Studies of the muon-induced neutron background in LSM: detector concept and status of the installation

A good particle candidate for Cold Dark Matter (CDM) is the supersymmetric neutralino or more generally a weakly interacting massive particle (WIMP). The expected interaction rate of WIMPs with the detector medium in the direct detection experiments is below 0.01 events/kg/day. This makes a good knowledge of the background conditions highly important, especially with ever increasing sensitivity of the detectors. One of the background components is related to cosmic muons and in particular to muon-induced neutrons. Detailed studies carried out by the Edelweiss collaboration in this respect are presented. This activity includes GEANT4 simulations with full event topology as well as a dedicated measurement with a new neutron counter installed in the fall of 2008 in LSM (Laboratoire Souterrain de Modane, France). This counter is incorporated into the existing muon veto system thus allowing to monitor neutrons in coincidence with the incoming muons.

💡 Research Summary

The EDELWEISS collaboration operates a direct dark‑matter search experiment deep underground at the Laboratoire Souterrain de Modane (LSM) in the French Alps. The experiment uses cryogenic germanium bolometers operated at ~20 mK to measure both heat and ionisation from nuclear recoils, providing excellent discrimination between electron‑recoil backgrounds (β, γ) and potential WIMP signals. However, neutrons—whether produced by natural radioactivity or induced by cosmic‑ray muons—can generate nuclear recoils indistinguishable from WIMPs, making their precise characterization essential, especially as the field moves toward tonne‑scale detectors such as EURECA.

The paper presents two complementary approaches to quantify the muon‑induced neutron background. First, a detailed GEANT4 Monte‑Carlo model of the entire EDELWEISS setup was developed. The simulation reproduces the muon flux specific to LSM (≈10⁻⁴ cm⁻² s⁻¹) and includes all relevant materials: the 50 cm polyethylene neutron moderator, the lead γ‑shield, the surrounding rock, and the plastic scintillator muon veto. By tracking muon interactions, the model predicts that a small fraction of muons that escape the veto can still generate neutrons that reach the Ge crystals. The simulated rate of muon‑bolometer coincidences (≈0.03 events kg⁻¹ day⁻¹) matches the measured value, but the rarity of such coincidences limits statistical power, prompting the need for an independent measurement.

To obtain that measurement, the collaboration designed and installed a dedicated neutron counter in September 2008. The detector consists of a 1 m³ volume of Gd‑loaded liquid scintillator (Bicron BC525) surrounded by 16 photomultiplier tubes (eight 8‑inch PMTs for neutron capture signals and six 2‑inch PMTs for the much brighter muon signals). Gadolinium has a very high thermal‑neutron capture cross‑section; each capture releases a cascade of γ‑rays with a total energy of about 8 MeV, which the scintillator efficiently converts to light. A 10 cm thick layer of lead bricks placed beneath the scintillator acts as an enhanced neutron‑production target, increasing the yield by roughly an order of magnitude compared with the surrounding rock. A plastic scintillator module identical to the existing muon veto sits on top of the detector to tag incoming muons. The whole assembly is housed in a plexiglass vessel filled with paraffin for the PMTs, encased in an aluminum safety container, and surrounded by iron plates to reflect escaping neutrons back into the scintillator.

Operational features include an LED‑based optical calibration system (eight 425 nm LEDs distributed throughout the volume) that allows regular monitoring of scintillator light yield and PMT gain, and a comprehensive safety‑monitoring suite. Sensors for vapor, temperature, and liquid leaks feed into the LSM safety infrastructure, while a Linux‑based LabVIEW application (the Karlsruhe Control of Safety, KA‑CS) reads the data via a NI‑6221 DAQ card and issues email alerts if thresholds are exceeded.

Based on the measured muon flux at LSM, the expected muon‑induced neutron count in the new detector is a few events per day. This rate, although low, provides a statistically meaningful sample over several weeks to validate and fine‑tune the GEANT4 predictions. The authors emphasize that such validation is crucial for extrapolating background estimates to future large‑scale experiments, where even a small residual neutron background could mimic a WIMP signal.

In conclusion, the paper demonstrates that the combination of high‑fidelity simulations and a purpose‑built neutron counter offers a robust pathway to quantify muon‑induced neutrons in deep‑underground dark‑matter experiments. The detector’s design addresses both detection efficiency (via Gd loading and lead target) and operational safety (through redundant sensors and automated monitoring). The successful commissioning of this system at LSM marks an important step toward the background control required for next‑generation, tonne‑scale dark‑matter searches.