Li2100deplMoO4 Scintillating Bolometers for Rare-Event Search Experiments

We report on the development of scintillating bolometers based on lithium molybdate crystals that contain molybdenum that has depleted into the double-β active isotope 100Mo (Li2100deplMoO4). We used two Li2100deplMoO4 cubic samples, each of which consisted of 45-millimeter sides and had a mass of 0.28 kg; these samples were produced following the purification and crystallization protocols developed for double-β search experiments with 100Mo-enriched Li2MoO4 crystals. Bolometric Ge detectors were utilized to register the scintillation photons that were emitted by the Li2100deplMoO4 crystal scintillators. The measurements were performed in the CROSS cryogenic set-up at the Canfranc Underground Laboratory (Spain). We observed that the Li2100deplMoO4 scintillating bolometers were characterized by an excellent spectrometric performance (∼3–6 keV of FWHM at 0.24–2.6 MeV γs), moderate scintillation signal (∼0.3–0.6 keV/MeV scintillation-to-heat energy ratio, depending on the light collection conditions), and high radiopurity (228Th and 226Ra activities are below a few µBq/kg), which is comparable with the best reported results of low-temperature detectors that are based on Li2MoO4 using natural or 100Mo-enriched molybdenum content. The prospects of Li2100deplMoO4 bolometers for use in rare-event search experiments are briefly discussed.

Lithium is also an element that has had a long-standing interest for use in rare-event search experiments. The dominant isotope, i.e., 7 Li (92% of natural lithium), is a light nucleus with a non-zero spin, thus making it a viable candidate for probing spin-dependent dark matter interactions [33,34]. Moreover, 7 Li is a good target for searching for solar axions via a resonant absorption of an axion by 7 Li and its subsequent γ de-excitation [11,[35][36][37][38]. Last but not least, presence of 6 Li (8% of natural lithium) allows neutron detection via a 6 Li(n, t)α reaction, which is characterized by a high cross-section to thermal neutrons; furthermore, enrichment in 6 Li (of up to 95%) is feasible and can enhance the detection efficiency. Thus, Li-containing detectors are of special interest for neutron flux monitoring in rare-event search experiments [39,40].
In this work, we present the development and investigation of scintillating bolometers based on a new type of the Li 2 MoO 4 compound, which is produced from molybdenum that is depleted in 100 Mo (Li 2 100depl MoO 4 ). The low concentration of the double-β active isotope 100 Mo (the transition energy is Q ββ = 3084 keV; the half-life is ∼7 × 10 18 yr [16,41]) significantly reduces the related background counting rate, which is on the level of 10 mBq/kg in a Li 2 100 MoO 4 crystal (an order of magnitude lower for the natural one), and it can be a dominant internal background of this material. Therefore, Li 2 100depl MoO 4 appears to be more suitable than natural or 100 Mo-enriched Li 2 MoO 4 crystals for dark matter and axion search experiments using 7 Li, for double-β decay searches in 92 Mo and 98 Mo, and for 6 Li-based neutron detection. Moreover, Li 2 100depl MoO 4 can be used in bolometric experiments to search for double-β decay in 100 Mo as a complementary detector for better understanding the background model. Therefore, a goal of this work is to study the prospects of Li 2 100depl MoO 4 low-temperature detectors for rare-event search applications.

Crystals Production and Construction of Detectors
In this study, we used two Li 2 100depl MoO 4 scintillation crystals with a size of 45 × 45 × 45 mm and a mass of around 0.28 kg each, which are identical to the Li 2 100 MoO 4 crystals that were produced for the CROSS experiment [26]. The samples were cut from the same crystal boule, which was grown using the low-thermal-gradient Czochralski technique as detailed in [42]. A 4N purity lithium carbonate, which was selected for the LUMINEU and CUPID-Mo crystals production [13,14,17], and molybdenum oxide that was depleted in 100 Mo (∼0.01% of 100 Mo; i.e., 1000 times lower than in natural Mo) have been used as the starting materials. The 100depl MoO 3 material purification, solid-phase synthesis of the Li 2 100depl MoO 4 compound, and crystal growth in a Pt crucible using the low-thermal-gradient Czochralski method (double crystallization approach) were realized while following the protocols of Mo-containing crystals production that was developed by LUMINEU [13,14,43] and adopted by CUPID-Mo [17]. A large crystal boule (the cylindrical part is around 60 × 100 mm) has been grown to have a crystal yield of about 80% from the initial charge [42] and two twin cubic samples (with a few millimeter chamfers) were produced.
