Limits on WIMP dark matter using scintillating CaWO4 cryogenic detectors with active background suppression

TL;DR

This paper reports the first significant WIMP limits achieved with scintillating CaWO4 cryogenic detectors that combine phonon and light measurements to suppress non-nuclear backgrounds. Two 300 g prototype modules were operated in LNGS for a total exposure of about 20.5 kg days, yielding 16 nuclear-recoil candidates in the 12–40 keV window, with a rate of 0.87 ± 0.22 events per kg day that is consistent with neutron background. By exploiting quenching factors, especially a high tungsten recoil factor (Q ≈ 40), the analysis isolates tungsten recoils and obtains a strong coherent WIMP scattering limit, including a zero-event observation in the tungsten region for the better module. The resulting limits are competitive with existing experiments in the 60 GeV/c^2 mass region, and the study highlights the potential of recoil-nucleus identification to further suppress backgrounds and verify a potential WIMP signal in future upgrades, including added neutron shielding and improved light-resolution detectors.

Abstract

We present first significant limits on WIMP dark matter by the phonon-light technique, where combined phonon and light signals from a scintillating cryogenic detector are used. Data from early 2004 with two 300 g CRESST-II prototype detector modules are presented, with a net exposure of 20.5 kg days. The modules consist of a CaWO4 scintillating ``target'' crystal and a smaller cryogenic light detector. The combination of phonon and light signals leads to a strong suppression of non-nuclear recoil backgrounds. Using this information to define an acceptance region for nuclear recoils we have 16 events from the two modules, corresponding to a rate for nuclear recoils between 12 and 40 keV of (0.87 +- 0.22) events/(kg day). This is compatible with the rate expected from neutron background, and most of these events lie in the region of the phonon-light plane anticipated for neutron-induced recoils. A particularly strong limit for WIMPs with coherent scattering results from selecting a region of the phonon-light plane corresponding to tungsten recoils, where the best module shows zero events.

Figures (11)

• Figure 1: Schematic representation of the detector for coincident phonon and light measurement. It consists of two cryogenic detectors enclosed in a highly reflective housing, read out by tungsten superconducting phase transition thermometers.
• Figure 2: Schematic thermal and electrical connections to the thermometer in the phonon channel together with elements of the external readout and heater circuits. (See text)
• Figure 3: Thermal and electrical connections to the sensor on the light detector. Aluminum/tungsten phonon collectors surround the tungsten thermometer. Connections shown are the aluminum bond wires for electrical connection to the SQUID and heater circuits and the gold bond wire for thermal contact to the heat bath (copper holder).
• Figure 4: The measured pulse height for heater test pulses on light detector BE13 as a function of time during the entire dark matter run. The energy scale on the right side refers to effective electron energy in the CaWO$_4$ crystal (see text). The actual energy seen by the light detector is about a factor of 100 smaller. The detector is seen to be stable within resolution.
• Figure 5: Peak with (3.2 $\pm$ 0.5) counts/day, of width 1.0 keV (FWHM) appearing in the phonon channel (Daisy) during the 53 day dark matter run. We attribute this peak to an $^{210}$Pb contamination in the vicinity of the detectors.