Galactic foreground contributions to the 5-year Wilkinson Microwave Anisotropy Probe maps

We compute the cross-correlation between intensity and polarization from the 5-year Wilkinson Microwave Anisotropy Probe (WMAP5) data in different sky regions with respect to template maps for synchrotron, dust and free–free emission. We derive the frequency dependence and polarization fraction for all three components in 48 different sky regions of HEALPix (N_(side)= 2) pixelization. The anomalous emission associated with dust is clearly detected in intensity over the entire sky at the K (23-GHz) and Ka (33-GHz) WMAP bands, and is found to be the dominant foreground at low Galactic latitudes, between b =−40° and +10°. The synchrotron spectral index obtained from the K and Ka WMAP bands from an all-sky analysis is β_s=−3.32 ± 0.12 for intensity and β_s=−3.01 ± 0.03 for polarized intensity. The polarization fraction of the synchrotron emission is constant in frequency and increases with latitude from ≈5 per cent near the Galactic plane up to ≈40 per cent in some regions at high latitudes; the average value for |b| 20°, it is 19.3 ± 0.8 (stat) ± 0.5 (sys) per cent. Anomalous dust and free–free emissions appear to be relatively unpolarized. Monte Carlo simulations showed that there were biases of the method due to cross-talk between the components, at up to ≈5 per cent in any given pixel, and ≈1.5 per cent on average, when the true polarization fraction is low (a few per cent or less). Nevertheless, the average polarization fraction of dust-correlated emission at the K band is 3.2 ± 0.9 (stat) ± 1.5 (sys) per cent or less than 5 per cent at 95 per cent confidence. When comparing real data with simulations, eight regions show a detected polarization above the 99th percentile of the distribution from simulations with no input foreground polarization, six of which are detected at above 2σ and display polarization fractions between 2.6 and 7.2 per cent, except for one anomalous region, which has 32 ± 12 per cent. The dust polarization values are consistent with the expectation from spinning dust emission, but polarized dust emission from magnetic-dipole radiation cannot be ruled out. Free–free emission was found to be unpolarized with an upper limit of 3.4 per cent at 95 per cent confidence