Poker Face of Inelastic Dark Matter: Prospects at Upcoming Direct Detection Experiments
5/28/2010
26 citations. Part of the composite inelastic dark matter program.
The Problem
The DAMA/LIBRA experiment had reported an 8.9σ annual modulation signal consistent with dark matter for over a decade, but every other direct detection experiment saw nothing. Inelastic dark matter, where the dark matter scatters into an excited state split by roughly 100 keV, could explain both observations simultaneously: the kinematics favor scattering off heavy nuclei like iodine (in DAMA’s NaI crystals) while suppressing scattering off lighter targets used by other experiments. By 2010, XENON100 and CRESST were about to come online with the sensitivity to directly test this explanation. The question was whether the predictions were sharp enough to be falsifiable, or whether uncertainties in the quenching factor, the dark matter velocity distribution, and the interaction structure left too much room to maneuver.
The Key Idea
The paper systematically maps how predictions for XENON100 and CRESST depend on the three main sources of uncertainty: quenching factor measurements (which translate nuclear recoil energy into the observable signal), the details of the dark matter-Standard Model interaction (which affect the relative rates at different targets), and the halo velocity distribution (which shifts the kinematics of inelastic scattering). When all uncertainties are accounted for, the number of expected events at CRESST and XENON100 varies by an order of magnitude. The inelastic dark matter explanation for DAMA was testable, but not as cleanly as optimistic projections suggested. The title, “Poker Face,” captures the point: the upcoming experiments would constrain the scenario, but the model had enough flexibility to absorb a range of outcomes without being definitively killed. This kind of honest uncertainty quantification, identifying exactly how much wiggle room a model has, was characteristic of the composite inelastic dark matter program’s approach to the DAMA anomaly.
Impact
The paper’s predictions were tested within a year. XENON100 published its first results in 2011, and CRESST followed shortly after. The inelastic dark matter explanation for DAMA survived the first round of data, as the paper had anticipated it might given the uncertainty ranges, but subsequent runs with larger exposures progressively squeezed the viable parameter space. By the mid-2010s, the standard inelastic dark matter scenario for DAMA was effectively excluded by a combination of XENON, LUX, and CRESST constraints, though the DAMA modulation signal itself persisted (and still persists, unresolved). The paper’s contribution was methodological: it established the framework for asking “how testable is this model, really?” rather than making overconfident predictions in either direction. The approach of mapping uncertainties in quenching factors, halo models, and interaction types became standard practice in the direct detection community’s handling of anomalies. The broader composite inelastic dark matter program, of which this was the experimental-facing component, fed into the development of the SIMP paradigm, where the idea that dark matter has internal structure and rich dynamics carried forward even after the specific DAMA application was ruled out.