A Deep Search for Extended Radio Continuum Emission from Dwarf Spheroidal Galaxies: Implications for Particle Dark Matter

We present deep radio observations of four nearby dwarf spheroidal (dSph) galaxies, designed to detect extended synchrotron emission resulting from weakly interacting massive particle (WIMP) dark matter annihilations in their halos. Models by Colafrancesco et al. (CPU07) predict the existence of angularly large, smoothly distributed radio halos in such systems, which stem from electron and positron annihilation products spiraling in a turbulent magnetic field. We map a total of 40.5 deg² around the Draco, Ursa Major II, Coma Berenices, and Willman 1 dSphs with the Green Bank Telescope (GBT) at 1.4 GHz to detect this annihilation signature, greatly reducing discrete-source confusion using the NVSS catalog. We achieve a sensitivity of σ_sub <~ 7 mJy beam^-1 in our discrete source-subtracted maps, implying that the NVSS is highly effective at removing background sources from GBT maps. For Draco we obtained approximately concurrent Very Large Array observations to quantify the variability of the discrete source background, and find it to have a negligible effect on our results. We construct radial surface brightness profiles from each of the subtracted maps, and jackknife the data to quantify the significance of the features therein. At the ~10' resolution of our observations, foregrounds contribute a standard deviation of 1.8 mJy beam^-1 <= σ_ast <= 5.7 mJy beam^-1 to our high-latitude maps, with the emission in Draco and Coma dominated by foregrounds. On the other hand, we find no significant emission in the Ursa Major II and Willman 1 fields, and explore the implications of non-detections in these fields for particle dark matter using the fiducial models of CPU07. For a WIMP mass M _χ = 100 GeV annihilating into b\bar{b} final states and B = 1 μG, upper limits on the annihilation cross-section for Ursa Major II and Willman I are log (langσvrang_χ, cm³ s^-1) <~ -25 for the preferred set of charged particle propagation parameters adopted by CPU07; this is comparable to that inferred at γ-ray energies from the two-year Fermi Large Area Telescope data. We discuss three avenues for improving the constraints on langσvrang_χ presented here, and conclude that deep radio observations of dSphs are highly complementary to indirect WIMP searches at higher energies.