Pith. sign in

REVIEW 3 major objections 5 minor 22 references

Very Late Afterpulses and Search for the Neutron Echo in IceCube

T0 review · 3 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read IceCube's delayed-light background has a fitted lifetime of about 194 microseconds, nearly matching the 217-microsecond neutron echo it would mask, so any Neutron Echo search must model this background.

desk verdict A useful, honest calibration study whose main result—a 194 µs late background component—is likely real but may be an LED property, so the neutron-echo overlap claim needs a light-source-free check before it carries weight. read the letter →

arxiv 2507.07042 v1 pith:5SIMLAR5 submitted 2025-07-09 astro-ph.HE astro-ph.IMhep-ex

Kaustav Dutta , Sebastian Böser , Martin Rongen (for the IceCube Collaboration) This is my paper · ORCID
classification astro-ph.HEastro-ph.IMhep-ex
keywords neutronechoafterpulsesIceCubedelayedphotonbackgroundPMTluminescenceneutrinoflavordiscriminationdeadtime-freereadout
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper establishes that the main delayed background in IceCube's optical modules (DOMs) consists of late photomultiplier afterpulses plus luminescence from glass and LED materials, and that this background decays with a long exponential lifetime of about 194 microseconds. That is close to the 217-microsecond lifetime expected for the neutron-capture echo, so a future Neutron Echo search cannot rely on timing alone to separate the physics signal from the background. The authors use a newly available deadtime-free readout to record pulses out to a millisecond that the standard IceCube acquisition would miss, and they fit the delayed spectrum as two exponentials plus a skewed Gaussian. If the result transfers to real neutrino events, the promised flavor and interaction-channel discrimination from neutron yields will depend on separating two templates that are almost degenerate in time.

What carries the argument

The carrying object is the delayed-photon time spectrum in a single IceCube optical module, recorded in a deadtime-free fADC-only readout that skips the ATWD digitizers and removes the roughly 30 microsecond per-channel deadtime of the standard acquisition. The background is modeled as $f(t) = A_1 e^{-t/\tau_1} + A_2 e^{-t/\tau_2} + A_g\,\mathrm{SG}(t;\mu,\sigma,\alpha)$, where $\tau_2$ is the late exponential lifetime that comes out near 194 microseconds and $\mathrm{SG}$ is a skewed Gaussian with mean near 68 microseconds; this model is what turns raw flasher waveforms into the claim that background timing overlaps the neutron echo. The signal it threatens is the neutron echo itself: Cherenkov light from Compton electrons produced by the 2.2 MeV gamma emitted when thermalized neutrons are captured on hydrogen in the ice, with a characteristic capture lifetime near 217 microseconds.

What would settle it

A decisive check is a laboratory measurement of a bare IceCube-type photomultiplier with no glass housing and no LED: if the ~194 microsecond exponential tail persists under fast laser pulses, it is intrinsic PMT behavior; if it vanishes, the tail belongs to the housing or light source and the flasher-based background model must be rebuilt for real neutrino events.

Watch

Extended reading notes

Core claim

The central claim is that the late-pulse background in IceCube DOMs is not a negligible tail but a structured, prompt-correlated component that lands directly on the neutron echo window. In flasher data from in-situ DOMs, the delayed-light intensity scales with the prompt signal, and the time spectrum is well described by an early exponential tail, a late exponential tail, and a skewed Gaussian centered near 68 microseconds. The brightness- and wavelength-averaged late lifetime is about 194 microseconds, which the paper states closely matches the expected Neutron Echo lifetime of about 217 microseconds. The authors conclude that the signal and background templates are hard to discriminate, and that any future search must account for this component.

Load-bearing premise

The flasher runs on a dozen in-situ modules are assumed to reproduce the delayed light that will accompany real neutrino-induced showers, even though the calibration LEDs differ in spectrum, brightness, and geometry from Cherenkov light.

