betaFeedback

WorksPhysics and Astronomy

Improved sensitivity in the search for rare decays of Gd isotopes

Abstract

Gadolinium is widely used in multiple low-background experiments, making its isotopes accessible for rare decay searches both in-situ and through radiopurity screening data. This study presents an improved search for rare alpha and double-beta decay modes in 152^{152}Gd, 154^{154}Gd, and 160^{160}Gd isotopes using ultra-low background HPGe detectors at the Boulby Underground Screening (BUGS) facility. A total exposure of 6.7 kg⋅\cdotyr of natural gadolinium was achieved using gadolinium sulfate octahydrate (Gd2(SO4)3⋅8H2O)(\text{Gd}_2(\text{SO}_4)_3 \cdot 8\text{H}_2\text{O}) samples, originally screened for radiopurity prior to their deployment in the Super-Kamiokande neutrino experiment. Due to the detection methodology, only decays into excited states accompanied by gamma-ray emission were accessible. A Bayesian analysis incorporating prior experimental results was employed, leading to new lower half-life limits in the range of 1019−102110^{19} - 10^{21} years - an improvement of approximately two orders of magnitude over previous constraints. No statistically significant decay signals were observed. These results demonstrate the effectiveness of repurposing large-scale radiopurity screening campaigns for fundamental physics research.

Introduction

Neutrinos are key particles in astrophysics. Although they were theoretically predicted in 1930 by W. Pauli [1] and experimentally detected for the first time in 1956 [2], a comprehensive understanding of their properties remains elusive. The most significant unresolved issues concern the neutrino’s mass and its nature [3, 4]. Among the various nuclear and electroweak processes involving neutrinos, neutrinoless double-beta decay offers the only practical means of obtaining definitive answers to both of these questions.

Neutrinoless double-beta decay is a hypothetical second-order nuclear process that is forbidden within the Standard Model of particle physics. However, if observed, it would establish the electron neutrino as a Majorana particle, implying that the neutrino possesses a non-zero effective Majorana mass term [5, 6, 7]. This conclusion arises because this process is only possible if the neutrino is its own antiparticle. Furthermore, the existence of Majorana-type neutrinos could enable mechanisms that would explain the observed matter-antimatter asymmetry in the Universe [8, 9]. Consequently, extensive worldwide experimental efforts are dedicated to this fundamental topic, with numerous international collaborations searching for neutrinoless double-beta decay in various isotopes [3, 10].

The current experimental sensitivity in this field has reached a half-life value of 102610^{26} years, yet the process remains undetected. Consequently, next-generation experiments aim to increase their sensitivity by employing an extended active detector mass in the range of 1–10 tons and by reducing the background level below 10−4−10−510^{-4} - 10^{-5} counts/keV·kg·yr [11]. To achieve this exceptionally low internal background counting rate, all potential background components must be meticulously controlled and mitigated. Therefore, all future ultra-low-background (ULB) experiments will necessitate complex assays of all materials and reagents used, which must adhere to stringent radiopurity protocols specific to each individual experiment. Typically, such screening campaigns incorporate a combination of low-background measurements of material samples using high-purity germanium (HPGe) γ\gamma-spectrometers located deep underground, analysis of radon emanation, and assays using ICP-MS analysis (such as [12, 13]).

In a different area of low-background, high-energy physics, the large-scale neutrino oscillation experiment, Super Kamiokande (SK), has dissolved 40 tons of a gadolinium salt (gadolinium sulfate octahydrate, Gd2(SO4)3⋅8H2O\mathrm{Gd}_{2}(\mathrm{SO}_{4})_{3} \cdot8\mathrm{H}_{2}\mathrm{O}) into their ultra-pure water Cherenkov detector to expand their scientific programmes and distinguish neutrinos from antineutrinos originating from supernova explosions. Consequently, before loading 13 tons of Gd2(SO4)3⋅8H2O\mathrm{Gd}_{2}(\mathrm{SO}_{4})_{3} \cdot8\mathrm{H}_{2}\mathrm{O} into SK in 2020 [14] and an additional 26 tons in 2022 [16], samples representing a portion of the gadolinium salt underwent radiopurity screening at Boulby Underground Laboratory (UK). The raw material was required to be substantially radiopure to maintain SK’s sensitivity to solar neutrinos and the diffuse supernova neutrino background (DSNB) signals. The requirements of <5< 5 mBq/kg early-chain 238U{}^{238}\mathrm{U}, <0.5< 0.5 mBq/kg late-chain 238U{}^{238}\mathrm{U}, <0.05< 0.05 mBq/kg 232Th{}^{232}\mathrm{Th}, and <30< 30 mBq/kg 235U{}^{235}\mathrm{U} were largely achieved, although some batch-to-batch variation occurred [15]. The experimental data acquired within this radioassay campaign provides a unique opportunity to search for rare nuclear processes that may occur in natural gadolinium isotopes at an unprecedented sensitivity level.

Therefore, this manuscript presents a new search for α\alpha and 2β2\beta decay of Gd isotopes to excited levels of their daughter nuclides using radioassay data with a total exposure of 6.7 kg·yr of natural Gd. Improved experimental sensitivity compared to the recent study [17] was achieved by using high-purity Gd-containing samples and increasing the sample exposure. However, as this experimental approach is best suited to studying rare decay modes with the emission of relatively high-energy de-excitation γ\gamma-rays (above 100 keV), the 2EC2\mathrm{EC} process in 152Gd{}^{152}\mathrm{Gd} is not considered here. Instead, the focus is on α\alpha and 2β2\beta decay of 152Gd{}^{152}\mathrm{Gd}, 154Gd{}^{154}\mathrm{Gd}, and 160Gd{}^{160}\mathrm{Gd} to excited states of their respective daughters. The decay schemes of all considered decay modes are shown in Figure 1.

Decay schemes for investigated decay modes in this work

Figure 1. Decay schemes for investigated decay modes in this work.

Experimental Setup and Sample

The Gd2(SO4)3⋅8H2O\mathrm{Gd}_{2}(\mathrm{SO}_{4})_{3} \cdot8\mathrm{H}_{2}\mathrm{O} raw material was produced by the Nippon Yttrium Company in approximately half-ton batches between 2019 and 2022 [14, 16, 15]. Two distinct production methods were employed, differentiated by their treatment of elemental radium. The second production method aimed to reduce radium concentration and thereby improve the late-chain thorium radiopurity. To verify the radiopurity of the produced batches, samples were sent to three underground screening laboratories worldwide, including the Boulby UnderGround Screening (BUGS) laboratory.

Following batch production, the manufacturer initially packed the samples in EVOH bags, each containing 5 kg (±0.5%) of material. These were shipped by aeroplane from the manufacturer in Japan to England. Upon arrival at the University of Sheffield, the samples were packed into type 448-G Marinelli beakers and transported to Boulby Underground Laboratory for HPGe screening.

