[Paper Review] Extended reionization in models beyond $Λ$CDM with Planck 2018 data

D. Paoletti, Dhiraj Kumar Hazra+2 more|arXiv (Cornell University)|May 25, 2020

Cosmology and Gravitation Theories4 citations

TL;DR

This paper updates constraints on extended reionization histories using Planck 2018 CMB anisotropy data, employing the flexible Poly-reion model with piecewise cubic Hermite interpolation to model ionization fraction evolution. It finds a reduced reionization duration (Δz^Reion = 4.59_{-2.45}^{+1.67}), a disfavored early onset (z_xe=0 = 18.18_{-10.89}^{+1.61}), and significantly reduced degeneracies with beyond-ΛCDM parameters, driven by improved low-multipole E-mode polarization measurements.

ABSTRACT

We provide an update on the constraints on extended reionization histories with the Planck 2018 cosmic microwave background anisotropy data. The Planck 2018 data on large angular scales improve the measurement of the $E$-mode polarization reionization bump at low multipoles providing the possibility to improve our previous results. Using a minor modification to the original Poly-reion model for the reionization history, we find that the Planck 2018 data significantly improve all our previous results: we find as optical depth of $τ=0.0572_{-0.0075}^{+0.0064}$ at 68% CL, that early onsets of reionization are strongly disfavoured, i.e. redshift when the reionization begins, $z_{xe=0}=18.18_{-10.89}^{+1.61}$ at 68% CL,and that reionization duration (defined between 10% and 99% reionization) is significantly reduced, i.e. $Δ_z^{Reion}=4.59_{-2.45}^{+1.67}$ at 68% CL. We explore possible correlations between reionization histories and cosmological parameters, including important extensions beyond $Λ$CDM. We find that the degeneracy between reionization and scalar spectral index,neutrino mass sum, spatial curvature, dark matter annihilation and other non-standard models are significantly reduced.The reduction of the error bars and the degeneracies, together with the shift towards lower values of the optical depth that we observe in the Poly-reion model are mainly driven by the new low-$\ell$ polarization likelihood of Planck 2018 baseline based on the HFI data. This is confirmed also by the results derived without this likelihood and the ones with different alternatives to the baseline that are presented for a subset of models.

Motivation & Objective

• To improve constraints on extended reionization histories beyond the standard ΛCDM model using updated Planck 2018 CMB data.
• To assess how the new low-multipole E-mode polarization likelihood reduces uncertainties and degeneracies in reionization parameters.
• To investigate the impact of reionization modeling flexibility on cosmological parameter degeneracies, especially with neutrino mass, spatial curvature, dark matter annihilation, and primordial tensor modes.
• To validate results across multiple likelihood combinations, including alternative HFI and LFI polarization data.

Proposed method

• The Poly-reion model uses piecewise cubic Hermite interpolating polynomials (PCHIP) to model the ionization fraction x_e(z) with four redshift nodes, including fixed nodes at z=0 and z=5.5, and flexible nodes at z_int and z_xe=0.
• The model allows free variation of the ionization fraction amplitude and redshift at intermediate nodes, enabling a more flexible reconstruction of reionization history than the standard hyperbolic tangent model.
• Constraints are derived using Planck 2018 baseline likelihoods, including the new low-ℓ polarization likelihood from HFI data, and compared with alternative likelihoods such as LFI joint (T,Q,U) likelihood.
• Degeneracy analysis is performed between reionization parameters and extensions to ΛCDM, including neutrino mass sum, spatial curvature, dark matter annihilation, lensing amplitude A_L, and primordial tensor contributions.
• The analysis includes results from reprocessed HFI data using the SROLL2 map-making algorithm to test robustness of findings.
• Statistical inference is performed using Markov Chain Monte Carlo methods to derive posterior distributions and credible intervals at 68% CL.

Experimental results

Research questions

• RQ1 What are the updated constraints on the optical depth τ and reionization duration Δz^Reion when using the Poly-reion model with Planck 2018 data?
• RQ2 How does the inclusion of the new low-ℓ E-mode polarization likelihood from HFI data affect the precision and degeneracy structure of reionization parameters?
• RQ3 To what extent do extended cosmological models (e.g., varying neutrino mass, spatial curvature, dark matter annihilation) alter the inferred reionization history and its uncertainties?
• RQ4 How do the results from the Poly-reion model compare with the standard hyperbolic tangent model in terms of optical depth and reionization timing?
• RQ5 What is the impact of alternative likelihoods (e.g., LFI or reprocessed HFI) on the robustness of the reionization constraints?

Key findings

• The optical depth is constrained to τ = 0.0572_{-0.0075}^{+0.0064} at 68% CL, consistent with the standard hyperbolic tangent model but with improved precision.
• The reionization duration, defined between 10% and 99% ionization, is significantly reduced to Δz^Reion = 4.59_{-2.45}^{+1.67} at 68% CL.
• Early onsets of reionization are strongly disfavored, with the redshift at which reionization begins constrained to z_xe=0 = 18.18_{-10.89}^{+1.61} at 68% CL.
• Degeneracies between reionization and beyond-ΛCDM parameters—particularly with neutrino mass sum, spatial curvature, and primordial tensor modes—are significantly reduced due to improved low-ℓ polarization data.
• The new low-ℓ E-mode polarization likelihood from Planck 2018 HFI data is the primary driver behind the improved constraints and reduced error bars.
• The Poly-reion model reduces the pull toward lower optical depth in extended models, shifting lensing amplitude and spatial curvature estimates closer to ΛCDM values, except for residual degeneracy with A_L and spatial curvature.