NUCLEUS Experiment: Reactor CEνNS Studies

• NUCLEUS experiment is a next-generation reactor neutrino initiative designed to measure coherent elastic neutrino–nucleus scattering (CEνNS) using advanced low-threshold cryogenic detectors.
• It leverages gram-scale calorimeters with multi-layer background suppression to achieve 20 eV nuclear recoil thresholds, enabling precise Standard Model tests and probes of new physics.
• Its phased deployment at the Chooz reactor facility plans to scale from initial 10g arrays to kg-scale arrays, benefiting reactor monitoring, dark matter searches, and astroparticle studies.

The NUCLEUS experiment is a next-generation reactor-neutrino initiative designed to measure coherent elastic neutrino–nucleus scattering (CEνNS) with gram-scale cryogenic calorimeters featuring nuclear recoil thresholds at the 20 eV scale. Its primary objective is twofold: the first direct detection and subsequent precision measurement of CEνNS from reactor antineutrinos, and the establishment of low-mass, low-threshold technology with broad impact on neutrino physics, dark matter searches, and new physics probes. The experiment leverages recent advances in ultra-low energy calorimetry, sophisticated multi-layer background suppression, and rigorous offline calibration and analysis, aiming for operation at the Chooz reactor complex in France.

1. Scientific Motivation and Physics Reach

CEνNS is a neutral-current process in which a neutrino scatters off an entire nucleus, leading to a cross-section enhancement proportional to the square of the neutron number ($\sigma_{\rm CE\nu NS} \propto N^2$). This coherence at low momentum transfer (few MeV-scale antineutrinos, $q\lesssim$ tens of MeV) results in observable sub-keV nuclear recoils. Measuring CEνNS serves several complementary goals:

• Standard Model Tests: CEνNS measurements explicitly test the predicted weak charge and probe the electroweak couplings at momentum transfers $\mathcal{O}(10\,\rm MeV)$, including the weak mixing angle.
• Physics Beyond the Standard Model: Sensitivity to non-standard neutrino interactions (NSI), possible new mediator particles, neutrino magnetic moments, and sterile neutrinos via deviations from the predicted recoil spectrum.
• Dark Matter and Astroparticle Physics: The CEνNS signal forms an irreducible background (“neutrino floor”) for direct dark matter detection, making characterization essential for next-generation rare-event searches (collaboration et al., 2022, Angloher et al., 2019).
• Reactor Monitoring and Safeguards: The observed event rates can be used for non-intrusive, real-time reactor monitoring applications, providing a path toward compact, robust neutrino detectors for safeguards (Angloher et al., 2019, Strauss et al., 2017).

Technically, the NUCLEUS experiment is configured to probe below the previously unexplored $10-100\,\rm eV$ recoil regime, enabling access to new parameter spaces.

2. Experimental Configuration

Site and Target

NUCLEUS will be installed at the Chooz B power station in France, exploiting proximity to two 4.25 GW$_{\mathrm{th}}$ reactor cores with a neutrino flux of $\sim1.7\times 10^{12}$ s$^{-1}$ cm$^{-2}$ at distances of 72 m and 102 m (collaboration et al., 2022, Angloher et al., 2019). The detectors are located in a 25 m$^2$ underground room in the administrative building’s basement ($\sim$3 m.w.e. overburden).

Detector Technology

• Primary Detector Modules: Gram-scale absorbers composed of CaWO$_4$ (calcium tungstate) and Al$_2$O$_3$ (sapphire) instrumented with tungsten transition-edge sensors (W-TES). Each crystal acts as a cryogenic calorimeter, converting recoil energy into a measurable temperature rise ($\Delta T = \Delta E/C$) at $T<30$ mK (collaboration et al., 2022, Strauss et al., 2017).
• Phased Mass Deployment: The initial configuration ("NUCLEUS-10g") comprises a few tens of grams; future upgrades target the kg-scale range ("NUCLEUS-1kg") (Angloher et al., 2019, collaboration et al., 2022).

