We explore the effect of cosmic radiative feedback from the sources of reionization on the thermal evolution of the intergalactic medium. We find that different prescriptions for this feedback predict quite different thermal and reionization historie
s. In spite of this, current data can not discriminate among different reionization scenarios. We find that future observations both from 21-cm and CMB experiments can be used to break the degeneracy among model parameters provided that we will be able to remove the foreground signal at the percent (or better) level.
We compute the imprints left on the CMB by two cosmic reionization models consistent with current observations but characterized by alternative radiative feedback prescriptions (suppression and filtering) resulting in a different suppression of star
formation in low-mass halos. The models imply different ionization and thermal histories and 21 cm background signals. The derived Comptonization, u, and free-free distortion, y_B, parameters are below current observational limits for both models. However, the value of u = 1.69 * 10^-7 (9.65 * 10^-8) for the suppression (filtering) model is in the detectability range of the next generation of CMB spectrum experiments. Through the dedicated Boltzmann code CMBFAST, modified to include the above ionization histories, we compute the CMB angular power spectrum (APS) of the TT, TE, and EE modes. For the EE mode the differences between these models are significantly larger than the cosmic and sampling variance over the multipole range l = 5-15, leaving a good chance of discriminating between these feedback mechanisms with forthcoming/future CMB polarization experiments. The main limitations come from foreground contamination: it should be subtracted at per cent level in terms of APS, a result potentially achievable by novel component separation techniques and mapping of Galactic foreground.
