How is the reionization epoch defined?

Abstract

We study the effect of a prolonged epoch of reionization on the angular power spectrum of the cosmic microwave background. Typically reionization studies assume a sudden phase transition, with the intergalactic gas moving from a fully neutral to a fully ionized state at a fixed redshift. Such models are at odds, however, with detailed investigations of reionization, which favour a more extended transition. We have modified the code cmbfast to allow the treatment of more realistic reionization histories and applied it to data obtained from numerical simulations of reionization. We show that the prompt reionization assumed by cmbfast in its original form heavily contaminates any constraint derived on the reionization redshift. We find, however, that prompt reionization models give a reasonable estimate of the epoch at which the mean cosmic ionization fraction was ≈50 per cent, and provide a very good measure of the overall Thomson optical depth. The overall differences in the temperature (polarization) angular power spectra between prompt and extended models with equal optical depths are less than 1 per cent (10 per cent).

Cosmic microwave background polarization and reionization: Constraining models with a double reionization

“…Kaplinghat et al (2003) modeled a two-step reionization, while Holder et al (2003) simulated reionization histories based on physical models of formation and evolution of the first ionizing sources. At variance with the double reionization model by Cen (2003) and with models considered here, almost all models discussed in these works display a monotonic increase of the ionized fraction with time.…”

“…A similar 2-step reionization was also schematically considered by Kaplinghat et al (2003), for its effects on CMB. Still different reionization patterns were treated by Bruscoli et al (2002), Hu & Holder (2003 ), Naselsky & Chiang (2004) and Colombo (2004), confirming the interest in the relation between CMB data and early reionization, although often specifying no link with early object formation models. The present work is based on the dependence of the low-l behavior of the C E l spectra on the whole reionization history, and aims to predict the actual detectability of τ, as well as of the reionization redshift(s), the ionization rate(s), etc., with data on large angular scales.…”

“…Double reionization models were suggested, even before WMAP, by Cen (2003), Wyithe et al (2003 ), Sokasian et al (2004), Ricotti & Ostriker (2004a) and others. In these models the ionization fraction attains a value x e ≈ 1, at some high z, then partial recombination occurs, while a second reionization takes place at z = 6-8 and then x e = 1 (for discussion purposes, we neglect helium ionization).…”

“…The duration of the first ionized period is fixed once by assuming that, between the two ionized eras, x e = 1/3. This category of reionization histories is meant to approach the pattern indicated by Cen (2003). We consider a grid of models spanning a large portion of the parameter space, and take into account temperature, polarization and temperature-polarization cross-correlation spectra.…”

Model-independent reionization observables in the CMB

“…Attempts to do so can result in biases and conflicting results. For example, using an ionization history derived from a particular numerical prescription [5] it was found that the partially ionized epoch was undetectable in the CMB, while analyses using semi-analytic models [6,7] showed that the predicted multiple epochs of reionization and partial recombination can be detected [8,9]. As another example, by assuming an overly simplistic step function model for reionization, the total optical depth τ in the semi-analytic models can be mis-estimated by up to a few 10 −2 [9]; while not a large amount it is already approaching the statistical errors in the first year WMAP data.…”