Boosted Dark Matter Directionality in Large Liquid Scintillator Detectors

Kenny C. Y. Ng; Koun Choi; Qishan Liu; Samuel S. H. Tse; Yu-Feng Li

REVIEW

Boosted Dark Matter Directionality in Large Liquid Scintillator Detectors

Samuel S. H. Tse, Qishan Liu, Kenny C. Y. Ng, Koun Choi, Yu-Feng Li

High Energy Physics - Phenomenology · Submitted 22 Oct 2026

Pith's verdict: A clean, conservative null result on neutron directionality for CRBDM in liquid scintillators, paired with a sensitivity claim whose margin over the free-proton channel is unquantified.

T0 · AI review

How this review was made: An AI referee (deepseek/deepseek-v4-flash) read the paper on October 23, 2026 and answered a fixed list of questions about it. T0 is the starting level every paper gets. The levels describe how deeply Pith has checked the paper, never the authors.

T0 AI review An AI read the paper and answered a fixed list of questions.
T1 Audit Automatic checks that the references and files the paper relies on exist and match.
T2 Compute The paper's code or calculations re-run with the same results.
T3 Formal statement The claims written out in a formal math language a computer can read.
T4 Proof Those claims proved, and the proof checked by computer.

How Pith checks papers →

· 3 major objections · 6 minor · Ask about this paper · The compiled record · Read the full review →

Challenge this review Comment (0) Read on arXiv ↗

Cite ▾

arXiv ID2610.31666 (version 2)

Pith IDYGVLS3ME, Pith's permanent record of this paper and its review

CitationKenny C. Y. Ng, Koun Choi, Qishan Liu, Samuel S. H. Tse, Yu-Feng Li. "Boosted Dark Matter Directionality in Large Liquid Scintillator Detectors." arXiv:2610.31666 (2026). https://pith.science/paper/YGVLS3ME

Cite this reviewBibTeX for Pith's review of this paper

Claim paper

Why Pith says this

Highlight any text to comment on it.

Editor's note

read the letter

Two things to know.

The durable result is negative: neutron-capture directionality in a large liquid scintillator will not materially improve CRBDM searches. The authors show this with a clean hierarchy — perfect BDM-direction reconstruction gives ~50% sensitivity gain at 95% CL, perfect neutron direction ~7%, realistic reconstruction ~3%. The knockout scattering angle kills the direction, not diffusion or vertex resolution. Both of their modeling assumptions (no BDM deflection during Earth attenuation, no FSI deflection of the neutron) inflate the directional contrast, so the null is conservative. This is a genuinely useful result for the field, and the appendix argument — even a perfect BDM direction map only buys 50% — is the right way to make it stick.

The headline sensitivity claim is the soft part. That 100 kt·yr of neutron-coincidence data probes new parameter space versus the free-proton channel of Ref. [1] rests on two stated assumptions: 100% tagging of single-neutron-capture events (Sec. III.A) and PSD/veto cuts that leave only atmospheric NCQES background (Sec. IV.A). No detector response or waveform simulation backs these. The stress-test scaling is fair: 70% tagging with NCQES-level contamination shifts the limit by roughly 1.7, 50% tagging by about 2.0. The paper never quantifies the margin between its neutron curve and the Ref. [1] free-proton curve in Fig. 7, so it is genuinely unclear whether the comparison survives realistic efficiency and background leakage. A referee should push on exactly this: state the margin, scan tagging efficiency and residual contamination.

Credit where due: the machinery is assembled from published inputs, the equations are internally consistent, the 3.6x neutron-to-free-proton target ratio checks out, and the assumptions are stated, not buried. The directionality conclusion is anchored by the perfect-direction upper bound and does not depend on the fragile sensitivity assumptions. No code or data are released, so the Geant4 retention fractions and transfer functions are not independently checkable — minor for this genre.

This paper is for CRBDM phenomenologists and LLSD experimentalists weighing neutron-tagging channels. It deserves a serious referee, with the margin and efficiency scan as the main revision points. I would send it out.

