Development of the NUCLEUS Detector to Explore Coherent Elastic Neutrino-Nucleus Scattering ^†

Abstract

The NUCLEUS experiment, currently being commissioned at the Technical University of Munich, is designed to observe coherent elastic neutrino-nucleus scattering ($\mathrm{CE}\nu \mathrm{NS}$) from reactor neutrinos and measure its cross-section with a percent-level precision at recoil energies below $100 \mathrm{eV}$. As a Standard Model process, $\mathrm{CE}\nu \mathrm{NS}$ provides a unique probe into neutrino properties, potential new physics, and background suppression techniques relevant to dark matter experiments. The experiment utilizes gram-scale cryogenic calorimeters operating at $10 \mathrm{mK}$ with an energy threshold of $20 \mathrm{eV}$. Situated at a shallow overburden of 3 m of water equivalent, the experimental site necessitates an advanced shielding strategy combining active vetoes and passive layers to reduce background rates to approximately $100\mathrm{counts}/(\mathrm{kg}·\mathrm{day}·\mathrm{keV})$, as confirmed by full setup simulations. The commissioning phase has successfully demonstrated the stable operation of the cryogenic target detectors, achieving baseline resolutions below $10 \mathrm{eV}$, and the integration of the various shielding systems. Following this milestone, the experiment is set to transition to the EdF Chooz B nuclear reactor in France in 2025, where it will enable precise measurements of $\mathrm{CE}\nu \mathrm{NS}$, contributing to the understanding of neutrino interactions and advancing the field of astroparticle physics.

1. $\mathrm{CE}\mathbf{\nu }\mathrm{NS}$ from Reactor Antineutrinos

Coherent elastic neutrino-nucleus scattering ($\mathrm{CE}\nu \mathrm{NS}$) is a Standard Model process that offers a valuable probe of the Standard Model validity and new physics. The first observation of $\mathrm{CE}\nu \mathrm{NS}$ was conducted in 2017 by the COHERENT collaboration using a $14.6 \mathrm{kg}$ CsI[Na] scintillation crystal at the Spallation Neutron Source [1] which provides a pulsed neutrino beam with a flux of $4.3·{10}^7$$\nu$/$(\mathrm{s}·{\mathrm{cm}}^2)$ at $20 \mathrm{m}$ from the source, with neutrino energies up to $53 \mathrm{MeV}$. A subsequent COHERENT measurement using argon [2] and germanium [3] probed the $N^2$-dependence of the $\mathrm{CE}\nu \mathrm{NS}$ cross-section.

Nuclear reactors, with a neutrino flux of $\mathcal{O}$(${10}^{20}$) $\nu$/($\mathrm{s}·{\mathrm{GW}}_{\mathrm{th}}$), are promising for $\mathrm{CE}\nu \mathrm{NS}$ studies at lower energies. Reactor antineutrinos are produced in fission beta-decays and have energies up to $10 \mathrm{MeV}$, which place them fully within the coherent mode. $\mathrm{CE}\nu \mathrm{NS}$ from reactor neutrinos allows researchers to measure the weak neutral current and the Weinberg angle at low momentum transfer, while deviations could indicate new physics [4].

Despite ongoing efforts in various collaborations, detecting $\mathrm{CE}\nu \mathrm{NS}$ at nuclear reactors remains challenging [5]. As of December 2024, the only detection of $\mathrm{CE}\nu \mathrm{NS}$ at a reactor was reported at the Dresden-II facility [6]; however, this finding conflicts with more sensitive searches conducted by the CONUS experiment [7]. The main challenges in observing $\mathrm{CE}\nu \mathrm{NS}$ from nuclear reactors are achieving sub-keV detection thresholds of nuclear recoils and managing background levels in shallow-depth reactor environments [5]. The NUCLEUS experiment addresses both challenges by combining gram-scale cryogenic calorimeters with energy thresholds of around $20 \mathrm{eV}$ with a target background rate of $100 \mathrm{counts}/(\mathrm{kg}·\mathrm{day}·\mathrm{keV})$ [8,9].