The assembly of detectors have been carried out in an ISO class 4 clean room at IJCLab (Orsay, France). The first Li 2 100depl MoO 4 sample (LMO-depl-1), corresponding to the upper part of the boule, was mounted inside a Cu housing using polytetrafluoroethylene (PTFE) supporting elements and Cu screws, as shown in Figure 1 (left). The holder design is rather similar to one used in bolometric measurements, using the same size of Li 2 100 MoO 4 crystal [23] and slightly larger TeO 2 sample [44]. The second Li 2 100depl MoO 4 sample (LMO-depl-2), corresponding to the bottom part of the boule, was assembled using Cu frames and columns and PTFE pieces, as seen in Figure 1 (middle). This sample was a part of a twelve-crystal array, in which all of the other crystals were Li 2 100 MoO 4 [25]. The lateral side of the LMO-depl-2 crystal was surrounded by a reflective film (Vikuiti™) to improve the light collection, while the Cu housing of the LMO-depl-1 detector served as a reflective cavity; however the aperture reduces the direct sight of the light detector. Thus, the light collection is sub-optimal both due to the poor reflectivity of copper and low scintillation photons collection efficiency of the light detector(s). depl-2 (middle). Both crystals have two epoxy-glued sensors; the left one is an NTD Ge thermistor, while the right one is a P-doped Si heater. Each scintillator was accompanied by a circular bolometric Ge light detector, as can be seen in the transparent area of the crystal in the middle panel. An additional square-shaped Ge light detector (right) was used for the LMO-depl-2 sample. All light detectors were instrumented with an NTD Ge sensor.
To detect particle interactions, each Li 2 100depl MoO 4 crystal was instrumented with a neutron transmutation-doped Ge [45] thermistor (NTD). A sensor with a size of 3 × 3 × 1 mm was epoxy-glued onto the crystal's surface using six spots of a bi-component glue (Araldite ® Rapid). The temperature-dependent resistance of the NTDs can be approximated as R(T) = R 0 · e (T 0 /T) 0.5 using the parameters R 0 ∼ 1 Ω and T 0 ∼ 3.7 K. NTD sensors with similar irradiation parameters have been used in the CUPID-Mo experiment [17] and in CUPID-related R&D tests [22][23][24][25]. Moreover, a P-doped Si heater [46] was glued on each crystal using a veil of the epoxy glue. This heating element was exploited to be injected using the Jules effect power, which can be used for, e.g., the stabilization of the thermal gain [47], optimization of the detector working point, heating of a device if necessary (as in CUPID-0 [48]), or pile-up simulations [49]. To provide electrical contacts, the NTDs and heaters were wire-bonded using 25 µm Au wires.
To allow for the detection of Li 2 100depl MoO 4 scintillation (with the emission maximum at ∼590 nm at low temperatures [12]), we accompanied crystals with bolometric detectors that were based on electronic-grade purity Ge wafers, which were supplied by Umicore (Belgium) [50]. Two of them had a circular shape (with a size of 45 × 0.18 mm each), while the third device was square-shaped (45 × 45 × 0.30 mm). All of the Ge disks were coated on both sides with a 70 nm SiO layer, which was aimed at reducing the light reflection [17,48,51]. Smaller NTD sensors (3 × 1 × 1 mm or 3 × 0.7 × 1 mm) were attached to the Ge wafers using a veil of epoxy glue. The Ge disks were PTFE-clamped in the Cu structure. A single circular light detector (LD-1-c) was coupled to the LMO-depl-1 crystal, while both the circular (LD-2-c) and square-shaped (LD-2-s) bolometric photodetectors viewed the LMO-depl-2 crystal. The mounted Ge light detectors are shown in Figure 1 (middle and right).

