Axion-Like Particle Search with a Hybrid Cherenkov-Scintillation Detector
Listen
Radio episode about this paper
Transcript
Introduction to the show: ident: High Energy Physics: Experiment. The day in high-energy experiment: detectors, searches and limits.
Kai: We're looking at the paper: "Axion-Like Particle Search with a Hybrid Cherenkov-Scintillation Detector".
Mira: Here's the gist of the paper. This analysis presents the first proof-of-concept search for axion-like particles (ALPs) using a hybrid Cherenkov-scintillation detector at a beam-dump facility.
Kai: The work demonstrates the potential of hybrid optical detection to improve signal discrimination in a high-rate beam dump environment.
Mira: Experimental Setup and Detector Capabilities The study utilizes a 10-ton liquid argon light collection detector, the Coherent CAPTAIN-Mills (CCM) experiment, located at Los Alamos National Laboratory.
Kai: The CCM200 detector is instrumented with 200 photomultiplier tubes (PMTs), providing approximately 50% photocathode coverage.
Title and authors: Kai: So we're looking at this paper, "Axion-Like Particle Search with a Hybrid Cherenkov-Scintillation Detector," and they didn't find an excess for axions in that ten to three MeV mass range.
Mira: Right. But the real thing here is how they proved that combining prompt Cherenkov light and delayed scintillation light gives us a much cleaner way to look for these particles in high-rate environments.
Lev: I agree with Mira. That six fold improvement in background rejection shows that this method can be practical for real hardware, even if it adds some complexity to the system design.
Kai: It really does suggest that we can expect future experiments to use these hybrid detection techniques, not just the single-mode scintillation light they used in their previous work.
Mira: That’s what I think. This approach provides a solid path forward for probing these lower energy axion masses that are harder to constrain otherwise.
Lev: It means if we build systems with this kind of timing and light sorting, the overhead lands in terms of complexity, but it pays off in getting a purer signal.
Kai: So to wrap up on the "Axion-Like Particle Search with a Hybrid Cherenkov-Scintillation Detector" paper, they didn't see an excess for ten to three MeV axions, but they proved that the improved discrimination capability is real.
Mira: It showed that this hybrid approach can provide comparable or even better ALP sensitivity across that same mass range given the exposure.
Lev: I just want to mention their method of using neutron events as a data-driven cross-check for the electromagnetic selection, because that’s a robust way to validate those rejection capabilities at higher energies.
Kai: So we’re done with this paper, and it really shows how combining prompt Cherenkov radiation with scintillation light can be a powerful tool when you're looking for rare interactions in high-rate environments.
Mira: It's a solid piece of work that shows the physics potential for this type of hybrid detector technology in beam dump facilities.
Lev: The method they used to check those selections against neutron events is just a very robust way to validate those rejection capabilities at higher energies.
The paper's summary: Kai: So, we've seen that even without finding an excess in that ten to three MeV range, this hybrid approach is genuinely useful for testing axions.
Mira: Exactly. The paper points out that the main improvement isn't just finding a signal, but actually getting a better handle on what your background is doing.
Lev: They detail how combining those four observables—the uncoated hits, the angular spread metric C, the pulse shape ratio, and spatial spread—actually allows for a much tighter selection of events.
Kai: That means they can cut down on that high-rate beam dump noise significantly, reducing the prebeam sample size to just about zero point four nine percent of what you started with.
Mira: That reduction in background is huge because it lets them see signals that would otherwise be completely drowned out by the constant background noise in a real detector.
Lev: From an error correction viewpoint, that improved purity means the actual computational overhead for running the analysis is manageable, even on real hardware.
Kai: They also focus on improving how they reconstruct positions using a GraphNeT machine learning framework with a DeepIce transformer backbone.
Mira: That’s pretty cool because it gives them an improved three dimensional position resolution of about seven point nine centimeters at one sigma.
Lev: That is roughly four fold better than the previous chargeweighted reconstruction method they used in those older CCM120 analyses.
Kai: And they tied that spatial reconstruction into their energy calculation using a position dependent calibration factor derived from charge collected within the first ninety nanoseconds of an event.
Mira: So, when you combine where the event happened with how much charge was collected early on, you get a more accurate estimate of the deposited energy where things vary so much.
