WorksPhysics and Astronomy
High sensitivity characterization of an ultra-high purity NaI(Tl) crystal scintillator with the SABRE proof-of-principle detector
Abstract
We present new results on the radiopurity of a 3.4-kg NaI(Tl) crystal scintillator operated in the SABRE proof-of-principle detector setup. The amount of potassium contamination, determined by the direct counting of radioactive K, is found to be ppb, lowest ever achieved for NaI(Tl) crystals. With the active veto, the average background rate in the crystal in the 1-6 keV energy region-of-interest (ROI) is counts/day/kg/keV, which is a breakthrough since the DAMA/LIBRA experiment. Our background model indicates that the rate is dominated by Pb and that about half of this contamination is located in the PTFE reflector. We discuss ongoing developments of the crystal manufacture aimed at the further reduction of the background, including data from purification by zone refining. A projected background rate lower than 0.2 counts/day/kg/keV in the ROI is within reach. These results represent a benchmark for the development of next-generation NaI(Tl) detector arrays for the direct detection of dark matter particles.
The existence of dark matter is widely accepted [1], yet the particle nature of dark matter is still an open fundamental question. Over the last 25 years, a series of experimental efforts have been made to search for the interaction of dark matter particles with ordinary matter in underground laboratories [1]. However, almost all efforts yielded null results despite the impressive progress in background reduction techniques and detector technologies [8, 9] [2, 3, 4, 5, 6, 7]. Experimental efforts have mainly focused on the search for the so-called weakly-interacting massive particles (WIMPs) introduced in 1985 [10]. At present, the best sensitivity has been obtained with a ton-scale liquid xenon time projection chamber: the spin-independent WIMP-nucleon scattering cross section is estimated to be less than for a 30-GeV WIMP [2].
As an alternative way to probe particle dark matter, it was shown that the motion of the Earth around the Sun in the dark matter halo can induce an annual modulation in the dark matter interaction rate [11]. The modulation, which is of the order of a few % in amplitude, has a specific phase that allows to discriminate against other non-modulating and modulating backgrounds. This approach has been exploited by the DAMA experiment (short for DAMA/NaI and DAMA/LIBRA) [12], which has been observing a clear annual modulation in its array of 250-kg extremely radiopure NaI(Tl) crystals. The annual modulation is consistent with the dark matter hypothesis, but in the standard WIMP framework, it is in tension with other sensitive direct detection experiments [8, 13] [2, 3, 4, 6]. Due to a potential target-dependence of WIMP-nucleon interactions, a verification using the same target material is indispensable.
To test the DAMA annual modulation, ANAIS [14] and COSINE [15] are currently operating 112.5 kg and 106 kg of NaI(Tl) crystal scintillators, respectively. However, they could not provide a definite answer to the longstanding controversy due to their signal-to-noise ratio diminished by constant radioactive backgrounds several times higher than that observed in the DAMA crystals [16, 17]. Large fractions of the backgrounds come from radioactive contaminants in the crystal, most notably , and . This indicates that the definitive test of the annual modulation claim has to be addressed by next-generation experiments using NaI(Tl) crystals with radiopurity similar to or below the DAMA level, such as in the proposed SABRE experiment [18].
In this article, we report a detailed study of the background components of an ultra-high radiopurity crystal grown for the SABRE experiment. While preliminary measurements on the same crystal had been performed on the surface [19] and underground inside a passive shielding [20], this study was carried out underground in the SABRE proof-of-principle (PoP) liquid scintillator active veto shielding (Figure 1) in the Hall C of Laboratori Nazionali del Gran Sasso (LNGS), Italy [18].

Figure 1. Technical drawing of the PoP system. From the outer to the inner region: water tanks (blue), HDPE layers (white), veto vessel (light grey), and crystal enclosure (brown).
