Resonant Scattering between Dark Matter and Baryons: Revised Direct Detection and CMB Limits

Abstract: Traditional dark matter models, eg. WIMPs, assume dark matter is weakly coupled to the standard model so that elastic scattering between dark matter and baryons can be described perturbatively by Born approximation. Most direct detection experiments are analyzed according to that assumption. We show that when the fundamental DM-baryon interaction is attractive, dark matter-nucleus scattering is non-perturbative in much of the relevant parameter range. The cross section exhibits rich resonant behavior with a highly non-trivial dependence on atomic mass; furthermore, the extended rather than point-like nature of nuclei significantly impacts the cross sections. The repulsive case also requires full numerical calculation. These non-perturbative effects change existing constraints. Near a resonance value of the parameters, the cross section has non-trivial velocity dependence rather than the usual $\sigma \sim v^0$; we take the velocity dependence into account. (However doing so has little impact on current constraints.) We report the corrected exclusion regions superseding previous limits from XQC, CRESST Surface Run, CMB power spectrum and extensions with Lyman-$\alpha$ and Milky Way satellites, and Milky Way gas clouds. Some limits become weaker than previous bounds in the literature, while others become stronger. Gaps which open by correct treatment of some particular constraint can sometimes be closed using a different constraint. We also discuss the dependence on mediator mass and give approximate expressions for the velocity dependence near a resonance. Sexaquark ($uuddss$) DM with mass around 2 GeV, which exchanges QCD mesons with baryons, remains unconstrained for most of the parameter space of interest. A statement in the literature that a DM-nucleus cross section larger than $10^{-25}\,{\rm cm}^2$ implies dark matter is composite, is corrected.