Erasing the variable: Empirical foreground discovery for global 21 cm spectrum experiments

TL;DR

This paper presents an empirical method to identify and remove foreground contamination in 21 cm cosmology experiments by exploiting spatial fluctuations, improving the detection of the cosmological signal amidst bright foregrounds.

Contribution

It introduces a novel empirical foreground discovery technique that uses spatial fluctuations to identify contaminated spectral modes, enhancing signal extraction in 21 cm experiments.

Findings

Foreground modes vary across the sky, enabling their empirical discovery.

Instrument passband stability of ~10^{-4} is crucial for effective foreground removal.

Controlling polarization leakage is important for rejecting Faraday rotation effects.

Abstract

Spectral measurements of the 21 cm monopole background have the promise of revealing the bulk energetic properties and ionization state of our universe from z ~ 6-30. Synchrotron foregrounds are orders of magnitude larger than the cosmological signal, and are the principal challenge faced by these experiments. While synchrotron radiation is thought to be spectrally smooth and described by relatively few degrees of freedom, the instrumental response to bright foregrounds may be much more complex. To deal with such complexities, we develop an approach that discovers contaminated spectral modes using spatial fluctuations of the measured data. This approach exploits the fact that foregrounds vary across the sky while the signal does not. The discovered modes are projected out of each line-of-sight of a data cube. An angular weighting then optimizes the cosmological signal amplitude estimate by giving preference to lower-noise regions. Using this method, we show that it is essential for the passband to be stable to at least ~10^{-4}. In contrast, the constraints on the spectral smoothness of the absolute calibration are mainly aesthetic if one is able to take advantage of spatial information. To the extent it is understood, controlling polarization to intensity leakage at the ~10^{-2} level will also be essential to rejecting Faraday rotation of the polarized synchrotron emission.