Each batch of Gd2(SO4)3⋅8H2O\mathrm{Gd}_{2}(\mathrm{SO}_{4})_{3} \cdot8\mathrm{H}_{2}\mathrm{O} contained excess water from the production process. The amount of extra water in each batch was monitored by SK during Gd loading, averaging 4.4±1.0%4.4 \pm1.0\% w/w across the samples included in this analysis. Therefore, the mass of each sample in this study is taken to be 4.78±0.064.78 \pm0.06 kg of Gd2(SO4)3⋅8H2O\mathrm{Gd}_{2}(\mathrm{SO}_{4})_{3} \cdot8\mathrm{H}_{2}\mathrm{O}.

This study utilised 30 individual sample measurements from the SK screening effort, resulting in 15.9 kg·yr of Gd2(SO4)3⋅8H2O\mathrm{Gd}_{2}(\mathrm{SO}_{4})_{3} \cdot8\mathrm{H}_{2}\mathrm{O} exposure, or 6.7 kg·yr of natural Gd, on ULB HPGe detectors at Boulby. This represents the largest exposure of ULB gadolinium to date for conducting a search for rare Gd decays.

The BUGS facility houses several ULB HPGe detectors [18]. These include ”Merrybent” and ”Belmont,” both p-type coaxial detectors with masses of 2.0 kg and 3.2 kg, and relative efficiencies of 110% and 160%, respectively. Their background levels and resulting minimum detectable activities are reported in [19].

The full energy peak efficiency and true coincidence summing factors for Merrybent and Belmont are calculated using a custom Geant4 [20] simulation of the detectors, shielding, and sample geometry. This simulation, described in [21], is validated against a LabSOCS [22] efficiency simulation by measuring an IAEA-385 Irish Sea Sediment standard reference material on each detector. Both the Geant4 and LabSOCS calculated efficiencies and coincidence summing factors are applied to the IAEA-385 measurement and compared with the certified values. Good agreement is achieved across all analysed nuclides after accounting for coincidence summing and half-life corrections of the certified reference activities. Figure 2 and Figure 3 demonstrate that the ratio of calculated reference source activities to the certified values is within 1σ1\sigma of unity across all analysed nuclides. Consequently, the efficiencies calculated by the Geant4 simulation agree with those calculated using LabSOCS.

Results of a 3-day measurement of an IAEA-385 standard reference material on the Merrybent detector

Figure 2. Results of a 3-day measurement of an IAEA-385 standard reference material on the Merrybent detector. The activities of various nuclides, derived using two independent calculations of detection efficiency and true coincidence summing factors, are compared with the certified values from the IAEA.

Results of a 3-day measurement of an IAEA-385 standard reference material on the Belmont detector

Figure 3. Results of a 3-day measurement of an IAEA-385 standard reference material on the Belmont detector. The activities of various nuclides, derived using two independent calculations of detection efficiency and true coincidence summing factors, are compared with the certified values from the IAEA.

The Geant4 simulation package for Belmont and Merrybent was used to calculate the full energy peak efficiency for the expected gammas from the double beta decay of 160Gd{}^{160}\mathrm{Gd} to the 160Dy{}^{160}\mathrm{Dy} 02+0^{+}_{2}, 01+0^{+}_{1}, and 21+2^{+}_{1} levels, the alpha decay of 152Gd{}^{152}\mathrm{Gd} to the 148Sm{}^{148}\mathrm{Sm} 21+2^{+}_{1} level, and the alpha decay of 154Gd{}^{154}\mathrm{Gd} to the 150Sm{}^{150}\mathrm{Sm} 21+2^{+}_{1} level. The excited daughter states were generated, and the detected energy spectrum was analysed to determine the combined effect of detection efficiency and coincidence summing factors. The total efficiency factors for Belmont and Merrybent are reported in Table 5.

Nuclide (decay)Daughter (level) (JπJ^\pi keV)γ\gamma-lines [keV]Det. eff. ϵ\epsilon [%]σres\sigma_{\mathrm{res}} [keV]T1/2T_{1/2} (90% C.I.) [yr] old [17]T1/2T_{1/2} (90% C.I.) [yr] newT1/2T_{1/2} (90% C.I.) [yr] comb.
160Gd{}^{160}\mathrm{Gd} (0ν/2νββ0\nu/2\nu\beta\beta)160Dy{}^{160}\mathrm{Dy} (2+2^{+} 966.2)879.41.23, 0.840.84, 0.95>9.7×1019> 9.7 \times10^{19} (*)>1.4×1021> 1.4 \times10^{21}>1.5×1021> 1.5 \times10^{21}
160Gd{}^{160}\mathrm{Gd} (0ν/2νββ0\nu/2\nu\beta\beta)160Dy{}^{160}\mathrm{Dy} (01+0^{+}_{1} 1279.9)1193.22.06, 1.370.90, 1.03>8.2×1019> 8.2 \times10^{19}>2.7×1021> 2.7 \times10^{21}>2.7×1021> 2.7 \times10^{21}
160Gd{}^{160}\mathrm{Gd} (0ν/2νββ0\nu/2\nu\beta\beta)160Dy{}^{160}\mathrm{Dy} (02+0^{+}_{2} 1456.8)1369.91.89, 1.240.94, 1.08>5.0×1019> 5.0 \times10^{19}>4.2×1021> 4.2 \times10^{21}>4.3×1021> 4.3 \times10^{21}
152Gd{}^{152}\mathrm{Gd} (α\alpha)148Sm{}^{148}\mathrm{Sm} (2+2^{+} 550.3)550.32.56, 1.820.70, 0.78>3.4×1017> 3.4 \times10^{17}>9.0×1018> 9.0 \times10^{18}>8.4×1018> 8.4 \times10^{18}
154Gd{}^{154}\mathrm{Gd} (α\alpha)150Sm{}^{150}\mathrm{Sm} (2+2^{+} 333.9)333.92.64, 1.930.60, 0.67>9.6×1018> 9.6 \times10^{18}>9.7×1019> 9.7 \times10^{19}>1.0×1020> 1.0 \times10^{20}

Table 5. Lower half-life limits for the investigated decay modes in gadolinium isotopes. Columns 3-5 display the gamma lines used in the fit, along with their corresponding detection efficiencies and energy resolutions. The latter two columns are separated for the Belmont and Merrybent detectors.

Cosmic Ray Activation of Natural Gd

During its time on the Earth’s surface and in transit from Japan to England, the Gd2_2(SO4_4)3_3·8H2_2O material was exposed to high-energy cosmic rays. Potential long-lived radioisotope production was investigated using an ACTIVIA [23] simulation of cosmogenic neutrons on natural Gd. While the activation of each measured batch varied, it generally consisted of several weeks of processing, storage, and transport at ground level, approximately one day in air transportation, and then several additional weeks on the surface before being transported underground. For modelling purposes, we simulated one month of activation at sea level, one day at 35,000 ft, and one additional month at sea level. Finally, a seven-day cooling-off period was simulated to represent the time spent underground at Boulby before screening commenced. ACTIVIA provides the corresponding neutron flux spectra as a function of atmospheric depth. After the activation and cooling-off periods, the remaining long-lived unstable products are isotopes of Eu, Gd, and Tb, as well as other isotopes produced through spallation and decay processes. Many of the most significant expected activation or spallation products with detectable gammas from the simulation are shown in Table 1, with those having significant expected gammas indicated.