Shielding and Veto Infrastructure

• Passive Shielding: Graded layers of low-activity lead, copper, and borated polyethylene surround the cryostat to attenuate external γ and neutron backgrounds (Angloher et al., 2019).
• Active Vetoes:
  • Muon veto: Hermetic cube of 5 cm BC-408 plastic scintillator panels, instrumented with WLS fibers and SiPMs; additional sub-Kelvin plastic scintillator disc inside the cryostat closes gaps (first of its kind) (Wagner et al., 2022, Erhart et al., 2022).
  • Outer cryogenic veto (COV): High-purity Ge detectors near the target region (multi-kg planned).
  • Inner veto: Instrumented Si structures for fiducialization and rejection of surface or holder-origin backgrounds.
• Mechanical Integration: Target modules are suspended via spring-coupled copper cages to mitigate vibrational noise. Support structures are optimized for minimal microphonics and radiopurity (collaboration et al., 2022, Abele et al., 4 Aug 2025).

Data Acquisition and Calibration

• Dual DAQ Systems: Veto channels (muon, COV) read out via STRUCK SIS3316 ADC, physical detectors with the VDAQ2 system.
• Optical Calibration: Photon-statistics-based optical system with UV LEDs ($\sim$5 eV/photon, $\sim$248 nm), with light delivered by fiber-optic switches and mirror wafers. This approach avoids TES saturation inherent to conventional radioactive source calibration and allows full automation (Castello, 2023).
• Analysis Frameworks: The DIANA software (adapted from CUORE) enables automated preprocessing and pulse shape fitting. Pulse reconstruction—including in saturation—uses an average pulse template and three-parameter least-squares fits; optimum filters further enhance SNR (Castello, 2023).

3. Backgrounds and Mitigation

Background Characterization

Given the modest 3 m.w.e. overburden, cosmic muons, neutron-induced events, and environmental γ-radiation are the dominant backgrounds (Angloher et al., 2019, collaboration et al., 2022). On-site surveys at Chooz measured a muon attenuation factor $a_\mu\simeq 0.71$ compared to surface, corresponding to an overburden of $2.9-3$ m.w.e. and a muon rate near 325 Hz for the full veto system (Angloher et al., 2019, Wagner et al., 2022). Neutron flux was reduced by a factor of $\sim$8 relative to surface levels through the combined action of the site’s overburden and dedicated borated PE shielding.

Active Veto Efficiency

Simulations and prototype runs demonstrate $>99$\% muon identification with the external veto, with the sub-Kelvin scintillator closing the open region at the top of the shielding with negligible light-yield loss and maintained pulse shape at $T\sim850$ mK (Wagner et al., 2022, Erhart et al., 2022). This configuration ensures dead-time remains at $<2$\% for veto time windows of 50 μs given the anticipated muon flux.

In-situ Commissioning

A reduced version of the experiment was commissioned at the Technical University of Munich (TUM), confirming the stability and efficiency of detector response, veto operation, and background modeling. Agreement between measured spectra and Geant4-based background simulations is within 20% above $\sim$60 keV (γ, μ, radioimpurity dominated), with low-energy excesses (LEE) at $<2$ keV still under investigation and requiring further refinement (Abele et al., 4 Aug 2025).

4. Calibration, Energy Reconstruction, and Analysis

Absolute and In-situ Calibration

• Photon-statistics Calibration: Employing controlled UV LED pulses, responsivity $r$ is extracted via Poissonian amplitude statistics: using $\mu=\langle$signal$\rangle$ and variance $\sigma^2$ for $N_{\rm ph}$ photons per pulse, $r=\mu/N_{\rm ph}$ and $\sigma^2=r\cdot\mu+\sigma_0^2$ allows calibration without explicit knowledge of $N_{\rm ph}$ (Castello, 2023).
• Energy Reconstruction: Offline reconstructions employ average pulse templates, fit to acquired waveforms [truncated in the case of TES saturation], recovering the linear (non-saturated) detector response across a wide dynamic range. Agreement with model-dependent hand-inspection analyses is better than 2% (Castello, 2023, Abele et al., 4 Aug 2025).
• Trigger and Time Resolution: Trigger efficiency as function of energy $E$ is parametrized as $$\epsilon_{\text{trig}}(E) = (A/2)\left[\mathrm{erf}((E-E_0)/\sqrt{2(\sigma_0^2+E\epsilon_\gamma)})+1\right]$$, with $A$ a normalization and $E_0$ the $50$\% efficiency threshold (Abele et al., 4 Aug 2025).

