Ricochet: Reactor Neutrino CEvNS Measurement

• Ricochet is a reactor neutrino experiment designed for precise measurement of coherent elastic neutrino–nucleus scattering (CEvNS) using cryogenic solid-state detectors.
• It employs a hybrid detection system combining Ge bolometers (CryoCube) and TES-based calorimeters (Q-Array) to achieve sub-100 eV nuclear recoil thresholds.
• The experiment prioritizes robust background rejection and spectral analysis to explore Standard Model parameters and probe beyond-the-Standard-Model effects.

Ricochet is a short-baseline reactor-neutrino observatory designed to perform a precision measurement of coherent elastic neutrino–nucleus scattering, $\mathrm{CE}\nu\mathrm{NS}$, with cryogenic solid-state detectors operated at millikelvin temperatures near a research reactor core. In its current form, the experiment is installed at the Institut Laue–Langevin (ILL) in Grenoble, France, with the detector center $8.8\,\mathrm{m}$ from a $58\,\mathrm{MW}$ research reactor, and it combines kilogram-scale target mass, recoil thresholds in the sub-$100\,\mathrm{eV}$ range, and event-by-event rejection of electromagnetic backgrounds (Collaboration et al., 2021, Augier et al., 2023, Chen et al., 2023). The collaboration’s program couples detector R&D, background characterization, and commissioning at the reactor site to a broader goal: a low-energy, high-precision $\mathrm{CE}\nu\mathrm{NS}$ spectrum measurement sensitive to Standard Model parameters and to beyond-the-Standard-Model effects such as non-standard neutrino interactions, light mediators, sterile neutrinos, and neutrino electromagnetic properties (Augier et al., 2022, Collaboration et al., 2021).

1. Scientific basis and measurement goals

Ricochet is built around $\mathrm{CE}\nu\mathrm{NS}$, a neutral-current process in which a neutrino scatters elastically from an entire nucleus. For sufficiently small momentum transfer, the interaction is coherent and the cross section scales roughly as $N^2$, where $N$ is the neutron number; this makes kilogram-scale targets viable, but reactor antineutrino energies of a few MeV push the nuclear recoil spectrum down to tens of eV (Chen et al., 2023). That recoil scale sets the central instrumental requirement: the experiment must detect nuclear recoils down to about $50\,\mathrm{eV}$ if it is to exploit the very high reactor flux (Chen et al., 2023).

The physics case is explicitly spectral rather than merely counting-based. Design papers describe Ricochet as targeting a precision measurement of the $\mathrm{CE}\nu\mathrm{NS}$ rate and spectrum, with percent-level precision identified as the long-term objective for the reactor measurement (Augier et al., 2023). The low-energy spectrum is where several benchmark beyond-the-Standard-Model effects are most visible, including non-standard neutrino interactions, new light mediators, sterile neutrinos, and neutrino magnetic moments (Augier et al., 2022, Collaboration et al., 2021). The same recoil spectrum also constrains nuclear form factors at low momentum transfer (Chen et al., 2023).

For the ILL configuration and a $8.8\,\mathrm{m}$0 threshold, the expected $8.8\,\mathrm{m}$1 rates quoted in the site and background studies are $8.8\,\mathrm{m}$2 events/kg/day for Ge and $8.8\,\mathrm{m}$3 events/kg/day for Zn (Augier et al., 2022). This scale is high enough that the limiting problem is not signal statistics alone, but the joint control of thresholds, detector response, and irreducible nuclear-recoil backgrounds, especially cosmogenic neutrons (Augier et al., 2022).

2. Site, reactor environment, and experimental architecture

Ricochet is located at the ILL H7 site in Grenoble, a shallow-overburden reactor hall beneath a water channel that provides about $8.8\,\mathrm{m}$4 shielding against cosmic rays (Augier et al., 2022, Collaboration et al., 2021). At the detector position, the reactor antineutrino flux is quoted as approximately $8.8\,\mathrm{m}$5 in one study and $8.8\,\mathrm{m}$6 in another, both for the $8.8\,\mathrm{m}$7 baseline configuration (Augier et al., 2022, Collaboration et al., 2021). The reactor itself is a $8.8\,\mathrm{m}$8–$8.8\,\mathrm{m}$9 research reactor, and the combination of compact core and short baseline is central to the experiment’s high-statistics design (Augier et al., 2022, Collaboration et al., 2021).

The present Ricochet payload is hybrid. It combines a Ge-based subsystem, CryoCube, and a TES-based subsystem, Q-Array, within a common cryostat and shielding design (Augier et al., 2023, Collaboration et al., 2021). Earlier work on the broader Ricochet concept also included a site-study phase at the MIT Research Reactor, where neutron backgrounds were unfolded with a $58\,\mathrm{MW}$0He detector and Bonner-cylinder-style moderation using Markov Chain Monte Carlo; later, after in situ studies of vibration and particle backgrounds, the collaboration selected ILL as the experiment site (Leder et al., 2017, Collaboration et al., 2021).

