First‐Year Wilkinson Microwave Anisotropy Probe ( WMAP ) Observations: Foreground Emission

Abstract

The WMAP mission has mapped the full sky to determine the geometry, content, and evolution of the universe. Full-sky maps are made in five microwave frequency bands to separate the temperature anisotropy of the cosmic microwave background (CMB) from foreground emission, including diffuse Galactic emission and Galactic and extragalactic point sources. We define masks that excise regions of high foreground emission, so CMB analyses can be carried out with minimal foreground contamination. We also present maps and spectra of the individual emission components, leading to an improved understanding of Galactic astrophysical processes. The effectiveness of template fits to remove foreground emission from the WMAP data is also examined. These efforts result in a CMB map with minimal contamination and a demonstration that the WMAP CMB power spectrum is insensitive to residual foreground emission. We use a maximum entropy method to construct a model of the Galactic emission components. The observed total Galactic emission matches the model to less than 1%, and the individual model components are accurate to a few percent. We find that the Milky Way resembles other normal spiral galaxies between 408 MHz and 23 GHz, with a synchrotron spectral index that is flattest (β s ∼ -2.5) near star-forming regions, especially in the plane, and steepest (β s ∼ -3) in the halo. This is consistent with a picture of relativistic cosmic-ray electron generation in star-forming regions and diffusion and convection within the plane. The significant synchrotron index steepening out of the plane suggests a diffusion process in which the halo electrons are trapped in the Galactic potential long enough to suffer synchrotron and inverse Compton energy losses and hence a spectral steepening. The synchrotron index is steeper in the WMAP bands than in lower frequency radio surveys, with a spectral break near 20 GHz to β s < -3. The modeled thermal dust spectral index is also steep in the WMAP bands, with β d ≈ 2.2. Our model is driven to these conclusions by the low level of total foreground contamination at ∼60 GHz. Microwave and Hα measurements of the ionized gas agree well with one another at about the expected levels. Spinning dust emission is limited to ≲5% of the Ka-band foreground emission, assuming a thermal dust distribution with a cold neutral medium spectrum and a monotonically decreasing synchrotron spectrum. A catalog of 208 point sources is presented. The reliability of the catalog is 98%; i.e., we expect five of the 208 sources to be statistically spurious. The mean spectral index of the point sources is α ∼ 0 (β ∼ -2). Derived source counts suggest a contribution to the anisotropy power from unresolved sources of (15.0 ± 1.4) × 10 -3 μK 2 sr at Q band and negligible levels at V band and W band. The Sunyaev-Zeldovich effect is shown to be a negligible "contamination" to the maps.