What would settle it

A calibration experiment could settle the directionality claim directly: place a neutron source with known emission energies (for example AmBe or $^{252}\text{Cf}$) in a large liquid scintillator detector, reconstruct emission and capture vertices, and compare the measured $\theta_{0.683}$ versus neutron energy with the values used here; materially smaller angles would mean the ~3% directional gain is pessimistic, and larger angles would confirm it. The sensitivity claim could be tested by measuring the real single-neutron tagging efficiency with the prompt–delayed coincidence on known inverse-$\beta$-decay events, since that channel's purported advantage over the free-proton search disappears if tagging efficiency falls well below 100%.

Referee report

Significance

If the sensitivity estimate holds, the paper establishes the neutron-coincidence channel as a competitive CRBDM search in LLSDs and provides the first quantitative demonstration that neutron directionality cannot separate this signal from atmospheric NCQES backgrounds. The directionality null result is robust: the Appendix B upper-bound hierarchy (perfect BDM direction ~50%, perfect neutron direction ~7%, realistic ~3%) anchors the conclusion in the kinematics of the knockout rather than in the detector model, and Section V explicitly flags the two assumptions that could only overestimate directional contrast. The paper is also transparent in method: it is a forward projection with no fitted parameters, it uses stated inputs (CR flux data, form factors, RFG spectral function, published GENIE background from Refs. [74,75]), and it employs machine-checked Geant4 transport with stated physics lists and Birks quenching. The principal weakness is that the headline claim of new parameter space over Ref.

Major objections (3)

[Section IV.A (sentence b129); Fig. 7] The headline claim that the neutron channel probes new parameter space relative to the free-proton channel of Ref. [1] rests on the assumption that every single-neutron-capture event is tagged as a signal candidate with 100% efficiency (sentence b129) combined with the assumption that only atmospheric NCQES remains as background. In a background-limited search the 95% CL limit scales roughly as sqrt((1+f)/epsilon) with tagging efficiency epsilon and residual contamination f; epsilon = 0.7 with f = 1 worsens the limit by ~1.7 and epsilon = 0.5 with f = 1 by ~2.0. The paper does not state the factor by which the w/o-dir neutron curve lies below the Ref. [1] curves in Fig. 7, whose upper ceilings are omitted, so the reader cannot tell whether the 'new parameter space' claim survives such a shift. Please quote the ratio of 95% CL limits at representative m_chi for both mediator masses, and add a robustness curve or table for epsilon < 1 and f > 0.
[Section IV.A (sentence b188); Fig. A.4] The background model assumes that PSD fully removes IBD and CCQE events and that muon veto plus fiducial cuts fully remove spallation, leaving only atmospheric NCQES on carbon (sentence b188), while Section III.A states that no detector configuration or detector-response modeling is performed (sentence b111). Fig. A.4 places the NCQES background at ~367.5 per year against signal counts from ~8 to ~2 x 10^4 per year over the sigma_n range shown, so even a percent-level residual contamination or an energy-dependent PSD leakage shifts the neutron-channel band by an amount comparable to the claimed margin over Ref. [1]. Since the authors state these choices explicitly, this is not a request for a full detector simulation, but the sensitivity should be presented as a function of the residual contamination fraction, or the tolerable contamination that the claimed margin can withstand should be stated.
[Section II.A (Eqs. (4)-(5), sentence b48)] The BDM flux input is assembled by splicing two CR data sets at ad hoc kinetic energies (1.5 GeV for protons, 0.9 GeV for helium) because the two curves do not align smoothly (sentence b48), with low-energy cutoffs at 22 MeV and 10 MeV taken from Ref. [1]. The absolute signal rate is sensitive to the low-energy part of the CR spectrum, so a short robustness check, for example varying the splice energies by +/-20% and recomputing the w/o-dir sensitivity, would strengthen the absolute sensitivity claim. This does not affect the directionality null result, which is the paper's more robust conclusion.