This crystal—NaI-33—was grown using ultra-high purity NaI powder after a series of R&D activities [21, 22]. The preparation of the powder and the crucible were carried out at Princeton [19, 22] and the crystal growth was done at Radiation Monitoring Devices (RMD) in Massachusetts using the vertical Bridgman method [19, 23]. The crystal has a mass of 3.4 kg, and is wrapped with ten layers of PTFE tape (about 1 mm) and optically coupled directly to 3-inch Hamamatsu R11065 photomultiplier tubes (PMTs) that features high quantum efficiency and low radioactivity [19]. The crystal-PMT assembly is sealed inside a 2-mm-thick, high-purity copper enclosure [20], whose inner volume is constantly flushed with nitrogen gas to purge moisture and radon. The enclosure is deployed vertically via a 2-mm-thick, 121-cm-long blind-end copper tube into a 1.3-m-diameter, 1.5-m-long stainless steel veto vessel [18]. The veto vessel is filled with 2 tonnes of pseudocumene (PC) liquid scintillator and 2.86 g/L of 2,5-diphenyloxazole (PPO) wavelength shifter, and is instrumented with ten Hamamatsu R5912 8-inch PMTs to serve as active veto to enhance background rejection [21, 22]. The inner surface of the veto vessel is lined with reflective Lumirror foils to improve light collection [22]. The light yield of the active veto is measured to be 0.52±0.01 photoelectrons/keV (phe/keV) at 2615 keV using a Th calibration source. The veto vessel is further shielded from the cavern by an inner high-density polyethylene (HDPE, 10 cm on the top and bottom and at least 40 cm on the sides) and an outer gamma shielding (∼90 cm of water on the sides and top, and 15-cm of lead on the bottom) [18]. A PTFE tube is used to position wire-mount calibration sources next to the copper enclosure.
The DAQ system consists of a CAEN V1495 custom FPGA trigger and two CAEN V1720 ADC boards (12-bit, 250 MS/s) [22]. The data acquisition is triggered by the logical AND between the two PMTs coupled to the crystal with a 125-ns coincidence window irrespective of the status of the veto detector [22]. Upon trigger, waveforms in the subsequent 3.5 μs window in all PMTs are digitized and read out by a dedicated DAQ software [24].
The light yield and FWHM resolution of NaI-33 crystal scintillator, measured by fitting the 59.5-keV photopeak of an Am source, is determined to be 12.1 ± 0.2 phe/keV and 13.5%, respectively. The light yield is slightly higher than that measured in [20]. To determine the background rate in the 1-6 keV ROI, data were taken between August 9, 2020 and September 5, 2020 for a total exposure of 26.4 days. To reject coincident backgrounds, events with energy larger than 50 keV in the veto are rejected with 42% veto rejection power in the ROI. In addition, to reduce noise, the following cuts were applied:
;
Trigger time delay ;
No. of clusters in each PMT ;
;
;
.
where is the amplitude at time in ns, and is the pulse area between and in ns. These variables are described in detail in [20]. Figure 2 shows the final acceptance rate as a function of energy after these cuts. The average event acceptance in the ROI is estimated to be 77.6% [25]. The energy spectrum after cuts and efficiency correction is shown in Figure 3 (black dots) up to 100 keV, and in Figure 4 up to 20 keV. The measured rate in 1-6 keV (1-10 keV) is 1.20 ± 0.05 cpd/kg/keV (1.09 ± 0.04 cpd/kg/keV).

Figure 2. Acceptance of crystal events after noise rejection cuts as a function of energy, evaluated using a calibration source.

Figure 4. Energy spectrum from NaI-33 below 20 keV with the spectral fit. The spectrum is shown after selection cuts and efficiency-corrected.

Figure 3. Energy spectrum from NaI-33 up to 100 keV with a spectral fit. The spectrum is shown after selection cuts and efficiency correction.
A spectral analysis was performed to quantify the contribution of different background components [25]. Predicted spectral shapes from different background sources were calculated by Monte Carlo [26]. In the fit, the activities of the following components were treated as free or semi-free parameters: K, Pb, H, Ra, Th, I, Te, Te, and a flat component which includes Rb and other internal and external background contributions. In addition, Pb from the PTFE reflector wrapping the crystal was included to reproduce the peak at ∼12 keV due to X-rays from Pb. In the spectral fit, a Gaussian penalty was applied to K with mean equal to 0.14±0.01 mBq/kg corresponding to the prediction from ICP-MS measurements [19]. Assuming secular equilibrium for and chain segments, Gaussian penalties (() Bq/kg for and () Bq/kg for ) were implemented based on measurements from - and - time-correlated events, respectively [20].