IsotopeHalf-Life (days)Residual Activity (mBq/kg)Significant GammasObserved Activity (mBq/kg)
156Eu{}^{156}\mathrm{Eu}15.20.33811.8 keV (9.7%)< 0.16
149Gd{}^{149}\mathrm{Gd}9.280.27346.7 keV (23.7%)< 0.10
156Tb{}^{156}\mathrm{Tb}5.40.20534.3 keV (67%)< 0.02
148Eu(∗){}^{148}\mathrm{Eu}(*)54.50.12550.3 keV (99%), 630.0 keV (71.9%)0.08 ± 0.01
160Tb{}^{160}\mathrm{Tb}72.40.08879.4 keV (30%)< 0.08
145Eu{}^{145}\mathrm{Eu}5.930.08893.7 keV (66%)< 0.02
143Pm{}^{143}\mathrm{Pm}2650.06741.9 keV (38.5%)0.06 ± 0.02
131Ba{}^{131}\mathrm{Ba}11.50.04496.3 keV (48%)< 0.03
146Eu{}^{146}\mathrm{Eu}4.510.03747.2 keV (98%), 633-634 keV (sum: 81.8%)0.04 ± 0.01
7Be{}^{7}\mathrm{Be}53.30.01477.6 keV (10.4%)0.39 ± 0.09
154Eu{}^{154}\mathrm{Eu}29930.0091274.4 keV (34.8%)0.09 ± 0.03
105Ag{}^{105}\mathrm{Ag}41.30.008344.6 keV (42%)< 0.05
144Pm{}^{144}\mathrm{Pm}3630.007618.0 keV (98%), 696.5 keV (99.5%)< 0.03
113Sn{}^{113}\mathrm{Sn}1150.006391.7 keV (65%)< 0.02
119Te{}^{119}\mathrm{Te}4.680.0051212.7 keV (66%)< 0.02
152Eu{}^{152}\mathrm{Eu}48910.0051408.0 keV (15%), 964.1 keV (10.5%)< 0.19
85Sr{}^{85}\mathrm{Sr}64.80.005514.0 keV (96%)< 0.02
88Zr{}^{88}\mathrm{Zr}83.40.004392.9 keV (97.3%)< 0.03
99Rh{}^{99}\mathrm{Rh}150.004528.2 keV (37.9%), 353.1 keV (34.5%)< 0.11
95Tc{}^{95}\mathrm{Tc}610.004582.1 keV (30%), 835.1 keV (26.6%)0.29 ± 0.05
106Ag{}^{106}\mathrm{Ag}8.50.0031045.8 keV (29.6%)< 0.08
148mPm(∗){}^{148m}\mathrm{Pm}(*)41.30.003550.3 keV (94.9%), 630.0 keV (89%)0.09 ± 0.01
54Mn{}^{54}\mathrm{Mn}3120.001834.8 keV (99.98%)0.08 ± 0.01

Table 1. A selection of long-lived radioisotopes produced by cosmic ray activation on natural Gd, as simulated by ACTIVIA. The residual specific activity for each nuclide is shown, considering one month of activation at sea level, one day at 35,000 ft, one additional month at sea level, and a final seven-day cooling-off period underground. Only products with long half-lives and significant gamma rays above 300 keV [25] are included, as these are most likely to affect HPGe measurements of rare Gd decays. The final column shows the observed activity in the Belmont < 100 d sample. Possible interferences, such as near 835 keV (54Mn{ }^{54}\mathrm{Mn}, 95Tc{}^{95}\mathrm{Tc}, 228Ac{}^{228}\mathrm{Ac}), have not been factored into the observed activities.

Activation products with gammas in the regions of interest for this rare Gd decay search are particularly problematic. Notably, 148Eu{}^{148}\mathrm{Eu} and 148mPm{}^{148m}\mathrm{Pm} have strong gammas (550.3 keV, Iγ>95%I_\gamma>95\%, T1/2=54.5T_{1/2}=54.5 d and T1/2=41.29T_{1/2}=41.29 d, respectively [24]) that interferes with the 152Gd{}^{152}\mathrm{Gd} alpha decay search. The half-lives of these isotopes are similar in magnitude to the typical time between upper-atmosphere exposure to cosmic radiation and sample measurement on underground, ULB HPGe detectors. The residual activity of 148Eu{}^{148}\mathrm{Eu} following simulated activation indicates the potential for medium-lived gamma interference within these measured Gd2(SO4)3⋅8H2O\mathrm{Gd}_2(\mathrm{SO}_4)_3\cdot8\mathrm{H}_2\mathrm{O} samples. This interference is most significant in samples screened soon after being brought underground. Both 550 keV and 630 keV gamma peaks are observed in the combined detector spectra (Figure 4), with relative intensities consistent with literature values. A possible interference at 550 keV from 220Rn{}^{220}\mathrm{Rn} is dismissed as the peak area would imply a 232Th{}^{232}\mathrm{Th}-chain activity which is two orders of magnitude greater than that observed from 208Tl{}^{208}\mathrm{Tl}.

Spectra showing cosmic activation peaks

Figure 4. This figure illustrates the interference of cosmic activation peaks with the rare event search. 148Eu{}^{148}\mathrm{Eu} (T1/2=54.5T_{1/2} = 54.5 d) and 148mPm{}^{148m}\mathrm{Pm} (T1/2=41.29T_{1/2} = 41.29 d) exhibit visible gamma lines at 550.3 keV and 630.0 keV, respectively. These interferences can be effectively eliminated by selecting data from samples that have spent more than 100 days underground.

Other identified activation products in the datasets include 146Eu{}^{146}\mathrm{Eu}, 154Eu{}^{154}\mathrm{Eu}, 54Mn{}^{54}\mathrm{Mn}, 95Tc{}^{95}\mathrm{Tc}, and 143Pm{}^{143}\mathrm{Pm} (see Table 1). Although sulfate activation was not simulated with ACTIVIA, the light spallation product 7Be{}^{7}\mathrm{Be} was observed in the <100<100 d datasets. Possible interferences around 835 keV between 95Tc{}^{95}\mathrm{Tc}, 54Mn{}^{54}\mathrm{Mn}, and 228Ac{}^{228}\mathrm{Ac} are currently under investigation.

