The Inelastic Frontier: Discovering Dark Matter at High Recoil Energy

There exist well motivated models of particle dark matter which predominantly scatter inelastically off nuclei in direct detection experiments. This inelastic transition causes the DM to up-scatter in terrestrial experiments into an excited state up to 550 keV heavier than the DM itself. An inelastic transition of this size is highly suppressed by both kinematics and nuclear form factors. We extend previous studies of inelastic DM to determine the present bounds on the scattering cross section, and the prospects for improvements in sensitivity. Three scenarios provide illustrative examples: nearly pure Higgsino DM; magnetic inelastic DM; and inelastic models with dark photon exchange. We determine the elastic scattering rate as well as verify that exothermic transitions are negligible. Presently, the strongest bounds on the cross section are from xenon at LUX-PandaX (\delta < 160 keV), iodine at PICO (160 < \delta < 300 keV), and tungsten at CRESST (when \delta > 300 keV). Amusingly, once \delta > 200 keV, weak scale (and larger) DM - nucleon scattering cross sections are allowed. The relative competitiveness of these experiments is governed by the upper bound on the recoil energies employed by each experiment, as well as strong sensitivity to the mass of the heaviest element in the detector. Several implications, including sizable recoil energy-dependent annual modulation, and improvements for future experiments are discussed. We show that the xenon experiments can improve on the PICO results, if they were to analyze their existing data over a larger range of recoil energies, i.e., 20-500 keV. We also speculate about several reported events at CRESST between 45-100 keV, that could be interpreted as inelastic DM scattering. Future data from PICO, CRESST and xenon experiments can test this with anaylses of high energy recoil data.