Well motivated theoretical models predict the annihilation of dark matter (DM) into standard model particles, a phenomenon which could be a significant source of photons in the gamma-ray sky. With its unprecedented sensitivity and its broad energy range (20 MeV to more than 300 GeV) the main instrument on board the Fermi satellite, the Large Area Telescope (LAT), might be able to detect an indirect signature of DM annihilations. In this work we revisit several interesting claims of extended dark matter emission made from analyses of Fermi-LAT data: First, based on three years of Fermi Large Area Telescope (LAT) gamma-ray data of the Virgo cluster, evidence for an extended emission associated with dark matter pair annihilation in the bb̄ channel has been reported by Han et al. (arxiv:1201.1003). After an in depth spatial and temporal analysis, we argue that the tentative evidence for a gamma-ray excess from the Virgo cluster is mainly due to the appearance of a population of previously unresolved gamma-ray point sources in the region of interest. These point sources are not part of the LAT second source catalogue (2FGL), but are found to be above the standard detection significance threshold when three or more years of LAT data is included.

Second, we confirm the detection of a spatially extended excess of 2-5 GeV gamma rays from the Galactic Center (GC), consistent with the emission expected from annihilating dark matter or an unresolved population of about 10³ milisecond pulsars. However, there are significant uncertainties in the diffuse galactic background at the GC. We have performed a revaluation of these two models for the extended gamma ray source at the GC by accounting for the systematic uncertainties of the Galactic diffuse emission model. We also marginalize over point source and diffuse background parameters in the region of interest. We show that the excess emission is significantly more extended than a point source. We find that the DM (or pulsars population) signal is larger than the systematic errors and therefore proceed to determine the sectors of parameter space that provide an acceptable fit to the data. We found that a population of order a 10³ MSPs with parameters consistent with the average spectral shape of Fermi-LAT measured MSPs was able to fit the GC excess emission. For DM, we found that a pure τ⁺τ⁻ annihilation channel is not a good fit to the data. But a mixture of τ⁻τ⁻ and bb̄ with a (σν) of order the thermal relic value and a DM mass of around 20 to 60 GeV provides an adequate fit.

We also consider the possibility that the GeV excess is due to nonthermal bremsstrahlung produced by a population of electrons interacting with neutral gas in molecular clouds. The millisecond pulsars and dark matter alternatives have spatial templates well fitted by the square of a generalized Navarro-Frenk-White (NFW) profile with inner slope γ = 1.2. We model the third option with a 20-cm continuum emission Galactic Ridge template. A template based on the HESS residuals is shown to give similar results. The gamma-ray excess is found to be best fit by a combination of the generalized NFW squared template and a Galactic Ridge template. We also find the spectra of each template is not significantly affected in the combined fit and is consistent with previous single template fits. That is, the generalized NFW squared spectrum can be fit by either of order 10³ unresolved MSPs or DM with mass around 30 GeV, a thermal cross section, and mainly annihilating to bb̄ quarks. While the Galactic Ridge continues to have a spectrum consistent with a population of nonthermal electrons whose spectrum also provides a good fit to synchrotron emission measurements. We also show that the current DM fit may be hard to test, even with 10 years of Fermi-LAT data, especially if there is a mixture of DM and MSPs contributing to the signal, in which case the implied DM cross section will be suppressed.