Ingunn Kathrine Wehus, Unni Fuskeland, Hans Kristian Eriksen
We study four particularly bright polarized compact objects (Tau A, Virgo A,
3C273 and Fornax A) in the 7-year WMAP sky maps, with the goal of understanding
potential systematics involved in estimation of foreground spectral indices. We
estimate the spectral index, the polarization angle, the polarization fraction
and apparent size and shape of these objects when smoothed to a nominal
resolution of 1 degree FWHM. Second, we compute the spectral index as a
function of polarization orientation, alpha. Because these objects are
approximately point sources with constant polarization angle, this function
should be constant in the absence of systematics. However, computing it for the
K- and Ka-band WMAP data we find strong index variations for all four sources.
For Tau A, we find a spectral index beta=-2.59+-0.03 for alpha=30 degrees, and
beta=-2.03+-0.01 for alpha=50 degrees. On the other hand, the spectral index
between Ka and Q band is found to be stable. The most likely cause of this
effect is beam asymmetries, which effectively results in a leakage of the
foreground amplitude into the spectral index. A simple toy model of an
asymmetric beam produces qualitatively very similar results. Based on these
findings, we therefore conclude that estimation of spectral indices with the
WMAP K-band polarization data at 1 degree scales is not robust. Second, we
argue that a model with a common spectral index for the Stokes' Q and U
parameters is likely to produce better results than a model with independent
indices. The methods used in this paper are likely to be useful to identify
systematics in other experiments as well.
View original:
http://arxiv.org/abs/1201.6348
No comments:
Post a Comment