Studying the Multi-frequency Angular Power Spectrum of the Cosmic Dawn 21-cm Signal

TL;DR

This paper investigates the multi-frequency angular power spectrum (MAPS) of the cosmic dawn 21-cm signal, analyzing its physical implications and measurement prospects with upcoming radio telescopes like HERA, NenuFAR, and SKA-Low.

Contribution

It introduces the use of MAPS to study the 21-cm signal, highlighting its advantages over the 3D power spectrum and providing error predictions for future observations.

Findings

HERA can measure MAPS at ≥3σ for ℓ ≤ 1000 with 100h observation.

SKA-Low can reach this sensitivity up to ℓ ≤ 3000 due to better sensitivity.

NenuFAR's measurements are limited to ℓ ≤ 600 at 2σ with 1000h, but improve with longer observations and channel combination.

Abstract

The light-cone (LC) anisotropy arises due to cosmic evolution of the cosmic dawn 21-cm signal along the line-of-sight (LoS) axis of the observation volume. The LC effect makes the signal statistically non-ergodic along the LoS axis. The multi-frequency angular power spectrum (MAPS) provides an unbiased alternative to the popular 3D power spectrum as it does not assume statistical ergodicity along every direction in the signal volume. Unlike the 3D power spectrum which mixes the cosmic evolution of the 21-cm signal along the LoS $k$ modes, MAPS keeps the evolution information disentangled. Here we first study the impact of different underlying physical processes during cosmic dawn on the behaviour of the 21-cm MAPS using simulations of various different scenarios and models. We also make error predictions in 21-cm MAPS measurements considering only the system noise and cosmic variance for mock observations of HERA, NenuFAR and SKA-Low. We find that $100~{\rm h}$ of HERA observations will be able to measure 21-cm MAPS at $\geq 3\sigma$ for $\ell \lesssim 1000$ with $0.1\,{\rm MHz}$ channel-width. The better sensitivity of SKA-Low allows reaching this sensitivity up to $\ell \lesssim 3000$. Note that due to the difference in the frequency coverage of the various experiements, the CD-EoR model considered for NenuFAR is different than those used for the HERA and SKA-Low predictions. Considering NenuFAR with the new model, measurements $\geq 2\sigma$ are possible only for $\ell \lesssim 600$ with $0.2\,{\rm MHz}$ channel-width and for a ten times longer observation time of $t_{\rm obs} = 1000~{\rm h}$. However, for the range $300 \lesssim \ell \lesssim 600$ and $t_{\rm obs}=1000~{\rm h}$ more than $3\sigma$ measurements are still possible for NenuFAR when combining consecutive frequency channels within a $5 ~{\rm MHz}$ band.