Lev: That helps ground things for running this on real hardware because it’s a data-driven way to handle the non-uniform detector response.
Kai: They also used neutron-dominated events as a control sample to validate their electromagnetic selection at higher energies, which is a smart move.
Mira: Using those neutron events as a cross check gives them confidence that the signal rejection capabilities they built are actually working correctly across different energy regimes.
Lev: It’s like having an internal test of your error correction code using known background events to make sure it's actually robust enough for real hardware deployment.
Kai: The main implication is that we can expect future searches to use these hybrid detection techniques, not just the single-mode scintillation light they used before.
Mira: It suggests that the physics potential here is proving that combining prompt Cherenkov and scintillation light can give us a much cleaner look at these rare interactions.
Lev: So, for anyone building hardware right now, this tells you that the complexity of adding more layers of discrimination might be worth it if you need that kind of background rejection.
Kai: The next thing we’re looking at is how those specific results—the mass range and the local significance—fit into the broader constraints on axion models like KSVZ QCD axions.
Mira: We can look at how those exclusion limits change when you consider different theoretical models for how these particles couple to photons.
The paper's improvements: Kai: So, we're looking at how they actually improved their detection setup on this paper about axions.
Mira: Right. The authors talk about building this sophisticated selection algorithm using four distinct observables to sort the data better than before.
Lev: They are talking about "Number of Hits on the Uncoated PMTs," "Spread in the Directionality" using that metric C, "Pulse Shape Ratio," and "Spatial Spread of Events."
Kai: That means they can cut down on that high-rate beam dump noise significantly, reducing the prebeam sample size to just about zero point four nine percent of what you started with.
Mira: That reduction in background is huge because it lets them see signals that would otherwise be completely drowned out by the constant background noise in a real detector.
Lev: From an error correction viewpoint, that improved purity means the actual computational overhead for running the analysis is manageable, even on real hardware.
Kai: They also focus on improving how they reconstruct positions using a GraphNeT machine learning framework with a DeepIce transformer backbone.
Mira: That’s pretty cool because it gives them an improved three dimensional position resolution of about seven point nine centimeters at one sigma.
Lev: That is roughly four fold better than the previous chargeweighted reconstruction method they used in those older CCM120 analyses.
Kai: And they tie that spatial reconstruction into their energy calculation using a position dependent calibration factor derived from charge collected within the first ninety nanoseconds of an event.
Mira: So, when you combine where the event happened with how much charge was collected early on, you get a more accurate estimate of the deposited energy where things vary so much.
Lev: That helps ground things for running this on real hardware because it’s a data-driven way to handle the non-uniform detector response.
Kai: They also used neutron-dominated events as a control sample to validate their electromagnetic selection at higher energies, which is a smart move.
Mira: Using those neutron events as a cross check gives them confidence that the signal rejection capabilities they built are actually working correctly across different energy regimes.
Lev: It’s like having an internal test of your error correction code using known background events to make sure it's actually robust enough for real hardware deployment.
Kai: The main implication is that we can expect future searches to use these hybrid detection techniques, not just the single-mode scintillation light they used before.
Mira: It suggests that the physics potential here is proving that combining prompt Cherenkov and scintillation light can give us a much cleaner look at these rare interactions.
Lev: So, for anyone building hardware right now, this tells you that the complexity of adding more layers of discrimination might be worth it if you need that kind of background rejection.
Kai: The next thing we’re looking at is how those specific results—the mass range and the local significance—fit into the broader constraints on axion models like KSVZ QCD axions.
Mira: We can look at how those exclusion limits change when you consider different theoretical models for how these particles couple to photons.
Conclusion: Kai: So we've looked at the full paper on "Axion-Like Particle Search with a Hybrid Cherenkov-Scintillation Detector," and they didn't find an excess for axions in that ten to three MeV range.
Mira: Right, but the real thing here is how they proved that combining prompt Cherenkov light and delayed scintillation light gives us a much cleaner way to look for these particles in high-rate environments.
Lev: I agree with Mira. That six fold improvement in background rejection shows that this method can be practical for real hardware, even if it adds some complexity to the system design.
Kai: It really does suggest that we can expect future experiments to use these hybrid detection techniques, not just the single-mode scintillation light they used in their previous work.
