Electron and proton acceleration efficiency by merger shocks in galaxy clusters

Radio relics in galaxy clusters are associated with powerful shocks that (re)accelerate relativistic electrons. It is widely believed that the acceleration proceeds via diffusive shock acceleration. In the framework of thermal leakage, the ratio of the energy in relativistic electrons to the energy in relativistic protons should be smaller than K<inf>e/p</inf> ~ 10^-2. The relativistic protons interact with the thermal gas to produce γ -rays in hadronic interactions. Combining observations of radio relics with upper limits from γ -ray observatories can constrain the ratio K<inf>e/p</inf>. In this work, we selected 10 galaxy clusters that contain double radio relics, and derive new upper limits from the stacking of γ -ray observations by Fermi.We modelled the propagation of shocks using a semi-analytical model, where we assumed a simple geometry for shocks and that cosmic ray protons are trapped in the intracluster medium. Our analysis shows that diffusive shock acceleration has difficulties in matching simultaneously the observed radio emission and the constraints imposed by Fermi, unless the magnetic field in relics is unrealistically large (≫ 10 μG). In all investigated cases (also including realistic variations of our basic model and the effect of re-acceleration), the mean emission of the sample is of the order of the stacking limit by Fermi, or larger. These findings put tension on the commonly adopted model for the powering of radio relics, and imply that the relative acceleration efficiency of electrons and protons is at odds with predictions of diffusive shock acceleration, requiring K<inf>e/p</inf> ≥ 10 - 10^-2.