Due to the inconsistent cosmic ray irradiation and ”cooling-down” periods of the Gd2(SO4)3⋅8H2O\mathrm{Gd}_2(\mathrm{SO}_4)_3\cdot8\mathrm{H}_2\mathrm{O} samples before HPGe screening, the screening results were separated based on whether the screening commenced less than or greater than 100 days after being transported underground and shielded from cosmic rays. This cutoff was chosen to minimise the impact of interfering gammas while retaining a significant portion of the sample exposure for the 152Gd{}^{152}\mathrm{Gd} alpha decay search. Figure 4 shows the relevant region for 148Eu{}^{148}\mathrm{Eu} gammas for the Belmont <100<100 d, Belmont >100>100 d, Merrybent <100<100 d, and Merrybent >100>100 d datasets. The 100-days cutoff effectively eliminates the small but significant gamma peaks from both strong gamma emitters.

Sample Characterization

The concentration of chemical impurities in a subset of the Gd2(SO4)3⋅8H2O\mathrm{Gd}_2(\mathrm{SO}_4)_3\cdot8\mathrm{H}_2\mathrm{O} samples was determined using a quadrupole inductively coupled plasma mass spectrometer equipped with collision cell (Agilent model 7850). The results are presented in Table 2. A semiquantitative analysis was performed, calibrating the instrument with a single standard solution containing 1 ppb of Li, Co, Y, Ce, and Tl. The uncertainty in the measured concentrations is approximately 25%. The contamination of K was measured by high-resolution inductively coupled plasma mass spectrometry (HR-ICP-MS, Thermo Fisher Scientific ELEMENT2) in “cool plasma” mode in order to enhance the sensitivity.

Element190502190804220471210601190705210811210711190904
K380295200635440425300210
Ba6786181788
La14760< 10258< 10< 2035
Ce85< 5< 313< 2< 54
Nd40< 10< 20< 20120< 50< 50< 30
Pr1227240112
Sm185230< 206215<20< 50220
Eu95200656851213170
Tb135140701379613011097
Lu< 3000< 3000< 3000< 3000< 3000< 3000< 3000< 3000
Th< 1< 1< 1< 1< 1< 1< 1< 1
U< 1< 1< 1< 1< 1< 1< 1< 1

Table 2. Concentrations, in units of parts-per-billion, of selected chemical impurities in Gd2(SO4)3⋅8H2O{\mathrm{Gd}_{2}(\mathrm{SO}_{4})_{3}\cdot8\mathrm{H}_{2}\mathrm{O}} samples, as determined by HR-ICP-MS analysis at LNGS. The uncertainty in the measured concentrations is approximately 25%. Limits are 68% C.L. .

For each sample, 150 mg of Gd-containing material was placed in a plastic vial with 5 ml of ultrapure water and 0.5 ml of nitric acid. This mixture was then placed in an ultrasonic bath at 60°C until complete sample decomposition. The resulting solutions were diluted with ultrapure water to a total volume of 10 ml (a dilution factor of approximately 2000) in preparation for ICP-MS analysis.

The results in Table 2 demonstrate the high chemical purity of the Gd-containing samples, confirming the effectiveness of the purification process employed during production. For example, the purification is highly effective at removing U and Th, with upper limits below 1 ppb established. The sensitivity of the ICP-MS measurements for these two elements could be further improved by employing a pre-concentration technique using anion-exchange resins [26, 27].

The high detection limit for Lu (3000 ppb) is caused by significant interference of signals from this element with the signal from the matrix (primarily GdOH+ ions). Similarly, the signal from Tb is affected by interference from GdH+ ions.

The radiopurity of all Gd2(SO4)3⋅8H2O\mathrm{Gd}_2(\mathrm{SO}_4)_3\cdot8\mathrm{H}_2\mathrm{O} powder samples used in this analysis was previously reported by SK [15, 16]. For comparison with the ICP-MS measurements performed in this study, a subset of ULB HPGe-measured radiopurity values is reproduced in Table 3.

ChainNuclide190502190804220471210601190705210811210711190904
238U{}^{238}\mathrm{U}234Th{}^{234}\mathrm{Th}<4.5< 4.5<4.8< 4.8<6.8< 6.8<5.8< 5.89±59 \pm5<4.1< 4.1<8.2< 8.2<5.0< 5.0
214Pb{}^{214}\mathrm{Pb}<0.20< 0.20<0.21< 0.210.86±0.230.86 \pm0.23<0.27< 0.27<0.30< 0.300.97±0.210.97 \pm0.21<0.23< 0.23<0.25< 0.25
235U{}^{235}\mathrm{U}235U{}^{235}\mathrm{U}<0.14< 0.14<0.27< 0.27<0.45< 0.45<0.10< 0.10<0.24< 0.24<0.18< 0.18<0.12< 0.12<0.26< 0.26
227Th{}^{227}\mathrm{Th}<0.62< 0.62<0.88< 0.88<0.75< 0.75<0.44< 0.441.4±0.51.4 \pm0.5<0.59< 0.59<0.58< 0.58<0.73< 0.73
232Th{}^{232}\mathrm{Th}228Ac{}^{228}\mathrm{Ac}<0.41< 0.410.70±0.260.70 \pm0.260.85±0.340.85 \pm0.34<0.17< 0.170.30±0.180.30 \pm0.18<0.34< 0.34<0.40< 0.400.77±0.280.77 \pm0.28
208Tl{}^{208}\mathrm{Tl}<0.37< 0.370.43±0.160.43 \pm0.160.96±0.280.96 \pm0.28<0.34< 0.340.53±0.160.53 \pm0.16<0.30< 0.30<0.22< 0.220.81±0.220.81 \pm0.22
176Lu{}^{176}\mathrm{Lu}0.34±0.080.34 \pm0.085.1±0.35.1 \pm0.30.12±0.060.12 \pm0.060.69±0.090.69 \pm0.091.8±0.21.8 \pm0.20.29±0.060.29 \pm0.060.37±0.060.37 \pm0.067.1±0.47.1 \pm0.4
137Cs{}^{137}\mathrm{Cs}<0.04< 0.04<0.05< 0.05<0.05< 0.05<0.02< 0.02<0.08< 0.08<0.08< 0.08<0.04< 0.04<0.09< 0.09
138La{}^{138}\mathrm{La}0.12±0.070.12 \pm0.07<0.13< 0.13<0.07< 0.07<0.09< 0.09<0.14< 0.14<0.14< 0.14<0.10< 0.10<0.13< 0.13
40K{}^{40}\mathrm{K}<2.7< 2.7<3.5< 3.5<6.3< 6.3<0.98< 0.98<1.7< 1.7<1.8< 1.8<1.4< 1.4<4.0< 4.0

Table 3. Specific activities of naturally occurring and cosmogenically activated radionuclides in the selected Gd2_2(SO4_4)3⋅_3 \cdot8H2_2O powder samples, as measured by gamma-ray spectrometry.

The upper detection limit for 40K{}^{40}\mathrm{K}, less than (1–6) mBq/kg, corresponds to a concentration of natural potassium of less than 32–192 ppb. These are in slight disagreement with the concentrations of natural potassium reported by ICP-MS. The BUGS laboratory is located within the potash layer of the Boulby Mine, so the potassium background is greater and sensitivities worse compared to other underground sites around the world.

