The bispectra of galactic CMB foregrounds and their impact on primordial non-Gaussianity estimation

TL;DR

This paper measures the bispectra of various galactic foregrounds in Planck data, assesses their potential confusion with primordial signals, and demonstrates that foreground contamination can be effectively identified and mitigated in CMB analyses.

Contribution

It introduces a method to quantify galactic foreground bispectra and evaluates their impact on primordial non-Gaussianity estimation using Planck data.

Findings

Galactic foreground bispectra peak in the squeezed configuration.

Dust bispectrum detected in raw 143 GHz map with expected amplitude.

No significant galactic residuals found in cleaned CMB map.

Abstract

We use the binned bispectrum estimator to determine the bispectra of the dust, free-free, synchrotron, and AME galactic foregrounds using maps produced by the Commander component separation method from Planck 2015 data. We find that all of these peak in the squeezed configuration, allowing for potential confusion with in particular the local primordial shape. Applying an additional functionality implemented in the binned bispectrum estimator code, we then use these galactic bispectra as templates in an $f_\mathrm{NL}$ analysis of other maps. After testing and validating the method and code with simulations, we show that we detect the dust in the raw 143 GHz map with the expected amplitude (the other galactic foregrounds are too weak at 143 GHz to be detected) and that no galactic residuals are detected in the cleaned CMB map. We also investigate the effect of the mask on the templates and the effect of the choice of binning on a joint dust-primordial $f_\mathrm{NL}$ analysis.