ON THE PHOTOIONIZATION OF THE INTERGALACTIC MEDIUM BY QUASARS AT HIGH-REDSHIFT

We discuss the reionization of the intergalactic medium (IGM) by quasa rs at high redshift. We compute the integrated UV background from obse rved QSOs, taking into account the hydrogen opacity associated with in tervening Lyalpha clouds and Lyman limit systems. We note that the pub lished data appear to indicate a significant underdensity of absorptio n systems in the Lyalpha forest with column densities N(HI) > 10(15) c m-2. ThiS deficit, if real, would result in a reduction of the opacity of the universe at high redshift by a factor of 1.5-3 relative to pre vious estimates. The QSO contribution over the range z = 3-5 to the me tagalactic flux at 912 angstrom may then be as large as 6[(1 + z)/4.5] 0.5 x 10(-22) ergs cm-2 s-1 Hz-1 sr-1 for q0 = 0, and 3 x 10(-22) ergs cm-2 s-1 Hz-1 sr-1 for q0 = 0.5. We show that this ionizing flux is c onsistent with current 1 sigma upper limits to the Gunn-Peterson optic al depth tau(TG) at z congruent-to 3 if the density of the diffuse com ponent of the IGM satisfies OMEGA(D)h50(2) less than or similar to 0.0 2. Models based on QSO photoionization can generate a value of tau(GP) < 0.1 at z greater than or similar to 3.8 only in the case of a highl y clumped medium. We argue that the Lyalpha forest may contain a large fraction of the baryons in the universe and estimate the correspondin g mass density parameter to be in the range 0.002 < OMEGA(Lyalpha) h50 < 0.05. We also set constraints on an alternative scenario in which t he ''intercloud'' medium is collisionally ionized, and QSOs photoioniz e the Lyalpha forest alone. We conclude that, within the uncertainties , the observed QSOs can provide the required number of ionizing photon s at early epochs in several of the models examined. If quasars turn o n suddenly, the universe will be completely photoionized within DELTAz less than or similar to 0.5 of the turn-on redshift. If instead, QSOs turn on gradually, the associated H II regions may not have fully per colated until z approximately 5-5.5. We argue that a new class of abso rption systems, arising from intervening patches of neutral unprocesse d material which have not yet been engulfed by an H II region, could b e observed in the spectra of QSOs just beyond this breakthrough epoch. We estimate their typical H I column densities, sizes, and velocity w idths to be 10(18) - 10(20) cm-2, 0.1-1 Mpc, and 100-1000 km s-1. The patches are expected to persist as distinct absorption features over a n extended redshift interval of width DELTAz approximately 0.5-1. Even tually, the increase of the equivalent widths of the lines with redshi ft results in their collective formation of a Gunn-Peterson absorption trough shortward of the Lyalpha emission line. We find that, for most successful models, the patches appear when tau(GP) is still less than unity. Because of the difficulty of measuring such small optical dept hs at high z, due to the presence of the Lyalpha forest, neutral patch es along the line of sight may provide the first unambiguous detection of a diffuse, primordial IGM. The appearance of these systems would b e a sensitive probe of the epoch of reionization. We also discuss the radio signatures of the lines. Although our emphasis is on the photoio nization of the IGM by QSOs, the formalism we develop, including the e ffects of UV attenuation by intervening absorption systems, as well as many of our results, applies to any discrete source of photoionizing radiation of known number density, intensity, and spectrum.