Pith. sign in

REVIEW 3 major objections 5 minor 20 references

A semi-analytical formula predicts the timing resolution of plastic scintillator detectors with WLS-fiber and SiPM readout, passing sub-percent toy Monte Carlo validation and matching full optical-photon simulations in trend.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

A semi-analytical model predicts timing resolution of scintillator+WLS-fiber+SiPM detectors from component parameters, validated at sub-percent level by toy MC and within ~20% by Geant4.

T0 review reviewed 2026-08-02 challenge →

load-bearing objection Solid engineering model with a real internal validation, but the sub-10% predictive claim at N_pe=50–200 is an extrapolation — worth a careful referee, not a desk reject. the 3 major comments →

arxiv 2607.14290 v1 pith:CUBTLQUR submitted 2026-07-15 physics.ins-det hep-ex

Comprehensive study of timing resolution in plastic scintillator detectors with wavelength-shifting fiber and silicon photomultiplier readout

Haohui Che , Guang Yang This is my paper
classification physics.ins-det hep-ex PACS 29.40.Mc
keywords timing resolutionplastic scintillatorWLS fibersilicon photomultiplier (SiPM)order statisticsleading-edge timingdetector optimizationphotoelectron yield
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

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 sets out to show that the timing resolution of a plastic-scintillator detector read out through a wavelength-shifting fiber and a silicon photomultiplier can be computed, not just measured, from a short list of component parameters. Its central move is to treat the event timestamp as the k-th photoelectron arrival time: the variance of that order statistic, computed from the folded single-photon detection-time PDF, is added in quadrature to fiber transit spread and electronics quantization. The formula is validated against a toy Monte Carlo at 80 points (ratio 0.9997 ± 0.0015) and against full optical-photon simulations that reproduce the WLS, position, boundary, and topology trends at the 0.84–0.96 level for N_pe≈14. A sympathetic reader would care because the result turns component specifications into design maps: it identifies WLS re-emission time, not SiPM speed, as the dominant timing lever, and it gives detector builders a direct path from measured or simulated photoelectron yield to expected timing resolution.

Core claim

The paper's central discovery is captured in Eq. (2.12): the leading-edge timing resolution is σ_t² = Var[T_(k:N)] + σ_transit² + σ_elec². Var[T_(k:N)] is the order-statistic variance of N independent samples from f_det, the convolution of the bi-exponential scintillator emission (Exp(τ_r) ⊗ Exp(τ_d)), the single-exponential WLS re-emission (Exp(τ_WLS)), and the Gaussian SiPM single-photon time resolution. The authors demonstrate that this order-statistic calculation accounts for the sub-percent agreement with toy Monte Carlo and for the main behaviors seen in full optical simulation—WLS-time scaling, position dependence, mirror neutrality on long bars, and the √2 double-ended gain. They als

What carries the argument

The machine that carries the argument is the constructed single-photon detection-time PDF f_det(t) = f_scint(t) ⊗ f_WLS(t) ⊗ G(0, σ_SPTR), with f_scint written as Exp(τ_r) ⊗ Exp(τ_d). Because the paper uses exponential building blocks, f_scint+WLS has the closed-form triple-exponential expression of Eq. (2.5), so the full f_det is obtained by one numerical Gaussian convolution on a fine grid. The other central object is the k-th order statistic: Eq. (2.9) produces the PDF of the k-th photoelectron time from N independent draws, and Eq. (2.8) integrates it for the variance. To this photon-statistics variance the model adds the uniform-mode-angle transit-time spread x·NA²/(2√12·c·n_core) and t

Load-bearing premise

The calculation collapses if a single WLS re-emission exponential with a fixed τ_WLS, added in quadrature with the uniform-mode-angle transit spread, does not adequately describe the true single-photon detection-time distribution at the photoelectron yields of interest (about 50–200).

What would settle it

On a short bar with one fiber and a single SiPM, record single-photon-resolved waveforms with a fast digitizer, measure first-photon arrival jitter versus N_pe over 5–200 photoelectrons at several injection positions, and compare the slope and intercept with the folded-order-statistic prediction. A deviation of more than about 10% at N_pe≈50—or a measured exponent outside 0.44–0.49—would falsify the central single-exponential-WLS/independent-sample assumption.