Similarly, the observed specific activity of 138La{}^{138}\mathrm{La} in sample 190502 (0.12 mBq/kg) corresponds well with the highest concentration of La (approximately 150 ppb) among all samples measured using ICP-MS. 176Lu{}^{176}\mathrm{Lu} decays through beta decay with gamma emission. The detected activity of this radioisotope in all analysed samples, at the level of (0.12–7.11) mBq/kg, corresponds to a natural lutetium concentration of 2.3–137.0 ppb. This result agrees well with the established detection limit for Lu in the ICP-MS measurements (<3000 ppb). However, due to the strong interference

effects with chemical elements of the complex compound observed with ICP-MS, gamma-ray spectrometry provides a more precise evaluation of the Lu concentration in Gd2_2(SO4_4)3⋅_3 \cdot8H2_2O samples.

It is important to note that chemical reactions and transformations typically disrupt the secular equilibrium in natural radioactive decay chains. This disruption is primarily driven by the chemical properties of the daughter nuclides in the U/Th chains. Furthermore, the production technology for the Gd2_2(SO4_4)3⋅_3 \cdot8H2_2O material used in SK was modified at least twice, leading to different disruptions of the secular equilibrium in each production run5. This variation can be observed in the specific activities of the radionuclides from the natural U/Th decay chains.

The isotopic composition of gadolinium was determined through complementary measurements of the eight Gd2_2(SO4_4)3⋅_3 \cdot8H2_2O samples selected for comparative chemical and radiopurity studies. HR-ICP-MS analysis was conducted at LNGS, and the results are presented in Table 4. The measured isotopic composition of gadolinium in all eight samples is consistent with natural abundance, within the quoted uncertainties. Therefore, tabulated values for the natural isotopic composition of gadolinium were used in subsequent analyses.

Sample152Gd{}^{152}\mathrm{Gd}154Gd{}^{154}\mathrm{Gd}155Gd{}^{155}\mathrm{Gd}156Gd{}^{156}\mathrm{Gd}157Gd{}^{157}\mathrm{Gd}158Gd{}^{158}\mathrm{Gd}160Gd{}^{160}\mathrm{Gd}
1905020.18±0.010.18 \pm0.012.20±0.012.20 \pm0.0114.80±0.1014.80 \pm0.1020.49±0.1820.49 \pm0.1815.63±0.0815.63 \pm0.0824.77±0.0224.77 \pm0.0221.92±0.1221.92 \pm0.12
1908040.18±0.010.18 \pm0.012.18±0.022.18 \pm0.0214.80±0.0514.80 \pm0.0520.37±0.1020.37 \pm0.1015.50±0.0615.50 \pm0.0624.96±0.0524.96 \pm0.0521.97±0.1021.97 \pm0.10
2204710.20±0.010.20 \pm0.012.18±0.012.18 \pm0.0114.90±0.1114.90 \pm0.1120.43±0.2520.43 \pm0.2515.64±0.1315.64 \pm0.1324.66±0.1724.66 \pm0.1721.95±0.1221.95 \pm0.12
2106010.18±0.010.18 \pm0.012.18±0.012.18 \pm0.0114.70±0.0314.70 \pm0.0320.53±0.0820.53 \pm0.0815.79±0.0515.79 \pm0.0524.74±0.0224.74 \pm0.0221.84±0.1121.84 \pm0.11
1907050.20±0.010.20 \pm0.012.17±0.012.17 \pm0.0114.80±0.2514.80 \pm0.2520.50±0.0320.50 \pm0.0315.66±0.0615.66 \pm0.0624.82±0.3324.82 \pm0.3321.86±0.1521.86 \pm0.15
2108110.18±0.010.18 \pm0.012.20±0.022.20 \pm0.0214.70±0.1414.70 \pm0.1420.52±0.2620.52 \pm0.2615.72±0.1115.72 \pm0.1124.89±0.2324.89 \pm0.2321.82±0.0221.82 \pm0.02
2107110.20±0.010.20 \pm0.012.19±0.012.19 \pm0.0114.70±0.0414.70 \pm0.0420.51±0.0520.51 \pm0.0515.63±0.0515.63 \pm0.0525.03±0.1325.03 \pm0.1321.74±0.0721.74 \pm0.07
1909040.20±0.010.20 \pm0.012.18±0.022.18 \pm0.0214.70±0.1014.70 \pm0.1020.43±0.2120.43 \pm0.2115.64±0.1615.64 \pm0.1625.06±0.1925.06 \pm0.1921.78±0.2121.78 \pm0.21
Tabulated value0.20±0.010.20 \pm0.012.18±0.022.18 \pm0.0214.80±0.0914.80 \pm0.0920.47±0.0320.47 \pm0.0315.65±0.0415.65 \pm0.0424.84±0.0824.84 \pm0.0821.86±0.0321.86 \pm0.03

Table 4. The isotopic abundances of gadolinium in the eight Gd2(SO4)3⋅8H2O{\mathrm{Gd}}_{2}({\mathrm{SO}}_{4})_{3}\mathbin{\cdot}8{\mathrm{H}}_{2}\mathrm{O} samples, as determined by HR-ICP-MS analysis.

Datasets

Gd2_2(SO4_4)3⋅_3 \cdot8H2_2O samples were measured at BUGS from September 2019 to February 2024 (Table 6 in appendix). Each sample was screened until achieving a sensitivity of <0.5<0.5 mBq/kg for the 609.3 keV gamma ray from 214Bi{}^{214}\mathrm{Bi}, at a 95% confidence level. In some instances, such as during the first UK COVID-19 lockdown from March to May 2020, sample measurements continued beyond the time required to reach this sensitivity due to restricted underground access. The inherent radioactivity of each batch is reported in [15] and [16], relative to the SK requirements. This Gd2_2(SO4_4)3⋅_3 \cdot8H2_2O material was among the most radiopure ever screened at Boulby, necessitating additional corrections to the spectral analysis to account for the shielding of background sources by the large, dense samples [28].

