Final results of the EDELWEISS-I dark matter search with cryogenic heat-and-ionization Ge detectors

TL;DR

This paper reports the final EDELWEISS-I results from a 62 kg·d fiducial exposure using cryogenic Ge detectors that measure heat and ionization. The analysis achieves a recoil threshold below 13 keV across three detectors and derives spin-independent WIMP–nucleon cross-section limits without background subtraction, noting residual neutron and surface-electron backgrounds. The study validates detector calibration, efficiency modeling, and data-combination across runs, and it informs the design and goals of the next phase, EDELWEISS-II, including a larger detector array, improved shielding, and surface-event tagging technologies. Collectively, the results refine the sensitivity landscape for WIMP searches in the early 2000s and demonstrate preparation for scale-up to higher exposures.

Abstract

The final results of the EDELWEISS-I dark matter search using cryogenic heat-and-ionization Ge detectors are presented. The final data sample corresponds to an increase by a factor five in exposure relative to the previously published results. A recoil energy threshold of 13 keV or better was achieved with three 320g detectors working simultaneously over four months of stable operation. Limits on the spin-independent cross-section for the scattering of a WIMP on a nucleon are derived from an accumulated fiducial exposure of 62 kg.d.

Figures (16)

• Figure 1: Example of filtered heat and ionization pulses for $\sim$ 10 keV$_{ee}$ signals (full lines) together with the template fit (dashed lines) for an ionization (center electrode) signal (a) and for the corresponding heat signal (b). In (c) is shown an example of a NTD event (see text) together with the template fits for a normal heat signal (dashed line) and for a NTD signal (dash-dotted line)
• Figure 2: Low-energy part of the spectrum recorded in the fiducial volume of the three detectors. The energy is calculated as the sum of the ionization and heat channels, weighted by their resolution squared. The peaks at 8.98 and 10.37 keV correspond to the de-excitation of the cosmogenic activation of $^{65}$Zn and $^{68}$Ge in the detectors, and the $^{71}$Ge activation that follows neutron calibrations. The lines correspond to a Gaussian fit with the indicated value of FWHM resolutions.
• Figure 3: Measurement of the efficiency as a function of recoil energy for the detector GGA3 in the run 2003p configuration, using neutron-coincidence data from a $^{252}$Cf calibration. Top: spectrum as a function of energy with different cuts. Dotted: minimum bias (selection based only on the presence of a neutron in the other two detectors); dot-dashed: adding the condition that the heat signal is above threshold; dashed: adding the 2.5 keV$_{ee}$ ionization cut; full line: adding the $\pm$ 1.65$\sigma$ and $<$$-$3.29$\sigma$ nuclear and electron recoil requirements (see text). Bottom: resulting efficiency as a function of energy. The maximum value is not 90 % because the data are not corrected for the effect of neutron-$\gamma$ coincidences.
• Figure 4: Baseline FWHM resolution on the heat channel of the three detectors in the run 2003i as a function of time in hours since the beginning of the run. The resolution is evaluated by 3-hours intervals centered on each hour. The dotted lines represent times when the cryostat was re-filled with liquid He. The full lines represent FWHM cuts.
• Figure 5: Baseline FWHM resolution on the heat channel of the three detectors in the run 2003p as a function of time in hours since the beginning of the run. The resolution is evaluated by 3-hours intervals centered on each hour.