Watch this falsifier. Get emailed when new claim-graph text bears on it.

If this is right

  • A detector builder who knows N_pe for a geometry can read the expected timing resolution directly from the paper's tables and maps, without running a new full simulation for each configuration.
  • Because the fitted scaling is σ_t ≈ A·N_pe^(−0.44 to −0.49), doubling the photoelectron yield improves timing by only 26–29%—diminishing returns that shift optimization toward fiber choice.
  • At fixed N_pe, replacing a slow WLS fiber (≈7 ns re-emission) with a fast one (≈1.5 ns) improves resolution by a factor of 1.7–2.2, roughly equivalent to quadrupling the light yield.
  • Double-ended readout delivers the √2 statistical gain and nearly position-independent resolution up to about 100 cm; a far-end mirror is beneficial only for bars shorter than about 20 cm.
  • Electronics enter the budget only when the TDC bin exceeds roughly 2 ns at N_pe≈50; below that, finer digitization adds essentially nothing.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Extension: the model's clean split between N_pe calibration and timing physics suggests a two-stage design flow—measure N_pe with a cheap prototype, then use the analytical maps to choose fiber, topology, and electronics—so detector optimization could become faster and cheaper than iterating full simulations.
  • Extension: the paper's finding that first-photon (k=1) triggering is optimal for these convolved PDFs implies that constant-fraction discriminators add little for fast-WLS/SiPM systems; this is directly testable with waveform data and, if confirmed, would simplify front-end electronics.
  • Extension: the paper explicitly leaves pile-up, afterpulsing, crosstalk, temperature drift, and fiber bending outside its scope; each of these perturbs the effective N_pe or adds late correlated photons, so a natural next step is to extend the order-statistic model to non-Poissonian or time-varying N_pe.
  • Extension: the steeper Geant4 power-law exponent at low N_pe hints that the single-exponential WLS/attenuation description misses some late or mode-dependent components; measuring τ_WLS as a function of injection distance would show whether this is a real effect or a simulation artifact.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. The manuscript presents a semi-analytical model for the timing resolution of plastic scintillator detectors with wavelength-shifting (WLS) fiber and SiPM readout. The model decomposes the detection chain into scintillation emission (bi-exponential), WLS re-emission (single exponential), SiPM single-photon time resolution (Gaussian), order-statistic timing for leading-edge triggering on the k-th photoelectron, fiber transit-time spread, and electronics quantization. The authors validate the numerical implementation with a high-statistics toy Monte Carlo and compare the timing model with full Geant4 optical simulations at the Geant4-measured photoelectron yield. They also benchmark against published measurements from SuperFGD, SciBar, MINOS, and FNAL, and provide extensive design maps and lookup tables covering scintillators, fibers, SiPMs, electronics, readout topologies, and boundary conditions.

Significance. If the model is quantitatively predictive, it would provide a practical design tool for scintillator timing detectors, allowing the timing resolution to be estimated from measurable parameters such as N_pe, tau_WLS, and readout electronics. The analytical framework is transparent, the toy Monte Carlo convincingly checks the order-statistic mathematics, and the Geant4 setup is a serious physics-based test. The authors are also explicit about several limitations, notably the N_pe range covered by the Geant4 validation and the need to calibrate N_pe from simulation or measurement. However, the central quantitative claim — sub-10% predictability of sigma_t for N_pe = 50–200 — is not established by the evidence presented. The only physics-based validation reaches N_pe ≈ 34 and shows systematic 4–16% deviations at N_pe ≈ 14, with a fitted Geant4 power-law exponent that differs from the analytical exponent. The paper is therefore more convincing as a qualitative design guide than as a quantitatively validated predictive model in the range emphasized by its own tables.