Mira: That’s what I think. This approach provides a solid path forward for probing these lower energy axion masses that are harder to constrain otherwise.
Lev: It means if we build systems with this kind of timing and light sorting, the overhead lands in terms of complexity, but it pays off in getting a cleaner signal.
Kai: So to wrap up on the "Axion-Like Particle Search with a Hybrid Cherenkov-Scintillation Detector," they didn't see an excess for ten to three MeV axions, but they proved that the improved discrimination capability is real.
Mira: It showed that this hybrid approach can provide comparable or even better ALP sensitivity across that same mass range given the exposure.
Lev: I just want to mention their method of using neutron events as a data-driven cross-check for the electromagnetic selection, because that’s a robust way to validate those rejection capabilities at higher energies.
Kai: So we’re done with this paper, and it really shows how combining prompt Cherenkov radiation with scintillation light can be a powerful tool when you're looking for rare interactions in high-rate environments.
Mira: It's a solid piece of work that shows the physics potential for this type of hybrid detector technology in beam dump facilities.
Lev: The method they used to check those selections against neutron events is just a very robust way to validate those rejection capabilities at higher energies.
Bartoszek Engineering · Columbia University New York State University of Florida Los Alamos National Laboratory Massachusetts Institute of Technology Universidad Nacional Autónoma de México University of New Mexico State University Texas A&M University Element Aero Fermi National Accelerator Laboratory Northwestern University
hep-ex
Submitted: 2026-08-12
Updated: 2026-10-27
Comments: Published in Physical Review D, Phys. Rev. D 114, 072008
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Importance score: 82/100
The gist: This analysis presents the first proof-of-concept search for axion-like particles (ALPs) using a hybrid Cherenkov-scintillation detector at a beam-dump facility.
Key concepts
- Hybrid Optical Detection
- This method uses a detector with both liquid argon scintillation light and Cherenkov radiation signals. By coating some photomultiplier tubes (PMTs) with wavelength-shifting material, researchers can selectively detect different types of light, which helps separate the desired ALP signal from background noise in high-rate environments.
- Discriminating Observables
- Four specific measurements are used to distinguish between different types of interactions. These include counting hits on uncoated PMTs (for Cherenkov light), measuring directionality spread, analyzing pulse shape ratios, and quantifying the spatial spread of events around the interaction point.
- Cherenkov Radiation
- Cherenkov radiation is a specific type of light emitted when charged particles travel through a medium faster than the speed of light in that medium. In this experiment, detecting this prompt Cherenkov signal helps identify potential ALP interactions during the initial part of an event.
- Axion-Like Particles (ALPs)
- ALPs are hypothetical particles that could interact with photons. The search looks for their signature, such as a diphoton decay (a -> γγ), which would appear as a specific pattern of light and timing in the detector, helping to confirm their existence.
Terminology
Summary
This analysis presents the first proof-of-concept search for axion-like particles (ALPs) using a hybrid Cherenkov-scintillation detector at a beam-dump facility. p. 0 The work demonstrates the potential of hybrid optical detection to improve signal discrimination in a high-rate beam dump environment. p. 3
Experimental Setup and Detector Capabilities
The study utilizes a 10-ton liquid argon light collection detector, the Coherent CAPTAIN-Mills (CCM) experiment, located at Los Alamos National Laboratory. The CCM200 detector is instrumented with 200 photomultiplier tubes (PMTs), providing approximately 50% photocathode coverage. p. 0 Optical discrimination is enabled by coating 80% of the PMTs with a wavelength-shifting material while the remaining 20% are left uncoated. p. 0 This configuration, combined with nanosecond-scale timing resolution, provides sensitivity to prompt Cherenkov radiation while maintaining efficient detection of liquid argon scintillation light. p. 0
Signal Modeling and Observables
The analysis constructs four observables that exploit the Cherenkov emission and event topology expected from ALP-induced electromagnetic interactions. The signal model focuses on ALPs coupling predominantly to photons via the effective Lagrangian given in Eq. p. 2 (1). The production mechanisms considered include the Primakoff process, where an incoming photon converts into an ALP through coherent scattering off the nuclear Coulomb field. p. 2 For heavier ALPs, the dominant experimental signature is the diphoton decay a → γγ. p. 2
Event Discrimination Techniques
Four discriminating observables are constructed to exploit complementary signatures of electromagnetic interactions. These observables include:
-