The result of the fit is shown in Figure 3 and in Figure 4. The -value of the fit is equal to 0.26 with . Table I summarises the breakdown of the background components determined from the fit. The background rate in the ROI, dominated by in the bulk of the crystal and on the surface of the PTFE reflector, is found to be conservatively equal to cpd/kg/keV. The activity is estimated to be mBq/kg. This value is consistent with an independent measurement of using the coincidence be-
tween 3.2-keV X-ray/auger electrons in the crystal and 1.46 MeV gamma ray in the active veto, which yielded mBq/kg ( ppb, or ppb at 90% CL) despite limited statistics.
| Source | Activity in NaI-33 [mBq/kg] | Rate in ROI in NaI-33 [cpd/kg/keV] | Projected rate in ROI [cpd/kg/keV] |
| (bulk) | |||
| — | |||
| — | |||
| (PTFE) | |||
| Other | |||
| total |
Table 1. Background components in NaI-33 from the spectral fit, current rate in ROI (1-6 keV), and projected rate in ROI for future crystals. The future rate assumes underground crystal production with zone-refining purification and improved reflector radiopurity. The activity of in the reflector is normalized to the crystal mass for comparison with bulk activities. Upper limits are given as one-sided 90% CL. Rates are conservatively calculated using upper limits.
The activity of bulk is determined to be mBq/kg, consistent with earlier measurements using counting of [19, 20]. Although this value is about one order of magnitude larger than that in DAMA, it is smaller than those in ANAIS and COSINE, where on average the activity is about 1 mBq/kg [27, 28].
The activity of in the reflector is measured to be mBq, with the first 4 m from the crystal surface mostly responsible for background events. Although the commercial PTFE tape was carefully cleaned by acid leach [22], this analysis indicates that the cleaning procedure is not very effective in removing impurities. The radioactivity of PTFE has been studied by other experiments: the special PTFE used in the CUORE-0 experiment has a activity µBq/kg (90% CL) [29, 30]; in the DarkSide-50 experiment [31], activity in PTFE is measured to be mBq/kg (90% CL) by -spectroscopy [32] and mBq/kg by counting [33]. These considerations imply that the background due to the PTFE reflector can be reduced to a secondary component by custom or special manufacturing of the PTFE tapes.
The activation rate of in NaI is measured to be 83 27 cpd/kg by ANAIS [34]. NaI-33 has undergone a 9-month surface exposure at sea level and a 12-month cooling underground. At the predicted activation rate, NaI-33 would have seen a activity of (37 12) µBq/kg. The actual activity is estimated to be (12 7) µBq/kg. Although limited by statistics, the production rate in NaI-33 seems to be lower. One possibility is that the previously reported rate is not entirely due to cosmogenic , and instead varying amount of is introduced into the crystal as NaOH impurity depending on the powder drying and crystal growth method [22].
To further improve radiopurity of crystals to be grown in the future, we have tested the zone refining of ultra-high purity NaI powder. Table II, reproduced from [35], shows that many impurities are greatly reduced in the first three samples (about 50% of the ingot), and in particular and are reduced to negligible levels.
| Isotope | Impurity concentration (ppb) | |||||
| Powder | ||||||
| 7.5 | <0.8 | <0.8 | 1 | 16 | 460 | |
| <0.2 | <0.2 | <0.2 | <0.2 | <0.2 | 0.7 | |
| 1.0 | 0.4 | 0.4 | <0.4 | 0.5 | 0.5 | |
| 14 | 10 | 8 | 6 | 7 | 140 | |
| 44 | 0.3 | 0.2 | 0.5 | 3.3 | 760 | |
| 9 | 0.1 | 0.2 | 1.4 | 19 | 330 |
Table II. In the zone refining test, 53 zone passes were performed on 744 g of ultra-high purity NaI powder, and five samples were taken from different ingot locations and analyzed to determine the purification efficiency [35]. A few other elements are included to show effectiveness for other metallic impurities.