Minor objections (6)

[Section II.A, Eq. (5)] The plus/minus sign branch in Eq. (5) deserves one sentence of explanation, since the branch choice for K_chi > 2 m_i versus K_chi < 2 m_i is easy for a reader to misapply.
[Section III.C] The mapping of the 68.3% quantile angle theta_0.683 onto the Gaussian smearing sigma is a heuristic; a sentence noting that this is an approximation sufficient for the paper's purpose, given the Appendix B upper bound, would be helpful.
[Section IV.B, Eqs. (23)-(25)] Eq. (25) is evaluated with k_i fixed to the background expectation; the paper should state explicitly that this is an Asimov or median sensitivity and that TS = 2.71 assumes the asymptotic chi-squared distribution of Ref. [78].
[Fig. 5 and Fig. A.4] The Galactic sky maps would be clearer with a logarithmic color scale and with explicit (b, l) axes; the outlier ratio ~6 x 10^5 in the top-right panel of Fig. 5 is created by a nearly vanishing pixel and deserves a parenthetical explanation at the figure caption.
[Abstract and Section VI] The phrase 'probe new parameter space compared to free proton signals' could be softened to reflect the stated assumptions, for example 'under the 100% tagging-efficiency and NCQES-only background assumptions of this work', so that the abstract does not overstate the robustness of the comparison to Ref. [1].
[Section VII and Eq. (1)] Minor textual issues: 'GRC grants' in the acknowledgments appears to be a typo for 'RGC grants', and the spin-independent cross section sigma_n is defined after Eq. (1) and again in Section IV.B; the notation should be unified at first use.

How Pith checks papers → · Full audit of this paper →

The paper

The paper claims

Cosmic-ray-boosted dark matter that knocks a neutron out of carbon leaves a taggable delayed gamma ray in liquid scintillator detectors — a channel that beats the proton-based search, though the displaced capture vertex adds only about 3%.

The reading

Cosmic rays can slam into dark matter in the Galactic halo and boost light dark-matter particles to relativistic energies, and this paper considers what happens when those boosted particles hit the carbon in a large liquid scintillator detector: the dark matter knocks a neutron out of the nucleus, and the neutron's delayed capture gamma ray, arriving about 210 μs later, tags the event. The paper argues that with 100 kt·yr of exposure this neutron-coincidence channel probes dark-matter–nucleon cross sections that the free-proton channel of earlier work cannot reach, and that it roughly matches the projected sensitivity of a next-generation water-Cherenkov detector at a light mediator mass. It then asks a second question: does the vector from the neutron's emission point to its capture point preserve enough of the Galactic-center direction to separate signal from background? The answer is that the quasi-elastic knockout itself spreads the neutron over a large scattering angle, so the directional information improves the 95% confidence limit by only about 3% at the benchmark masses. The practical upshot is that liquid scintillator detectors should run the neutron channel for boosted dark matter, but neutron directionality is not the tool that will sharpen it.

What carries the argument

The load-bearing object is the displaced neutron-capture vertex. A boosted dark-matter particle scatters quasi-elastically off a bound neutron in $^{12}\text{C}$ (restricted to momentum transfers $|q| > 350$ MeV, with the bound-neutron kinematics described by a relativistic Fermi-gas spectral function, Pauli blocking, and a final-state-interaction optical potential), the knocked-out neutron propagates through the liquid scintillator, and it is captured about 210 μs later, predominantly on hydrogen, with a 2.2 MeV gamma ray. Reconstructing the initial vertex from the earliest secondary scintillation light (a 2 ns time cut on the prompt signal) and the capture vertex from the delayed gamma ray yields a displacement vector; the angular smearing of that vector relative to the true neutron emission direction is quantified by the 68.3% quantile angle $\theta_{0.683}$ in each neutron-energy bin, which is then used to smooth the 768-pixel signal sky map. The sensitivity estimate couples that map to a Monte Carlo transport simulation that supplies the prompt-energy transfer function and the single-neutron-capture retention fraction, and applies a Poisson log-likelihood ratio over 768 direction bins times 15 energy bins, with atmospheric-neutrino neutral-current quasi-elastic scattering on carbon as the only indistinguishable background.