2. The NUCLEUS Experiment

The NUCLEUS experimental site, referred to as the “Very Near Site” (VNS) at the Chooz B Nuclear Power Plant in France, occupies a $25 {\mathrm{m}}^2$ room in the basement of an administrative building, positioned $102 \mathrm{m}$ and $72 \mathrm{m}$ from the two $4.25 {\mathrm{GW}}_{\mathrm{th}}$ reactor cores. This proximity ensures a high average electron antineutrino flux of approximately $1.7·{10}^{12}$ ${\overline{\nu }}_e$/(s $·$ cm²) [9]. However, the low overburden of 3 m of water equivalent (m.w.e.) necessitates a robust cosmic-ray-induced background mitigation strategy. To achieve the desired background level of $100 \mathrm{counts}/(\mathrm{kg}·\mathrm{day}·\mathrm{keV})$, multiple layers of passive and active shielding are required [9]. Within the region of interest (ROI), spanning the range from $20 \mathrm{e}\mathrm{V}$ to $100 \mathrm{e}\mathrm{V}$, the NUCLEUS experiment anticipates a total counting rate of 30 ${\overline{\nu }}_e$/ (kg $·$ day) above the background [10]. The shielding strategy was studied and optimized by Monte Carlo simulations based on measurements at the VNS [11].

2.1. The NUCLEUS Target Detectors: Gram-Scale Cryogenic Calorimeters

The NUCLEUS experiment uses gram-scale cryogenic calorimeters of Calcium Tungstate (${\mathrm{CaWO}}_4$) and Sapphire (${\mathrm{Al}}_2{\mathrm{O}}_3$), with a total target mass of $10 \mathrm{g}$. When a particle induces a nuclear recoil in the target crystal, the generated lattice vibrations propagate through the crystal via phonons. These phonons are detected by a tungsten-based transition edge sensor (W-TES) operating at the transition temperature of the tungsten film at around $10 \mathrm{mK}$ [9]. The slight increase in temperature from the particle interaction results in a significant change in the TES resistance, which is read out by Superconducting QUantum Interference Devices (SQUIDs) [12]. This technology was developed within the CRESST experiment, which focuses on dark matter searches [13] and was further advanced by the NUCLEUS experiment for $\mathrm{CE}\nu \mathrm{NS}$ detection. A prototype utilizing $0.5 \mathrm{g}$ of ${\mathrm{Al}}_2{\mathrm{O}}_3$ demonstrated the proof of principle for this detection method [8,12]. To take advantage of the $N^2$-dependence of the $\mathrm{CE}\nu \mathrm{NS}$ cross-section, the experiment employs a multi-target approach, featuring an array of nine ${\mathrm{CaWO}}_4$ crystals ($6 \mathrm{g}$) for $\mathrm{CE}\nu \mathrm{NS}$ measurements and nine ${\mathrm{Al}}_2{\mathrm{O}}_3$ crystals ($4 \mathrm{g}$) for in situ background assessment [9]. The ${\mathrm{CaWO}}_4$ detectors are expected to capture the $\mathrm{CE}\nu \mathrm{NS}$ signal above the background; in contrast, the ${\mathrm{Al}}_2{\mathrm{O}}_3$ detectors will have a suppressed neutrino scattering rate, providing an in situ measurement of the background. This background measurement with ${\mathrm{Al}}_2{\mathrm{O}}_3$ allows us to constrain the background model more precisely, thereby enhancing our sensitivity to $\mathrm{CE}\nu \mathrm{NS}$ [9].

2.2. The NUCLEUS Inner Detector Module for the First Physics Run at Chooz

An inner detector module consisting of nine target detectors and a dedicated holding structure is currently being developed for NUCLEUS’ first physics run at the nuclear power plant in Chooz. The first version of the module was built with detectors equipped with a single TES, while ongoing research investigates detectors with two TESs as an additional background-mitigation tool [14] which might be implemented in future versions.

