Cosmic Reionization Study : Principle Component Analysis After Planck

TL;DR

This paper employs principal component analysis to reconstruct the cosmic reionization history from multiple observational datasets, revealing that the ionization fraction is mostly consistent with an instantaneous model but shows slight deviations at certain epochs.

Contribution

It introduces a model-independent PCA approach to analyze reionization history using combined observational data, highlighting potential deviations from the instantaneous reionization assumption.

Findings

Reconstructed ionization fraction aligns with instantaneous reionization.

Slight deviations in ionization history are observed at specific epochs.

Stronger constraints are achieved by removing noisy modes in PCA.

Abstract

The study of reionization history plays an important role in understanding the evolution of our universe. It is commonly believed that the intergalactic medium (IGM) in our universe are fully ionized today, however the reionizing process remains to be mysterious. A simple instantaneous reionization process is usually adopted in modern cosmology without direct observational evidence. However, the history of ionization fraction, $x_e(z)$ will influence cosmic microwave background (CMB) observables and constraints on optical depth $\tau$. With the mocked future data sets based on featured reionization model, we find the bias on $\tau$ introduced by instantaneous model can not be neglected. In this paper, we study the cosmic reionization history in a model independent way, the so called principle component analysis (PCA) method, and reconstruct $x_e (z)$ at different redshift $z$ with the data sets of Planck, WMAP 9 years temperature and polarization power spectra, combining with the baryon acoustic oscillation (BAO) from galaxy survey and type Ia supernovae (SN) Union 2.1 sample respectively. The results show that reconstructed $x_e(z)$ is consistent with instantaneous behavior, however, there exists slight deviation from this behavior at some epoch. With PCA method, after abandoning the noisy modes, we get stronger constraints, and the hints for featured $x_e(z)$ evolution could become a little more obvious.