Annual modulation results from three-year exposure of ANAIS-112

ANAIS (annual modulation with NaI scintillators) is a dark matter direct detection experiment consisting of 112.5 kg of NaI(Tl) detectors in operation at the Canfranc Underground Laboratory (LSC), in Spain, since August 2017. ANAIS’ goal is to confirm or refute in a model independent way the DAMA/LIBRA positive result: an annual modulation in the low-energy detection rate having all the features expected for the signal induced by dark matter particles in a standard galactic halo. This modulation, observed for about 20 years, is in strong tension with the negative results of other very sensitive experiments, but a model-independent comparison is still lacking. By using the same target material, NaI(Tl), such a comparison is more direct and almost independent in dark matter particle and halo models. Here, we present the annual modulation analysis corresponding to three years of ANAIS data (for an effective exposure of $313.95\mathrm{kg}\times \mathrm{y}$), applying a blind procedure, which updates the one developed for the 1.5 years analysis, and later applied to 2 years. The analysis also improves the background modeling in the fitting of the region of interest rates. We obtain for the best fit in the [1–6] keV ([2–6] keV) energy region a modulation amplitude of $-0.0034\pm 0.0042\mathrm{cpd}/\mathrm{kg}/\mathrm{keV}$ ($0.0003\pm 0.0037\mathrm{cpd}/\mathrm{kg}/\mathrm{keV}$), supporting the absence of modulation in our data, and incompatible with the DAMA/LIBRA result at 3.3 (2.6) $\sigma$, for a sensitivity of 2.5 (2.7) $\sigma$. Moreover, we include two complementary analyses: a phase-free annual modulation search and the exploration of the possible presence of a periodic signal at other frequencies. Finally, we carry out several consistency checks of our result, and we update the ANAIS-112 projected sensitivity for the scheduled 5 years of operation.

Figure 1

Left panel: Artistic view of ANAIS-112 setup showing the 9 NaI(Tl) modules inside a shielding made of lead, antiradon box, active muon vetoes, polyethylene, and water tanks. Right panel: the black line represents the accumulated exposure since August 2017 with the nine ANAIS-112 modules. For comparison, the red line corresponds to a 100% live time for 112.5 kg detection mass.

Figure 2

The ANAIS-112 trigger rate from August 2017 until August 2020. Total rate is shown in black, and the rate of events having more than one peak in every PMT detected by the peak-finding algorithm is shown in red.

Figure 3

Evolution of the PMT gain provided by the SER for PMT-0 (upper panel) and PMT-1 (lower panel) coupled to each module. The HV setting of D4 and D5’s PMTs was modified after the first year of data taking.

Figure 4

Evolution of the light collection in ANAIS-112 modules, shown separately for each PMT: PMT-0 values are shown in red and PMT-1 values in blue (lower panels); total values are shown in black (upper panels). In all the figures shown in this article, the data from the nine modules will be shown in the same way: upper panels from left to right correspond to modules D0, D1, and D2; middle row from left to right D3, D4, and D5, lower row from left to right D6, D7, and D8.

Figure 5

Evolution during the three years of data taking of the position of the three calibration lines (determined in $\mathrm{mV}\times \mathrm{ns}$ units) from ${}^{109}\mathrm{Cd}$ in the ANAIS-112 modules, shown as relative deviation with respect to the average values. The relative position of the lines 11.9, 22.6, and 88.0 keV are shown in black, red, and blue solid dots, respectively. Slow drifts are observed in most of the modules, but periodical calibration allows for correcting those drifts. Only the modules D4 and D5 showed strong instability during the first year of operation. Then the PMT operation high voltage (HV) was changed just before starting the second year of measurement; because of that, for D4 and D5 in this figure, different average values have been considered for the periods before and after the HV change.

Figure 6

Evolution of the muon rate on each side of the ANAIS-112 scintillator veto system (top panel) and the rate of coincidences on two sides (lower panel) on a monthly basis since August 2017. N, S, W and E stand for north, south, west, and east sides of the setup.