Quantitatively, the reduction of K and Rb depends on the segregation coefficient and the fraction of the purified material to be reserved [22, 35]. The segregation coefficient of K is estimated to be 0.57 while for Rb it is < 0.59 at 90% CL [35]. Based on this, Table III lists the reduction of impurity concentration for different combinations of number of zone passes and fraction of the material reserved (see Fig. 9 in [35]). Based on this, the purity can be further improved by a factor of 10 with 25 zone passes (1-week processing time) and at the cost of 20% of the initial material, or at the same cost of material but with 50 zone passes (2-week processing time), the average purity can be improved by a factor of 25. Therefore, for future NaI crystals, the background due to and can be made subdominant compared to the amplitude of the modulation.
| No. of passes | Impurity reduction | ||||
| Fraction of material retained | |||||
| 50% | 60% | 70% | 80% | 90% | |
| 10 | 0.3 | 0.36 | 0.42 | 0.48 | 0.56 |
| 25 | 0.03 | 0.050 | 0.07 | 0.11 | 0.25 |
| 50 | 0.001 | 0.0026 | 0.0086 | 0.037 | 0.18 |
Table III. Reduction of K for different numbers of passes and when different fractions of the ingot is reserved.
Pb is more or less uniformly reduced by a factor of without following the typical impurity distribution of zone refining [35]. Although one can still exploit this factor of 3, zone refining is not as efficient in removing . Thus we have started new R&D activity to explore alternative purification methods and have achieved preliminary progress. Since the analysis of data on the removal of Pb is still in progress, in this article, we adopt the conservative reduction factor for obtained from zone refining alone.
With the aforementioned measures, a projected background rate of 0.21 cpd/kg/keV can be practically achieved in the future SABRE crystals (see Table I). In this projection, we have assumed the contamination level of NaI-33 for and bulk scaled by the reduction factor of zone refining in Table III. For in the reflector, we have assumed a contamination equal to the upper limit for the PTFE measured in DarkSide-50. To suppress cosmogenic backgrounds to a negligible level, we are also investigating the possibility of establishing an underground crystal growth facility at the Canfranc Laboratory in Spain [36].
To quantify the sensitivity to the dark matter-induced annual modulation, the figure-of-merit (FoM) defined in [37] can be used where a small FoM indicates a high sensitivity. In this framework, the DAMA/LIBRA phase II [38] has FoM = dkgkeV. A similar value of FoM can be obtained with a 60 kg5 yr exposure and the predicted background rate of 0.21 cpd/kg/keV in Table I. With the same exposure and background rate, the minimum detectable rate [39] is expected to be cpd/kg/keV at 90% CL, which is about 5 times less than DAMA modulation amplitude of cpd/kg/keV in 1-6 keV ROI [38]. On the other hand, for the current generation of NaI(Tl) detector arrays, several decades are needed to obtain the same FoM as DAMA/LIBRA phase II. Therefore, the present results obtained from NaI-33 together with the planned improvements discussed above will be fundamental for the next generation NaI-based dark matter detectors.
The growth of NaI-33 and its characterization in the SABRE detector are a breakthrough in the technology to support next-generation NaI experiments to probe the long-standing DAMA result: the background rate in the – keV ROI is determined to be cpd/kg/keV, several times lower than all presently running NaI-based experiments except DAMA. At present, the dominant backgrounds are due to bulk and in the PTFE reflector. With a combination of zone refining, special low-radioactivity PTFE tape, and underground growth of crystals, the background in the ROI in future crystals can be further reduced to cpd/kV/keV. This result marks a significant improvement in the quest for testing the dark matter-induced annual modulation, and paves the road to the next-generation NaI-based experiments with higher sensitivities.
Acknowledgments
This work has been supported by INFN funding and National Science Foundation under the award number PHY-1242625, PHY-1506397 and PHY-1620085. We thank Ezio Previtali, Andrea Giuliani, and Monica Sisti for discussions about PTFE radiopurity in CUORE. We thank Fausto Ortica and Aldo Romani from Perugia University for support in photospectrometric measurements. We thank the Gran Sasso Laboratory for the support during the installation of the SABRE PoP setup.
