Cosmogenic Neutron Production in Water at SNO+

The SNO+ collaboration measured a cosmogenic neutron yield in ultra-pure water at high muon energies, finding agreement with FLUKA simulations over GEANT4 and revealing a lower yield than in heavy water, which underscores the critical influence of target material composition on neutron production for future underground experiments.

The Gist

Imagine the Earth is a giant, thick blanket protecting us from the "cosmic rain" of high-energy particles coming from deep space. Most of this rain gets stopped by the atmosphere, but the toughest drops—called cosmic muons—can punch right through the ground, all the way down to deep underground laboratories like the one at SNOLAB in Canada.

This paper is about a team of scientists (the SNO+ Collaboration) who wanted to understand a specific side effect of this cosmic rain: neutrons.

Here is the story of their discovery, broken down into simple concepts:

1. The Setup: A Giant Underwater Camera

Think of the SNO+ detector as a massive, crystal-clear swimming pool buried 2 kilometers underground. It's lined with thousands of ultra-sensitive cameras (photomultiplier tubes) that can see the faintest glimmer of light.

For a while, this pool was filled with ultra-pure water. The scientists wanted to see what happens when a cosmic muon (a fast-moving particle) swims through this water.

2. The Problem: The "Ghost" Neutrons

When a muon smashes into the water, it doesn't just pass through; it creates a chaotic splash. This splash knocks loose tiny particles called neutrons.

Why do the scientists care? Because these neutrons are like ghosts. They are invisible, but when they eventually get caught by a hydrogen atom in the water, they let out a tiny, specific flash of light (a gamma ray).

• The Analogy: Imagine you are trying to hear a whisper in a quiet library. If someone drops a heavy book (the muon), it makes a huge noise. But if that book knocks over a stack of papers that then makes a tiny, specific "pop" (the neutron capture), you need to be very careful to distinguish that "pop" from the background noise.

These "ghost" neutrons are a major problem for scientists looking for other rare events (like dark matter or neutrino decay) because they can fake the signal scientists are actually looking for.

3. The Experiment: Counting the Ghosts

The scientists spent about a year watching their water pool. They used a clever trick to find the neutrons:

1. Spot the Muon: First, they identified a muon passing through the pool (it leaves a bright trail of light).
2. Wait for the Pop: They waited a fraction of a second to see if a "pop" (neutron capture) happened nearby.
3. The Result: They counted 1,412 of these neutron "pops" caused by 13,690 muons.

From this, they calculated a "yield": For every gram of water a muon travels through, it produces a specific number of neutrons. Their result was 3.38 × 10⁻⁴.

4. The Big Surprise: The Computer Models Were Wrong

This is where the story gets exciting. Scientists use computer programs to predict how physics works. Two famous programs are GEANT4 and FLUKA.

• The Expectation: The scientists asked the computers, "How many neutrons should we see?"
• The Reality: The GEANT4 program predicted they would see about 30% fewer neutrons than they actually found. It was like the computer saying, "You'll catch 7 fish," but the scientists actually caught 10.
• The Winner: The FLUKA program, however, predicted the number almost perfectly.

Why does this matter? It tells us that our current understanding of how particles smash into water (specifically the "GEANT4" model) needs an update. If we are building future experiments to find dark matter, we need to know exactly how many "ghost" neutrons to expect, or we might mistake a ghost for a real discovery.

5. The "Heavy Water" Comparison

The SNO+ experiment is the successor to the famous SNO experiment, which used Heavy Water (water with a special, heavier type of hydrogen called deuterium).

• The Comparison: The new experiment used normal water; the old one used heavy water. Both were at the same depth, so the "cosmic rain" was identical.
• The Finding: The heavy water experiment found more neutrons than the normal water experiment.
• The Lesson: This proves that the ingredients of the water matter. Just like a cake made with almond flour behaves differently than one made with wheat flour, neutrons are produced differently depending on whether the water has "heavy" hydrogen or "normal" hydrogen. This teaches us that the atomic structure of the material is a key player in how these particles interact.

The Bottom Line

This paper is a "quality control" check for the universe. By measuring exactly how many neutrons are created when cosmic rays hit water, the scientists have:

1. Proven that one of our main computer models (GEANT4) is underestimating the chaos caused by cosmic rays.
2. Shown that the type of water (normal vs. heavy) changes the outcome.
3. Provided a better map for future experiments to avoid getting tricked by "ghost" neutrons when they hunt for the secrets of the universe.

In short: They went deep underground, counted the tiny flashes of light, and realized our computer simulations need a software update to match reality.

Technical Summary

1. Problem Statement

Cosmogenic muons penetrating underground laboratories produce neutrons through spallation and secondary interactions. These neutrons constitute a significant background for low-energy physics searches, including neutrinoless double beta decay, dark matter detection, and solar neutrino studies.