Figure 7

Total efficiency, $ϵ(E,d)$, for event detection in all the ANAIS-112 modules as a function of energy. It has been obtained by combining the trigger efficiency, the asymmetry cut efficiency, and the pulse shape cut efficiency, all of them recalculated for the full analyzed exposure corresponding to three years by following the same procedure established in [22].

Figure 8

Time evolution of the rate of events corresponding to ${}^{40}\mathrm{K}$ (upper panel) and ${}^{22}\mathrm{Na}$ (lower panel) at low energy identified by the coincidence with the corresponding HE gamma in a second module.

Figure 9

ANAIS-112 full exposure background after unblinding in the low (upper panel) and high (bottom panel) energy regions. Inset in upper panel shows the background in the ROI. Data correspond to M1 events surviving our filter protocols and corrected with the efficiency. Background model estimate is depicted in green.

Figure 10

Time evolution of the rate corresponding to M1 events above the ROI (upper panel) and M2 events in the ROI (lower panel). Both rates show exponential decays in time with very different “effective” lifetimes (fit is shown in red). Estimates of our background model are also shown as green solid lines, conveniently normalized by a factor, f, displayed in both panels.

Figure 11

Evolution in time predicted by our background model in the ROI, [1–6] keV energy region, for every ANAIS-112 module.

Figure 12

Upper panels: ANAIS-112 fit results for three years of data in the [1–6] keV (left) and [2–6] keV (right) energy regions, both in the modulation (blue) and null hypothesis (red) when the background is described by Eq. (4). Lower panels: same, but using the background described by Eq. (5). Best fits for $S_m$, ${\chi }^2$ and p values are also shown.

Figure 13

Results of the fit of the nine modules data using Eq. (6) in the [1–6] keV energy region, in the modulation (blue) and null hypothesis (red). Best fits for $S_m$, ${\chi }^2$ and p values are also shown.

Figure 14

Same as Fig. 13 for [2–6] keV energy region.

Figure 15

Comparison between ANAIS-112 results on annual modulation using three years of data and DAMA/LIBRA modulation best fit. Estimated sensitivity is shown at different confidence levels as colored bands: green at $1\sigma$, yellow at $2\sigma$, and cyan at $3\sigma$.

Figure 16

Best fit values for the three methods in the [1–6] keV (upper panel) and [2–6] keV (lower panel) energy regions for different choices of the time binning. Circles, squares, and triangles correspond to fits to Eqs. (4), (5), and (6), respectively.

Figure 17

Black points: best fit results for three years of data in [1–6] keV (upper panels) and [2–6] keV (lower panels) energy regions in the ($S_m$, $t_0$) plane for the three fitting procedures (left, middle, and right panels correspond to best fits to functions (4), (5), and (6), respectively). Exclusion contours at 1, 2, and $3\sigma$ are depicted as blue solid, dotted, and dashed lines. DAMA/LIBRA results are shown for comparison (red cross). 1 and $2\sigma$ biased contours extracted from the MC simulation for each fit are shown as solid green and yellow regions.

Figure 18

Distribution of the best fit modulation amplitudes for a large set of MC experiments generated for [1–6] keV with $S_m=0$ and fitted to Eq. (6) leaving free both, $S_m$ and $t_0$. In red, mean value of every lobe of the bimodal distribution.

Figure 19

${\chi }_0^2-{\chi }^2$ periodograms for three years of data in [1–6] keV (upper panels) and [2–6] keV (lower panels) energy regions. Left, middle, and right panels correspond to best fits to functions (4), (5), and (6), respectively. Local and global significances are calculated as described in the text.

Figure 20

ANAIS-112 sensitivity to the DAMA/LIBRA signal in $\sigma$ C.L. units (see text) as a function of real time in the [1–6] keV (lower panel) and [2–6] keV (upper panel) energy regions. The black dots are the sensitivities measured experimentally. The blue bands represent the 68% C.L. DAMA/LIBRA uncertainty.