Operation at Canfranc Underground Laboratory
The Li 2 100depl MoO 4 scintillating bolometers were operated in the CROSS cryogenic setup (C2U) [23,52] at the Canfranc Underground Laboratory (LSC, Spain), which provided a substantial reduction in the cosmic muon flux thanks to the rock overburden [53]. The detectors were assembled as parts of scintillating bolometer arrays and were installed inside the cryostat, as illustrated in Figure 2. The facility exploits the HEXA-DRY dilution fridge by CryoConcept (France), which is equipped with the Ultra-Quiet Technology™ [54] to decouple a pulse tube (Cryomech PT415) from the dilution unit, thus reducing vibrations [55]. To further improve the noise conditions, the detector arrays were spring-suspended from the cold plate of the cryostat. The detector volume inside the cryostat is shielded on top by a 13 cm thick disk made of interleaved lead and copper (partially seen in Figure 2), while the outer vacuum chamber is surrounded by a 25 cm thick layer of low-radioactivity lead. In addition, a deradonized air (∼1 mBq/m 3 of Rn [56]) flow around the cryostat was supplied the whole time during the experiment with the LMO-depl-2 detector. After reaching the base temperature of the cryostat (∼10 mK), we regulated the detector plate temperature at 18 mK and then 12 mK for measurements using the LMO-depl-1 bolometer; the plate temperature was regulated at 14 mK for the LMO-depl-2 bolometer operation. The control and readout of the bolometers was performed with the help of lownoise, room temperature, DC front-end electronics, which were restyled from the Cuoricino experiment [63]. The data acquisition (DAQ) system was composed of two 12-channel boards with an integrated 24-bit ADC and a programmable 6-pole Bessel-Thomson anti-aliasing filter (the cut-off frequency of the low-pass filter was set at 300 Hz) [64,65].
In order to find an optimal working point of the detectors that represents the best signal-to-noise ratio, we spanned the bolometric response with respect to the current across an NTD [66]. We used heaters to inject thermal pulses to both Li 2 100depl MoO 4 bolometers, while LED generated photons, which were transmitted from a room temperature LED (the emission maximum was at ∼880 nm) through an optic fiber, were exploited for the light detectors. The heater/LED signal injection was performed with the help of a wave function generator (Keysight 33500B). As a result of the optimization, we polarized the NTDs of the detectors under a current strength of a few nA, which reduced the NTD resistances from hundreds of MΩ (at low power) to a few MΩ (at the working point).
For each operational temperature, we performed measurements using a removable 232 Th γ source, which was made of a thoriated tungsten wire; we also performed measurements without the source (data are referred to as the calibration and background, respectively). Even if the γ source was primarily conceived for the calibration of the Li 2 100depl MoO 4 bolometers, we also used it to evaluate the energy scale of the light detectors, similar to [17,44].
The continuous data of each channel were acquired at a sampling rate of 2 kS/s and stored on a disk for the offline analysis. We processed the data with the help of a MATLAB-based analysis tool [67], which implements the signal processing using the Gatti-Manfredi optimum filter [68] to maximize the signal-to-noise ratio. In order to apply the filter, we used data-based information about the signal shape (represented by an average signal of high-energy events in the order of tens) and measured noise (represented by 10,000 waveforms with no signal). The data were triggered using a threshold corresponding to 5σ of the filtered noise. For each triggered signal, we collected information about its amplitudes (i.e., energy) and about several pulse-shape parameters. The results of the detectors characterization are presented in the next section.