As depicted in Figure 1, these nine target detector crystals are secured by two TES-instrumented silicon plates that serve as an inner veto system for background rejection, targeting events related to surface interactions on the one hand and to mechanical stress-relaxation on the other hand. The latter is a potential candidate for the low energy excess (LEE), a phenomenon observed across low-threshold experiments, characterized by a steep rise in event rate below a few hundred eV [15]. Simultaneous readouts from both the target detector and the holder can identify events originating from their interface. The inner veto consists of a rigid $0.45 \mathrm{mm}$ thick and a flexible $0.25 \mathrm{mm}$ thin silicon plate. To minimize the contact area between the target crystals and the support structure and thereby prevent signal phonons from escaping detection in the TES of the target crystals, the contact points are implemented by silicon pyramids. These pyramids with a height of $0.2 \mathrm{mm}$ are machined by the wet chemical etching of the inner veto plates and can be seen in the right insert of Figure 1. The detectors and inner veto plates are secured by two $1 \mathrm{mm}$ thick silicon holding plates that provide all necessary electrical and thermal connections for the operation of the detectors and the inner veto system. Furthermore, the contact between the non-instrumented holding plates and the inner veto parts is facilitated by $1 \mathrm{mm}$ thick ${\mathrm{Al}}_2{\mathrm{O}}_3$ spheres, ensuring electrical and thermal isolation. The height of the brass separation shells defines the force bending the thin inner veto silicon wafer, securing the target detectors. For the first physics run at Chooz, we target to deploy one inner detector module from ${\mathrm{CaWO}}_4$ detectors ($6 \mathrm{g}$) and one from ${\mathrm{Al}}_2{\mathrm{O}}_3$ detectors ($4 \mathrm{g}$), representing the NUCLEUS-$10g$ target detector module.

2.3. Active and Passive Shielding Layers for Background Mitigation

The NUCLEUS experiment employs a multi-layered shielding strategy, which is shown in Figure 2, combining active and passive components to effectively suppress background signals and ensure the high sensitivity required for $\mathrm{CE}\nu \mathrm{NS}$ detection.

The cryogenic outer veto (COV) combines six high-purity germanium detectors with a total mass of $4 \mathrm{kg}$ and a thickness of $2.5 \mathrm{cm}$, operated as ionized semiconductor detectors at around $10 \mathrm{mK}$. The six COV crystals are held in place by a copper structure arranged to supply a $4\pi$ coverage around the two inner detector modules to significantly reduce ambient gamma-ray backgrounds. In addition, three layers of passive shielding are arranged inside and outside the cryostat for $4\pi$ coverage to suppress remaining critical background components. The first layer represents the boron carbide (${\mathrm{B}}_4\mathrm{C}$) shield enclosed within the cryostat and surrounding the COV, which serves as the primary barrier against low-energy neutron backgrounds. The second layer, consisting of borated polyethylene, provides a barrier against neutrons. The final third layer consists of lead and offers an additional attenuation of gamma and neutron backgrounds. To address cosmic-ray muons, the outermost veto comprises $5 \mathrm{cm}$ thick plastic scintillator plates equipped with silicon photomultipliers (SiPMs) and Wavelength Shifting fibers, achieving a detection threshold of $5 \mathrm{MeV}$ and a simulated efficiency for muon identification of >$99\%$ [16]. This muon veto system detects high-energy muons, which generate spallation neutrons in the lead shielding, a significant background mimicking $\mathrm{CE}\nu \mathrm{NS}$-like nuclear recoils. The muon veto system and the lead and polyethylene shieldings are deployed both outside and inside the cryostat to ensure complete coverage around the cryogenic detectors. The structure inside the cryostat includes a cryogenic muon veto disk, a lead disk, and several polyethylene disks, mimicking the external shielding structure. The cryogenic muon veto disk has been developed [17] to be deployed inside the cryostat above the boron carbide shield and thermalized to $800 \mathrm{mK}$. Monte Carlo simulations showed the ability of the nearly $4\pi$-covering active and passive shielding layers to mitigate the cosmic and radiogenic background components to meet the goal of $100 \mathrm{counts}/(\mathrm{kg}·\mathrm{day}·\mathrm{keV})$ in the ROI [10,11,18].

