CMB Polarization and Temperature Power Spectra Estimation using Linear Combination of WMAP 5-year Maps

TL;DR

This paper presents a model-independent method to estimate CMB polarization and temperature power spectra from WMAP 5-year maps, achieving smaller error bars and consistent results with previous analyses.

Contribution

It introduces a new linear combination approach in harmonic space for foreground removal that does not rely on diffuse foreground templates or noise models.

Findings

Smaller error bars on polarization spectra compared to WMAP results.

Bias in the estimation procedure is minimal, confirmed by simulations.

Quadrupole power estimate is significantly higher than previous WMAP estimate.

Abstract

We estimate CMB polarization and temperature power spectra using WMAP 5-year foreground contaminated maps. The power spectrum is estimated by using a model independent method, which does not utilize directly the diffuse foreground templates nor the detector noise model. The method essentially consists of two steps, (i) removal of diffuse foregrounds contamination by making linear combination of individual maps in harmonic space and (ii) cross-correlation of foreground cleaned maps to minimize detector noise bias. For temperature power spectrum we also estimate and subtract residual unresolved point source contamination in the cross-power spectrum using the point source model provided by the WMAP science team. Our $TT$, $TE$ and $EE$ power spectra are in good agreement with the published results of the WMAP science team. The error bars on the polarization power spectra, however, turn out to be smaller in comparison to what is obtained by the WMAP science team. We perform detailed numerical simulations to test for bias in our procedure. We find that the bias is small in all cases. A negative bias at low $l$ in $TT$ power spectrum has been pointed in an earlier publication. We find that the bias corrected quadrupole power ($l(l+1)C_l/2\pi)$ is 532 $\mu K^2$, approximately 2.5 times the estimate (213.4 $\mu K^2$) made by the WMAP team.