QUASISTELLAR OBJECTS, ABSORPTION LINES BRADLEY M. PETERSON Because of their high luminosities, quasars are detectable at very large distances. Except for a handful of extraordinarily bright emission line galaxies, they are the only luminous objects that have been detected at redshifts ***1/2. Quasars thus provide one of the few direct probes of the history of the universe. Light from distant quasars that is received at the Earth now was emitted when the universe was only a fraction of its current age, and it has been traveling through the universe ever since. As light from these distant sources travels to us, there is a reasonable chance that it may be intercepted by an intervening galaxy or intergalactic gas cloud along the way. The signature of such an occurrence will be an absorption spectrum, produced by the intervening gas, superposed on the quasar spectrum (continuum plus broad emission lines). The absorption lines will be at a lower redshift than the emission lines of the quasar itself because the intervening object is closer to us and has a lower cosmological recession velocity. Quasars can thus be used simply as background sources against which we can observe less luminous objects which themselves may be at very large distances or "look-back" times. A fundamental difficulty in using absorption lines to study cosmological evolution is that one does not know precisely what kind of object is producing the absorption lines. Indeed, one cannot always clearly distinguish between truly intervening gas (in which the physical properties of the gas have nothing to do with the background quasar) and absorbing gas in the immediate vicinity of the quasar. This point historically has been a source of great controversy, although most researchers now agree that only a minority of the observed absorption features can be attributed to material associated with the quasars themselves. The strongest absorption lines seen in quasars are generally hydrogen Ly-* *1216, C ** **1548, 1550, and Mg ** **2795, 2802. Whether or not a specific line is detected in any absorption system depends on the redshift of the absorber, which determines whether the redshifted line will lie in the observable spectral window of a given detector, as well as on the physical parameters of the absorbing cloud (column density, elemental abundances, state of ionization, and velocity dispersion). The absorption lines seen in quasar spectra generally can be placed into three categories: (1) heavy-element systems, (2) Ly-* forest systems, and (3) broad absorption line (BAL) systems. In the first two types, the absorption lines are so narrow that they are unresolved in optical spectra. The heavy-element systems are characterized by lines of ionized and neutral metals (meaning elements heavier than helium), and, of course, Ly-* when it is accessible. Neutral hydrogen column densities in these systems are typically 10**-10** cm**, and metals are probably slightly underabundant relative to solar values. The velocity dispersions determined from curve-of-growth analyses are typically tens of kilometers per second for these systems, which is much too large to be attributed to purely thermal motions in gas that is not highly ionized. A high-redshift quasar may have up to a few such systems at different absorption redshifts in the observed spectral window. Important subclasses of such systems are the damped Ly-* systems and the Lyman-limit systems, both of which are characterized by very high column densities. This is shown in the former case by the damping wings on the Ly-* absorption line, and in the latter case by the high optical depth in the Lymann continuum, which results in an abrupt disappearance of the quasar continuum at wavelenghts shortward of redshifted 912 *. In a few cases, H**21-cm absorption has been detected, but molecular absorption (H*) has been detected in only one case. There is no clear evidence for dust absorption in any of these systems. At wavelengths shortward of the Ly-* emission line in quasars, the quasar spectrum is riddled with strong, narrow absorption features, nearly all of which are attributable to Ly-* absorption by clouds at lower redshift. The name "Ly-* forest" has been applied to this spectral region on account of the high density of high-contrast absorption features. These systems have low neutral hydrogen column densities, 10**-10** cm**, and velocity dispersions of a few tens of kilometers per second. These systems are apparently devoid of corresponding metal features, although metal lines should be detectable only in the larger column density systems. In a few cases, it has been shown that the metal abundances in the Ly-* clouds must be less than 0.001 the solar value. In a minority of quasars (-5%), broad absorption features are seen in the short-wavelength wings of the ultraviolet resonance lines. These are almost certainly due to gas