major comments (3)
  1. [§5.3, §5.4, Fig. 26] The Geant4 N_pe scan yields a fitted power-law exponent alpha ≈ 0.84, while the analytical model gives alpha ≈ 0.44–0.49. The manuscript itself states that the Geant4 validation spans N_pe ≈ 2–34 (Sec. 5.4), whereas Tables 1–4 and the design maps emphasize N_pe = 50–200. Because the order-statistic variance (Eq. 2.10) is controlled by the leading edge of f_det, a shape discrepancy that already produces G4/analytical ratios of 0.84–0.96 at N_pe ≈ 14, together with a steeper N_pe scaling, implies that the discrepancy is N_pe-dependent and cannot be assumed to vanish at higher N_pe. The authors should either extend the Geant4 validation into the N_pe = 50–200 range or explicitly restrict the quantitative predictive claims to the validated range.
  2. [§5.2, Fig. 21b] The G4/analytical ratio of 0.84–0.96 at N_pe ≈ 14 is attributed to 'realistic optical transport effects ... that produce a slightly sharper leading edge'. This is a load-bearing admission: Eq. (2.10) is precisely the leading-edge-sensitive order-statistic variance, so a 10–16% bias in sigma_t at N_pe ≈ 14 indicates that the folded triple-exponential-plus-Gaussian f_det of Eq. (2.5)–(2.6) is not the correct single-photon detection PDF. The paper should either quantify a modified f_det (e.g., an effective faster rise time) and propagate it through the order-statistic integral, or present the Geant4 comparison as a qualitative validation of trends rather than as a quantitative confirmation of the analytical timing predictions.
  3. [§2.2, §2.7, §5.4] The model assumes a single-exponential WLS re-emission time (Eq. 2.3), a uniform mode-angle distribution for the transit-time spread (Eq. 2.11), and a single-exponential attenuation law for N_pe(x) (Eq. 2.7). The Geant4 simulation itself finds an effective attenuation length of about 150 cm versus the tabulated 350 cm (Sec. 5.4), which shows that the transport model is not reproducing the simulated fiber physics. This is partly masked because the timing comparison is evaluated at the Geant4-measured N_pe, but it matters for the position-dependent and length-dependent predictions in Sec. 3.3 and for the design tables. The manuscript should state more prominently that the transport-related assumptions are not independently validated and that predictions for long bars and low-N_pe regions carry additional uncertainty.
minor comments (5)
  1. [Abstract and §5.4] The abstract says the analytical predictions are 'confirmed' by Geant4 without mentioning the N_pe ≈ 2–34 range or the 16–20% spread. The range limitation should be stated in the abstract or introduction so that readers do not overinterpret the validation.
  2. [Fig. 27b] The aggregate ratio plot reports mean = 1.01 ± 0.21. It would be useful to give the median and quartiles, or to split the ratio distribution by N_pe, because the low-N_pe points dominate the scatter and the high-N_pe behavior is the relevant regime for the lookup tables.
  3. [Eq. (2.5)] The degenerate cases tau_r = tau_WLS or tau_d = tau_WLS are handled by 'infinitesimal perturbations'. For reproducibility, it would be better to provide the limiting forms explicitly or to state the numerical prescription used.
  4. [Eq. (2.13)] The weighted-mean formula for double-ended readout is stated without defining the weights. Since the weights determine the position uniformity claim in Sec. 3.3, a short definition would be helpful.
  5. [Table 6] The footnote states that BCF-XL series is single-clad, while the text in Sec. 5.1 describes a double-clad fiber model. Please clarify whether the BCF simulations in Sec. 5 use single- or double-clad geometry and whether the trapping efficiency values in Table 6 are consistent with that geometry.

Circularity Check

0 steps flagged

No significant circularity: toy-MC agreement is explicitly a numerical consistency check; Geant4 and published benchmarks provide independent physics comparisons.

full rationale

I traced the derivation chain from Eqs. (2.1)–(2.12). None of the load-bearing quantities is defined in terms of the target σ_t: the single-photon detection PDF is assembled from independently reported material parameters (τ_r, τ_d, τ_WLS from refs. [9],[17]; SPTR from refs. [7],[18]), the order-statistic variance is a direct integral of Eq. (2.8), and transit-time and electronics terms are separate quadrature additions. The toy-MC check of Sec. 4 is honest about its role: the paper states it demonstrates that the analytical order-statistic calculation and the event-by-event Toy MC are numerically consistent when the assumptions are matched, so the 0.9997±0.0015 ratio verifies the numerical implementation, not the physical assumptions. The Geant4 comparison evaluates the analytical model at the Geant4-measured N_pe, which is a conditional test of the timing model and not a circular fit; the paper explicitly discloses the validation range (N_pe≈2–34), the G4/analytical spread (≈0.84–0.96 at N_pe≈14; aggregate 1.01±0.21), and the fact that design tables emphasize N_pe≈50–200. The benchmark comparisons with SuperFGD, SciBar, MINOS, and FNAL use published N_pe values and measured resolutions, with ratios 0.87–0.96 (excluding the FNAL electronics-dominated floor). No parameter is fitted to the target predictions, and no load-bearing conclusion reduces to a self-citation chain. The limited validation at high N_pe and the model's simplifying assumptions are legitimate scientific limitations, but they are not circularity.