References
[?] G. Bertone and D. Hooper, Reviews of Modern Physics 90, 045002 (2018).
[?] E. Aprile, J. Aalbers, F. Agostini, M. Alfonsi, L. Althueser, F. Amaro, M. Anthony, F. Arneodo, L. Baudis, et al., Physical Review Letters 121, 111302 (2018).
[?] D. Akerib, S. Alsum, H. Araújo, X. Bai, A. Bailey, J. Balajthy, P. Beltrame, E. Bernard, A. Bernstein, T. Biesiadzinski, et al., Physical Review Letters 118, 021303 (2017).
[?] X. Cui, A. Abdukerim, W. Chen, X. Chen, Y. Chen, B. Dong, D. Fang, C. Fu, K. Giboni, F. Giuliani, et al., Physical Review Letters 119, 181302 (2017).
[?] P. Agnes, I. Albuquerque, T. Alexander, A. Alton, G. Araujo, M. Ave, H. Back, B. Baldin, G. Batignani, K. Biery, et al., Physical Review D 98, 102006 (2018).
[?] G. Angloher, M. Bauer, I. Bavykina, A. Bento, C. Bucci, C. Ciemniak, G. Deuter, F. von Feilitzsch, D. Hauff, P. Huff, et al., The European Physical Journal C 72, 1 (2012).
[?] A. Aguilar-Arevalo, D. Amidei, D. Baxter, G. Cancelo, B. C. Vergara, A. Chavarria, J. D’Olivo, J. Estrada, F. Favela-Perez, R. Gaior, et al., Physical Review Letters 125, 241803 (2020).
[8] R. Agnese, T. Aralis, T. Aramaki, I. Arnquist, E. Azadbakht, W. Baker, S. Banik, D. Barker, D. Bauer, T. Binder, et al., Physical Review Letters 121, 051301 (2018).
[9] R. Agnese, T. Aralis, T. Aramaki, I. Arnquist, E. Azadbakht, W. Baker, S. Banik, D. Barker, D. Bauer, T. Binder, et al., Physical Review Letters 122, 069901 (2019).
[10] M. W. Goodman and E. Witten, Physical Review D 31, 3059 (1985).
[11] A. K. Drukier, K. Freese, and D. N. Spergel, Physical Review D 33, 3495 (1986).
[12] R. Bernabei et al., Prog. Part. Nucl. Phys. 114, 103810 (2020).
[13] P. Agnes, I. Albuquerque, T. Alexander, A. Alton, G. Araujo, D. M. Asner, M. Ave, H. O. Back, B. Baldin, G. Batignani, et al., Phys. Rev. Lett. 121, 081307 (2018).
[14] J. Amaré, S. Cebrián, I. Coarasa, C. Cuesta, E. García, M. Martínez, M. Oliván, Y. Ortigosa, A. O. de Solórzano, J. Puimedón, et al., The European Physical Journal C 79, 1 (2019).
[15] G. Adhikari, P. Adhikari, E. B. de Souza, N. Carlin, S. Choi, W. Choi, M. Djamal, A. Ezeribe, C. Ha, I. Hahn, et al., The European Physical Journal C 78, 1 (2018).
[16] J. Amaré, S. Cebrián, I. Coarasa, C. Cuesta, E. García, M. Martínez, M. Oliván, Y. Ortigosa, A. O. de Solórzano, J. Puimedón, et al., Physical Review Letters 123, 031301 (2019).
[17] G. Adhikari, P. Adhikari, E. B. de Souza, N. Carlin, S. Choi, M. Djamal, A. Ezeribe, C. Ha, I. Hahn, E. Jeon, et al., Physical Review Letters 123, 031302 (2019).
[18] M. Antonello, E. Barberio, T. Baroncelli, J. Benziger, L. Bignell, I. Bolognino, F. Calaprice, S. Copello, D. D’angelo, G. D’imperio, et al., The European Physical Journal C 79, 1 (2019).
[19] B. Suerfu, M. Wada, W. Peloso, M. Souza, F. Calaprice, J. Tower, and G. Ciampi, Physical Review Research 2, 013223 (2020).
