Search for a Dark Matter-Induced Annual Modulation Signal in NaI(Tl) with the COSINE-100 Experiment

We present new constraints on the dark matter-induced annual modulation signal using 1.7 years of COSINE-100 data with a total exposure of 97.7 kg yr. The COSINE-100 experiment, consisting of 106 kg of NaI(Tl) target material, is designed to carry out a model-independent test of DAMA/LIBRA’s claim of WIMP discovery by searching for the same annual modulation signal using the same NaI(Tl) target. The crystal data show a $2.7\text{cpd/kg/keV}$ background rate on average in the 2–6 keV energy region of interest. Using a $\chi$-squared minimization method we observe best fit values for modulation amplitude and phase of $0.0092\pm 0.0067\text{cpd/kg/keV}$ and $127.2\pm 45.9\mathrm{d}$, respectively.

Figure 1

COSINE-100’s environmental parameters as a function of time. (a) Detector room and near-crystal temperature. (b) Relative humidity for the detector room and the top volume of acrylic box, at the top of the LS. Note that the measurement taken at the top of the LS began on day 450. (c) The radon level in the detector room air. (d) Rate of muons passing through the detector over time. Here, the rate is binned in 30-day intervals.

Figure 2

Efficiency-corrected and time-integrated energy spectra for the five crystals used in this analysis between 2–20 keV (top panels) and signal selection efficiency evaluated using ${}^{60}\mathrm{Co}$ calibration data (bottom panel). The efficiencies at 2 keV are $>60\%$ for all crystals. The primary sources of background in the crystals are ${}^{210}\mathrm{Pb}$ and ${}^{40}\mathrm{K}$, which are lower for Crystal 6 and Crystal 7. These spectra are obtained using the full dataset considered in this analysis.

Figure 3

Rate vs time for Crystals 2, 3, 4, 6, and 7 from October 21, 2016 to July 18, 2018 for the 2–6 keV energy region binned in 15-day intervals. The histograms show the result of the fit described in the text. Solid blue arrows indicate the peak date in the modulation as reported by DAMA/LIBRA [12]. Data taking was suspended for calibrations at the end of 2016 as indicated by the shaded region.

Figure 4

The COSINE-100 best fit and 68.3%, 95.5%, and 99.7% C.L. contours as functions of modulation amplitude and phase relative to January 1, for a fixed period of 365.25 d. A Feldman-Cousins technique is used as a crosscheck and resultant 68.3% C.L. is shown. The amplitude and phase reported by DAMA/LIBRA in the 2–6 keV energy interval with statistical uncertainties (blue cross) and the phase expected from the standard halo model (June 2) are overlaid for comparison. Top and side panels show the dependence of $\mathrm{\Delta }{\chi }^2$ on phase and amplitude, respectively, along with two-sided significance levels.

Figure 5

Modulation amplitude as a function of energy in 1 keV bins for the 1.7 year COSINE-100 single-hit (red closed circle) and multiple-hit (orange open circle) events. DAMA/LIBRA phase 1 (blue) and phase 2 (green) from Ref. [12] are also shown for reference. The period and phase are fixed at 365.25 and 152.5 d. Horizontal error bars represent the width of the energy bins used for the analysis. Vertical error bars are $\pm 1\sigma$ errors on the binned modulation fit amplitudes.