Compatibility of DAMA/LIBRA dark matter detection with other searches

TL;DR

This study reevaluates the DAMA/LIBRA annual modulation signal in light of null results from other direct-detection experiments, using the 36-bin DAMA data and incorporating ion-channeling effects within a standard isothermal halo model. Employing three statistical approaches—global likelihood, mass-by-mass raster scans, and a goodness-of-fit test—the authors map out the viable WIMP parameter space for spin-independent, spin-dependent, and mixed couplings, and contrast these with external constraints from CDMS, CoGeNT, CRESST I, TEXONO, XENON10, and SuperK. They find that, while the best-fit DAMA regions are largely incompatible with null results at high significance, there exist low-mass windows (and, for SD-proton and mixed couplings, some high-mass regions if indirect detection limits are relaxed) where DAMA can be reconciled with other data, especially when channeling is included. The analysis highlights the critical roles of detector effects, channeling, and external calibrations (notably XENON10 L_eff) in determining compatibility, and suggests that some parameter space remains testable with future data and improved modeling, including Migdal effects.

Abstract

The DAMA/NaI and DAMA/LIBRA annual modulation data, which may be interpreted as a signal for the existence of weakly interacting dark matter (WIMPs) in our galactic halo, are examined in light of null results from other experiments. We use the energy spectrum of the combined DAMA modulation data given in 36 bins, and include the effect of channeling. Several statistical tools are implemented in our study: likelihood ratio with a global fit and with raster scans in the WIMP mass and goodness-of-fit (g.o.f.). These approaches allow us to differentiate between the preferred (global best fit) and allowed (g.o.f.) parameter regions. It is hard to find WIMP masses and couplings consistent with all existing data sets. For spin-independent (SI) interactions, the best fit DAMA regions are ruled out to the 3$σ$ C.L., even with channeling taken into account. However, for WIMP masses of ~8 GeV some parameters outside these regions still yield a moderately reasonable fit to the DAMA data and are compatible with all 90% C.L. upper limits from negative searches, when channeling is included. For spin-dependent (SD) interactions with proton-only couplings, a range of masses below 10 GeV is compatible with DAMA and other experiments, with and without channeling, when SuperK indirect detection constraints are included; without the SuperK constraints, masses as high as ~20 GeV are compatible. For SD neutron-only couplings we find no parameters compatible with all the experiments. Mixed SD couplings are examined: e.g. ~8 GeV mass WIMPs with a_n = +/- a_p are found to be consistent with all experiments. In short, there are surviving regions at low mass for both SI and SD interactions; if indirect detection limits are relaxed, some SD proton-only couplings at high masses also survive.

Figures (23)

• Figure 1: Contraints on spin-independent (SI) scattering cross-sections for various experiments with null results. Cross-sections below each line are excluded by the given experiment at the 90% (solid), 3$\sigma$ (dashed), and 5$\sigma$ (dotted) confidence levels. The same coloring scheme will be used in later figures.
• Figure 2: DAMA modulation amplitude data as well as best fit spectra for various cases. Boxes represent the 1$\sigma$ bounds of the measured modulation amplitude for each bin: pink for the full 36 bin data set and blue for a 2 bin data set (2--4 & 6--14 keVee). Also shown are the $S_m$ spectra for the best fit SI cross-sections (in the SI only coupling case) at a WIMP mass of 4 GeV for the 2 bin data (blue) and 36 bin data (red), with (dashed) and without (solid) inclusion of the ion channeling (IC) effect. The $\chi^2$ values (over the degrees of freedom) are indicated in the legend. For comparison, the best (SI only) fit at any mass, which occurs around 80 GeV, is shown (green).
• Figure 3: Most likely parameters to produce the DAMA signal as determined using 2 bins (2--4 & 6-14 keVee; blue) and 36 bins (pink) for the case of SI scattering. The dark/light solid regions correspond to 90%/3$\sigma$ confidence regions; these regions are surrounded by 5$\sigma$ and 7$\sigma$ contours. The 2 bin data set provides very little constraint on the WIMP mass, while the 36 bin constrains the WIMP mass to two regions around 12 GeV and 80 GeV, corresponding to predominantly scattering off of Na and I, respectively. Null experiment constraints as described in Figure \ref{['fig:SIExpts']} are shown at 90% exclusion levels.
• Figure 4: Same as Figure \ref{['fig:Binning']}, but including the channeling effect for DAMA. The 36 bin low mass region around 12 GeV is predominantly due to scatters off of I that undergo the channeling effect, as opposed to scatters off of Na as in the no ion channeling case shown in Figure \ref{['fig:Binning']}; ordinary Na scatters and channeled I scatters coincidentally provide good fits at similar WIMP masses.
• Figure 5: Experimental constraints and DAMA best fit parameters for SI only scattering. The DAMA best fit regions are determined using the likelihood ratio method with (green) and without (orange) the channeling effect.