Search for 2{\beta} decay of 116Cd with the help of enriched 116CdWO4 crystal scintillators

Cadmium tungstate crystal scintillators enriched in $^{116}$Cd to 82% ($^{116}$CdWO$_4$, total mass of $\approx$1.2 kg) are used to search for 2$\beta$ decay of $^{116}$Cd deep underground at the Gran Sasso National Laboratory of the INFN (Italy). The radioactive contamination of the $^{116}$CdWO$_4$ crystals has been studied carefully to reconstruct the background of the detector. The measured half-life of $^{116}$Cd relatively to 2$\nu$2$\beta$ decay is $T^{2\nu2\beta}_{1/2}$ = [2.8 $\pm$ 0.05(stat.) $\pm$ 0.4(syst.)] $\times$ 10$^{19}$ yr, in agreement with the results of previous experiments. The obtained limit on the 0$\nu$2$\beta$ decay of $^{116}$Cd (considering the data of the last 8696 h run with an advanced background 0.12(2) counts/yr/kg/keV in the energy interval 2.7-2.9 MeV) is $T_{1/2} \ge 1.0 \times 10^{23}$ yr at 90% C.L. The sensitivity of the experiment to the $0\nu2\beta$ process is $\lim T_{1/2} = 3 \times 10^{23}$ yr at 90% C.L. over 5 years of the measurements and it can be advanced (by further reduction of the background by a factor 3-30) to the level of $\lim T_{1/2} = (0.5-1.5) \times 10^{24}$ yr for the same period of the data taking.


Introduction
The rarest nuclear decay among observed -double beta decay (2β) -is a powerful tool to check physics beyond the Standard Model (SM) of particle physics. Allowed in the SM two neutrino double beta decay (2ν2β) has been registered for only 11 nuclides with a typical half-lives in the range of 10 18 − 10 24 yr (see e.g. in [1,2]). Neutrinoless double beta decay (0ν2β) -forbidden in the SM process -should exist, if neutrino is a Majorana particle, thanks to recently observed non-zero masses of the neutrino [3]. The observation of the 0ν2β decay will testify the lepton number non-conservation, the Majorana nature of the neutrino, an absolute value and hierarchy of the neutrino masses. The 0ν2β decay is also sensitive to an existence of the right-handed currents, hypothetical particles (majorons) and other effects beyond the SM (see the recent reviews [4][5][6][7][8][9] and references therein).
One of the most promising candidates for the 2β decay search is 116 Cd thanks to high 2β energy release (Q 2β = 2813.50 (13) keV [10]), large natural isotopic abundance (δ = 7.49% [11]), availability of enrichment by relatively cheap centrifugal method [12], and promising theoretical a e-mail: poda@kinr.kiev.ua estimations (see e.g. [5]). Moreover, existence of excellent detector, cadmium tungstate (CdWO 4 ) scintillator, already used for 2β decay searches [13][14][15] allows to realize "source = detector" experiment with a high detection efficiency. In addition, detector based on cadmium tungstate scintillator has several features (see e.g. [15,16]) which are very important for such kind of studies: low level of intrinsic radioactivity, good scintillation properties, particle identification ability, relatively low cost of the scintillation material, stability for long term operation.
Recently an enriched 116 CdWO 4 crystal scintillator with a large mass (1.87 kg) has been developed from deeply purified raw materials, and a new high sensitive 2β experiment has been started in February 2011 at the Gran Sasso National Laboratories (LNGS) of the INFN (Italy) [17]. The results of the last run performed from October 2012 till October 2013 will be reported here.

