Sensitivity to sub-GeV dark matter in forthcoming spallation-source neutrino experiments

This paper demonstrates that upcoming low-threshold coherent elastic neutrino-nucleus scattering experiments at major spallation sources (ESS, J-PARC, and CSNS) will significantly probe previously unexplored parameter space for sub-GeV thermal dark matter produced via neutral pion decays and mediated by vector portals.

The Gist

Imagine the universe is a giant, bustling city. We know about the people we can see and touch (stars, planets, you, me), but we suspect there's a massive, invisible crowd of "ghosts" living among us, making up most of the city's weight. We call these ghosts Dark Matter.

For decades, scientists have been trying to catch these ghosts. The usual strategy has been to build giant, ultra-sensitive traps (like huge tanks of liquid xenon) and wait for a ghost to bump into a regular atom. But these traps have a problem: they are like fishing nets with holes that are too big. They can catch the "heavy" ghosts (which are rare), but they completely miss the "light" ghosts (which are very common but very hard to feel).

This paper is about a new, clever way to catch those light, sub-GeV dark matter ghosts using facilities that were originally built for something else: neutrino factories.

The Setting: The "Spallation" Factory

Think of a spallation neutron source (like the ones in Sweden, Japan, and China) as a massive particle cannon.

1. The Cannon: They shoot a super-fast beam of protons (tiny particles) into a heavy metal target (like mercury or tungsten).
2. The Explosion: When the protons hit the target, it's like a billiard ball hitting a rack of balls. It creates a chaotic shower of new particles.
3. The Main Product: The factory's main job is to make neutrons for scientific research.
4. The Byproduct: But in this chaos, they also create a huge number of neutral pions (let's call them "pi-zeroes"). These are unstable particles that usually vanish instantly, turning into light (photons).

The New Idea: The "Dark Photon" Portal

The authors of this paper propose a "What if?" scenario. What if, instead of turning into light, some of these pi-zeroes turn into a Dark Photon?

• The Analogy: Imagine a pi-zero is a magician. Usually, it pulls a rabbit out of a hat (light). But in this theory, sometimes it pulls out a Dark Photon instead.
• The Escape: Light gets absorbed by the metal walls of the factory immediately. But Dark Photons are "ghostly"—they pass right through the walls without stopping.
• The Transformation: Once outside, the Dark Photon decays into two Dark Matter particles. These are the ghosts we want to catch.

The Hunt: Listening for a Whisper

The scientists plan to place special detectors near these factories. These detectors are designed to catch neutrinos (another ghostly particle) by feeling the tiny "kick" (recoil) when a neutrino hits an atom.

The authors realized that if Dark Matter is created in the factory, it will fly out and hit these same detectors.

• The Signal: When a Dark Matter particle hits an atom in the detector, it gives the atom a tiny kick, just like a neutrino does.
• The Trick: The key is timing. Neutrinos arrive in a steady stream or a specific pattern. Dark Matter, coming from the immediate decay of the pi-zeroes, arrives in a very specific, ultra-fast burst. By looking at exactly when the hits happen, the scientists can filter out the background noise and isolate the Dark Matter signal.

The Comparison: Two Ways to Predict the Chaos

To know what to look for, the scientists had to predict exactly how many pi-zeroes are created in the explosion. They used two methods:

1. The Supercomputer Simulation (GEANT4): A detailed, complex 3D simulation of every single collision. This is like simulating every single billiard ball in the rack.
2. The Shortcut Formula (Sandford-Wang): A mathematical shortcut based on past experiments. This is like using a rule of thumb to guess the outcome.

The Result: They found that both methods give almost the same answer. The shortcut is good enough, which is great news because it saves a lot of computing time!

The Verdict: A New Hunting Ground

The paper concludes that these upcoming facilities (in Europe, Japan, and China) are perfect hunting grounds for light Dark Matter.

• Why? They have incredibly powerful beams (lots of pi-zeroes) and very sensitive detectors that can feel the tiniest kicks.
• The Impact: They will be able to test regions of the "Dark Matter map" that no one has ever explored before. They might finally catch the light, elusive ghosts that the big, heavy traps have missed for years.

In a Nutshell

Scientists are taking advantage of the "collateral damage" from massive particle cannons. By watching for tiny, perfectly timed kicks in their detectors, they hope to prove that the invisible "ghosts" of the universe are actually light, fast particles that have been hiding in plain sight all along. It's like realizing that while you were listening for a lion's roar, the real mystery was the whisper of a mouse you couldn't hear until now.

Technical Summary

1. Problem Statement

The nature of Dark Matter (DM) remains unknown. While Weakly Interacting Massive Particles (WIMPs) in the GeV–TeV range are well-studied, sub-GeV thermal dark matter remains challenging to probe.

• Limitations of Direct Detection: Conventional direct detection experiments (e.g., multi-ton liquid xenon) have recoil-energy thresholds ($\sim$1 keV$_{nr}$) that render them "blind" to nuclear recoils induced by sub-GeV DM.
• Limitations of Colliders: High-energy colliders often lack the sensitivity for feebly interacting light mediators.
• The Opportunity: Accelerator-based fixed-target experiments, specifically spallation neutron sources, offer a unique environment. High-intensity proton beams striking dense targets produce abundant neutral pions ($\pi^0$). In models with a "vector portal," $\pi^0$ can decay into a dark vector boson ($A'$), which subsequently decays into DM pairs. These DM particles can escape the target and be detected via Coherent Elastic Dark Matter-Nucleus Scattering (CE$\nu$NS), mimicking the neutrino signal but distinguishable via timing and energy spectra.

