Comprehensive study of cosmogenic neutron production in large liquid scintillator detectors

TL;DR

This work tackles the problem of accurately modeling cosmogenic neutrons produced by muons in large liquid scintillator detectors. It combines GEANT4 simulations with three hadronic lists and a TALYS-based cross-section reweighting to calibrate $Y_n$ and neutron-multiplicity observables against Daya Bay data, finding substantial normalization improvements but persistent discrepancies in single-neutron events. The study reveals that BIC-based GEANT4 models align better with data than BERT-based ones, while TALYS adjustments reduce the total-yield mismatch to near-percent levels for BIC and to about 6% for BERT, highlighting remaining gaps in intranuclear cascade and deexcitation modeling. It establishes a reproducible benchmarking framework and a practical two-step refinement strategy—channel-specific cross-section tuning followed by cascade- and deexcitation-parameter exploration—that can inform background predictions for JUNO, Hyper-Kamiokande, and DUNE.

Abstract

Neutrons produced by cosmic ray muons constitute a significant background for underground experiments investigating neutron oscillations, neutrinoless double beta decay, dark matter, and other rare event signals. This work benchmarks measured neutron yields and neutron multiplicities--with a focus on data from the Daya Bay Reactor Neutrino Experiment--against comprehensive simulations using three GEANT4 hadronic physics lists. These simulations are further refined via a TALYS-based adjustment of hadronic cross sections. For the BERT-based models, the adjustment reduces the discrepancy in the total neutron yield from about 20% to approximately 6%, while for the BIC-based models it improves the agreement from roughly 13% to the sub-percent level (~0.3%), indicating a markedly better consistency of the BIC-based models with the experimental data. Nevertheless, a clear tension persists: simulations systematically underproduce single-neutron events while overproducing multi-neutron events. The study establishes a reproducible benchmark for cosmogenic neutron modeling and proposes a targeted refinement strategy--including channel-specific reweighting and intranuclear cascade parameter tuning--to guide future model development.

Figures (10)

• Figure 1: Energy spectrum and the angular distributions of the simulated muons. Left panel: normalized energy spectrum of muons, shown as a probability density with a logarithmic y-axis. Right panel: angular distributions of the muon directions. The black curve shows the probability density as a function of $\cos(zenith)$ (bottom x-axis), while the blue curve shows the azimuth angle distribution (top x-axis). Both angular distributions are normalized to unit area.
• Figure 2: Schematic illustration of the detector geometry used in the simulation, showing the GdLS, LS, and mineral oil regions. The geometry of the water pool surrounding the LS detector is also modeled, though it is not illustrated here.
• Figure 3: Model composition and energy ranges of the three GEANT4 hadronic physics lists used in this work: FTFP_BERT_HP, QGSP_BERT_HP, and QGSP_BIC_HP. The shaded regions indicate smooth transitions between cascade and string models.
• Figure 4: Comparison of inclusive inelastic cross sections ($\sigma_{in}(E)$) for $n$, $p$, and $\gamma$ interacting with $^{12}$C, obtained from TALYS and GEANT4 hadronic models (FTFP_BERT_HP, QGSP_BERT_HP, and QGSP_BIC_HP). Note that charged pions are not included in this comparison as they are not supported by the TALYS framework.
• Figure 5: Comparison of the dominant neutron-multiplying inelastic reaction-channel cross sections for $n$, $p$, and $\gamma$ projectiles on $^{12}\mathrm{C}$ below 200 MeV. For each projectile, the reaction channel in which the total $n$ number in the final state exceeds that of the incident particle and has the highest occurrence probability in GEANT4FTFP_BERT_HP is identified, and the corresponding cross sections are compared with those from QGSP_BERT_HP, QGSP_BIC_HP, and TALYS.