Low-background experiment
The experiment has been performed deep underground (≈ 3600 m w.e.) at the LNGS (Italy). The description of the detector and low background set-up at the previous stages of the experiment one can find in Refs. [17,18]. Here we outline the main features of the last stage.
Two 116 CdWO 4 crystal scintillators (586 g and 589 g) were fixed inside polystyrene light-guides (⊘70 × 194 mm). The cavities inside the light-guides were filled with the Ultima Gold AB (PerkinElmer) liquid scintillator (LS). Two high purity quartz light-guides (⊘70 × 200 mm) were glued to the polystyrene light-guide on both sides. The detectors were viewed from the opposite sides by two low radioactive 3" photomultipliers (PMTs, Hamamatsu R6233MOD). The detector modules, covered by the 3M foil to improve the light collection, were installed inside a low radioactive air-tight Cu box at the DAMA/R&D setup. Copper bricks 5 cm thick were used as an additional passive shield. The Cu box was continuously flushed with high purity nitrogen gas to remove the residual environmental radon. Outer passive shield consists of high purity Cu (10 cm thick), 15 cm of low radioactive lead, 1.5 mm of cadmium and 4 to 10 cm of polyethylene/paraffin. The whole shield is contained inside a Plexiglas box, also continuously flushed by high purity nitrogen.
An event-by-event data acquisition system based on a 1 GS/s 8 bit transient digitizer (Acqiris DC270) was used to record time of each event and scintillation pulseprofiles. The energy scale and the resolution of the detector was calibrated by using the reference 22 Na, 60 Co, 137 Cs, and 228 Th γ sources. The energy resolution of the detector was FWHM 4.3(4)% at Q 2β of 116 Cd.

Results and discussion
The data at this stage were collected over 8696 h. The applied data analysis is similar to the one described in [17]. The energy spectrum of γ(β) events accumulated with the 116 CdWO 4 detector and selected off-line by the pulseshape and the front edge analyzes is shown in Fig. 1. The background model was built from possible components like internal 40 K, 90 Sr-90 Y, 110m Ag, radionuclides from U/Th chains, external γ from the details of the set-up, 2ν2β decay of 116 Cd simulated by using the EGS4 [19] and De-cay0 [20] codes. The radiopurity of the 116 CdWO 4 crystals is reported in [17,21,22]. The radioactive contamination of the PMTs [23] and the sample of Ultima Gold AB liquid scintillator has been measured at the STELLA HPGe facility of the LNGS (Italy). The radiopurity of Cu has been derived from the fit of the background spectrum and it is comparable with the results of the low background measurements with the Cu samples (see e.g. [24]). Table 1 contains radiopurity level of the main sources of the background.
Fit of the background spectrum and the main components of the background are also presented in Fig.1. The fit gives the number of the 2ν2β events of 116 Cd as S 2ν2β = (34927 ± 614) counts for the total distribution. The signal to background ratio in the energy range of 1.6 − 2.4 MeV is ≈ 2:1. Taking into account the efficiency of the applied pulse-shape discrimination procedure (η PS D = 0.92), selection of the 212 Bi-Po events (η BiPo = 0.99), detection efficiency to the process searched for  Counts / 20 keV T 2ν2β 1/2 = ln 2 · η PS D · η BiPo · η 2ν2β · N 116 · t / S 2ν2β . The derived half-life T 2ν2β 1/2 = [2.8 ± 0.05(stat.) ± 0.4(syst.)] × 10 19 yr is in agreement with the results of the previous experiments (see Table 2) and the recent world average value (2.8 ± 0.2) × 10 19 yr [31].
In case of the 0ν2β decay of 116 Cd to the ground state of 116 Sn we should see in the experimental spectrum a peak at the energy ≈ 2.8 MeV. There is no such peculiarity in the spectrum, therefore only lower half-life limit on the process can be set according to formula: lim T 0ν2β 1/2 = ln 2 · η PS D · η BiPo · η 0ν2β · N 116 · t/ lim S . Here η 0ν2β is the detection efficiency for the 0ν2β process (equal to 0.947), lim S is the number of 0ν2β events which can be excluded at a given confidence level (C.L.), and other parameters are described above. The fit gives the area of the 0ν2β peak as (2.3 ± 4.4) counts. It corresponds to lim S = 9.2 counts at 90% C.L. according to the Feldman-Cousins procedure [32]. Substituting all parameters, one can obtain the limit T 0ν2β 1/2 ≥ 1.0 × 10 23 yr at 90% C.L., which is on the level of the Solotvina experiment [15] (see Table 3).
One can derive the similar result for the 0ν2β decay of 116 Cd by using also the "1σ approach". For example, there are N = 28 counts in the region of interest (ROI) 2.7 − 2.9 MeV, where the detection efficiency for the 0ν2β process is 0.849. The value lim S in such a case can be taken as International Workshop on Radiopure Scintillators RPSCINT 2013  1.64· √ N = 8.7 counts at 90% C.L. So, the calculated halflife limit T 0ν2β 1/2 ≥ 0.97 × 10 23 at 90% C.L. is very close to the one obtained from the fit of the background spectrum.
The background 0.12(2) counts/yr/kg/keV in the ROI allows us to reach a sensitivity to the half-life of 0ν2β decay of 116 Cd equal to lim T 1/2 = 3 × 10 23 yr at 90% C.L. over 5 years of measurements. In a case of further reduction of the background by a factor 3−30, the sensitivity can be advanced to the level of lim T 1/2 = (0.5 − 1.5) × 10 24 yr (it corresponds to the effective neutrino mass m ν = 0.4 − 1.4 eV).
The main components of the background in the ROI remain 2615 keV γ quanta of 208 Tl from contamination of the set-up by 232 Th, and 208 Tl from internal contamination of the 116 CdWO 4 crystals by 228 Th. The background could be suppressed by removing of one PMT from each detector module, by installation of an additional light-guide (quartz) between the crystal and the PMT, by replacing the PMT by another one with a higher level of radiopurity (e.g. Hamamatsu R11065SEL) and by recrystallization of the 116 CdWO 4 crystals to decrease thorium thanks to observed low segregation of Th and Ra by cadmium tungstate [21]. The first two modifications have already been applied in October, 2013. In addition, the Ultima Gold LS polluted by 40 K has been replaced by the Borexino LS and the new measurements are in progress.