Extended reading notes

Core claim

The central claim has two parts. First, quasi-elastic scattering of cosmic-ray-boosted dark matter off bound neutrons in $^{12}\text{C}$ produces a clean, taggable signal in a liquid scintillator detector — a prompt energy deposit from the recoil and secondary particles, followed roughly 210 μs later by the characteristic 2.2 MeV gamma ray of neutron capture on hydrogen — and the paper estimates that 100 kt·yr of such tagged events would constrain the spin-independent dark-matter–nucleon cross section $\bar\sigma_n$ more strongly than the free-proton channel of the same exposure considered in Ref. [1]. Second, the displacement vector between the reconstructed emission vertex and the capture vertex preserves only partial directional information about the incoming dark matter, and incorporating it improves the 95% CL limit by only about 3% (for $m_\chi = 1$ MeV, $m_\phi = 3$ GeV). The paper shows this is not a detector-resolution failure: even with the true neutron emission direction known perfectly, the improvement would be only about 7%, and with the true dark-matter direction known perfectly, about 50%. The conclusion is that the smearing comes from the nucleon scattering angle itself, so it is intrinsic to any knocked-out nucleon signal, protons included.

Load-bearing premise

The central estimate assumes every single-neutron-capture event can be tagged with 100% efficiency and that muon spallation, inverse beta decay, and charged-current backgrounds are all removed by the veto and pulse-shape arguments, so that only atmospheric-neutrino neutral-current scattering remains; the directionality result is safer, since it is driven by the modeled knockout scattering angle rather than by tagging or background rates.

Editorial extensions

If this is right

With 100 kt·yr of exposure, the neutron-knockout channel yields stronger 95% CL limits on $\bar\sigma_n$ than the free-proton channel of the same detector in Ref. [1], opening parameter space that proton-based searches did not cover.
At mediator mass $m_\phi = 1$ GeV the neutron channel reaches roughly the sensitivity projected for a next-generation water-Cherenkov detector with five years of data, while the long-baseline neutrino detector projection compared in Ref. [1] remains the strongest.
Directional analysis improves the limit by only ~3% at the $m_\chi = 1$ MeV, $m_\phi = 3$ GeV benchmark, and even perfect reconstruction of the neutron's initial direction would give only ~7%, so the small gain is intrinsic to the knockout kinematics rather than to vertex resolution.
Because the directional improvement costs nothing extra — it uses the low-energy vertex reconstruction already available in such detectors — the authors recommend that liquid scintillator experiments include it despite the modest gain.
For a flux more sharply concentrated at the Galactic center, such as dark-matter annihilation into boosted particles or neutrinos, whose line-of-sight integral scales as $\rho^2$, the directional gain is expected to be larger: about 20% on the cross-section limit for a fixed annihilation rate in the case the authors examine.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

The paper's diagnosis that the knockout scattering angle — not detector resolution — destroys directionality implies that proton-directionality schemes for the same signal (for example first-photon timing techniques) would face the same intrinsic ceiling, since any knocked-out nucleon inherits the same angular spread.
The ~3% figure is computed at benchmark masses; the structural argument that the scattering-angle spread is comparable to or larger than the Galactic-center angular signal suggests the conclusion persists across the model range, but the authors do not map the directional gain over the full ($m_\chi$, $m_\phi$) plane, so a dedicated scan would be a natural check.
One testable extension suggested by the analysis: a detector located where Earth attenuation displaces the brightest sky patch away from the Galactic center (an effect the paper notes) could in principle discriminate attenuation models by comparing the predicted and observed direction of the signal peak in the neutron map.
The 2 ns early-light time cut developed here to pin down the neutron emission vertex is a general-purpose tool: any liquid scintillator measurement that needs a precise prompt vertex, such as supernova-neutrino pointing, could reuse it independently of dark matter.

Original abstract

Light dark matter can be boosted by cosmic-ray-dark matter interactions, which opens the possibility of direct-detection searches with large neutrino detectors. We consider the signal produced by boosted dark matter undergoing quasi-elastic scattering off carbon and knocking out a neutron in a large liquid scintillator detector. A distinctive delayed gamma ray from neutron capture can be reconstructed after $\sim 210$ $\mu$s, which is paired up with the prompt signal during neutron propagation. We find that 100 kt$\cdot$yr of data from the neutron-coincidence channel will be able to probe new parameter space compared to free proton signals found in previous work. We further consider the possibility of utilizing directional information from the displaced vertex associated with the capture of the final-state neutron. However, we find that directional information does not materially improve the sensitivity due to the large nucleon scattering angle.