flowing outward from the quasars at velocities up to -30,000 km s**. It is not known whether all quasars have BAL regions that cover only a fraction of the sky as seen from the quasar or only a small fraction of quasars have BALs when seen from any direction. It has been claimed that there are some differences between the emission spectra of quasars with BALs and other quasars. The true situation is very unclear, however, because the very presence of BALs severely alters the appearance of a QSO spectrum. It has long been supposed that the narrow quasar absorption lines arise in the disks or extended halos of galaxies. The required galaxy cross sections for producing absorption lines stronger than a given equivalent width can be computed by using the Holmberg radius-luminosity relation for galaxies (*** L***), the Schechter luminosity function, and the measured incidence of absorption systems per unit comoving path length. It is found that the sizes of galaxies must exceed their Holmberg radii by a factor of 2-5 to account for all observed narrow absorption lines with rest equivalent widths larger than -300 m*. Furthermore, the absence of absorption in the spectra of many quasars where our line of sight falls close to a known galaxy indicates that the gas would have to be distributed in a very patchy fashion, and thus extend even farther out than a few times the optical radius of the galaxy. The absorbing clouds are thus probably small and far more numerous than galaxies. However, they must somehow be associated with galaxies, because deep searches of the fields of quasars with relatively low-redshift absorption lines have turned up galaxies near the absorption redshift in a few cases. Moreover, the very fact that heavy elements are observed argues for some association with galaxies, at least in the case of the heavy-element systems. Sizes of the absorbers can be estimated from the incidence of identical absorption features in the spectra of pairs of quasars that appear close together on the sky, or in the different images of a gravitationally lensed quasar. Typical sizes derived this way are greater than or of order 10 kpc, though it is clear that there must be a considerable range in sizes. The patchiness of the absorbers supports the idea that the absorption occurs in small clouds within larger structures. The multicomponent structure of many of the heavy-element systems as well as their large velocity dispersions is consistent with such a picture. If the clouds producing the narrow absorption lines are distributed uniformly per unit comoving volume and their cross sections for producing lines larger than a given equivalent width are constant, the number of detected absorption systems per unit redshift will be *************************, where n* and ** are the space density and cross section, respectively, of clouds at the current epoch. This expression is often approximated as *****************, where y=1 for **=0 or y=1/2 for q*=1/2. Thus, in any plausible cosmology, *>1 would imply evolution in that the number of absorbers or their cross sections were larger in the past than they are today. Analysis of known absorption systems containing Mg **,lines (typically at z<1) gives y=1.5*O.5 (although values exceeding 2 have also been reported), and the data on the Lyman-limit systems (z>2.7) give y=0.7*0.5. However, the Ly-* forest lines show stronger evidence for evolution, with *>2. In any given quasar, however, the density of Ly-* forest lines decreases close to the quasar redshift, an effect known as the "inverse" or "proximity" effect. This is likely due to the absence of neutral gas near the quasar, which is capable of photoionizing all of the diffuse gas within several megaparsecs. The spectral density of C** absorption lines (observed at z>1) by contrast seems to show an overall decrease with redshift. It has been speculated that this is due to low cosmic metal abundance at high redshift. There appears to be, however, an enhancement of C** absorbers near the emission line redshift in steep-spectrum, radio-loud quasars, though not in other types of quasar. Additional Reading Bergeron, J.(1988). Metal-rich absorption-line systems. In The Post-Recombination Universe, N. Kaiser and A.N. Lasenby, eds. Kluwer Academic Publishers, Dordrecht, p. 202. Blades, J.C., Turnshek, D.A., and Norman, C.A., eds.(1988). QSO Absorption lines: Probing the Universe. Cambridge University Press, Cambridge. Peterson, B.A.(1986). QSO absorption lines: heavy elements and Lyman-* clouds. In Quasars, G. Swarup and V.K. Kapahi, eds. Kluwer Academic Press, Dordrecht, p. 555. Turnshek, D.A.(1986). Broad absorption line QSOs. In Quasars, G. Swarup and V.K. kapahi, eds. Kluwer Academic Publishers, Dordrecht, p. 317. Weymann, R.J., Carswell, R.F., and Smith, M.G. (1981). Absorption lines in the spectra of quasi-stellar objects. Ann. Rev. Astron. Ap. 19 41.