SampleDetectorMeasurement Time (days)Underground Time (days)Start Date
190502Belmont35.2< 1002019/09/24
190604Belmont30.3> 1002020/08/07
190704Belmont37.3< 1002019/11/29
190706Belmont27.4< 1002020/01/09
190804Belmont31.0< 1002020/02/14
190902Belmont98.4< 1002020/03/18
190902_2Belmont7.4> 1002021/06/29
190902_3Belmont18.3> 1002021/08/13
190904Belmont28.4> 1002020/07/09
210601 (A)Belmont54.1> 1002022/02/18
210711 (C)Belmont31.7< 1002021/12/16
210712 (D)Belmont37.7< 1002022/12/20
210713_2 (E)Belmont19.5> 1002023/02/27
210811 (F)Belmont25.2> 1002022/04/14
211106_2 (K)Belmont37.1> 1002023/04/04
220241 (M)Belmont40.2> 1002023/06/09
220242 (N)Belmont25.7> 1002023/07/20
220251 (O)Belmont36.9> 1002023/09/15
220352 (Q)Belmont20.4> 1002023/11/01
220361 (S)Belmont15.2> 1002023/11/22
220471 (U)Belmont20.4> 1002024/01/25
190501Merrybent12.7< 1002019/09/10
190606Merrybent28.5< 1002019/10/24
190705Merrybent44.1< 1002019/12/19
190705_2Merrybent24.9> 1002021/07/01
190802Merrybent56.9> 1002020/07/09
190806Merrybent45.1< 1002020/03/11
210601 (A)Merrybent9.2< 1002021/09/02
210711 (C)Merrybent93.0> 1002022/11/17
210811 (F)Merrybent61.3> 1002023/06/09
210821 (G)Merrybent82.7< 1002021/12/16
210822 (I)Merrybent20.2> 1002023/11/01
211201_2 (L)Merrybent20.3> 1002023/11/22
220482 (W)Merrybent24.4> 1002024/01/04
220582 (Y)Merrybent12.3> 1002024/01/31
220691 (Z)Merrybent10.5> 1002024/02/13

Table 6. Constituent datasets used in the current analysis.

To account for drift in the absolute ADC calibration, fluctuating backgrounds, and external factors affecting detector resolution, each measured spectrum was individually calibrated. This calibration procedure involved identifying significant peaks attributable to common naturally occurring radioactive material (NORM) gamma full energy peaks. Depending on the Gd batch, these peaks could include 201.83 keV and 306.82 keV from 176Lu{}^{176}\mathrm{Lu}, 295.22 keV and 351.93 keV from 214Pb{}^{214}\mathrm{Pb}, 609.31 keV, 1120.29 keV, 1764.49 keV, and 2204.21 keV from 214Bi{}^{214}\mathrm{Bi}, 911.20 keV and 968.96 keV from 228Ac{}^{228}\mathrm{Ac}, 583.19 keV and 2614.51 keV from 208Tl{}^{208}\mathrm{Tl}, and 1173.21 keV and 1332.49 keV from 60Co{}^{60}\mathrm{Co}. A linear ADC calibration was evaluated for each spectrum based on the locations of these peaks and the literature gamma energy values. The Gaussian width parameter of each fitted peak was also used to estimate the detector energy resolution, σ(E)\sigma(E), with the functional form: σ(E)=a+bE\sigma(E)=a+b\sqrt{E}, where aa and bb are free parameters. The resolution for each detector remained stable throughout the screening programme. The corresponding values for the gamma energies of interest for this analysis are shown in Table 5.

Detector stability was monitored throughout the Gd measurement period using several methods. The inherent background spectrum, measured multiple times during the screening programme, was checked for consistency. Periods exhibiting higher-than-normal background levels were investigated and generally attributed to external factors, such as N2{\mathrm{N}}_{2} purge failures. Additionally, periodic measurements of a check source containing 3737 kBq of both 155Eu{}^{155}\mathrm{Eu} and 22Na{}^{22}\mathrm{Na} were conducted to monitor the full energy peak width of the characteristic gamma rays and the dead layer thickness. The latter was determined by examining the ratio of the count rate in the 86.586.5 keV and 105.3105.3 keV peak areas. Datasets exhibiting fluctuations in detector background or full energy peak resolution indicative of suboptimal detector conditions were excluded.

The calibrated spectra from each detector were rebinned and combined according to the <100<100 day and >100>100 day classifications. Figure 5 displays the full spectra in 1010 keV bins for all four combined datasets. Statistical analysis of the rare event search employed 0.50.5 keV binned spectra.

Spectra of all four datasets

Figure 5. This figure presents the spectra of all four datasets used in this analysis, categorized by detector (Belmont and Merrybent) and the time elapsed between the samples being brought underground and the commencement of measurements (<100 days and >100 days).

Analysis

This analysis builds upon the work presented in [17], employing the same peak search procedure based on Bayesian statistics (see Eqs. (5), (6), and (7) from that manuscript) and incorporating the results from the same study as prior input. For this study, as in the previous one, an independent analysis is performed for each decay mode. All four datasets are combined using fits with a single inverse half-life parameter (T1/2)−1(T_{1/2})^{-1} but independent nuisance parameters. To avoid interference from cosmogenically produced isotopes (148Eu{ }^{148}\mathrm{Eu} and 148mPm{ }^{148m}\mathrm{Pm}), only the two datasets with >100>100 days underground are used for the alpha decay of 152Gd{}^{152}\mathrm{Gd}. The marginalized posteriors of (T1/2)−1(T_{1/2})^{-1} from [17] serve as prior functions for (T1/2)−1(T_{1/2})^{-1} in this analysis.

The live times, TT, of the four datasets were as follows: (i) 0.819190.81919 yr for Belmont <100<100 d; (ii) 1.037811.03781 yr for Belmont >100>100 d; (iii) 0.608200.60820 yr for Merr ybent <100<100 d; and (iv) 0.887050.88705 yr for Merr ybent >100>100 d. The mass of the gadolinium sample was 4.784.78 kg for all samples and datasets, corresponding to 1.54×10221.54 \times10^{22}, 1.68×10231.68 \times10^{23}, 3.37×10243.37 \times10^{24} atoms of 152Gd{}^{152}\mathrm{Gd}, 154Gd{}^{154}\mathrm{Gd}, and 160Gd{}^{160}\mathrm{Gd}, respectively. The uncertainties associated with sample mass and isotopic abundance are subdominant and correlated to the uncertainty of the detection efficiency (10%10\%); therefore, they are neglected. The detection efficiencies and energy resolutions, which differ by detector but share the same prior input for the <100<100 d and >100>100 d division, are provided in Table 5. It is important to note that each dataset has its own fit parameters, allowing for systematic variations between datasets even with identical input values.

The spectral fits for all decay modes are shown in Figure 6–Figure 10 in the Appendix. The best-fit functions are shown in blue. All investigated rare decay modes are consistent with the assumption of no detected signal. The fit functions with the signal process set to the 90% credibility limit are shown in red.

Fit plots for Belmont and Merrybent samples

Figure 6. Fit of the 879.4 keV γ\gamma-line from 160Gd{}^{160}\mathrm{Gd} 0ν/2νββ0\nu/2\nu\beta\beta decay into the excited 2+2^{+} state (966.2 keV) of 160Dy{}^{160}\mathrm{Dy}.

Four-panel fit of the 333.9 keV γ-line

Figure 10. Fit of the 333.9 keV γ\gamma-line from 154Gd{}^{154}\mathrm{Gd} α\alpha decay into the excited 21+2^{+}_{1} state (333.9 keV) of 150Sm{}^{150}\mathrm{Sm}.

The 90% quantiles of the marginalized posteriors of T1/2−1T_{1/2}^{-1} yield the limits. These distributions are shown in red in Figure 11 in the Appendix. The blue curves represent the priors from [17], and the shaded areas indicate the 90% probability regions of the posterior.

