Experimental constraints on self-consistent reionization models

A self-consistent formalism to jointly study cosmic reionization and thermal history of the intergalactic medium (IGM) in a ΛCDM cosmology is presented. The model implements most of the relevant physics governing these processes, such as the inhomogeneous IGM density distribution, three different classes of ionizing photon sources [massive Population III (PopIII) stars, Population II (PopII) stars and quasi-stellar objects (QSOs)], and radiative feedback inhibiting star formation in low-mass galaxies. By constraining the model free parameters with available data on redshift evolution of Lyman-limit absorption systems, Gunn–Peterson and electron scattering optical depths, near-infrared background and cosmic star formation history, we select a fiducial model, whose main predictions are as follows. (i) Hydrogen was completely reionized at z≈ 15, while He ii must have been reionized by z≈ 12, allowing for the uncertainties in the ionizing photon efficiencies of stars. At z≈ 7, He iii suffered an almost complete recombination as a result of the extinction of PopIII stars, as required by the interpretation of the NIRB. (ii) A QSO-induced complete He ii reionization occurs at z= 3.5; a similar double H reionization does not take place due to the large number of photons with energies >13.6 eV from PopII stars and QSOs, even after all PopIII stars have disappeared. (iii) Following reionization, the temperature of the IGM corresponding to the mean gas density, T0, is boosted to 1.5 × 104 K; following that it decreases with a relatively flat trend. Observations of T0 are consistent with the fact that He is singly ionized at z≳ 3.5, while they are consistent with He being doubly ionized at z≲ 3.5. This might be interpreted as a signature of (second) He ii reionization. (iv) Only 0.3 per cent of the stars produced by z= 2 need to be PopIII stars in order to achieve the first hydrogen reionization. In addition, we get useful constraints on the ionizing photon efficiencies (which are a combination of the star-forming efficiency and the escape fraction of ionizing photons from collapsed haloes) of PopII and PopIII stars, namely, ePopII < 0.01, 0.002 < ePopIII < 0.03. Varying the efficiencies in these two ranges does not affect the scenario described above. Such a model not only relieves the tension between the Gunn–Peterson optical depth and WMAP observations, but also accounts self-consistently for all known observational constraints. We discuss how the results compare with recent numerical reionization studies and other theoretical arguments.