• The Gap: While neutron yields have been measured in various media (liquid scintillators, heavy water) and at various depths, there was a lack of precise measurements in ultra-pure water (UPW) at high average muon energies.
• The Specific Challenge: Detecting neutrons in water is difficult because the signal (neutron capture on hydrogen emitting a 2.2 MeV gamma ray) is low energy and easily masked by backgrounds. Furthermore, existing simulation models (GEANT4 vs. FLUKA) show discrepancies in predicting neutron yields, particularly at high energies.
• The Opportunity: The SNO+ experiment, located deep underground (6010 m.w.e.), offers a unique environment with one of the highest average cosmogenic muon energies (~364 GeV) and the capability to operate with ultra-pure water.

2. Methodology

The analysis utilized data from the SNO+ "water phase" (May 2017 – July 2019), where the detector was filled with 905 tonnes of ultra-pure water.

• Detector Setup:
  • An acrylic vessel (AV) containing the water, surrounded by 9,362 inward-facing PMTs and 91 outward-looking PMTs (OWLs) for muon tagging.
  • The OWLs detect Cherenkov light from muons passing through the external water shield, providing a robust muon trigger.
• Event Selection:
  • Muon Selection: Through-going muons were selected based on OWL hits, high PMT charge, and specific timing/charge distributions. A total of 13,690 cosmogenic muons were selected.
  • Neutron Selection: Neutrons were identified by their capture on hydrogen, producing a 2.2 MeV gamma ray. Selection criteria included:
    • Time window: 10 µs to 776 µs after the muon.
    • In-time PMT hits: 9 to 25.
    • Spatial constraints: Within 4.6 m of the muon track and <8 m radius from the detector center.
  • Result: 1,412 neutron follower events were identified.
• Simulation & Efficiency:
  • Monte Carlo simulations (using GEANT4 and the RAT framework) were used to model muon propagation and neutron production.
  • Efficiency Calibration: The neutron selection efficiency was calculated as a function of muon track length. Crucially, the simulation was corrected using data from an Am-Be calibration source and comparisons between data and MC for neutron capture distances and PMT hit counts.
  • Corrections: Two main corrections were applied to the simulation efficiency:
    1. Distance Correction: Scaling the neutron travel distance in MC to match data (factor 1.07).
    2. Hits Correction: Scaling the number of in-time PMT hits (factor 0.95 for cosmogenic neutrons).

3. Key Results

• Measured Neutron Yield:
  The study determined the cosmogenic neutron yield in ultra-pure water at SNO+ to be:
  $$Y_n = (3.38^{+0.23}_{-0.30}) \times 10^{-4} \, \text{cm}^2\text{g}^{-1}\mu^{-1}$$
  This measurement corresponds to an average muon energy of 364 GeV.

• Model Comparison (FLUKA vs. GEANT4):

  • FLUKA: The measured yield shows clear agreement with the FLUKA neutron production model.
  • GEANT4: The widely used GEANT4 model predicts a yield of $2.13 \times 10^{-4} \, \text{cm}^2\text{g}^{-1}\mu^{-1}$, which is approximately 30% lower than the measured value. The discrepancy is primarily driven by an excess of single-neutron events in the GEANT4 simulation compared to data.

• Material Composition Comparison (SNO vs. SNO+):

  • The SNO experiment previously measured neutron yields in heavy water (D$_2$O) under identical muon flux conditions.
  • The SNO+ result (in H$_2$O) is significantly lower than the SNO result.
  • Interpretation: This difference is attributed to nuclear structure and target composition. In heavy water, deuterium nuclei can release neutrons more readily than oxygen nuclei in ordinary water, leading to a higher yield in D$_2$O.

4. Significance and Implications

• Background Modeling: The result provides a critical benchmark for background modeling in future water-based underground experiments (e.g., Hyper-Kamiokande, DUNE). It demonstrates that relying solely on GEANT4 for high-energy muon-induced neutron backgrounds may lead to underestimations.
• Nuclear Physics Insight: The direct comparison between SNO (D$_2$O) and SNO+ (H$_2$O) under the same muon flux provides rare experimental evidence that target material composition significantly influences neutron production rates. This highlights the importance of nuclear structure in spallation processes.
• Simulation Validation: The study validates the FLUKA model for high-energy muon interactions in water, suggesting it should be preferred over GEANT4 for these specific high-energy regimes.
• Future Prospects: The SNO+ detector is currently filled with liquid scintillator. Repeating this measurement in scintillator will allow for a direct comparison of neutron yields across three different media (H$_2$O, D$_2$O, and Scintillator) using the same detector and muon flux, further refining our understanding of cosmogenic backgrounds.

Conclusion

This paper presents the first measurement of cosmogenic neutron yield in ultra-pure water at high average muon energies. It establishes a yield of $3.38 \times 10^{-4} \, \text{cm}^2\text{g}^{-1}\mu^{-1}$, reveals a ~30% deficit in the standard GEANT4 model compared to data, and confirms the strong dependence of neutron production on the target material's nuclear composition. These findings are essential for optimizing the design and background rejection strategies of next-generation underground physics experiments.