Priors and posteriors for $^{160}\mathrm{Gd}$ $0\nu/2\nu\beta\beta$ decay into the $2_{2}^{+}$ state

(a) 160Gd^{160}\mathrm{Gd} 0ν/2νββ0\nu/2\nu\beta\beta decay into the 22+2_{2}^{+} state

Priors and posteriors for $^{160}\mathrm{Gd}$ $0\nu/2\nu\beta\beta$ decay into the $0_{1}^{+}$ state

(b) 160Gd^{160}\mathrm{Gd} 0ν/2νββ0\nu/2\nu\beta\beta decay into the 01+0_{1}^{+} state

Priors and posteriors for $^{160}\mathrm{Gd}$ $0\nu/2\nu\beta\beta$ decay into the $0_{2}^{+}$ state

(c) 160Gd^{160}\mathrm{Gd} 0ν/2νββ0\nu/2\nu\beta\beta decay into the 02+0_{2}^{+} state

(d) 152Gd^{152}\mathrm{Gd} α\alpha decay into the 21+2_{1}^{+} state

Priors and posteriors for $^{154}\mathrm{Gd}$ $\alpha$ decay into the $2_{1}^{+}$ state

(e) 154Gd^{154}\mathrm{Gd} α\alpha decay into the 21+2_{1}^{+} state

Figure 11. Priors and posteriors of the Bayesian analysis for the inverse half-life variable. The shaded area show the 90% quantile.

The lower half life limits are presented in the last columns of Table 5, showing the previous limits from [17], the limits obtained using only data from this work (flat prior in T1/2−1T_{1/2}^{-1}), and the combined data as described above.

The data combination results in both higher and lower limits compared to using the new data alone but does not significantly influence the overall results. This outcome is attributed to the significantly greater amount of data and sensitivity in the present work, resulting in a relatively flat and uninformative prior. Random over- or under-fluctuations in the previous data only slightly inform the Bayesian parameter interpretation of the new data. Nevertheless, using the combined results is recommended, as they utilise more experimental data and avoid bias from selecting the best limit. Therefore, the reported posteriors in Figure 11 can serve as prior input for future searches.

Discussion

It is noteworthy that the sensitivity of modern experimental techniques, such as the ULB HPGe gamma spectrometers employed in this study, allows for the investigation of rare nuclear processes even within routine materials screening campaigns. This study demonstrates the feasibility of achieving experimental sensitivities of 1019−102110^{19} - 10^{21} years for various rare nuclear pro(*) the old half-life limit is based on a combined fit of two γ\gamma-lines

cesses in natural Gd isotopes. This was accomplished using a set of large-mass Gd-containing samples with a total exposure of 6.7 kg·yr of natural Gd, originally intended for radiopurity assays.

None of the investigated decay modes yielded an observable signal. 90% credibility limits were established using a Bayesian analysis that accounted for dominant systematic uncertainties. The recently established experimental limits from [17] were improved by approximately two orders of magnitude. However, these experimental limits for the alpha decays of 152{}^{152}Gd and 154{}^{154}Gd remain far from theoretical expectations (102510^{25} yr and 108010^{80} yr, respectively). Calculations for the expected half-lives of various double beta decay modes of 160{}^{160}Gd are still pending.

Further improvements in experimental searches for rare decays of 152{}^{152}Gd, 154{}^{154}Gd, and 160{}^{160}Gd are challenging within the ”source ≠\ne detector” approach. This limitation arises from two main disadvantages of this technique: (i) the decrease in detection efficiency with decreasing gamma energy due to absorption in surrounding materials (e.g., sample container, end cap, and Ge crystal holder); and (ii) the limited benefit of increasing sample mass due to increased self-absorption.

Therefore, innovative experimental approaches that significantly enhance detection efficiency, such as those described in [29], or increase the number of isotopes of interest through enrichment [30] are necessary. The latter option appears more promising, as the potential sensitivity enhancement would be the product of the enrichment factor and the improvement in detection efficiency achieved through reduced sample dimensions and self-absorption. This enhancement could range from a factor of 5 to several orders of magnitude.

A more substantial enhancement in detection efficiency for the processes of interest could be achieved by implementing the ”source = detector” approach. In this approach, the decaying Gd nuclei are embedded within the detector’s sensitive volume, which could be a liquid scintillator, a crystal scintillator, or a bolometer/scintillating bolometer. This configuration enables the detection of not only de-excitation gamma quanta but also short-range particles (alpha and beta particles) in the final reaction channel. This capability allows for further background rejection through coincidence measurements of different reaction products.

Furthermore, effective particle identification based on pulse-shape discrimination, variations in emitted scintillation light, or variations in the ratio of emitted light to phonon signal would enable the detection of alpha decay to ground states. This capability would also enhance the experimental sensitivity to decay modes that occur through transitions to excited states of daughter nuclei. Finally, this experimental approach accommodates the use of a much larger sample mass.

For example, the PIKACHU (Pure Inorganic scintillator experiment in KAMioka for CHAllenging Underground sciences) project was recently initiated to fabricate high-purity Ce-doped Gd3_3Ga5_5Al2_2O12_{12} (GAGG - Gadolinium Aluminium Gallium Garnet) single crystals for studying all modes of 160{}^{160}Gd double beta decay [31]. This experiment benefits from a combination of technological expertise and successful multi-ton Gd-salt purification for the SK experiment [15] and in-house growth of GAGG single crystals with strict control at each production stage. Moreover, crystal growth leverages the well-developed technology of GAGG single crystal production (see, for instance [32] and [33]), driven by commercial market demands. Consequently, high-quality, large-volume scintillating crystals with masses of a few kilograms are commercially available.

Initially, PIKACHU will employ two GAGG scintillating crystals, 6.5 cm in diameter and 14.5 cm in length, each containing 710 g of 160{}^{160}Gd. Due to the excellent light yield (approximately 50,000 photons/MeV) and pulse-shape discrimination capability of these crystals, a sensitivity level of 102210^{22} years is anticipated for the double beta decay modes of 160^{160}Gd. Subsequently, twenty large-volume, high-radiopurity GAGG crystals will be used to further enhance the experimental sensitivity. Therefore, the most stringent constraints on not only 160^{160}Gd double beta decay modes but also rare alpha decays of 152^{152}Gd and 154^{154}Gd will likely emerge from this experiment in the near future.

Conclusion

This study investigated rare nuclear decay modes in natural gadolinium isotopes using high-purity samples originally intended for the Super-Kamiokande neutrino experiment. Despite achieving a sensitivity of 101910^{19}–102110^{21} years, no evidence for alpha or double beta decay to excited states of daughter nuclides was observed. The resulting limits improve upon previous measurements by two orders of magnitude. Future searches employing enriched isotopes or the “source = detector” approach are necessary to reach the theoretical predictions for these decays. Projects such as PIKACHU, which leverage high radiopurity GAGG scintillating crystals, hold promise for achieving the required sensitivities and providing crucial insights into rare nuclear processes that may occur in Gd isotopes and neutrino properties.