3. Development of the Cryogenic Target Detectors for the Long Background Run to Commission the NUCLEUS Experiment at TUM

Over the course of 2023 and 2024, the various detector and shielding components were installed in the underground laboratory (UGL) at the Technical University of Munich (TUM) to commission the experiment. The commission phase involved several runs over the course of these two years to commission the sub-systems and concluded in a long background run (LBR) to demonstrate the stable operation of the full system. The major goal is to demonstrate the performance of the cryogenic target detectors in anti-coincidence with the COV and the muon veto systems and to validate our shielding strategy by comparing the measured to the simulated background data for the UGL site [19].

3.1. Cryogenic Target Detectors

While developing the cryogenic target detectors, we evaluated the performance of ${\mathrm{CaWO}}_4$ and ${\mathrm{Al}}_2{\mathrm{O}}_3$ as target materials. As of December 2024, we demonstrated the reliable performance of both materials in a series of measurements conducted in the customized NUCLEUS dry dilution refrigerator. These detectors consistently achieved baseline resolutions of less than 10 eV over several weeks of operation across multiple runs. Two ${\mathrm{Al}}_2{\mathrm{O}}_3$ target detectors were simultaneously operated in the NUCLEUS inner detector module, achieving similar baseline resolutions and no cross-talk [19]. Meanwhile, other ${\mathrm{CaWO}}_4$ and ${\mathrm{Al}}_2{\mathrm{O}}_3$ detectors were operated in simplified copper holders, as efforts are ongoing to develop the TES-instrumented silicon holding structures for anti-coincidence operation with the target detectors.

Multiple ${\mathrm{CaWO}}_4$ detectors demonstrated excellent energy resolution of less than 10 eV over several weeks in various runs, enabling the development and validation of calibration techniques for cryogenic particle detectors. Specifically, a ${\mathrm{CaWO}}_4$ detector was calibrated using nuclear recoils with the CRAB method [20] to investigate differences between nuclear and electron recoils. Another ${\mathrm{CaWO}}_4$ detector was calibrated with sub-keV X-ray fluorescence lines and LED pulses [21], allowing us to explore detector non-linearities.

In addition to a detector cube with a single TES, we investigated detectors with two TESs (=the double-TES approach), developed to test a plausible source of the LEE in TES-based detectors: events arising from stress at the interface between the TES and the crystal or within the TES structure itself. Detectors with two TESs were tested to enable simultaneous and independent signal readout. In this configuration, particle recoils originating from the crystal volume are expected to distribute their energy among the sensors, while signals from TES-related events occurring near a specific TES would predominantly or entirely be measured by the closest TES. This approach was validated by CRESST [22,23] and SPICE [24] and further advanced by the NUCLEUS collaboration [14,25], with double-TES detectors achieving sub-$100 \mathrm{eV}$-thresholds, enabling precise LEE measurements and event differentiation.

Finally, for the LBR, we deployed a ${\mathrm{CaWO}}_4$ single-TES detector cube with a $5 \mathrm{mm}$ side length and a mass of $0.75 \mathrm{g}$, as is shown in the left photograph of Figure 3, as well as a ${\mathrm{Al}}_2{\mathrm{O}}_3$ double-TES detector measuring $[5\times 5\times 7.5]{\mathrm{mm}}^3$ with the same mass, shown in the photograph in the middle of Figure 3. Both detectors were operated at their transition temperature of approximately $14 \mathrm{mK}$. The goal of using these two target detectors was to characterize the particle background in the different target materials and further evaluate the background mitigation potential of the double-TES approach.

