[Paper Review] Foregrounds and CMB Experiments: I. Semi-analytical estimates of contamination

F. R. Bouchet, R. Gispert|ArXiv.org|Mar 11, 1999

Galaxies: Formation, Evolution, Phenomena65 references117 citations

TL;DR

This paper introduces a semi-analytical framework based on generalized Wiener filtering to quantify foreground contamination in CMB experiments, enabling accurate assessment of CMB reconstruction errors. It shows that Planck's design achieves sub-6 μK reconstruction error for ℓ < 1000, significantly outperforming MAP’s ~40 μK, with high robustness to channel failures and noise variations.

ABSTRACT

As Cosmic Microwave Background (CMB) measurements are becoming more ambitious, the issue of foreground contamination is becoming more pressing. This is especially true at the level of sensitivity, angular resolution and for the sky coverage of the planned space experiments MAP and PLANCK. We present in this paper an indicator of the accuracy of the separation of the CMB anisotropies from those induced by foregrounds. Of course, the outcome will depend on the spectral and spatial characteristics of the sources of anisotropies. We thus start by summarising present knowledge on the spectral and spatial properties of Galactic foregrounds, point sources, and clusters of galaxies. This information comes in support of a modelling of the microwave sky including the relevant components. The accuracy indicator we introduce is based on a generalisation of the Wiener filtering method to multi-frequency, multi-resolution data. While the development and use of this indicator was prompted by the preparation of the scientific case for the \plancks satellite, it has broader application since it allows assessing the effective capabilities of an instrumental set-up once foregrounds are fully accounted for, with a view to enabling comparisons between different experimental arrangements. The real sky might well be different from the one assumed here, and the analysis method might not be in the end Wiener filtering, but this work still allow meaningful {\em comparative} studies. As a matter of examples, we compare the CMB reconstruction errors for the \maps and \plancks space missions, as well as the robustness of the \plancks outcome to possible failures of specific spectral channels or global variations of the detectors noise level across spectral channels.

Motivation & Objective

• To develop a quantitative method for assessing foreground contamination in CMB experiments, particularly for future space missions like MAP and Planck.
• To model the microwave sky using current knowledge of Galactic and extragalactic foregrounds, including dust, free-free, synchrotron, point sources, and Sunyaev-Zeldovich effects.
• To define a 'quality factor' based on Wiener filtering that evaluates the effective accuracy of CMB reconstruction after foreground separation.
• To enable comparative analysis of different experimental setups by estimating residual errors and effective resolution in the presence of foregrounds.
• To evaluate the robustness of Planck’s performance under realistic failure scenarios, such as loss of specific frequency channels or noise variations.

Proposed method

• Uses a multi-frequency, multi-resolution extension of Wiener filtering to optimally separate CMB anisotropies from foreground emissions based on their spectral and spatial characteristics.
• Applies spherical harmonic decomposition to represent temperature anisotropies, with power spectra $ C_\ell $ characterizing Gaussian fields.
• Constructs a sky model combining Galactic dust (single temperature, $ \nu^2 $ emissivity), free-free, synchrotron, extragalactic point sources, and SZ effects.
• Derives an effective point spread function (window function) and effective noise level that account for foreground residuals and detector noise.
• Computes the CMB reconstruction error as a function of multipole $ \ell $, distinguishing contributions from noise and incomplete foreground subtraction.
• Introduces a quality factor to quantify the degradation in CMB power spectrum estimation due to foregrounds, under the assumption of Gaussian foregrounds.

Experimental results

Research questions

• RQ1 How accurately can CMB anisotropies be reconstructed from multi-frequency, multi-resolution data when foregrounds are present?
• RQ2 What is the impact of foreground contamination on the effective resolution and noise level of CMB experiments like MAP and Planck?
• RQ3 How robust is the CMB reconstruction to failures in specific frequency channels or variations in detector noise levels?
• RQ4 To what extent do uncertainties in foreground modeling affect the final CMB power spectrum accuracy?
• RQ5 How do the performance characteristics of MAP and Planck compare in terms of residual reconstruction error and effective dynamic range?

Key findings

• The Planck mission is projected to achieve a CMB reconstruction error of less than 6 μK for multipoles ℓ < 1000, significantly better than MAP’s ~40 μK.
• The effective resolution of Planck is expected to extend to ℓ ~ 2300 (up to ℓ ~ 2500 for a Lambda CDM model), outperforming MAP’s limit of ℓ ~ 900–1000.
• The most critical frequency channels for CMB recovery are 143 and 217 GHz; losing them increases the residual rms by less than 10%.
• Losing all HFI channels except 217 GHz increases the residual rms by nearly 20%, highlighting the importance of maintaining sensitivity at this central channel.
• MAP is more sensitive to global noise level variations than Planck, with the latter showing a nearly linear response to such changes.
• The framework shows that even with imperfect foreground models, CMB recovery remains robust because the dominant contributions in key channels are the CMB and detector noise, not foregrounds.