[20] M. Antonello, I. Arnquist, E. Barberio, T. Baroncelli, J. Benziger, L. Bignell, I. Bolognino, F. Calaprice, S. Copello, L. Dafinei, et al., The European Physical Journal 81, 1 (2021).
[21] E. K. Shields, SABRE: A search for dark matter and a test of the DAMA/LIBRA annual-modulation result using thallium-doped sodium-iodide scintillation detectors, Ph.D. thesis, Princeton University (2015).
[22] Suerfu, Developing Ultra-Low Background Sodium-Iodide Crystal Detector for Dark Matter Searches, Ph.D. thesis, Princeton University (2018).
[23] P. Bridgman, in Proceedings of the American academy of arts and sciences, Vol. 60 (JSTOR, 1925) pp. 305–383.
[24] B. Suerfu, Journal of Instrumentation 13, T12004 (2018).
[25] A. Mariani, The Proof-of-Principle of the SABRE experiment for the search of galactic dark matter through annual modulation, Ph.D. thesis, Gran Sasso Science Institute (2021).
[26] M. Antonello, E. Barberio, T. Baroncelli, J. Benziger, L. Bignell, I. Bolognino, F. Calaprice, S. Copello, D. D’Angelo, G. D’Imperio, et al., Astroparticle Physics 106, 1 (2019).
[27] J. Amaré, S. Cebrián, I. Coarasa, C. Cuesta, E. García, M. Martínez, M. Oliván, Y. Ortigosa, A. O. De Solórzano, J. Puimedón, et al., The European Physical Journal C 79, 1 (2019).
[28] G. Adhikari, P. Adhikari, E. B. de Souza, N. Carlin,
References
- [8]bakht, W. Baker, S. Banik, D. Barker, D. Bauer, T. Binder, et al., Physical Review Letters 121, 051301 (2018).
- [9]R. Agnese, T. Aralis, T. Aramaki, I. Arnquist, E. Azadbakht, W. Baker, S. Banik, D. Barker, D. Bauer, T. Binder, et al., Physical Review Letters 122, 069901 (2019).
- [10]M. W. Goodman and E. Witten, Physical Review D 31, 3059 (1985).
- [11]A. K. Drukier, K. Freese, and D. N. Spergel, Physical Review D 33, 3495 (1986).
- [12]R. Bernabei et al., Prog. Part. Nucl. Phys. 114, 103810 (2020).
- [13]P. Agnes, I. Albuquerque, T. Alexander, A. Alton, G. Araujo, D. M. Asner, M. Ave, H. O. Back, B. Baldini, G. Batignani, et al., Phys. Rev. Lett. 121, 081307 (2018).
- [14]J. Amaré, C. Cebrián, I. Coarasa, C. Cuesta, E. García, M. Martínez, M. Oliván, Y. Ortigoza, A. O. de Solórzano, J. Puimedón, et al., The European Physical Journal C 79, 1 (2019).
- [15]G. Adhikari, P. Adhikari, E. B. de Souza, N. Carlin, S. Choi, W. Choi, M. Djamal, A. Ezeribe, C. Ha, I. Hahn, et al., The European Physical Journal C 78, 1 (2018).
- [16]J. Amaré, C. Cebrián, I. Coarasa, C. Cuesta, E. García, M. Martínez, M. Oliván, Y. Ortigoza, A. O. de Solórzano, J. Puimedón, et al., Physical Review Letters 123, 031301 (2019).
- [17]G. Adhikari, P. Adhikari, E. B. de Souza, N. Carlin, S. Choi, W. Choi, M. Djamal, A. Ezeribe, C. Ha, I. Hahn, E. Jeon, et al., Physical Review Letters 123, 031302 (2019).
- [18]M. Antonello, E. Barberio, T. Baroncelli, J. Benziger, L. Bignell, I. Bolognino, F. Calaprice, S. Copello, D. D’angelo, G. D’Imperio, et al., The European Physical Journal C 79, 1 (2019).
- [19]B. Suerfu, M. Wada, W. Peloso, M. Souza, F. Calaprice, J. Tower, and G. Ciampi, Physical Review Research 2, 013223 (2020).