Editorial extensions

If this is right

  • Any future Neutron Echo search in IceCube must include a delayed-background template with a late exponential lifetime near 194 microseconds, otherwise background will be misattributed to neutron captures.
  • The deadtime-free fADC-only readout is sufficient to record the full delayed spectrum out to 1 millisecond, making a background-subtracted echo search experimentally possible.
  • The skewed-Gaussian component centered near 68 microseconds is stable across brightness and wavelength, indicating a PMT-related origin that can be characterized once and then subtracted.
  • The wavelength dependence of the late lifetime, with longer lifetimes at 370 nm, supports a luminescence origin and implies the background is weaker for distant modules because ultraviolet Cherenkov light attenuates before reaching them.
  • If the neutron echo can be separated from this background, the neutron yield difference between shower types would improve neutrino flavor and interaction-channel classification over current cascade-only methods.

Reading between the lines

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

  • If the flasher background transfers to real showers, the practical limit of a neutron echo analysis may be template separation rather than statistics; the 23-microsecond gap between 194 and 217 microseconds is small enough that spatial or spectral information will be needed, not just timing.
  • One testable extension is to check whether the per-event delayed spectrum in high-energy starting events scales with prompt charge; if it does, the background dominates, and any residual excess decaying with the 217-microsecond timescale would be the first evidence of the echo.
  • The 194-microsecond average is taken over LED wavelengths and brightness settings; real Cherenkov light is broadband and modules see a mix of source distances, so a physics-event-weighted background lifetime must be measured before the overlap claim can be made quantitative.
  • A dedicated comparison of standard and high-quantum-efficiency DOMs under identical flashes could separate photocathode effects from glass luminescence, refining which component actually needs mitigation.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. This proceedings paper characterizes very late (30–1000 µs) delayed light signals recorded by IceCube DOMs using a newly developed deadtime-free fADC readout. LED flasher runs at several brightnesses and wavelengths are fit with a sum of two exponentials and a skewed Gaussian, and the authors report a brightness- and wavelength-averaged late exponential lifetime of about 194 µs, which is close to the ~217 µs neutron capture lifetime in ice. A PMT-associated Gaussian component near 68 µs is also identified. The paper argues that these delayed backgrounds will complicate a future neutron echo search and that the new DAQ mode enables the required measurements.

Significance. If the 194 µs component is shown to accompany neutrino-induced events, this is an important experimental result: it quantifies a background that temporally overlaps the neutron echo signal and must be modeled in any future flavor-composition or hadronic-content analysis. The development of the deadtime-free DAQ mode and the systematic in-situ flasher study across brightnesses and wavelengths are valuable contributions. The paper is also honest in attributing the exponential component to 'luminescence originating from the LED, glass, or both' and in framing the neutron echo search as ongoing work. The main weaknesses are the lack of a quantitative comparison between the fitted and expected lifetimes and the unproven transferability of the flasher-derived background to physics events.

major comments (3)
  1. [Section 3.3, Eq. (1)] The central quantitative claim that the brightness- and wavelength-averaged τ2 of 194 µs 'approaches' the neutron echo lifetime of ~217 µs is not supported by a statistical comparison. No uncertainty is given for the weighted mean τ̄2, no goodness-of-fit values are reported for the fits, and no test is shown for the difference between 194 µs and 217 µs. Given the DOM-to-DOM and wavelength scatter visible in Fig. 7, the reader cannot assess whether the two lifetimes are statistically compatible. Please report the uncertainty on τ̄2, the fit quality, and preferably a confidence interval or p-value for the difference.
  2. [Section 3.3 and Section 4] The late exponential is extracted from LED flasher data, and Section 4 states it is 'consistent with luminescence originating from the LED, glass, or both.' If the late tail is dominated by LED afterglow, the fitted 194 µs lifetime is a property of the calibration light source rather than of the DOM response to Cherenkov light, and the conclusion that this background will overlap the neutron echo in neutrino events does not follow. The HESE stacked distribution in Fig. 3 shows that some delayed artifact exists in real events, but the same functional form is not fitted to those data, so the 194 µs lifetime is not established for physics events. The paper should either fit the model to the HESE distribution or otherwise demonstrate that the flasher-derived lifetime applies to Cherenkov-induced signals, or explicitly restrict the claim to the calibration setup.
  3. [Section 3.3] The fit model f(t) has eight free parameters (A1, τ1, A2, τ2, Ag, μ, σ, α) and is fit over the interval 30–1000 µs, but the paper does not report the full best-fit results or the covariance/correlations among parameters. Since τ2 is the central quantity, correlated or poorly constrained fits could materially change the conclusion. Please provide the best-fit values and uncertainties for all parameters, at least for the representative DOM in Fig. 6, and justify the choice of the 30 µs lower fit bound, which excludes the early tail region where parameter correlations may be strongest.
minor comments (5)
  1. [Section 4] The text says the delayed background spectrum 'between 30 and 100 µs' was observed to scale with prompt signal intensity, but the fits in Section 3.3 are performed over 30–1000 µs; the range should be corrected.
  2. [Abstract and full text] There is a missing space in 'easily identifiedbutthedetector' in the abstract and in the full-text version; please fix the formatting.
  3. [Section 3.3] The fit interval is stated as '20 µs–1000 µs' in one sentence and '30 and 1000 µs' in the next; please reconcile these numbers.
  4. [Figure 7] The weighted means are shown but the definition of the weights is not given; please state how the DOM-specific uncertainties are combined.
  5. [Figure 3] The HESE stacked distribution is shown without statistical uncertainties or error bars; adding them would help the reader judge the significance of the delayed excess.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the 194 µs fitted background lifetime and the 217 µs neutron echo expectation are independent inputs.