Performance of Detectors
At first, we investigated the recorded bolometric signals in terms of the time constants of the signal shape. The rising part of a signal is commonly characterized by the rise time parameter, which is computed as the time required by the signal to increase from 10% to 90% of its amplitude. The descending part is described by the decay time, which is defined here as the time required to drop from 90% to 30% of signal amplitude. The rise and decay time parameters of the operated low-temperature detectors are summarized in Table 1. We found that the Li 2 100depl MoO 4 bolometers have signals with a rise time of ∼20 ms and decay time of ∼100 ms. These time constants are similar to the values reported for NTD-instrumented low-temperature detectors that were based on similarly sized Li 2 MoO 4 crystals produced from molybdenum with the natural isotopic abundance and molybdenum from the enriched amount in 100 Mo [12,13,17,22,23,25]. The Ge light detectors, being gram-scale bolometric devices equipped with smaller NTDs (i.e., reduced heat capacity compared with the Li 2 100depl MoO 4 bolometers), have an order of magnitude faster response, which is typical for such devices [13,17,22,23,25,48,69]. Then, using an amplitude distribution of the events recorded in the calibration runs, we calibrated the energy scale of the bolometers and evaluated their sensitivity, which was expressed as a voltage amplitude per unit of deposited energy (e.g., nV/keV); we also evaluated the energy resolution of the bolometers in the limit of zero amplitude (baseline noise) and at a mono-energetic radiation. In order to calibrate the Li 2 100depl MoO 4 bolometers, we used the most intense γ quanta emitted by the 232 Th source in the energy interval of 0.2-2.6 MeV, as illustrated in Figure 3 (left). In the background data, we also relied on the presence of γ peaks from environmental radioactivity (mainly the γ-active daughters of radon, 214 Pb, and 214 Bi; examples are given below). The sensitivities of the Li 2 100depl MoO 4 bolometers were measured to be ∼20-40 nV/keV (see Table 1); the LMO-depl-1 signal increases by a factor of two at a colder heat sink temperature. Taking into account that the chosen working points are characterized by relatively high NTD currents, the achieved sensitivities are not exceptional among Li 2 MoO 4 -based bolometers (the highest reported values are ∼100-150 nV/keV) [12,13,17,22,23,25]. Furthermore, the baseline noise was found to be rather low at ∼2-4 keV FWHM (Table 1), and it was similar to an early reported performance of NTD-instrumented Li 2 MoO 4 bolometers (the best performing detectors have an FWHM noise of ∼1 keV) [12,13,17,22,23,25]. To further improve the baseline noise (e.g., for dark matter search applications), one can reduce the absorber's volume (i.e., heat capacity) and/or use an advanced performance phonon sensor technology [34]. Despite not being an extraordinary noise resolution, both Li 2 100depl MoO 4 bolometers show a comparatively high energy resolution, as presented in Figure 3 (right). As it is also seen in Table 1, the energy resolution for high-energy γ quanta (1.8 and 2.6 MeV) is only a factor of 2-3 worse than the resolution at 0 energy, which is a good feature of Li 2 MoO 4 bolometers [13]. Consequently, the Li 2 100depl MoO 4 energy resolution at the 2615 keV MeV γs, listed in Table 1 and illustrated in Figure 3 (left, inset), is among the best reported for Li 2 MoO 4 low-temperature detectors [13,17,[22][23][24][25].
Aiming a permanent calibration during measurements, the LD-1-c was supplied by an 55 Fe X-ray source, which irradiated a Ge side that was opposite to the LMO-depl-1 crystal. This source emitted a doublet of Mn K α and K β X-rays with energies of 5.9 and 6.5 keV and intensities 25% and 3%, respectively; an example of the energy spectrum is presented in Figure 4 (left). To overcome the absence of permanent X-ray sources in the assembly of the LMO-depl-2 scintillating bolometer, we irradiated this detector with the 232 Th γ source to induce an X-ray fluorescence of the materials that were close to the light detectors, i.e., in the Cu structure and in the crystal. An illustration of the resulting spectrum is shown in Figure 4 (right). Thus, knowing the energy scale, we observed a good sensitivity for two light detectors (1.2-2.2 µV/keV), while the third device had a reduced value (0.4 µV/keV) due to a stronger polarization of the NTD, as exhibited by a lower NTD resistance (see Table 1). This performance is typical for these types of bolometric detectors with NTD thermistors; a further gain is also feasible by reducing the heat capacity of the sensor/absorber (e.g., see [13] and references therein) and/or by upgrading with a Neganov-Trofimov-Luke-effect-based signal amplification [66]. The less sensitive light detector had a comparatively modest noise resolution (about 300 eV FWHM), while the other two detectors demonstrated a rather low noise resolution of 60-100 eV FWHM (e.g., see [4]). The resolution of the 6 keV X-ray peak was found to be close to the baseline noise value, while a more broader 17 keV Mo X-ray peak was detected by both light detectors irrespective of the 5-times difference in the baseline noise. This effect can be explained by a position-dependent response by such thin bolometers [13].

