Cosmic reionization after Planck II: contribution from quasars

In the light of the recent Planck downward revision of the electron scattering optical depth, and of the discovery of a faint active galactic nuclei (AGN) population at z > 4, we reassess the actual contribution of quasars to cosmic reionization. To this aim, we extend our previous Markov Chain Monte Carlo based data-constrained semi-analytic reionization model and study the role of quasars on global reionization history. We find that the quasars can alone reionize the Universe only for models with very high AGN emissivities at high redshift. These models are still allowed by the recent cosmic microwave background data and most of the observations related to HI reionization. However, they predict an extended and early He II reionization ending at z ≳ 4 and a much slower evolution in the mean He II Ly-α forest opacity than what the actual observation suggests. Thus, when we further constrain our model against the He II Ly-α forest data, this AGN-dominated scenario is found to be clearly ruled out at 2σ limits. The data seems to favour a standard two-component picture where quasar contributions become negligible at z ≳ 6 and a non-zero escape fraction of ~10 per cent is needed from early-epoch galaxies. For such models, mean neutral hydrogen fraction decreases to ~10-4 at z = 6.2 from ~0.8 at z = 10.0 and helium becomes doubly ionized at much later time, z ~ 3. We find that these models are as well in good agreement with the observed thermal evolution of IGM as opposed to models with very high AGN emissivities.