full rationale

The central comparison is not a derivation. In Sec. 3.3, the background model f(t)=A1 exp(-t/tau1)+A2 exp(-t/tau2)+Ag SG(t;...) is fitted to in-situ LED flasher waveforms, producing a brightness- and wavelength-averaged tau2 of 194 us. The neutron echo lifetime of about 217 us is taken from Ref. [13], a Geant4 simulation of delayed photons in IceCube with stated assumptions, and is not an output of this fit. Equation (1) contains no neutron echo term, so tau2 cannot equal 217 us by construction; the reported closeness is empirical. No neutron echo detection or exclusion is claimed, and the background model is not used to predict the neutron echo. Self-citation of a Mainz thesis for the simulated echo expectation is not load-bearing because the simulation assumptions do not include the flasher fit and the cited result is independently falsifiable. The skeptic's concern that LED afterglow rather than DOM/PMT response may drive tau2 is an external validity or correctness issue, not circular reasoning, and the paper itself flags LED or glass luminescence as a possible origin.

Assumptions & free parameters 8 free parameters · 4 assumptions · 0 invented entities

The analysis relies on eight fitted empirical parameters describing the delayed background, plus external assumptions about the neutron capture time, the representativeness of LED flasher data, the chosen functional form, and the applicability of prior PMT and glass luminescence studies. No new physical entities are introduced.

free parameters (8)
  • A1, early exponential tail amplitude = varies with DOM, brightness, and wavelength
    Amplitude of the fast-decaying component in the background model f(t) fitted to waveforms between 30 and 1000 microseconds.
  • tau1, early exponential tail lifetime = order of magnitude shorter than tau2, not quoted globally
    Time constant of the early tail, reported to be about an order of magnitude shorter than the late tail at 370 nm.
  • A2, late exponential tail amplitude = varies with DOM, brightness, and wavelength
    Amplitude of the long-lived background component in the empirical fit.
  • tau2, late exponential tail lifetime = 194 microseconds, weighted mean across brightness and wavelength
    Key fitted parameter that the paper compares with the neutron echo lifetime of about 217 microseconds.
  • Ag, skewed Gaussian amplitude = varies
    Amplitude of the PMT-associated Gaussian component in the background model.
  • mu, skewed Gaussian mean position = about 68 microseconds
    Gaussian mean reported to be stable, near 68 microseconds, and independent of brightness and wavelength.
  • sigma, skewed Gaussian width = about 46 microseconds
    Gaussian width fluctuating around 46 microseconds across settings.
  • alpha, skewed Gaussian skewness = not reported in text
    Skewness parameter of the skewed Gaussian component, included in the model but not quoted with a fitted value.
assumptions (4)
  • domain assumption Neutron capture lifetime in ice is approximately 217 microseconds
    Used as the expected neutron echo timing from Ref. [13], a PhD thesis, without independent verification; introduced in Section 2.
  • domain assumption LED flasher measurements represent the delayed background in actual neutrino events
    The background model is built from flasher data on 5 to 12 DOMs and assumed to transfer to neutrino-induced showers; enters in Section 3.3.
  • ad hoc to paper The two-exponential plus skewed-Gaussian functional form adequately describes the background
    The paper states 'we model the full distribution as the sum of two exponentials and a skewed Gaussian function' without a goodness-of-fit test or comparison to alternative models.
  • domain assumption Glass luminescence spectra and PMT afterpulse properties from prior publications apply to the in-situ DOMs
    References [18], [19], [20], and [21] are used to interpret the measured delayed components; these external measurements are assumed relevant to the tested IceCube DOMs.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Very Late Afterpulses and Search for the Neutron Echo in IceCube." pith.science (2026). https://pith.science/paper/5SIMLAR5