Scintillation Detection and Particle Identification
A combination of both heat and scintillation channels of a scintillating bolometer can provide particle identification, which exploits the dependence of the light output on the energy loss mechanism (i.e., particle type) [4,70]. In order to find coincidences between signals in the Li 2 100depl MoO 4 bolometers and in the associated light detectors, the latter channels were processed using the trigger positions of Li 2 100depl MoO 4 events and accounting for a difference in the time response (see Table 1), similar to [71]. It is convenient to present such data by normalizing the light detector signal on the corresponding heat energy release, the so-called light-to-heat parameter (L/H) , in units of keV/MeV. The dependence of the L/H parameter on the energy and type of particles that impinged the Li 2 100depl MoO 4 detectors is illustrated in Figure 5. Several populations of events are clearly seen in each of the data presented in Figure 5. The most dominant one, which is mainly distributed below 3 MeV, is originated by γ(β) particles. We selected γ(β)s with energies above 2 MeV to evaluate the corresponding light-to-heat parameter (L/H γ(β) ) , which is reported in Table 2. The lowest L/H γ(β) value (∼0.3 keV/MeV) was obtained for the LMO-depl-1; this is expected due to the sub-optimal scintillation light collection conditions (e.g., the aperture between the crystal and the photodetector, the Cu surrounding instead of it being a reflective foil, the smaller light-detector area). Indeed, the twin detector (LMO-depl-2), which was surrounded by the reflective foil and coupled to the a photodetector with the same size, detected about 30% more scintillation energy (∼0.4 keV/MeV), while the square-shaped light detector allowed almost double the scintillation signal (∼0.6 keV/MeV). At higher energies, such as above ∼3 MeV, we observe populations of events that are characterized by the scintillation being reduced to ∼20% compared with the γ(β)s, as seen in Table 2. These events are originated by αs from either U/Th traces of detector bulk/surface contamination or a U source and α+t particles, the products of neutron capture on 6 Li. Despite the different light collection conditions of the characteristics of the Li 2 100depl MoO 4 scintillating bolometers, the L/H γ(β) and QF α listed in Table 2 are similar to the ones reported for detectors based on Li 2 MoO 4 crystals produced from molybdenum using the natural isotopic composition as well as molybdenum enriched in 100 Mo [13,17,[22][23][24]72]. 3.3. Radiopurity of Li 2 100depl MoO 4 crystals Thanks to efficient particle identification ( Figure 5) and comparatively long background measurements, we can investigate the radiopurity of the Li 2 100depl MoO 4 crystals with a high sensitivity to the α-active radionuclides of the U/Th decay chains. With this aim in mind, we selected α particles from the data of both detectors and re-calibrated the spectra to an alpha-energy scale for an analysis of the α contaminants; the resulting data are shown in Figure 6. It is seen in Figure 6 that the α spectra of the Li 2 100depl MoO 4 crystals are rather similar with the exception of the energy region below ∼4.7 MeV, which is populated by α particles from a 238 U/ 234 U source that was used in the set-up nearby the LMO-depl-2 detector. The spectra only contain two peak-like structures, which is a clear indication of a high internal radiopurity. Moreover, the first peak at ∼4.8 MeV is originated by the detection of products (α plus triton) of thermal neutron captures by 6 Li; this peak is detected by both bolometers at a similar rate of 1.8(2) counts/day. Furthermore, the doublet of 5.3 and 5.4 MeV peaks is a summed contribution of detector surface and crystal bulk contaminations by 210 Po; in the later case, an α particle and 206 Pb nuclear recoil (taking away 0.1 MeV of energy) were detected. The activity of 210 Po in both crystals is the same at ∼35 µBq/kg; it is originated by a residual 210 Pb contamination, which is typical for scintillators, particularly for Mo-containing compounds [13,17,[73][74][75].
