Constraints on perturbations to the recombination history from measurements of the CMB damping tail

TL;DR

This paper investigates how small-scale CMB measurements constrain the ionization history during recombination, revealing a persistent tension with the standard model that future data could clarify.

Contribution

It introduces a framework to constrain ionization fraction perturbations using principal components and applies it to 2011 and 2012 CMB data, highlighting ongoing tensions.

Findings

Standard recombination is consistent with data but shows tension similar to helium fraction variations.

Only two eigenmodes are constrained with current data, but future experiments will improve this.

The prior on eigenamplitudes is influenced by electron number conservation, affecting results.

Abstract

The primordial CMB at small angular scales is sensitive to the ionization and expansion history of the universe around the time of recombination. This dependence has been exploited to constrain the helium abundance and the effective number of relativistic species. Here we focus on allowed ionization fraction trajectories, $\Xe (z)$, by constraining low-order principal components of perturbations to the standard recombination scenario ($\Xe$-eigenmodes) in the circa 2011 SPT, ACT and WMAP7 data. Although the trajectories are statistically consistent with the standard recombination, we find that there is a tension similar to that found by varying the helium fraction. As this paper was in press, final SPT and ACT datasets were released and we applied our framework to them: we find the tension continues, with slightly higher significance, in the new 2012 SPT data, but find no tension with the standard model of recombination in the new 2012 ACT data. We find that the prior probabilities on the eigenamplitudes are substantially influenced by the requirement that $\Xe$ trajectories conserve electron number. We propose requiring a sufficient entropy decrease between posterior and prior marginalized distributions be used as an $\Xe$-mode selection criterion. We find that in the case of the 2011 SPT/ACT+WMAP7 data only two modes are constrainable, but upcoming ACTPol, Planck and SPTPol data will be able to test more modes and more precisely address the current tension.