The degeneracy between primordial non-Gaussianity and foregrounds in 21cm intensity mapping experiments

TL;DR

This paper investigates how foreground removal in 21cm intensity mapping experiments biases the measurement of primordial non-Gaussianity, highlighting the need for better foreground modeling to accurately probe early universe signals.

Contribution

First forecasts incorporating realistic foreground removal techniques demonstrate the degeneracy and bias in measuring primordial non-Gaussianity using 21cm intensity mapping.

Findings

Foreground cleaning causes significant bias in $f_\text{NL}$ estimates.

Rigid foreground models can recover unbiased results but are unrealistic.

Marginalizing over foreground parameters increases uncertainties significantly.

Abstract

Potential evidence for primordial non-Gaussianity (PNG) is expected to lie in the largest scales mapped by cosmological surveys. Forthcoming 21cm intensity mapping experiments will aim to probe these scales by surveying neutral hydrogen (HI) within galaxies. However, foreground signals dominate the faint 21cm emission, meaning foreground cleaning is required to recover the cosmological signal. The effect this has is to damp the HI power spectrum on the largest scales, especially along the line-of-sight. Whilst there is agreement that this contamination is potentially problematic for probing PNG, it is yet to be fully explored and quantified. In this work we carry out the first forecasts on $f_\text{NL}$ that incorporate simulated foreground maps that are removed using techniques employed in real data. Using an MCMC analysis, we demonstrate that foreground cleaned data recovers hugely biased values ($f_\text{NL} = -102.1_{-7.96}^{+8.39}$ [68% CL]) on our $f_\text{NL}=0$ fiducial input. Introducing a model with fixed parameters for the foreground contamination allows us to recover unbiased results ($f_\text{NL} = -2.94_{-11.9}^{+11.4}$). However, it is not clear that we will have sufficient understanding of foreground contamination to allow for such rigid models. Treating the main parameter $k_\parallel^\text{FG}$ in our foreground model as a nuisance parameter and marginalizing over it, still recovers unbiased results but at the expense of much larger errors ($f_\text{NL} = 0.75^{+40.2}_{-44.5}$), that can only be reduced by imposing the Planck 2018 prior. Our results show that significant progress on understanding and controlling foreground removal effects is necessary in order to study PNG with HI intensity mapping.