Constraining Extended Gamma-ray Emission from Galaxy Clusters
Cold dark matter models predict the existence of a large number of substructures within dark matter halos. If the cold dark matter consists of weakly interacting massive particles, their annihilation within these substructures could lead to diffuse GeV emission that would dominate over the annihilation signal of the host halo. In this work we search for GeV emission from three nearby galaxy clusters: Coma, Virgo and Fornax. We first remove known extragalactic and galactic diffuse gamma-ray backgrounds and point sources from the Fermi 2-year catalog and find a significant residual diffuse emission in all three clusters. We then investigate whether this emission is due to (i) unresolved point sources; (ii) dark matter annihilation; or (iii) cosmic rays (CR). Using 45 months of Fermi-LAT data we detect several new point sources (not present in the Fermi 2-year point source catalogue) which contaminate the signal previously analyzed by Han et al.(arxiv:1201.1003). Including these and accounting for the effects of undetected point sources, we find no significant detection of extended emission from the three clusters studied. Instead, we determine upper limits on emission due to dark matter annihilation and cosmic rays. For Fornax and Virgo the limits on CR emission are consistent with theoretical models, but for Coma the upper limit is a factor of 2 below the theoretical expectation. Allowing for systematic uncertainties associated with the treatment of CR, the upper limits on the cross section for dark matter annihilation from our clusters are more stringent than those from analyses of dwarf galaxies in the Milky Way. We rule out the thermal cross section for supersymmetric dark matter particles for masses as large as 100 GeV (depending on the annihilation channel).
💡 Research Summary
This paper investigates whether diffuse GeV‑scale gamma‑ray emission from three nearby galaxy clusters—Coma, Virgo, and Fornax—can be attributed to dark‑matter (DM) annihilation, cosmic‑ray (CR) interactions, or unresolved point sources. The authors begin by modeling and subtracting all known point sources from the Fermi‑LAT 2‑year catalog, together with the Galactic and extragalactic diffuse backgrounds, from 45 months of LAT data. After this cleaning, a residual extended emission remains in each cluster, reproducing the signal reported by Han et al. (2012).
A key advance of the present work is the identification of several new point sources that were not present in the 2‑year catalog. These sources lie near the cluster centers and, if ignored, would artificially inflate any extended emission. The authors therefore construct an updated source model that includes the newly detected objects and also statistically accounts for the contribution of still‑undetected sub‑threshold sources. This is achieved by Monte‑Carlo simulations of the LAT point‑source detection process, using the measured source‑count distribution (dN/dS) to assign a probability for a faint source to remain hidden. The resulting “source‑confusion” correction reduces the apparent extended flux to a level consistent with statistical fluctuations.
With the cleaned data set, the authors test two physical hypotheses. First, they model the gamma‑ray yield from WIMP annihilation for several canonical channels ($b\bar b$, $\tau^+\tau^-$, $W^+W^-$) and adopt a range of substructure boost factors derived from high‑resolution N‑body simulations (Aquarius, Via Lactea). Second, they compute the expected gamma‑ray emission from CR protons interacting with the intra‑cluster medium, using gas density profiles from X‑ray observations and CR pressure models calibrated on simulations of cluster formation.
The analysis yields no statistically significant detection of extended emission in any of the three clusters. Consequently, the authors set 95 % confidence upper limits on both the DM annihilation cross‑section and the CR‑induced gamma‑ray luminosity. For Fornax and Virgo, the CR limits are compatible with theoretical expectations, whereas for Coma the upper limit is roughly a factor of two below the predicted CR flux, suggesting either an over‑estimate in the CR model or a lower CR content in this particular cluster.
When the DM limits from the three clusters are combined, the resulting constraints on the velocity‑averaged annihilation cross‑section are more stringent than those obtained from dwarf‑spheroidal galaxy observations for WIMP masses up to ∼100 GeV, depending on the annihilation channel. In particular, the thermal relic cross‑section ($\langle\sigma v\rangle\sim3\times10^{-26},\mathrm{cm^{3},s^{-1}}$) is excluded for supersymmetric WIMP masses as high as 100 GeV.
The paper emphasizes that accurate treatment of unresolved point sources and realistic CR modeling are essential for any claim of extended gamma‑ray emission from clusters. It also points out that future LAT data releases, with longer exposure and improved event reconstruction, together with next‑generation ground‑based gamma‑ray observatories such as the Cherenkov Telescope Array, will further tighten these constraints and may eventually detect the faint DM‑induced signal if it exists.