Statistical Sensitivity Analysis

• Discovery Potential: Sensitivity studies employ binned and unbinned extended likelihood frameworks (via PyCEnNS and FORECAST), modeling the differential CEνNS spectrum and backgrounds to conduct likelihood ratio tests, with statistical discovery significance $Z$ computed using Asimov or Monte Carlo pseudo-experiments. Performance checks demonstrate $<0.005\%$ discrepancy between independent codes (Castello, 2023, collaboration et al., 2022).

5. Performance Benchmarks and Commissioning Results

• Stable Detector Operation: Two target detectors (CaWO$_4$ and Al$_2$O$_3$) were operated continuously over eight weeks at TUM, achieving $83\%$ duty cycle even with incomplete shielding and veto infrastructure (Abele et al., 4 Aug 2025).
• Muon Veto Efficiency: The muon veto achieved over $98\%$ tagging efficiency correlated with events in the cryogenic outer veto.
• Calibration Stability: Copper K$_\alpha$ X-ray lines and LED-based photon-statistics calibration demonstrated a $\sim$25\% systematic discrepancy in the absolute energy scale (mirror wafer + copper X-ray vs. LED), a dominant uncertainty to be resolved in upcoming runs.

Subsystem	Metric	Current Status
Target threshold	$E_{\rm th}^{\rm recoil}$	$\sim 20$ eV (prototype)
Muon veto eff.	Identification power	$> 99\%$
Background index	Counts/(kg·keV·d) in ROI	$<100$ (goal)
Calibration	Energy scale systematics	17.25–21.5 eV/mV
Signal expectation	CEνNS rate (10 g, Chooz)	$0.33$ events/ d

6. Planned Upgrades and Future Prospects

• Full Detector Array: Expansion to 18 targets (split CaWO$_4$, Al$_2$O$_3$), all six cryogenic veto (COV) modules, instrumented inner vetoes, and implementation of boron carbide neutron absorbers.
• DAQ and Readout: Transition to upgraded multi-channel systems and possible double-TES readout for improved discrimination of LEE.
• Technical Run at Chooz: The experiment is scheduled for technical deployment and a first “real” run in 2026 at the Chooz reactor complex, with full background rejection and predicted CEνNS event rates $\sim$0.14 d$^{-1}$ (aggregate array) (Abele et al., 4 Aug 2025).
• Long-Term Goals: Move to the “NUCLEUS-1kg” precision phase capable of sub-percent measurement uncertainties on the CEνNS cross-section, enabling sensitive Standard Model tests and novel constraints on new physics scenarios (Angloher et al., 2019, collaboration et al., 2022).

7. Comparative Context and Broader Impact

NUCLEUS provides complementary discovery and precision reach alongside accelerator-based CEνNS experiments (e.g., COHERENT (Akimov et al., 2022)) and other low-threshold, low-mass approaches such as CONNIE (Aguilar-Arevalo et al., 2019). The experiment is distinguished by its operation at the lowest thresholds, multi-material approach for cross-validation and background control, and modular, cryogenic shield design. The rapid cadence from first detection in 10 g arrays to “precision” kg-scale operation underpins extensive cross-disciplinary impact, including contributions to reactor monitoring, nuclear astrophysics, dark matter direct detection, and fundamental neutrino interactions (collaboration et al., 2022, Angloher et al., 2019, Castello, 2023).

In summary, the NUCLEUS experiment represents a technically ambitious and scientifically broad initiative to probe CEνNS with unprecedented precision, leveraging multi-layered background rejection, state-of-the-art calibration and analysis methodologies, and a progressive mass-scaling strategy. Its phased approach is expected to transition the field from first detection of reactor CEνNS to a regime of precision neutrino physics and sub-keV rare-event searches.