1970s and early 1980s when the combination of 4m telescopes (e.g., the AAT, KPNO, MMT, Palomar) and sensitive photon counting electronic detectors (e.g. the University College London's IPCS) first permitted quantitative spectroscopy on high redshift QSOs to be performed. Making use of the new technology the work by Sargent et al. (1980) set the stage for what for many years has been the standard picture of the Ly forest: Ly absorbers were found to be consistent with a new class of astronomical objects, intergalactic gas clouds, which are distinct from galaxies (and metal absorption systems) by their large rate of incidence (dN/dz) and their weak clustering. Upper limits on the gas temperature and estimates for the ambient UV flux and for the cloud sizes were found to be consistent with a highly ionized (n_HI / n_H 10^-4) optically thin gas kept at a temperature T ~ 3 × 10⁴K by photoionization heating. Sargent et al (1980) suggested that denser clouds in pressure equilibrium with a hotter (ie. more tenuous) confining intercloud medium (ICM) could explain the apparent lack of change of these objects with time. It was argued that this picture matches the inferred cloud properties better than clouds held together by gravity, and there were a number of other appealing features. In the wake of the dark matter-based structure formation scenarios, the pressure confined clouds have given way to models where Ly clouds arise as a natural immediate consequence of gravitational collapse. These results are discussed later.

In an earlier review, Weymann et al (1981) introduced a classification of absorption systems that is still useful, although some of the distinctions introduced have been blurred by the most recent research (Tytler et al 1995; Cowie et al 1995). In particular, the earlier review distinguished two classes of absorption systems, physically separated from the QSO environment, according to whether they do, or do not, show metal absorption lines in addition to the ubiquituous Ly. For most of the Ly clouds detectable with current technology (N(HI) 10¹² cm^-2) metal lines with metallicities common at high redshifts (Z 10^-2 Z) are simply below the detection threshold. Therefore this classification is simply an observational one. Rather than explore the nature of the division, if appropriate, between metal absorbers and Ly systems here we will concentrate on the low column density absorbers. The study of metal absorption systems, possibly of great relevance to galaxy formation, is left for future review.

Below we first discuss observational techniques and observed properties of Ly systems (using the terms absorption systems, absorbers, clouds interchangeably). Then we turn to various models of Ly absorbers, and finally we address some recent results, and speculate about future developments.