Conclusions
A low-background experiment using enriched 116 CdWO 4 scintillators (total mass of 1.16 kg, enrichment by 116 Cd to 82%) is in progress at the LNGS of the INFN (Italy). The exposure of the experiment at the last stage of the data taking (October, 2012 -October, 2013) is 1.15 kg×yr. The energy resolution of the 116 CdWO 4 detector is FWHM = 4.3% at Q 2β of 116 Cd, the background counting rate was reduced to 0.12(2) counts/yr/kg/keV in the region of interest 2.7 − 2.9 MeV, where the 0ν2β peak is expected. The main components of the background in this region remain γ quanta of 208 Tl from the contamination of the set-up and internal contamination of the 116 CdWO 4 crystals by 228 Th.
The half-life relatively to the 2ν2β decay of 116 Cd to the ground level of 116 Sn was measured as T 2ν2β 1/2 = [2.8 ± 0.05(stat.) ± 0.4(syst.)] × 10 19 yr, in agreement with the results of the previous experiments. The half-life limit on 0ν2β decay of 116 Cd to the ground state of 116 Sn was established as T 0ν2β 1/2 ≥ 1.0 × 10 23 at 90% C.L. The current level of the background restricts the sensitivity of the experiment to 0ν2β decay of 116 Cd to lim T 1/2 = 3 × 10 23 yr at 90% C.L. over 5 years of the measurements. The sensitivity can be advanced to the level of lim T 1/2 ≈ (0.5 − 1.5) × 10 24 yr (which corresponds to the effective neutrino mass m ν = 0.4 − 1.4 eV) after improvement of the background by factor 3 − 30. An R&D to improve the background conditions of the experiment is in progress.