@misc{pith2026250707042,
  author       = {Pith},
  title        = {Pith review of: Very Late Afterpulses and Search for the Neutron Echo in IceCube},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/5SIMLAR5}},
  note         = {Machine review of arXiv:2507.07042}
}
abstract

While high-energy astrophysical neutrinos are well-established, their flavor composition remains relatively unconstrained. In IceCube, long muon tracks from $\nu_\mu$ charged-current interactions are easily identified but the detector geometry does not allow sufficient resolution to distinguish cascade-type events. The Neutron Echo - a delayed light signal in the detector from neutron capture and de-excitation - can probe the shower's hadronic content and thus the underlying interaction. A significant background arises from late PMT afterpulses, which are temporally coincident with the physics signal. The traditional IceCube data acquisition system has a limited readout window with significant deadtime between triggers, which is insufficient to capture the late pulses. A recently developed deadtime-free readout mode, with an extended window, enables their detection. An observed excess in the delayed time spectrum over the background would be compatible with the Neutron Echo hypothesis. In this contribution, we summarize the physics scope of delayed signals, discuss the timing spectrum of signal and PMT background, highlight the capabilities of the new system for recording late pulses, and emphasize the potential of IceCube for particle identification through delayed signals.

Figures

Figures reproduced from arXiv: 2507.07042 by the authors.

Figure 1
Figure 1. Left: Timing spectrum of delayed photons incident on modules within 15–1000 𝜇s from Geant4 simulations of IceCube NC events, before applying detector effects such as deadtime; Right: Delayed photon counts from Geant4 simulations for shower-like events between 100–150 TeV. Both taken from [13]. The relative intensity of the neutron echo scales with the neutron abundance, which is highest in hadronic showers from NC e… view at source ↗
Figure 2
Figure 2. DOM launch rates averaged over 500 FBLED flashes at a brightness setting b25. Standard DAQ with ATWD-induced deadtime is shown in blue; deadtime-free mode in orange. Data acquisition by the fADC and ATWDs is synchronized. The fADC digitizes continuously over a 6.4 µs window, while the ATWD captures and sequentially digitizes each gain channel re￾quiring ∼30 µs per channel [14]. During digiti￾zation, the ATWD chip is… view at source ↗
Figure 3
Figure 3. Stacked time distribution of delayed pho￾tons from 1250 high energy starting events (HESE) [17] events recorded between 2015 and 2019, along￾side flasher run data from two in-situ DOMs. Independent measurements in 8-inch Hama￾matsu and Thorn PMTs also found unexpected pulses roughly 100 µs after the prompt pulse [18]. The excess signal was attributed to PMT-intrinsic effects since they were measured in bare PMTs, al… view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Emission spectrum measured through Ice￾Cube Benthos glass, corrected for detector response and glass transmission. The 265 nm excitation light is fully absorbed by the glass, and the detected signal arises from delayed re-emission within the glass. Certain other proces…
Figure 5
Figure 5. Figure 5: Left: DOM launch distributions with the deadtime-free DAQ mode from 505 nm LED emission at brightness settings b10, b25, and b35, where the brightness intensity increases toward b35. Right: DOM launch distributions for LEDs at 340 nm, 370 nm, 450 nm, and 505 nm, all at…
Figure 6
Figure 6. Figure 6: Average charge distribution from 5000 waveforms recorded for a module illuminated by a 405 nm flasher, with fit components and parameters. Each waveform corresponds to a single LED flash. The DOM launch distributions with the deadtime-free mode show that delayed pulse …
Figure 7
Figure 7. Figure 7: Fitted parameters with uncertainties for five in-situ DOMs are shown versus signal intensity and LED wavelength. Weighted means 𝜏¯1𝑤, 𝜏¯2𝑤, 𝜇ˆ𝑤, and 𝜎ˆ 𝑤 are shown as red dashed lines. Although a fully coherent interpretation is challenging, several trends are still ev…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

