Characterization of inverted coaxial 76 Ge detectors in GERDA for future double- β decay experiments

Abstract

Neutrinoless double- $\beta$ decay of ${}^{76}$ Ge is searched for with germanium detectors where source and detector of the decay are identical. For the success of future experiments it is important to increase the mass of the detectors. We report here on the characterization and testing of five prototype detectors manufactured in inverted coaxial (IC) geometry from material enriched to 88% in ${}^{76}$ Ge. IC detectors combine the large mass of the traditional semi-coaxial Ge detectors with the superior resolution and pulse shape discrimination power of point contact detectors which exhibited so far much lower mass. Their performance has been found to be satisfactory both when operated in vacuum cryostat and bare in liquid argon within the Gerda setup. The measured resolutions at the Q-value for double- $\beta$ decay of ${}^{76}$ Ge ( $Q_{\beta \beta }$ = 2039 keV) are about 2.1 keV full width at half maximum in vacuum cryostat. After 18 months of operation within the ultra-low background environment of the GERmanium Detector Array (Gerda) experiment and an accumulated exposure of 8.5 kg $·$ year, the background index after analysis cuts is measured to be $4.9_{-3.4}^{+7.3}\times {10}^{-4}\text{counts}/(\text{keV}·\text{kg}·\text{year})$ around $Q_{\beta \beta }$ . This work confirms the feasibility of IC detectors for the next-generation experiment Legend.

Fig. 1

Left: Main IC detector features. Middle: ADL calculation of the weighting potential. Right: Electric field strength in kV/cm. The minimum required electric field is 200 V/cm (dark blue). The black dashed lines show electron drift paths ending at the n+ contact while white solid lines are the hole drift paths reaching the signal contact

Fig. 2

Configurations used for detector characterization. Left: Setup for depletion voltage estimation with a mixed 1.5 kBq source of ${}^{60}$Co–${}^{137}$Cs–${}^{241}$Am; bias and readout circuits are indicated. Middle: Setup for PSD studies with a flood top and side 13 kBq ${}^{228}$Th source and with a collimated 250 kBq ${}^{228}$Th source for lateral scans. The vacuum cryostat (cyan) and detector holder (orange), both made of aluminum, are added here for illustration. Right: Setup for scans with the collimated 4.3 MBq ${}^{241}$Am source: lateral at 3 azimuthal angles (dashed lines), 2 orthogonal directions on top (solid lines) and a circular one (dotted lines)

Fig. 3

Waveform examples of SSEs (top) and MSEs (bottom) of detector 50A after applying a moving window average. The amplitude of the maximum current A and of energy E are explicitly shown. The boundaries of the rise time are estimated at 0.5 and 90% of the maximum charge amplitude

Fig. 4

Spectrum taken with detector 50A and the mixed source of ${}^{60}$Co, ${}^{137}$Cs, ${}^{241}$Am for the determination of the nominal bias voltage. The inset shows the fit to the ${}^{60}$Co 1333 keV $\gamma$ line. The MCA module was used for this measurement using a Gaussian energy filter, thus explaining the significant tail of the $\gamma$ line from ballistic deficit

Fig. 5

Energy resolution FWHM of the ${}^{60}$Co 1333 keV line as a function of the applied bias voltage. The arrows show depletion voltages and resolutions reported by the manufacturer. The statistical uncertainties are less than the widths of the markers

Fig. 6

Energy resolution FWHM as a function of $\gamma$ ray energy; data are taken at the bias voltage recommended by the manufacturer. The dashed lines show fits to the data, performed for each detector separately. The statistical uncertainties are less than the widths of the markers

Fig. 7

Example of ${}^{241}$Am FEP data recorded with the highly collimated 4.3 MBq source, positioned above the upper surface of detector 50A. The model (red line) is shown together with its decomposition into a Gaussian, a tail and a shoulder functions

Fig. 8

Results of the lateral (left) and top radial (right) scans for the 60 keV ${}^{241}$Am $\gamma$ line obtained with detector 50A. Statistical uncertainties are less than the widths of the markers. The dashed gray lines on the left(right) show the expected holder ditch (well) position (see Fig. 2)

Fig. 9

${}^{228}$Th spectrum taken with detector 50A and the source in lateral position. The main $\gamma$ lines, ${}^{208}$Tl double escape (DEP), ${}^{208}$Tl single escape (SEP) and ${}^{208}$Tl and ${}^{212}$Bi full energy (FEP) peaks, used in the pulse analysis, are emphasized together with the $Q_{\beta \beta }$ $\pm 35$ keV Compton continuum region. The inset shows a fit to the 2615 keV $\gamma$ line

Fig. 10

Correlation of A/E with the signal rise time [0.5–90%] of detector 50A for indicated heights H of the lateral collimator. The p+ contact is at $H=0$ mm. The gray dashed lines show the A/E cut position. These four datasets correspond to ${}^{208}$Tl DEP events from the highly collimated ${}^{228}$Th source

Fig. 11

Left (right): Rise time (A/E) distribution of ${}^{208}$Tl DEP events obtained from the lateral ${}^{228}$Th source flood measurement with detector 50A. The lateral, short and long analysis datasets selection are emphasized

Fig. 12

Spectrum measured with the IC detectors at the exposure of 8.5 kg$·$year in Gerda prior to and after indicated analysis cuts. The inset shows a zoom in the background analysis window. The only surviving event at 2058.9 keV was recorded with detector 74A on October 9, 2018, 01:09:14 (UTC)