We found no clear evidence of other α-active radionuclides from U/Th chains in the α spectra of the Li 2 100depl MoO 4 detectors; thus, we set upper limits on their activities in the crystal bulk using the Feldman-Cousins approach [76]. As a signal, we took all events in the range of ±25 keV around the Q-value of a radionuclide that we searched for; meanwhile, the background estimate was performed in the neighbor energy interval out of the Q-values of the U/Th αs, the process of which is detailed in [73]. The results of the study of the Li 2 100depl MoO 4 crystals' radiopurity are summarized in Table 3. The limits on the U/Th activity were obtained on the level of a few µBq/kg; therefore, the radiopurities of the two Li 2 100depl MoO 4 samples were similar to the purity level of the Li 2 100 MoO 4 crystals produced for the LUMINEU [13,15], CUPID-Mo [17,72] and CROSS [23,25] experiments, thanks to the same purification and crystallization protocols applied.

Background Reconstruction Capability of Li 2 100depl MoO 4 Bolometers
As mentioned in Section 1, 100 Mo presents a great interest for double-β decay studies. However, the two-ν double-β decay of 100 Mo -which is characterized by the fastest halflife among all double-β active isotopes [41]-is an important source of the background in ν-less double-β decay search experiments. Indeed, it generates a 10 mHz rate in a 1 kg 100 Mo-enriched lithium molybdate crystal, and it is a dominant background component in a wide energy interval [13,19,77]. However, even crystals with natural Mo content have a non-negligible internal activity of 100 Mo (∼1 mBq/kg). The impact of 100 Mo radioactivity can be seen in Figure 7, where the energy spectra accumulated with 100 Mo-enriched (up to ∼97%)/depleted Li 2 MoO 4 bolometers in a common measurement at the C2U facility are shown. It is worth noting that the background of the Li 2 100 MoO 4 bolometer was spoiled by a β-component of the used external α source with a notable activity, which was around an order of magnitude higher than that of the 100 Mo two-ν double-β decay; this is evident in Figure 7. A clear γ background reduction was exhibited by the LMO-depl-2 bolometer in comparison to the LMO-depl-1 data, which was achieved thanks to using a deradonized air flow around the cryostat shielding. Both Li 2 100depl MoO 4 detectors have a similar high background of below 0.7 MeV, which is explained by a 210 Bi activity that is induced by the 210 Pb contamination of the lead shield. Moreover, the residual γ(β) activity inside the experimental volume of the set-up, which was detected by both Li 2 100depl MoO 4 bolometers above ∼1 MeV, is higher than, e.g., the CUPID-Mo [19,77], CUPID-0 [78,79], and CUORE [80,81] experiments. Thus, the difference between the 100 Mo two-ν double-β decay distribution and the background data of the Li 2 100depl MoO 4 detectors, which were acquired in not fully optimized background conditions, is not remarkable as it may have been in a better shielded set-up; see Figure 7.