Number of Hits on the Uncoated PMTs, which is used to identify Cherenkov light during a prompt time window. p. 3
-
Spread in the Directionality, calculated using the metric C, which evaluates the angular uniformity of triggered PMTs. p. 4
-
Pulse Shape Ratio (PSR), which is constructed as the ratio of charge collected during early pulse to total prompt charge. p. 5
-
Spatial Spread of Events, quantified by the chargeweighted root-mean-square (RMS) distance of PMT hits from the reconstructed interaction vertex. p. 5
Results and Conclusion
No statistically significant excess above the expected background prediction is observed for 10−3 MeV <ma < 10 MeV. p. 0 p. 0 Nevertheless, the improved background rejection enabled by hybrid Cherenkov-scintillation detection allows this analysis to surpass the sensitivity of the previous CCM120 search [3] despite less exposure and demonstrate the physics potential for hybrid Cherenkov-scintillation detectors. p. 0 The best-fit signal yields a decrease in the negative log-likelihood of 0. p. 994 relative to the backgroundonly hypothesis, with a local significance of 0. p. 986σ. No evidence for ALP production is found. p. 9 The allowed parameter space for the KSVZ QCD axion model is denoted by the dashed black lines [7, 8]. p. 0 The sensitivity of the current search is optimized for low-energy signals, reflecting the detector calibration based on the 22Na source presented in Ref. p. 3 p. 3 [2]. The improved background rejection compensates for the reduced exposure, resulting in comparable or improved ALP sensitivity across the presented mass range. p. 0 The overall signal selection efficiency as a function of ALP mass shows that efficiency reaches a maximum of roughly 25% near 6 MeV before decreasing sharply at higher masses. p. 7 The resulting 90% confidence level exclusion limits on the ALP parameter space obtained in this analysis show no statistically significant evidence for an ALP signal. No statistically significant excess above the expected background prediction is observed for 10−3 MeV <ma < 10 MeV. The final event selection reduces the prebeam sample to 0. p. 649% of its original size, with a factor of ∼6 improvement compared to CCM120. p. 6 The detector upgrades allowing use of the Cherenkov signal enhanced the steady state background rejection capabilities by approximately six-fold. p. 6 The observed event rate, integrated over the full exposure of 1. p. 623 × 1021 POT, is consistent with a timeindependent background. p. 6 The fit with a constant function yields a background rate of 11. p. 682±0.17 events/ns. p. 1 The result is consistent with a uniform background rate and shows no statistically significant residual time dependence following the event selection. p. 7 The observed energy spectrum of the selected prebeam sample shows most events cluster near 1 MeV, consistent with low-energy electromagnetic backgrounds that satisfy the Cherenkov-enhanced event selection. p. 7 Most events cluster near 5 MeV, one possible explanation is gamma-ray emission following neutron capture on argon through the reaction 40Ar(n, γ) 41Ar. p. 7 The reconstructed energy spectrum of the selected prebeam sample is presented in Fig. p. 7 8. The reconstructed energy resolution improves from approximately 12. p. 75% at 1 MeV to 7. p. 05% at 10 MeV. p. 0 The agreement between data and simulation demonstrates that the positiondependent energy calibration accurately models the detector response and provides a reliable energy estimator for the low-energy events considered in the ALP search. p. 1 The reconstructed position is then used together with the measured charge deposition to estimate the event energy. p. 5 The reconstruction achieves an approximately isotropic resolution of ∼5 cm in each coordinate, corresponding to a three-dimensional position resolution of 7. p. 19 cm at the 1σ level. p. 9 The energy reconstruction is validated using the 22Na calibration source. p. 1 The reconstructed energy peaks at approximately 1. p. 32 MeV in both data and simulation, with spectral shape differences of less than 10%. p. 3 The corresponding event start-time projection is shown in Fig. p. 9 12. The background expectation is uniform across the analysis window, whereas the signal model exhibits the characteristic triangular timing profile produced by the proton beam pulse. p. 9 The observed timing distribution shows no statistically significant preference for this signal-like structure. p. 9 The best-fit point occurs at ma = 0. p. 918 MeV and gaγ = 1. p. 963 × 10−4 GeV−1. p. 1 The profiled nuisance parameters remain close to their nominal values, yielding a POT scaling of 0. p. 999, a beam start time of-584. p. 952 ns, and a background rate of 11. p. 367 events/ns. p. 1 The observed 90% confidence level exclusion limits on the ALP parameter space obtained in this analysis show no statistically significant evidence for an ALP signal. p. 8 No statistically significant excess above the expected background prediction is observed for 10−3 MeV <ma < 10 MeV. The sensitivity of the current search is optimized for low-energy signals, reflecting the detector calibration based on the 22Na source presented in Ref. [2].
Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.
Improvements for AI systems
-
Improved signal discrimination in high-rate beam dump environments by combining four complementary observables, including
Number of Hits on the Uncoated PMTs,
Spread in the Directionality,
Pulse Shape Ratio,
andSpatial Spread of Events.
The improved AI system can perform an event selection that reduces the prebeam sample to 0.49% of its original size while enhancing sensitivity. -
Enhanced spatial resolution for event reconstruction by utilizing a GraphNeT machine learning framework with a DeepIce transformer backbone. This allows the improved AI system to achieve a three-dimensional position resolution of approximately 7.9 cm at the one sigma level, representing
roughly a four-fold improvement over the chargeweighted reconstruction previously used in CCM120 analyses.
-
Robust energy estimation for low-energy events by incorporating reconstructed position into the energy calculation using a position dependent calibration factor derived from charge collected within the first 90 ns of an event. This enables accurate estimation of deposited energy where
the detector response varies from approximately 80 PE/MeV at the center to 200 PE/MeV near the edges.
-
Optimized search parameter fitting by employing a frequentist scan over a grid of 104 ALP parameter points within the two-dimensional space of reconstructed energy and event start time. The improved AI system can perform this fit while profiling three nuisance parameters to determine the best-fit signal, yielding a
local significance of 0.86 sigma.
-
Improved background characterization through data-driven control samples by using neutron-dominated events as a validation test for the electromagnetic event selection at higher energies. This allows the AI system to use these events to provide a
data-driven cross-check of the selection
and validate signal rejection capabilities.
Sources
- First Event-by-Event Identification of Cherenkov Radiation from Sub-MeV Particles in Liquid Argon
- Measurement of the Liquid Argon Scintillation Pulse Shape Using Differentiable Simulation in the Coherent CAPTAIN-Mills Experiment
- Prospects for detecting axionlike particles at the Coherent CAPTAIN-Mills experiment
- An Improved Experimental Limit on the Electric Dipole Moment of the Neutron
- Measurement of the permanent electric dipole moment of the neutron
- The Search for Feebly-Interacting Particles
- Search for Axion Dark Matter from 1.1 to 1.3 GHz with ADMX
- A new upper limit on the axion-photon coupling with an extended CAST run with a Xe-based Micromegas detector
- Physics potential of the International Axion Observatory (IAXO)
- Dark Matter Axion Search with HAYSTAC Phase II
- Search for Dark Photons between 16.96--19.52 $\mu$eV with the HAYSTAC Experiment
- DMRadio-Core: A new approach for GUT-scale axion searches
- Search for hadronic decays of feebly-interacting particles at NA62
- Photoproduction of axion-like particles in the NA64 experiment
- Search for Axionlike and Scalar Particles with the NA64 Experiment
- Dark Sectors at the Fermilab SeaQuest Experiment
- Axion-like Particle Searches at DarkQuest
- New Constraints on Axion-Like Particles with the NEON Detector at a Nuclear Reactor
- MINER Reactor Based Search for Axion-Like Particles Using Sapphire (Al2O3) Detectors
- The Borexino detector at the Laboratori Nazionali del Gran Sasso
Related papers
- DrSai: an agentic AI system toward autonomous physics analysis workflows at BESIII
- CANTON- mu Proposal: A Next-Generation Muon g-2 Measurement at Sub-0.1 ppm Precision
- Search for dark photon dark matter near 1,meV using a superconducting mixer
- ANUBIS detector and its sensitivity to neutral long-lived particles
- How to scale your HEP ML models: A recipe for robust architecture comparisons at scale
- Search for the 16 O(ppp) to 13 C pi+ pi+ e+ Decay Mode in Super-Kamiokande Using Machine Learning Techniques