22 extracted references · 22 canonical work pages

  1. [1]

    C. L. Cowan and et al.,Science124 (1956) 103–104

    work page 1956

  2. [2]

    4, (2009) 320–328

    Super-Kamiokande Collaboration,Astroparticle Physics 31no. 4, (2009) 320–328

    work page 2009

  3. [3]

    Daya Bay Collaboration,Phys. Rev. Lett. 108no. 17, (2012) 171803

    work page 2012

  4. [4]

    Double Chooz Collaboration,Phys. Rev. Lett. 108 no. 13, (2012) 131801

    work page 2012

  5. [5]

    Super-Kamiokande Collaboration,Astropart. Phys. 60(2015) 41–46

    work page 2015

  6. [6]

    IceCube Collaboration,Phys. Rev. Lett. 114 no. 17, (2015) 171102

    work page 2015

  7. [7]

    IceCube Collaboration,Phys. Rev. Lett. 132 (Apr, 2024) 151001

    work page 2024

  8. [8]

    S. W. Li, M. Bustamante, and J. F. Beacom,Phys. Rev. Lett. 122 no. 15, (2019) 151101

    work page 2019

Show all 22 references
  1. [9]

    6160, (2013) 1242856

    IceCube Collaboration,Science 342 no. 6160, (2013) 1242856

    2013

  2. [10]

    IceCube Collaboration,Phys. Rev. D 99 no. 3, (2019) 032004

    2019

  3. [11]

    IceCube Collaboration,Phys. Rev. Lett. 125 (Sep, 2020) 121104

    2020

  4. [12]

    J. Kopp, J. Liu, and X.-P. Wang,J. High Energy Phys. 2015 no. 04, (2015) 105

    2015

  5. [13]

    A. Steuer. PhD thesis, Johannes Gutenberg-Universität Mainz, 2018

    2018

  6. [14]

    03, (2017) P03012

    IceCube Collaboration,JINST 12no. 03, (2017) P03012

    2017

  7. [15]

    Nachweis verzögerter Neutronen in IceCube

    A. Meinusch, “Nachweis verzögerter Neutronen in IceCube.” Bachelorarbeit, Johannes Gutenberg-Universität Mainz, 2017

    2017

  8. [16]

    IceCube Collaboration,JINST 5 (2010) P03009

    2010

  9. [17]

    ICRC 2019

    IceCube Collaboration,PoS(ICRC2019)358(2019) 1004. ICRC 2019

    2019

  10. [18]

    R. V. Poleshchuk and et al.,Nucl. Instrum. Methods Phys. Res. A 695 (2012) 362–364

    2012

  11. [19]

    Method and apparatus for detecting faults in periodic structures,

    B. Schlesinger, “Method and apparatus for detecting faults in periodic structures,” September, 1974

    1974

  12. [20]

    V. A. Morozov, N. V. Morozova, and P. Budzyński,Nucl. Instrum. Methods Phys. Res. A 1053 (Aug,

  13. [21]

    D. Jin, R. Connally, and J. Piper,Journal of Physics D: Applied Physics 39no. 3, (Jan, 2006) 461

    2006

  14. [22]

    8 Very Late Afterpulses and Search for the Neutron Echo in IceCube Full Author List: IceCube Collaboration R

    IceCube Collaboration,Astroparticle Physics 35(2012) 615–624. 8 Very Late Afterpulses and Search for the Neutron Echo in IceCube Full Author List: IceCube Collaboration R. Abbasi16, M. Ackermann63, J. Adams17, S. K. Agarwalla39, a, J. A. Aguilar10, M. Ahlers21, J.M. Alameddine...

    2012

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.