It is evident in Figure 7 that the Li 2 100depl MoO 4 bolometer allows for a significantly improved reconstruction of the γ background, including low-intensity contributions, compared with the Li 2 100 MoO 4 detector, which has a dominant double-β (and β) decay events continuum. Therefore, Li 2 100depl MoO 4 low-temperature detectors with high spectrometric performance and high radiopurity can provide complementary information about the background model in double-β decay searches with Li 2 100 MoO 4 bolometers. A similar example is the CUPID-0 experiment with two natural and twenty-four 82 Se-enriched zinc selenide bolometers [48]); however the detectors using selenium at the natural isotopic concentration were not included into the background model analysis [78]. Moreover, a combination of Li 2 100depl MoO 4 and Li 2 100 MoO 4 bolometers of similar performance and purity would allow for the extraction of the half-life and spectral shape of the 100 Mo two-ν double-β decay. A similar approach has been used in several double-β searches, such as in 40 Ca-/ 48 Ca-enriched calcium fluoride scintillation detectors [82], a 78 Kr-enriched/depleted gas filled proportional counter [83], and a 136 Xe-enriched/depleted xenon gas-based time projection chamber [84]. However, the enrichment in 48 Ca, which is present in natural calcium at a ∼0.2% level, is not yet available in as large of quantities as for 100 Mo, 78 Kr, and 136 Xe [85]. Moreover, the very long half-lives of 78 Kr (10 22 yr) and 136 Xe (10 21 yr) require a huge exposure to collect reasonably high datapoints for double-β decay events; the rate of 100 Mo being more than two orders of magnitude faster is an advantage for this type of study.  Figure 7. Energy spectra of γ(β) events detected by bolometers based on 0.28 kg lithium molybdate crystals produced from molybdenum that was either depleted in 100 Mo (pink, LMO-depl-1, and black, LMO-depl-2) or enriched in 100 Mo (blue, LMO-enr data from [23]); the bolometers were operated in the C2U set-up at the Canfranc Underground Laboratory. The detectors LMO-depl-1 and LMO-enr were run together (see Figure 2), while the LMO-depl-2 detector was measured in the next cryogenic run, where the set-up was flushed with deradonized air. It is worth noting that the LMO-enr bolometer was irradiated by a close 238 U/ 234 U α source, which emits β particles, 234 Th (Q β = 0.27 MeV), and 234m Pa (Q β = 2.2 MeV). Thus, the difference between the acquired spectra is mainly explained by both the 100  In the case of a resonant absorption of solar axions to the first excited state of 7 Li along with its subsequent γ de-excitation, a background in the Li 2 100depl MoO 4 bolometer at ∼0.5 MeV that is an order of magnitude lower than the Li 2 100 MoO 4 one would provide a higher sensitivity to an expected peak at 478 keV [11,38]. At lower energies, the contribution of the 100 Mo two-ν double-β decay becomes negligible [23], causing no preference in 100 Mo content in the Li 2 MoO 4 crystals during the search of a spin-dependent dark matter interaction on 7 Li. At the same time, the presence of 6 Li allows for neutron detection (illustrated above in Figures 5 and 6); this can be exploited for neutron flux monitoring, which is particularly relevant for dark matter search applications.

Conclusions
In the present study, we demonstrated that scintillating bolometers composed of lithium molybdate crystals produced from molybdenum depleted in 100 Mo (Li 2 100depl MoO 4 ) show a high performance that is comparable with the devices that are based on crystals from molybdenum of the natural isotopic composition or molybdenum enriched in 100 Mo. Thanks to the strict purification and crystallization protocols, which were developed and already applied to high-sensitivity searches for 100 Mo ν-less double-β decay, the radiopurity of Li 2 100depl MoO 4 crystals is rather high and is comparable to 100 Mo-enriched crystals of the same production line. Thus, given an availability of molybdenum that is depleted in 100 Mo (a by-product of industrial enrichment in 100 Mo) for Li 2 100depl MoO 4 crystals production, in addition to having a high spectrometric performance, efficient scintillation-assisted particle identification capability, and high material radiopurity, Li 2 100depl MoO 4 scintillating bolometers show a high potential for applications in rare-event search experiments. In particular, such detectors represent a great interest for studies of 100 Mo two-ν double-β decay, searches for 7 Li axions and spin-dependent interactions of dark matter particles on 7 Li, and 6 Li-based neutron spectroscopy and γ(β) background control measurements in low-temperature, low-background experiments. Given the obtained results, we will use the available Li 2 100depl MoO 4 crystals in the CROSS experiment and propose exploiting such low-temperature detectors in CUPID.  Acknowledgments: Russian and Ukrainian scientists have provided and continue to provide crucial contributions to CROSS. For this reason, the CROSS collaboration is particularly sensitive to the current situation in Ukraine. The position of the collaboration leadership on this matter, which was approved by majority, is expressed at (accessed on 2 June 2023) https://a2c.ijclab.in2p3.fr/en/a2chome-en/assd-home-en/assd-cross/. Majority of the work described here was completed before 24 February 2022.

Conflicts of Interest:
The authors declare no conflict of interest.