Results from 730 kg days of the CRESST-II Dark Matter Search

TL;DR

CRESST-II reports results from 730 kg·days of data in a CaWO$_4$-based cryogenic detector, using a two-channel phonon-light readout to discriminate backgrounds and search for low-energy nuclear recoils from WIMPs. A comprehensive maximum likelihood analysis, accounting for $e/oldsymbol{ extgamma}$ leakage, $oldsymbol{ extalpha}$ background, neutron scatters, and Pb recoils, finds 67 events in the WIMP acceptance region whose distribution cannot be fully explained by these backgrounds alone, with a significance exceeding $4\sigma$ for a potential WIMP signal. The analysis yields two favored WIMP parameter sets: $m_\chi=25.3$ GeV with $\sigma_{\text{WN}}=1.6\times 10^{-6}$ pb, and $m_\chi=11.6$ GeV with $\sigma_{\text{WN}}=3.7\times 10^{-5}$ pb, and places them in context with other direct-detection limits. The authors outline concrete improvements for the next run to suppress dominant backgrounds, aiming to clarify the presence of a WIMP signal and to tighten the inferred parameter space.

Abstract

The CRESST-II cryogenic Dark Matter search, aiming at detection of WIMPs via elastic scattering off nuclei in CaWO$_4$ crystals, completed 730 kg days of data taking in 2011. We present the data collected with eight detector modules, each with a two-channel readout; one for a phonon signal and the other for coincidently produced scintillation light. The former provides a precise measure of the energy deposited by an interaction, and the ratio of scintillation light to deposited energy can be used to discriminate different types of interacting particles and thus to distinguish possible signal events from the dominant backgrounds. Sixty-seven events are found in the acceptance region where a WIMP signal in the form of low energy nuclear recoils would be expected. We estimate background contributions to this observation from four sources: 1) "leakage" from the e/γ-band 2) "leakage" from the α-particle band 3) neutrons and 4) Pb-206 recoils from Po-210 decay. Using a maximum likelihood analysis, we find, at a high statistical significance, that these sources alone are not sufficient to explain the data. The addition of a signal due to scattering of relatively light WIMPs could account for this discrepancy, and we determine the associated WIMP parameters.

Figures (13)

• Figure 1: Contributions of the three types of nuclei present in a CaWO$_4$ target to the total rate of WIMP interactions, as a function of the WIMP mass and for a cross-section of [1]pb, assuming coherent $\sim A^2$ interactions. The calculation assumes a detector with a sensitive energy range between 12 and [40]keV.
• Figure 2: Schematic drawing of a CRESST detector module, consisting of the target crystal and an independent light detector. Both are read out by transition edge sensors (TES) and are enclosed in a common reflective and scintillating housing.
• Figure 3: Illustration of background events due to surface contaminations with $^{210}$Po.
• Figure 4: Schematic drawing of the CRESST setup. A cold finger (CF) links the cryostat (CR) to the experimental volume, where the detectors are arranged in a common support structure, the so-called carousel (CA). This volume is surrounded by layers of shielding from copper (CU), lead (PB), and polyethylene (PE). The copper and lead shieldings are additionally enclosed in a radon box (RB). An active muon veto (MV) tags events which are induced by cosmic radiation.
• Figure 5: Data obtained with one detector module in a calibration measurement with an AmBe neutron source, with the source placed outside the lead shielding. The solid red lines mark the boundary of the calculated oxygen recoil band ([10]% of events are expected above the upper and [10]% below the lower boundary). The vertical dashed lines indicate the lower and upper energy bounds of the WIMP acceptance region as will be introduced in Section \ref{['sec:latestrun']}.