2. Methodology

The authors perform a comprehensive sensitivity analysis for future experiments at three major facilities:

• European Spallation Source (ESS) (Sweden)
• J-PARC (Japan)
• China Spallation Neutron Source (CSNS)

Theoretical Frameworks:
Two representative vector-portal models are considered:

1. Generic Dark Photon ($U(1)'$): Kinetic mixing between the dark photon and the Standard Model hypercharge. The mediator couples to electric charge (millicharge).
2. Baryophilic Vector ($U(1)_B$): The mediator couples to baryon number. This enhances interactions with nuclei (scaling with atomic mass number $A$ rather than charge $Z$), making it ideal for CE$\nu$NS detectors.

Simulation and Flux Calculation:
A critical component of the study is the accurate determination of the neutral pion ($\pi^0$) flux produced in the target. The authors compare two methods:

• GEANT4 Simulation: A full Monte Carlo transport simulation of proton-target interactions (using the FTFP_BERT physics list), accounting for intranuclear cascades and secondary interactions.
• Sandford-Wang (S&W) Parametrization: A semi-empirical analytical approach based on charged-pion cross-sections, adapted for $\pi^0$ via isospin symmetry.
• Validation: The GEANT4 simulations were validated against existing COHERENT collaboration data from the Spallation Neutron Source (SNS).

Detection and Statistical Analysis:

• Detectors: The study considers specific detector configurations:
  • ESS: 20 kg Xenon (Xe), 7 kg Germanium (Ge), 22.5 kg Cesium Iodide (CsI).
  • J-PARC: 20 kg Xe, 7 kg Ge, 44 kg CsI.
  • CSNS: 300 kg CsI.
• Signal: DM-induced nuclear recoils.
• Backgrounds:
  • CE$\nu$NS: Neutrino-induced nuclear recoils (the dominant background).
  • SSB: Steady-state backgrounds (cosmic rays).
  • pBRN: Prompt beam-related neutrons (significant for Ge at J-PARC).
• Timing Cuts: For J-PARC and CSNS, the short proton beam pulses allow for timing cuts (rejecting events $t > 1.5 \mu s$) to significantly suppress the CE$\nu$NS background. ESS has longer spills (2.8 ms), limiting timing discrimination.
• Statistical Tool: A $\chi^2$ analysis using a Poissonian likelihood function, incorporating nuisance parameters for signal and background normalization uncertainties.

3. Key Contributions

• Flux Comparison: The paper provides a rigorous comparison between GEANT4 simulations and the S&W parametrization for $\pi^0$ production. They find that while S&W reproduces the general spectral trends, it underestimates the total yield by $\sim$10–20% compared to GEANT4 due to missing subleading physics (e.g., soft pion production in secondary interactions). However, the resulting sensitivity projections differ by only $\sim$50%, validating the use of analytical approximations for quick estimates.
• Comprehensive Facility Analysis: This is one of the first detailed studies projecting the reach of the upcoming ESS, J-PARC, and CSNS facilities specifically for sub-GeV DM via the vector portal.
• Baryophilic Scenario: The study extends the analysis to the baryophilic mediator model, demonstrating that these facilities are exceptionally sensitive to this scenario due to the $A^2$ coherence enhancement in the scattering cross-section.
• Configuration Optimization: The authors quantify the impact of detector placement (on-axis vs. 90$^\circ$ off-axis), finding that on-axis configurations improve sensitivity by a factor of $\sim$2.5 due to the forward-peaked nature of $\pi^0$ decay.

4. Results

• Exclusion Reach: The projected 90% Confidence Level (CL) exclusion limits in the $(m_{A'}, Y)$ plane (where $Y = \epsilon^2 \alpha_D (m_\phi/m_{A'})^4$) show that these facilities can probe parameter space unexplored by current experiments (COHERENT, CCM, MiniBooNE, NA64).
• Optimal Mass Range: The highest sensitivity is achieved for mediator masses in the range $8 \text{ MeV} \lesssim m_{A'} \lesssim 100 \text{ MeV}$.
• Performance by Detector:
  • Germanium (Ge) detectors at ESS and J-PARC provide the strongest constraints, reaching $Y \sim 2 \times 10^{-12}$ for low mediator masses.
  • CsI detectors show strong potential, particularly with the larger 300 kg mass at CSNS, though higher thresholds and backgrounds slightly reduce sensitivity compared to Ge.
  • Xenon (Xe) detectors offer competitive sensitivity but are generally outperformed by Ge in the low-mass regime due to threshold effects.
• Thermal Relic Density: The experiments can probe regions below the thermal relic density target for scalar DM, meaning they can test scenarios where DM is under-abundant or requires additional components.
• Baryophilic Limits: For the baryophilic model, the constraints are significantly stronger due to the coupling to baryon number, potentially probing $\alpha_B$ values down to $10^{-12}$–$10^{-11}$.
• Comparison with Other Future Experiments: The projected sensitivities of ESS/J-PARC/CSNS are competitive with, and in some mass ranges superior to, other proposed experiments like DUNE-PRISM, LDMX, and SHiP, particularly for the specific $m_{A'}$ range accessible via $\pi^0$ decay.

5. Significance

• Complementarity: These facilities offer a unique, complementary approach to high-energy colliders and direct detection. They are uniquely sensitive to the "light" dark sector (MeV scale) where other methods fail.
• Background Discrimination: The ability to use timing information (at J-PARC and CSNS) to distinguish prompt DM signals from the neutrino background is a critical advantage, significantly enhancing the discovery potential.
• Validation of Tools: By validating the S&W parametrization against GEANT4, the paper provides a computationally efficient tool for future phenomenological studies while highlighting the necessity of full simulations for precision.
• Physics Potential: The results demonstrate that the next generation of spallation-source neutrino experiments will play a pivotal role in the global search for secluded dark sectors, potentially discovering sub-GeV dark matter or tightly constraining vector-portal models.