Subsystem	Readout concept	Representative design target
CryoCube	Ge bolometers with NTD heat readout and ionization electrodes	$58\,\mathrm{MW}$1 Ge crystals of $58\,\mathrm{MW}$2, $58\,\mathrm{MW}$3 threshold (Salagnac et al., 2021)
Q-Array	TES-based cryogenic calorimeters	$58\,\mathrm{MW}$4 modules, $58\,\mathrm{MW}$5 per detector, $58\,\mathrm{MW}$6 threshold (Chen et al., 2021, Chen et al., 2023)

The full installation couples these detectors to a dedicated cryostat, heavy passive shielding, and an outer muon veto (Collaboration et al., 2021, Armatol et al., 30 Jul 2025). This suggests a deliberately redundant strategy: low threshold and recoil-type discrimination at the detector level, plus shielding and reactor-on/off subtraction at the facility level.

3. CryoCube: germanium heat-and-ionization bolometers

CryoCube is the Ge branch of Ricochet. One design paper specifies a $58\,\mathrm{MW}$7 array of $58\,\mathrm{MW}$8 high-purity Ge crystals, each of mass about $58\,\mathrm{MW}$9, operated at roughly $100\,\mathrm{eV}$0–$100\,\mathrm{eV}$1 and instrumented with an NTD-Ge thermal sensor and aluminum electrodes for simultaneous heat and ionization readout (Salagnac et al., 2021). This dual-readout architecture is inherited from EDELWEISS/SuperCDMS-style bolometry and is central to event-by-event separation of electron recoils and nuclear recoils through the ionization-to-heat ratio (Augier et al., 2023, Salagnac et al., 2021).

The CryoCube electronics program is correspondingly demanding. A dedicated $100\,\mathrm{eV}$2 HEMT-based front-end was developed to deliver about $100\,\mathrm{eV}$3 RMS baseline resolution for the heat channel and $100\,\mathrm{eV}$4 RMS for the ionization channel, with high dynamic range and sufficient timing resolution for muon-veto coincidence (Baulieu et al., 2021). The motivation is explicit: these resolutions are required to maintain discrimination power between nuclear and electron recoils at the lowest energies (Baulieu et al., 2021). In parallel, a Python-based waveform-processing pipeline was developed to simulate continuous data streams, optimize triggering and matched filtering, and quantify reconstruction efficiencies and biases; this study concluded that the CryoCube array could achieve a $100\,\mathrm{eV}$5 energy threshold combined with electron/nuclear recoil discrimination down to about $100\,\mathrm{eV}$6 (Colas et al., 2021).

Prototype performance milestones preceded full installation. In an early Ge cryogenic sub-assembly, the collaboration reported the first demonstration of a $100\,\mathrm{eV}$7 baseline ionization resolution (RMS) in a cryogenic bolometer at $100\,\mathrm{eV}$8, obtained with dedicated HEMT-based front-end electronics (Augier et al., 2023). That same paper frames the result as a factor $100\,\mathrm{eV}$9–$\mathrm{CE}\nu\mathrm{NS}$0 improvement over earlier EDELWEISS and SuperCDMS ionization readout and as a concrete step toward the design goal of $\mathrm{CE}\nu\mathrm{NS}$1 RMS for the full CE$\mathrm{CE}\nu\mathrm{NS}$2NS program (Augier et al., 2023).

4. Q-Array and the TES-based calorimeter program

The TES branch of Ricochet, Q-Array, is designed as an array of nine cryogenic calorimeters with target masses of order $\mathrm{CE}\nu\mathrm{NS}$3 and a targeted threshold of $\mathrm{CE}\nu\mathrm{NS}$4 (Chen et al., 2021, Chen et al., 2023). The detector concept is modular: a macroscopic absorber is coupled by a thin Au collection layer and Au wire bond to a separate TES sensor chip, allowing absorber and sensor fabrication to be optimized independently (Chen et al., 2023). This modularity is also intended as a demonstrator path toward future neutrino experiments with thousands of TES-based calorimeters (Chen et al., 2023).

Prototype and modeling studies use several absorber materials. Design papers discuss Zn-based TES calorimeters in the Q-Array, while prototype studies have characterized Si and Ge absorbers to validate the thermal architecture and readout concept (Chen et al., 2021, Collaboration et al., 2023, Chen et al., 2023). In a first $\mathrm{CE}\nu\mathrm{NS}$5 Si prototype, the collaboration obtained a baseline RMS resolution below $\mathrm{CE}\nu\mathrm{NS}$6, specifically $\mathrm{CE}\nu\mathrm{NS}$7 RMS for absorber hits and $\mathrm{CE}\nu\mathrm{NS}$8 RMS for TES-chip hits, thereby demonstrating that the TES–gold-wire coupling scheme could already reach the noise scale required for Ricochet (Collaboration et al., 2023).

