Constraints on WIMP and Sommerfeld-Enhanced Dark Matter Annihilation from HESS Observations of the Galactic Center

We examine the constraints on models of weakly interacting massive particle (WIMP) dark matter from the recent observations of the Galactic center by the High Energy Spectroscopic System (HESS) telescope. We analyze canonical WIMP annihilation into Standard Model particle final states, including b/b-bar, t/t-bar and W^+ W^-. The constraints on annihilation into b/b-bar is within an order of magnitude of the thermal cross section at ~3 TeV, while the tau^+/tau^- channel is within a factor of ~2 of thermal. We also study constraints on Sommerfeld-enhanced dark matter annihilation models, and find that the gamma-ray observational constraints here rule out all of the parameter space consistent with dark matter annihilation interpretations of PAMELA and the Fermi-LAT e^+/e^- spectrum, in specific classes of models, and strongly constrains these interpretations in other classes. The gamma-ray constraints we find are more constraining on these models, in many cases, than current relic density, cosmic microwave background, halo shape and naturalness constraints.

💡 Research Summary

This paper investigates the constraints that can be placed on weakly interacting massive particle (WIMP) dark‑matter models, as well as on models featuring a Sommerfeld‑enhanced annihilation cross‑section, using recent very‑high‑energy gamma‑ray observations of the Galactic Center (GC) performed by the High Energy Stereoscopic System (HESS). The authors first describe the motivation for indirect dark‑matter searches: WIMPs that thermally freeze out in the early universe naturally yield a relic density consistent with cosmological measurements, and their annihilation into Standard Model particles (quarks, gauge bosons, leptons) produces energetic gamma rays through hadronization, final‑state radiation, and decay of neutral pions. HESS, with sensitivity from a few hundred GeV up to tens of TeV, is ideally suited to probe the high‑mass regime where other instruments (e.g., Fermi‑LAT) lose sensitivity.

The data set consists of 112 hours of live time collected with zenith angles below 30°, focused on the inner 1° around the GC. A reflected‑background technique is employed: the signal region is defined by a grid of 0.02° × 0.02° pixels within 1° of the GC, while background pixels are generated by rotating the signal region around the telescope pointing direction by 90°, 180°, and 270°. Regions contaminated by the Galactic plane (|b| < 0.3°) and the known source HESS J1745‑303 are masked. No statistically significant excess is observed in the signal region relative to the background.

To translate this null result into limits on the annihilation cross‑section, the authors adopt two dark‑matter density profiles that are commonly used in the literature: a non‑adiabatically contracted Navarro‑Frenk‑White (NFW) profile and an Einasto profile. Both profiles predict a higher line‑of‑sight integral of the density squared (the J‑factor) in the signal region than in the background region, which is essential for the reflected‑background method to be sensitive. The adopted J‑factors, taken from the HESS collaboration analysis, are ⟨J_NFW^s⟩ = 1604, ⟨J_NFW^b⟩ = 697, ⟨J_Einasto^s⟩ = 3142, and ⟨J_Einasto^b⟩ = 1535 (dimensionless after normalization). The authors note that if the Milky Way possessed a large constant‑density core (e.g., a Burkert or cored‑isothermal profile extending beyond ~450 pc), the signal and background J‑factors would be essentially equal and the method would lose constraining power. Current high‑resolution N‑body simulations, however, generally favor cuspy profiles, especially when baryonic contraction is taken into account, justifying the use of NFW/Einasto as a conservative baseline.

The gamma‑ray spectra for each annihilation channel are generated with PYTHIA 6.4. For two‑body final states (b b̄, t t̄, W⁺W⁻) the simulation includes parton showering, hadronization, and π⁰ decay, yielding a broad continuum extending up to the WIMP mass. For four‑lepton final states that arise when annihilation proceeds via a light dark‑force mediator (e.g., χχ → φφ → 4e or 4μ), the authors turn off initial‑state radiation and only allow photons from final‑state radiation and meson decays, which dramatically reduces the high‑energy photon yield. When the mediator mass is very low (m_φ < 0.5 GeV) PYTHIA’s low‑energy cutoffs become problematic; in those cases the authors adopt analytic formulas from the literature (e.g., Ref.