Axiom & Free-Parameter Ledger

4 free parameters · 8 axioms · 0 invented entities

No new physical entities are introduced. The ledger contains assumptions about material timing profiles, transport models, and statistical independence. The most burdensome inputs are the fitted/supplied WLS decay times and the geometry-dependent N_pe, which the paper deliberately treats as external calibration values.

free parameters (4)
  • WLS re-emission time τ_WLS = 7.10 (Y-11), 3.64 (YS-2), 2.15 (YS-4), 1.47 (YS-6), 12.00 (BCF-91AXL), 2.10 (BCF-92XL), 2.10 (BCF-9929AXL), 2.41 (BCF-99
    Taken as input from published measurements (Kodama et al. [9], Yamazumi et al. [17]). The central design conclusion depends on these values.
  • Geometry correction factor C_geom in N_pe(x) Eq. (2.7) = not specified
    A geometry-dependent factor in Eq. (2.7); the paper says it approximates light-collection effects. It is not derived from first principles; in practice it must be calibrated against simulation or a prototype.
  • Effective attenuation length Λ_eff = ≈150 cm (in the Geant4 simulation)
    The paper finds Λ_eff≈150 cm in the Geant4 optical transport, versus tabulated λ_att=350 cm. The model's N_pe(x) depends on this exponent; mismatch signals unmodeled optical effects (cladding-mode losses, wavelength-dependent absorption).
  • Power-law exponent α in σ_t = A·N_pe^-α = 0.44-0.49 (analytical); 0.84 (Geant4 fit, fig. 26)
    Exponent is computed by fitting the model output. The Geant4-derived value differs from the analytical value by a factor of ~1.8, a quantitative discrepancy the paper attributes qualitatively to low-N_pe Poisson effects and optical effects.
axioms (8)
  • domain assumption Scintillation emission follows the bi-exponential PDF Eq. (2.1) with parameters τ_r, τ_d.
    Standard modeling of plastic scintillators; the paper cites no measurement validating the bi-exponential form for all listed materials.
  • domain assumption WLS re-emission follows a single exponential with one decay constant τ_WLS (Eq. 2.3), independent of wavelength, injection distance, and mode.
    This is the weakest stated physical assumption; the paper's own Geant4 calibration finds a 20% offset and a shorter effective attenuation length, indicating that the real fiber transport is more complex.
  • domain assumption Timestamps are generated by a leading-edge trigger on the k-th detected photoelectron, with all N_pe samples independent and identically distributed from f_det.
    This is the mathematical core of the model; it excludes correlated effects such as afterpulsing, dark counts, optical crosstalk, and pulse pile-up, which the paper lists as limitations.
  • domain assumption Fiber mode angles are uniformly distributed over the guided range, giving the transit-time-spread formula Eq. (2.11).
    Analytical approximation; the paper never validates the uniform-mode assumption against the Geant4 mode-selective propagation in detail.
  • domain assumption The photoelectron count N_pe is Poission-distributed around mean N_pe(x) of Eq. (2.7), with a single-exponential attenuation.
    Standard; the paper observes that the effective attenuation length differs from the tabulated value, so the stated attenuation law is not exact.
  • domain assumption All noise terms add in quadrature with no correlations, and each SiPM channel is independent.
    Used for RT2 and RT3 averaging; toy MC and G4 verify independence to ~1%, but real detectors have crosstalk and common dark noise, unmodeled.
  • domain assumption Published values of PDE, SPTR, NA, trapping efficiency, scintillator light yields, and WLS decay times are correct and apply to the simulated geometries.
    The model's quantitative predictions are only as good as the material property tables (tables 5-8).
  • standard math The CDF and order-statistic integrals in Eq. (2.8)-(2.9) converge and are evaluated with sufficient numerical accuracy (trapezoidal rule, 10-ps grid).
    The toy MC validates the implementation at 0.2% level, so the numerical evaluation is not risky.