A later detailed thermal–electrical characterization of a prototype TES-based Ge detector used simultaneous fits to the complex impedance $\mathrm{CE}\nu\mathrm{NS}$9 and the average pulse shape from $\mathrm{CE}\nu\mathrm{NS}$0 internal $\mathrm{CE}\nu\mathrm{NS}$1 events (Chen et al., 2023). The model treated the target, gold pads, wire bonds, TES, silicon chip, meander, glue, and bath as coupled thermal nodes and inferred $\mathrm{CE}\nu\mathrm{NS}$2 and $\mathrm{CE}\nu\mathrm{NS}$3 for the TES transition, together with a much smaller-than-expected Au–TES conductance and a larger-than-expected glue-mediated TES–bath conductance (Chen et al., 2023). Those results directly motivated design changes, including larger Au–TES overlap and alternative mounting strategies such as clamping rather than gluing (Chen et al., 2023). In that sense, Q-Array is not merely an auxiliary branch of Ricochet; it is also a technology-development program in low-threshold cryogenic calorimetry.

5. Backgrounds, shielding, and neutron control

Background control in Ricochet is dominated by neutron-induced nuclear recoils, because these are experimentally indistinguishable from $\mathrm{CE}\nu\mathrm{NS}$4 in the absence of auxiliary handles (Augier et al., 2022). The collaboration therefore carried out a dedicated fast-neutron characterization of the ILL H7 site using a low-radioactivity $\mathrm{CE}\nu\mathrm{NS}$5He proportional counter, combined with Geant4 and CRY simulations for cosmogenic neutrons and an MCNP-based model for reactor-related neutrons (Augier et al., 2022). These studies showed that the raw reactogenic fast-neutron flux at H7 is larger than the cosmogenic one, but that after the planned shielding the residual nuclear-recoil background is dominated by cosmogenic neutrons rather than reactor neutrons (Augier et al., 2022).

The shielding design is correspondingly layered. The outer room-temperature shield comprises $\mathrm{CE}\nu\mathrm{NS}$6 of $\mathrm{CE}\nu\mathrm{NS}$7-borated polyethylene, $\mathrm{CE}\nu\mathrm{NS}$8 of lead, additional polyethylene above the experiment, and soft iron for magnetic shielding; inside the cryostat, further lead, polyethylene, and copper are deployed above the detectors, and a muon veto surrounds the installation (Augier et al., 2022). In simulation, passive shielding alone reduces reactogenic neutrons by about $\mathrm{CE}\nu\mathrm{NS}$9 and cosmogenic neutrons by about a factor $N^2$0, while the muon veto provides an additional factor of about $N^2$1 on cosmogenic neutrons (Augier et al., 2022).

These background studies are directly tied to projected $N^2$2 significance. For a $N^2$3 threshold and $N^2$4 effective detection efficiency, the signal-to-background ratio is quoted as about $N^2$5 without veto and about $N^2$6 with an ideal veto; for a single reactor cycle, the expected statistical significance ranges from $N^2$7 to $N^2$8, and from $N^2$9 to $N$0 when a conservative factor-$N$1 systematic on the neutron rate is included (Augier et al., 2022). These numbers clarify a recurring point in the Ricochet literature: the experiment is not reactor-neutron-limited after shielding, but cosmogenic-neutron-limited.

6. Commissioning at ILL and present experimental status

The first full commissioning of the Ricochet installation at ILL with physics detectors used a mini-CryoCube module consisting of three $N$2 Ge cryogenic calorimeters (Armatol et al., 30 Jul 2025). This campaign established the performance of the cryogenic station, passive shielding, muon veto, and synchronized data acquisition under both reactor-on and reactor-off conditions (Armatol et al., 30 Jul 2025). After setup improvements, the ionization baseline reached $N$3 electron equivalent, while the phonon-channel resolutions ranged from $N$4 to $N$5 of total phonon energy (Armatol et al., 30 Jul 2025).

The commissioning runs also yielded a first in situ validation of the background model. In the $N$6–$N$7 recoil region, a nuclear-recoil rate of $N$8 events/(kg day keV) was measured during reactor-off periods for events in coincidence with the muon veto, in agreement with the cosmogenic-neutron rate predicted by the Geant4 simulation (Armatol et al., 30 Jul 2025). After rejecting events in coincidence with the muon veto, the combined $N$9 C.L. limit on the nuclear-recoil background during reactor-on running was $50\,\mathrm{eV}$0 events/(kg day keV) in the same energy range, compatible with the model prediction of $50\,\mathrm{eV}$1 events/(kg day keV) (Armatol et al., 30 Jul 2025).

The sensitivity of this first analysis was nonetheless limited by surface-event contamination (Armatol et al., 30 Jul 2025). Commissioning thus identified a concrete next-step problem rather than a conceptual obstacle: upgraded detectors, including FID-style surface-event rejection, are being introduced to address the contamination (Armatol et al., 30 Jul 2025). A plausible implication is that Ricochet has now passed from site qualification and component R&D into a genuine science-phase transition, with the principal remaining limitations located in low-energy background rejection rather than in cryogenic operation or detector threshold alone.