Acknowledgements The authors express their sincere gratitude to the Super-Kamiokande Collaboration for providing the Gd samples used in this study. This work was supported by the U.K. Science and Technology Facilities Council [grant numbers ST/T00200X/1, ST/V002821/1, ST/V006185/1, ST/X002438/1]. We would also like to thank Ruben Saakyan who suggested looking for rare decays in gadolinium screening data and connected our small collaboration.

Appendix

Four-panel fit of the 1193.2 keV γ-line

Figure 7. Fit of the 1193.2 keV γ\gamma-line from 160Gd{}^{160}\mathrm{Gd} 0ν/2νββ0\nu/2\nu\beta\beta decay into the excited 01+0^{+}_{1} state (1279.9 keV) of 160Dy{}^{160}\mathrm{Dy}.

Four-panel fit of the 1369.9 keV γ-line

Figure 8. Fit of the 1369.9 keV γ\gamma-line from 160Gd{}^{160}\mathrm{Gd} 0ν/2νββ0\nu/2\nu\beta\beta decay into the excited 02+0^{+}_{2} state (1456.8 keV) of 160Dy{}^{160}\mathrm{Dy}.

Two-panel fit of the 550.3 keV γ-line

Figure 9. Fit of the 550.3 keV γ\gamma-line from 152Gd{}^{152}\mathrm{Gd} α\alpha decay into the excited 21+2^{+}_{1} state (550.3 keV) of 148Sm{}^{148}\mathrm{Sm}.

References

  1. [1]L. Brown, Phys. Today 31, 9 (1978) https://doi.org/10.1063/1.2995181
  2. [2]C. L. Cowan, F. Reines, F. B. Harrison, H. W. Kruse, A. D. McGuire, Science, 124, 3212 (1956) https://doi.org/10.1126/science.124.3212.103
  3. [3]M. J. Dolinski, A. W. P. Poon, W. Rodejohann, Annu. Rev. Nucl. Part. Sci. 69, 219, 51 (2019) https://doi.org/10.1146/annurev-nucl-101918-023407
  4. [4]R. N. Mohapatra, et al., Rep. Prog. Phys. 70, 1757 (2007) https://doi.org/10.1088/0034-4885/70/11/R02
  5. [5]E. Takasugi, Phys. Lett. B 149 (1984) https://doi.org/10.1016/0370-2693(84)90426-X
  6. [6]J. Schachter, J. W. F. Valle, Phys. Rev. D 25 (1982) https://doi.org/10.1103/PhysRevD.25.2951
  7. [7]F. T. Avignone, S. R. Elliott, J. Engel, Rev. Mod. Phys. 80 (2008) https://doi.org/10.1103/RevModPhys.80.481
  8. [8]M. Fukugita, T. Yanagida, Phys. Lett. B 174 (1986) https://doi.org/10.1016/0370-2693(86)91126-3
  9. [9]F. F. Deppisch, L. Graf, J. Harz, W.-C. Huang, Phys. Rev. D 98, 055029 (2018) https://doi.org/10.1103/PhysRevD.98.055029
  10. [10]B. Pritychenko, V.I. Tretyak, Atomic Data and Nuclear Data Tables 161, 101694 (2025) https://doi.org/10.1016/j.adt.2024.101694
  11. [11]A. Barabash, Universe 9, 6, 290 (2023) https://doi.org/10.3390/universe9060290
  12. [12]D. S. Akerib, et al., Eur. Phys. J. C 80, 1044 (2020) https://doi.org/10.1140/epjc/s10052-020-8420-x
  13. [13]E. Aprile, et al., Eur. Phys. J. C 82, 599 (2022) https://doi.org/10.1140/epjc/s10052-022-10345-6
  14. [14]K. Abe, et al., Nucl. Instr. Meth. A 1027, 166248 (2022) https://doi.org/10.1016/j.nima.2021.166248
  15. [15]K. Hosokawa, et al., Prog. Theo. Exp. Phys. 2023, 1, 013H01 (2023) https://doi.org/10.1093/ptep/ptac170
  16. [16]K. Abe, et al., Nucl. Instr. Meth. A 1065, 169480 (2024) https://doi.org/10.1016/j.nima.2024.169480
  17. [17]M. Laubenstein, B. Lehnert, S. S. Nagorny, S. Nisi, Eur. Phys. J. C 83, 1114 (2023) https://doi.org/10.1140/epjc/s10052-023-12218-y
  18. [18]P. R. Scovell, M. Thiesse, S. A. Maouloud, C. Ghag, J. Dobson, Front. Phys. 11, 1310146 (2023) https://doi.org/10.3389/fphy.2023.1310146
  19. [19]P. R. Scovell, et al., J. Instrum. 19, P01017 (2024) https://doi.org/10.1088/1748-0221/19/01/P01017
  20. [20]J. Allison et al., Nucl. Instrum. Meth. A 835 (2016) https://doi.org/10.1016/j.nima.2016.06.125
  21. [21]P. R. Scovell, et al., Astropart. Phys., 97 (2018) https://doi.org/10.1016/j.astropartphys.2017.11.006
  22. [22]LabSOCS, Canberra Industries.
  23. [23]J. J. Back, Y. A. Ramachers, Nucl. Instr. Meth. A 586, 2 (2008) https://doi.org/10.1016/j.nima.2007.12.008
  24. [24]N. Nica, Nuclear Data Sheets 117, 1-229 (2014) https://doi.org/10.1016/j.nds.2014.02.001
  25. [25]National Nuclear Data Center https://www.nndc.bnl.gov (2020).
  26. [26]S. Ito, et al., Prog. Theor. Exp. Phys., 2017, 11, 113H01, (2017) https://doi.org/10.1093/ptep/ptx145
  27. [27]S. Nisi, L. Copia, I. Dafinei, M. L. Di Vacri, Intl. J. Mod. Phys. A 32, 30 (2017) https://doi.org/10.1142/S0217751X17430035
  28. [28]M. Thiesse, P. R. Scovell, L. Thompson, App. Rad. Iso. 188, 110384 (2022) https://doi.org/10.1016/j.apradiso.2022.110384
  29. [29]S. Nagorny, M. Laubenstein and S. Nisi, J. of Instr. 16 P03027 (2021) https://doi.org/10.1088/1748-0221/16/03/P03027
  30. [30]J. W. Beeman, et al., Eur. Phys. J. C 75, 12 (2015) https://doi.org/10.1140/epjc/s10052-015-3822-x
  31. [31]T. Omori, et al., Prog. Th. Exp. Phys. 2024, 3, 033D01 (2024) https://doi.org/10.1093/ptep/ptae026
  32. [32]K. Kamada et al., J. Cryst. Growth 352, 88 (2012) https://doi.org/10.1016/j.jcrysgro.2011.11.085
  33. [33]K. Kamada et al., J. Cryst. Growth 452, 15 (2016) https://doi.org/10.1016/j.jcrysgro.2016.04.037

Paper details

vlnc:W_x69rehn8
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents