Simulation of Muon-induced Backgrounds for the Colorado Underground Research Institute (CURIE)

This paper presents a comprehensive Monte Carlo simulation framework using coupled \textsc{mute} and \textsc{geant4} tools to characterize muon-induced neutron backgrounds at the shallow-underground CURIE facility, providing site-specific flux predictions and a validated depth-intensity relation to guide experimental design for low-background physics.

The Gist

Imagine you are trying to listen to a single, incredibly quiet whisper in a room that is currently being pelted by a storm of hail. To hear that whisper, you need to go deep underground, where the rock above acts like a thick blanket, stopping most of the hail.

This paper is about building a perfect digital map of the "hail" that still gets through the blanket at a new underground lab called CURIE (Colorado Underground Research Institute).

Here is the breakdown of what the scientists did, using simple analogies:

1. The Problem: The "Ghost" Particles

Scientists studying the universe's biggest mysteries (like dark matter or how the universe began) need to build detectors that are incredibly sensitive. But even deep underground, they aren't safe from cosmic rays.

Think of cosmic rays as a constant rain of invisible particles from space. When these particles hit the Earth's atmosphere, they create a shower of "hail" (muons). Most of this hail stops at the surface, but some punches through the rock underground.

• The Danger: When these underground muons hit the rock walls of the lab, they don't just stop; they smash into the rock and create a secondary explosion of new particles (neutrons, gamma rays, electrons).
• The Analogy: Imagine throwing a tennis ball (the muon) at a brick wall. The ball stops, but it knocks loose a cloud of dust and tiny pebbles (the secondary particles). If your experiment is looking for a specific type of dust, this "knocked loose" dust creates a lot of noise, making it hard to find the real signal.

2. The Solution: A Two-Stage Digital Simulation

The scientists couldn't just wait around and count every single particle hitting the lab for years; that would take too long. Instead, they built a video game simulation to predict exactly what would happen.

They used a clever two-step process, like a relay race:

• Runner 1 (The "Mute" Framework): This part simulates the muons traveling from the surface, through miles of complex, jagged rock, down to the lab entrance. It calculates exactly how fast they are going and what angle they are coming from.
• Runner 2 (The "Geant4" Framework): This part takes the data from Runner 1 and simulates what happens when those muons hit the specific rock walls of the lab. It calculates the "secondary explosion" (the dust and pebbles) in extreme detail.

Why two runners? Simulating the whole journey in one go would be like trying to calculate every single raindrop in a hurricane and every ripple in a puddle simultaneously. It would crash the computer. By splitting the job, they got the best of both worlds: speed and precision.

3. The "Recipe" for Accuracy

To make sure their simulation was real, they didn't just guess the rock composition. They actually took real rock samples from the CURIE tunnels and analyzed them under microscopes.

• The Metaphor: It's like a chef trying to bake a cake. Instead of guessing how much flour and sugar is in the local soil, they went to the farm, weighed the exact ingredients, and then baked the cake. This ensured their "digital cake" tasted exactly like the real thing.

They also realized that simply using an "average" speed for the muons wasn't good enough.

• The Analogy: If you tell a driver, "The average speed limit is 55 mph," they might drive 55 mph the whole time. But in reality, some drivers go 30, some go 70, and some go 90. The paper showed that using the exact distribution of speeds (the "traffic pattern") changed the results significantly. Ignoring the fast drivers would have led to a wrong prediction of how much "dust" (background noise) would be created.

4. The Results: What Did They Find?

After running the simulation, they produced a "menu" of background noise for the lab:

• The Neutrons: These are the "heavy hitters." They are hard to stop and can mimic the signals scientists are looking for. The team calculated exactly how many neutrons would hit the lab walls every second.
• The Gamma Rays: These are the "loud noise." Surprisingly, the simulation showed that gamma rays (a type of high-energy light) are actually the most common background particle, outnumbering neutrons by a huge margin.
• The "Depth-Intensity" Rule: They created a new mathematical formula (a "Depth-Intensity Relation") that acts like a thermometer for underground labs. If you know how deep a lab is, this formula tells you exactly how much background noise to expect. This is useful for labs that are too deep for surface rules but too shallow for deep-mining rules.

5. Why This Matters

This paper is a blueprint for the future.

• For CURIE: It tells the scientists exactly how much shielding (like thick lead or plastic walls) they need to build to protect their experiments.
• For the World: The scientists made their simulation code publicly available. It's like open-sourcing the "recipe" so that other scientists building labs in different countries can use the same tools to compare their results fairly.

In a nutshell:
The scientists built a super-accurate digital twin of an underground lab to predict the "noise" created by space particles hitting the rock. They found that the noise is louder and more complex than previously thought, and they provided a new set of tools (the simulation and the depth formula) to help scientists everywhere build better, quieter experiments to listen to the whispers of the universe.

Technical Summary

1. Problem Statement

Low-background physics experiments (e.g., dark matter searches, neutrinoless double beta decay) require precise knowledge of background radiation sources to design effective shielding and veto systems. While deep-underground facilities (>1 km.w.e.) effectively suppress cosmic muons, shallow-underground facilities (<1 km.w.e.) like the Colorado Underground Research Institute (CURIE) are increasingly vital for R&D, material radioassay, and detector testing.

However, at shallow depths, the cosmogenic muon flux remains significant. These muons interact with the surrounding rock and facility materials to produce secondary particles (neutrons, gammas, electrons, positrons).

• The Challenge: Existing models often rely on simplified approximations (e.g., mean muon energy, isotropic flux) or generic rock compositions, which fail to capture the complex angular dependencies and local geological variations specific to a site.
• The Gap: There is a lack of high-precision, end-to-end simulation frameworks tailored for shallow-underground sites with complex overburden geometries, leading to uncertainties in background predictions for experimental design.

2. Methodology

The authors developed a two-stage, coupled Monte Carlo simulation framework using mute (v2.0.1) and geant4 (v11-02-01) to simulate muon-induced backgrounds for two specific CURIE laboratories: the Subatomic Particle Hideout (SPH) and Cryolab I (CLI).

A. Simulation Strategy

1. Stage 1: Muon Transport (mute)

   • Simulates the propagation of primary cosmic-ray muons from the surface through the specific, detailed overburden profiles of the Edgar Experimental Mine.
   • Generates site-specific, angular-dependent muon energy and angular distributions ($\Phi_\mu(E_\mu, \theta, \phi)$) at the boundary of a hemispheric rock volume surrounding the lab.
   • Innovation: Unlike standard approaches that use a mean energy approximation, this method preserves the correlation between arrival angle and energy, accounting for azimuthal anisotropy caused by the complex mountain topology.

2. Stage 2: Secondary Production (geant4)

   • Takes the muon distributions from mute as input.
   • Models the production and transport of secondary particles within a simplified hemispherical rock geometry (radius 8.4 m for SPH, 12.0 m for CLI) surrounding the cavern.
   • Tracks particles until they cross the "rock-cavern boundary" (the interface between the rock and the air-filled lab).
   • Physics Lists: Utilized the shielding_hp physics list (optimized for low-background simulations) with the neutron High Precision (HP) package.

B. Key Technical Refinements

• Geological Characterization: Rock samples from SPH and CLI were analyzed via Scanning Electron Microscopy (SEM) to determine precise chemical compositions (Si, O, Al, K, etc.) and densities, which were implemented in geant4.
• Shower Equilibrium Validation: The study verified that a minimum rock thickness of 4 meters is sufficient to reach "shower equilibrium," where secondary production stabilizes. This justifies the use of the simplified hemispherical geometry without simulating the entire mountain.
• Correction for Simulation Bias: Recognizing that geant4 historically underpredicts muon-induced neutron yields compared to experimental data (by factors of 1.4–2.8), the authors applied an empirical correction function (based on Mei & Hime) derived from liquid-scintillator data to scale the simulated neutron multiplicities.
• Systematic Uncertainties: A comprehensive error budget was established, accounting for rock density, muon energy, live-time calculations, and the geant4 model choice.

3. Key Contributions

1. First Comprehensive Simulation for CURIE: Provided the first quantitative, end-to-end background characterization for the new CURIE facility, establishing benchmarks for future experiments.
2. Angular-Dependent Sampling: Demonstrated that using angular-dependent muon energy spectra (rather than a mean energy approximation) is critical. The study found that mean-energy approximations overestimate neutron flux by ~7.6% and fail to reproduce the correct spectral shape, particularly at low and high energies.
3. New Depth-Intensity Relation (DIR): Developed a new parameterization for muon-induced neutron flux as a function of depth ($X$) that is valid for both shallow and deep underground facilities:
   $$\Phi_n(X) = A \cdot \left(\frac{X_0}{X}\right)^n \cdot e^{-X/X_0}$$
   This model improves upon previous extrapolations (like Mei & Hime) which showed deviations at shallow depths (<0.6 km.w.e.).
4. Open-Source Framework: The entire simulation framework, including geometry definitions and scripts, has been made publicly available to the low-background physics community to enable standardized inter-facility comparisons.

4. Key Results

A. Flux Predictions

The simulations predicted the total muon-induced secondary fluxes at the rock-cavern boundary for SPH and CLI:

Particle Type	SPH Flux ($m^{-2}s^{-1}$)	CLI Flux ($m^{-2}s^{-1}$)
Neutrons (Total)	$(8.52 \pm 1.30_{sys}) \times 10^{-3}$	$(8.86 \pm 1.62_{sys}) \times 10^{-3}$
Gamma Rays (Total)	$(5.54 \pm 0.70_{sys}) \times 10^{-1}$	$(6.51 \pm 1.06_{sys}) \times 10^{-1}$
Electrons	$(4.39 \pm 0.56_{sys}) \times 10^{-2}$	$(5.28 \pm 0.86_{sys}) \times 10^{-2}$
Positrons	$(1.08 \pm 0.14_{sys}) \times 10^{-2}$	$(1.29 \pm 0.21_{sys}) \times 10^{-2}$

• Dominance of Electromagnetic Backgrounds: Contrary to some assumptions where neutrons are the primary concern, electromagnetic components (gamma rays) dominate the total flux by nearly two orders of magnitude.
• Energy Spectra:
  • Neutrons: Show distinct regions: thermal peak ($\sim 2.5 \times 10^{-8}$ MeV), evaporation ($0.1-10$ MeV), and high-energy spallation ($\sim 100$ MeV extending to GeV). A specific absorption feature near $10^{-3}$ MeV was attributed to neutron capture on $^{23}$Na in the rock.
  • Gammas: Extend to GeV energies, significantly higher than the typical $\sim 3$ MeV from natural radioactivity. A peak at 511 keV is observed due to positron annihilation.

B. Validation

• The new Depth-Intensity Relation (DIR) aligns well with experimental data from facilities like SciBath and BUST, and with the Mei & Hime model at depths $>0.6$ km.w.e., but provides a more accurate fit for the shallow regime.
• Predictions for the Soudan Underground Laboratory matched existing measurements of fast-neutron flux within uncertainties.

5. Significance

• Experimental Design: The results provide essential input parameters for designing shielding (e.g., high-hydrogen materials for neutrons, dense materials for high-energy gammas) and active veto systems for experiments at CURIE.
• Methodological Advancement: The paper highlights that local geology and overburden geometry significantly influence secondary yields. It argues against the "one-size-fits-all" approach, emphasizing the need for site-specific simulations that account for angular anisotropy.
• Community Resource: By releasing the simulation framework, the authors enable other shallow-underground facilities to perform similar high-precision background characterizations, fostering better standardization in the low-background physics community.
• Bridging the Gap: The work effectively bridges the gap between surface-level facilities and deep-underground labs, validating that shallow sites can support sensitive R&D if backgrounds are accurately modeled and mitigated.