The proximity effect in the spectra of quasi-stellar objects and the evolution of the ultraviolet background from z = 4 to z = 0

Abstract

I present moderate resolution spectra for 39 Quasi-Stellar Objects (QSOs) at z ≈ 2 obtained at the Multiple Mirror Telescope (MMT). These are combined with spectra of comparable resolution of 60 QSOs from the literature with z > 1.7 to investigate the distribution of Lyman α (Ly-α) forest absorption lines in redshift and equivalent width. I find γ = 1.88 ± 0.22 for lines stronger than a rest equivalent width of 0.32 Å, where γ is the line redshift distribution parameter, in good agreement with some previous studies. These spectra are used to measure J(ν₀), the mean intensity of the extragalactic background radiation at the Lyman limit, using the proximity effect signature. I find J(ν₀) = 7.0⁺³·⁴₋₄.₄ x 10⁻²² ergs s⁻¹ cm⁻² Hz⁻¹ sr⁻¹ at 1.7 < z < 3.8. A sample of 151 QSO spectra from the Faint Object Spectrograph on the Hubble Space Telescope are used to measure J(ν₀) at low redshift. I find J(ν₀) = 6.5⁺³⁸₋₁.₆ x 10⁻²³ ergs s⁻¹ cm⁻² Hz⁻¹ sr⁻¹ at z < 1, and J(ν₀) = 1.0⁺³·⁸₋₀.₂ x 10⁻²² ergs s⁻¹ cm⁻² Hz⁻¹ sr⁻¹ at z > 1, indicating that J(ν₀) is evolving over 0.03 < z < 3.8. This work confirms that the evolution of the number density of Ly-alpha lines is driven by a decrease in the ionizing background from z ∼ 2 to z ∼ 0 as well as by the growth of structure in the intergalactic medium and the formation of galaxies from intergalactic gas. These measurements of J(ν₀) are in reasonable agreement with the predictions of models based on the integrated quasar luminosity function. I present simulated Ly-α forest spectra created using the lognormal approximation to the linear and mildly non-linear evolution of the density and velocity fields. The model spectra give a mean Ly-α forest flux decrement of 0.128 at < z >= 2.07, while the MMT data show < D >= 0.129. The photoionization effects of quasars placed in the simulated density fields on the surrounding intergalactic medium are incorporated into the synthetic spectra. This reasonably reproduces the proximity effect signature seen in the data, a 2-3σ deficit of absorption lines within 2h⁻¹ Mpc of quasars. I find that maximum likelihood methods reliably estimate the ionization rate from the UV background radiation if quasars do not preferentially occupy regions of high overdensity. I analyze the extent to which the clustering of mass around quasars and uncertainty in quasar redshifts will bias the measurement of the ionizing background. In both cases, the ionization rates are overestimated by a factor of ∼3.