The compiled record

The paper's own sentences, arranged, kept beside the review and separate from it. Each label is drawn from a closed vocabulary; the number beside it is the probability the classifier gave that label. Numbers in the tables are the authors', checked by code. No sentence in this record was written by a model.

Compiled by classifier-1.13.0 via TypeSafe API, question bank abstract-p1-20260923, 2026-10-23: 1 request, 4,929 input tokens.

What the abstract itself states, as probabilities: claims first 0.05 · states limitation 0.95 · human participants 0.01 · controlled comparison 0.19 · randomized assignment 0.01 · states quantitative result 0.93 · compares with prior results 0.94 · mentions code or data availability 0.02.

The abstract, sentence by sentence

arXiv listing abstract v1. Results and the main claim are underlined; show every role.

Light dark matter can be boosted by cosmic-ray-dark matter interactions, which opens the possibility of direct-detection searches with large neutrino detectors. We consider the signal produced by boosted dark matter undergoing quasi-elastic scattering off carbon and knocking out a neutron in a large liquid scintillator detector. A distinctive delayed gamma ray from neutron capture can be reconstructed after $\sim 210$ $\mu$s, which is paired up with the prompt signal during neutron propagation. We find that 100 kt$\cdot$yr of data from the neutron-coincidence channel will be able to probe new parameter space compared to free proton signals found in previous work. We further consider the possibility of utilizing directional information from the displaced vertex associated with the capture of the final-state neutron. However, we find that directional information does not materially improve the sensitivity due to the large nucleon scattering angle.

background contribution method result interpretation main claim

What the classifier reports for each sentence. The number is the probability it gave the label.
#	sentence	role	the authors' own work	states the main conclusion	strength
s0	Light dark matter can be boosted by cosmic-ray-dark matter interactions, which opens the possibility of direct-detection searches with large neutrino detectors.	background 1.00	0.06	0.05	hedged possibility 0.53
s1	We consider the signal produced by boosted dark matter undergoing quasi-elastic scattering off carbon and knocking out a neutron in a large liquid scintillator detector.	contribution 0.91	0.50	0.04	asserted 0.79
s2	A distinctive delayed gamma ray from neutron capture can be reconstructed after $\sim 210$ $\mu$s, which is paired up with the prompt signal during neutron propagation.	method 0.75	0.48	0.05	asserted 0.99
s3	We find that 100 kt$\cdot$yr of data from the neutron-coincidence channel will be able to probe new parameter space compared to free proton signals found in previous work.	result 0.94	0.92	0.39	asserted 0.93
s4	We further consider the possibility of utilizing directional information from the displaced vertex associated with the capture of the final-state neutron.	contribution 0.91	0.46	0.05	hedged possibility 0.90
s5	However, we find that directional information does not materially improve the sensitivity due to the large nucleon scattering angle.	result 0.91	0.96	0.34	asserted 1.00

What kind of paper the classifier reports this is: simulation study 0.49 · evidence from simulation or computation 0.87.

What was examined, and what was not

sources read

arXiv listing abstract v1

not examined

full text
figures and tables
references and citing sentences
relations to other papers

levels run

P abstract

not run

S sections
T body sentences
C citations
R cross-paper relations

Every quoted sentence on this record is the authors'. Every label is from Pith's closed vocabulary, with the probability shown beside it. Every number in the measurements table is the authors', checked by code. Record as JSON

Figures

**Figure 1.** FIG. 1. The BDM flux spectrum at different

**Figure 2.** FIG. 2. The detector’s orbit due to Earth’s rotation (dotted ring in purple) affects the depth

**Figure 3.** FIG. 3. The retention percentage of single-neutron-capture

**Figure 4.** FIG. 4. The 68.3% quantile angle as a function of the

**Figure 5.** FIG. 5. The angular distribution of signal in Galactic coordinates. As an example, we take

**Figure 6.** FIG. 6. Signal (left) and background (right) counts in LLSD for an exposure of 100 kt

**Figure 7.** FIG. 7. The estimated 95% CL sensitivity limits on DM-nucleon scattering cross section for a 100 kt

Figures reproduced from arXiv:2610.31666 by the authors.

References

Discussion

Disagree with the review, or know something it missed? Leave a comment, highlight any passage above to comment on it, or challenge this review formally.