Spatial variations in the spectral index of polarized synchrotron emission in the 9 yr WMAP sky maps

We estimate the spectral index, beta, of polarized synchrotron emission as observed in the 9 yr WMAP sky maps using two methods, linear regression ("T-T plot") and maximum likelihood. We partition the sky into 24 disjoint sky regions, and evaluate the spectral index for all polarization angles between 0 deg and 85 deg in steps of 5. Averaging over polarization angles, we derive a mean spectral index of beta_all-sky=-2.99+-0.01 in the frequency range of 23-33 GHz. We find that the synchrotron spectral index steepens by 0.14 from low to high Galactic latitudes, in agreement with previous studies, with mean spectral indices of beta_plane=-2.98+-0.01 and beta_high-lat=-3.12+-0.04. In addition, we find a significant longitudinal variation along the Galactic plane with a steeper spectral index toward the Galactic center and anticenter than toward the Galactic spiral arms. This can be well modeled by an offset sinusoidal, beta(l)=-2.85+0.17sin(2l-90). Finally, we study synchrotron emission in the BICEP2 field, in an attempt to understand whether the claimed detection of large-scale B-mode polarization could be explained in terms of synchrotron contamination. Adopting a spectral index of beta=-3.12, typical for high Galactic latitudes, we find that the most likely bias corresponds to about 2% of the reported signal (r=0.003). The flattest index allowed by the data in this region is beta=-2.5, and under the assumption of a straight power-law frequency spectrum, we find that synchrotron emission can account for at most 20% of the reported BICEP2 signal.