3.2. Development of the Target Detector Holders

Figure 4 shows the components of the detector holders of our cryogenic target detectors used for preparing the LBR. Each detector is completely encapsulated by a copper housing to block thermal radiation from the outside and secured with bronze clamps, as seen in the right picture of Figure 3. ${\mathrm{Al}}_2{\mathrm{O}}_3$ spheres are used for thermal and electrical isolation between the copper holder, the clamps, and the detector itself. The signal readout is facilitated by printed circuit board (PCB) traces that connect the TES wire bonds to external pin connections, where cables for the TES readout are attached. During the development phase, commercial PCBs with ${\mathrm{FR}}_4$-cores, as shown in the left picture of Figure 4, were initially used. However, they were deemed unsuitable for the LBR due to radioactive impurities that would dominate the signal observed in our detectors [11,18]. In contrast, the customized PCB, seen in the right picture of Figure 4, is made from copper and features copper-kapton-copper traces glued to the copper plate. To further reduce background from impurities, all copper parts were fabricated from electrolytic tough pitch (NOSV) copper [26] due to its high purity and low oxygen content, enhancing the intrinsic thermal conductivity. Additionally, all components of the holders undergo a rigorous cleaning procedure involving isopropyl alcohol and acetone, followed by etching with citric acid in a hydrogen peroxide solution to effectively remove any oxidation layers and improve thermal conductivity.

3.3. Status of the Commissioning

The LBR to commission the full NUCLEUS experiment at TUM concluded in September 2024, demonstrating the stable performance of the cryogenic target detectors with most of the active and passive shielding systems over more than a month. During the LBR, the full muon veto system, comprising 28 Si-PM panels and an additional cold muon veto disk, was successfully operated for approximately 40 days. The outer and inner shielding, consisting of lead and polyethylene layers, were deployed, while the ${\mathrm{B}}_4\mathrm{C}$ layer was dispensable due to the higher overburden at the UGL site compared to the Chooz site [11].

We implemented one ${\mathrm{CaWO}}_4$ detector with single-TES readout and one ${\mathrm{Al}}_2{\mathrm{O}}_3$ detector with double-TES readout, both calibrated using a customized LED calibration system [27]. Initial analyses confirmed the stable operation of both detectors over nearly 500 h each with approximately 80% of usable physics data with stable baseline resolutions below 20 eV, achieving the targeted performance metrics. Additionally, one of the six COV Ge crystals was deployed above the target detectors and successfully operated at the same time. Detailed analyses of the cryogenic detectors in anti-coincidence with the COV and muon veto systems, as well as investigations of the low-energy excess (LEE) and the comparison of the measured to simulated background level, are currently ongoing.

4. Conclusions and Outlook

The NUCLEUS experiment aims to observe $\mathrm{CE}\nu \mathrm{NS}$ from reactor neutrinos at the Chooz Nuclear Power Plant using gram-scale cryogenic calorimeters with a threshold of $20 \mathrm{eV}$. Operating at a shallow overburden of 3 m.w.e., the experimental site requires a robust shielding strategy combining active vetoes and passive layers to achieve the low background rate of approximately $100 \mathrm{counts}/(\mathrm{kg}·\mathrm{day}·\mathrm{keV})$ necessary for successful $\mathrm{CE}\nu \mathrm{NS}$ detection. The long background run to commission the NUCLEUS experiment at the underground laboratory of the Technical University of Munich demonstrates the stable operation of all systems over a one-month period, confirming the integration of the cryogenic target detectors and their radio-pure holding systems.

Following this milestone, the installation at the Chooz Nuclear Power Plant is planned for 2025, enabling the NUCLEUS experiment to investigate neutrino interactions. Efforts to address the LEE through innovative detector designs are ongoing. For the first physics run, a detector module with $10 \mathrm{g}$ of cryogenic detectors is under development to achieve an initial $\mathrm{CE}\nu \mathrm{NS}$ measurement. A second phase, featuring a larger detector with a target mass of approximately $1\mathrm{kg}$, is planned to enable precision measurements of the $\mathrm{CE}\nu \mathrm{NS}$ cross-section with several-percent accuracy [9].