- [20]M. Antonello, I. Arnquist, E. Barberio, T. Baroncelli, J. Benziger, L. Bignell, I. Bolognino, F. Calaprice, S. Copello, I. Dafinei, et al., The European Physical Journal C 81, 1 (2021).
- [21]E. K. Shields, SABRE: A search for dark matter and a test of the DAMA/LIBRA annual-modulation result using thallium-doped sodium-iodide scintillation detectors, Ph.D. thesis, Princeton University (2015).
- [22]Suerfu, Developing Ultra-Low Background Sodium-Iodide Crystal Detector for Dark Matter Searches, Ph.D. thesis, Princeton University (2018).
- [23]P. Bridgman, in Proceedings of the American academy of arts and sciences, Vol. 60 (JSTOR, 1925) pp. 305–383.
- [24]B. Suerfu, Journal of Instrumentation 13, T12004 (2018).
- [25]A. Mariani, The Proof-of-Principle of the SABRE experiment for the search of galactic dark matter through annual modulation, Ph.D. thesis, Gran Sasso Science Institute (2021).
- [26]M. Antonello, E. Barberio, T. Baroncelli, J. Benziger, L. Bignell, I. Bolognino, S. Copello, D. D’Angelo, G. D’Imperio, et al., Astroparticle Physics 106, 1 (2019).
- [27]J. Amaré, S. Cebrián, I. Coarasa, C. Cuesta, E. García, M. Martínez, M. Oliván, Y. Ortigoza, A. O. De Solórzano, J. Puimedón, et al., The European Physical Journal C 79, 1 (2019).
- [28]G. Adhikari, P. Adhikari, E. B. de Souza, N. Carlin, J. Choi, S. Choi, M. Djamal, A. Ezeribe, L. Franca, C. Ha, et al., arXiv preprint arXiv:2101.11377 (2021).
- [29]C. Alduino, K. Alfonso, D. Artusa, F. Avignone III, O. Azzolini, M. Balata, T. Banks, G. Bari, J. Beeman, F. Bellini, et al., Journal of Instrumentation 11, P07009 (2016).
- [30]C. Alduino, K. Alfonso, D. Artusa, F. Avignone, O. Azzolini, T. Banks, G. Bari, J. Beeman, F. Bellini, A. Bersani, et al., The European Physical Journal C 77, 1 (2017).
- [31]P. Agnes, T. Alexander, A. Alton, K. Arisaka, H. Back, B. Baldin, K. Biery, G. Bonfini, M. Bossa, A. Brigatti, et al., Physics Letters B 743, 456 (2015).
- [32]M. Laubenstein and P. Meyer, private communication (2020).
- [33]G. Zuzel, K. Pelczar, and M. Wójcik, Applied Radiation and Isotopes 126, 165 (2017).
- [34]J. Amaré, J. Castel, S. Cebrián, I. Coarasa, C. Cuesta, T. Dafni, J. Galán, E. García, J. Garza, F. Iguaz, I. Irastorza, G. Luzón, M. Martínez, H. Mirallas, M. Oliván, Y. Ortigosa, A. Ortiz de Solórzano, J. Puimedón, E. Ruiz-Chóliz, M. Sarsa, J. Villar, and P. Villar, Astroparticle Physics 97, 96 (2018).
- [35]B. Suerfuß, F. Calaprice, and M. Souza, “Zone refining of ultra-high purity sodium iodide for low-background detectors,” (2021), arXiv:2105.06431 [physics.ins-det].
- [36]A. Ianni, Journal of Physics: Conference Series 675, 012002 (2016).
- [37]I. Coarasa, J. Amaré, S. Cebrián, C. Cuesta, E. García, M. Martínez, M. Oliván, Y. Ortigoza, A. O. de Solórzano, J. Puimedón, et al., The European Physical Journal C 79, 1 (2019).
- [38]R. Bernabei, P. Belli, A. Bussolotti, F. Cappella, V. Caracciolo, R. Cerulli, C. Dai, A. d’Angelo, A. Di Marco, H. He, et al., Nuclear Physics and Atomic Energy 19, 307 (2018).
- [39]N. Tsoufanidis, Measurement and detection of radiation (CRC press, 2010).