reviewed 2026-08-02 · how reviews work

0 comments
Cite this review

Pith. "Pith review of Comprehensive study of timing resolution in plastic scintillator detectors with wavelength-shifting fiber and silicon photomultiplier readout." pith.science (2026). https://pith.science/paper/CUBTLQUR

@misc{pith2026260714290,
  author       = {Pith},
  title        = {Pith review of: Comprehensive study of timing resolution in plastic scintillator detectors with wavelength-shifting fiber and silicon photomultiplier readout},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/CUBTLQUR}},
  note         = {Machine review of arXiv:2607.14290}
}
Share X Bluesky LinkedIn Reddit HN
abstract

We present a comprehensive study of the timing resolution achievable in plastic scintillator detectors read out through wavelength-shifting (WLS) fibers coupled to silicon photomultipliers (SiPMs), combining a semi-analytical framework, toy Monte Carlo validation, and full Geant4 optical photon simulation. The analytical model traces the complete photon detection chain: scintillation emission, WLS fiber re-emission, optical transit time dispersion, SiPM single-photon time resolution, and electronics quantization. It expresses the timing resolution $\sigt$ as a function of the detected photoelectron yield $\Npe$, scintillator decay constants ($\taur$, $\taud$), WLS re-emission time ($\tauwls$), fiber numerical aperture, detector geometry, and readout electronics parameters. The analytical predictions are validated at two levels. First, toy Monte Carlo simulations ($2\times 10^5$ events per parameter point across 80 grid points spanning 8 fiber types and $\Npe$ from 5 to 200) achieve analytical-to-MC agreement of $0.9997 \pm 0.0015$. Second, full Geant4 optical photon simulations track the entire scintillation, wavelength-shifting, and fiber transport chain in realistic detector geometries, confirming the analytical timing predictions and providing first-principles photoelectron yield calibration. A comprehensive parameter scan covering 7 scintillator materials, 8 WLS fiber types, 5 SiPM models, 5 electronics configurations, 3 readout topologies, and 3 boundary conditions produces quantitative design maps and lookup tables for detector optimization.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

20 extracted references · 3 canonical work pages

  1. [1]

    Blondel et al.,The SuperFGD Prototype charged particle beam tests,JINST15(2020) P12003 [arXiv:2008.08861]

    A. Blondel et al.,The SuperFGD Prototype charged particle beam tests,JINST15(2020) P12003 [arXiv:2008.08861]

    Pith/arXiv arXiv 2020

  2. [2]

    Alekseev et al.,SuperFGD prototype time resolution studies,JINST18(2023) P01012 [DOI:10.1088/1748-0221/18/01/P01012]

    I. Alekseev et al.,SuperFGD prototype time resolution studies,JINST18(2023) P01012 [DOI:10.1088/1748-0221/18/01/P01012]

    work page doi:10.1088/1748-0221/18/01/p01012 2023

  3. [3]

    Aliaga et al

    L. Aliaga et al. [MINERvA Collaboration],Design, calibration, and performance of the MINERvA detector,Nucl. Instrum. Meth. A743(2014) 130 [arXiv:1305.5199]

    Pith/arXiv arXiv 2014

  4. [4]

    Michael et al

    D.G. Michael et al. [MINOS Collaboration],The Magnetized Steel and Scintillator Calorimeters of the MINOS Experiment,Nucl. Instrum. Meth. A596(2008) 190 [arXiv:0805.3170]

    Pith/arXiv arXiv 2008

  5. [5]

    Nitta et al.,The K2K SciBar detector,Nucl

    K. Nitta et al.,The K2K SciBar detector,Nucl. Instrum. Meth. A535(2004) 147 [arXiv:hep-ex/0406023]

    Pith/arXiv arXiv 2004

  6. [6]

    Denisov, V

    D. Denisov, V. Evdokimov, S. Lukić, P. Ujić,Test beam studies of the light yield, time and coordinate resolutions of scintillator strips with WLS fibers and SiPM readout,Nucl. Instrum. Meth. A848 (2017) 54 [arXiv:1611.03211]

    Pith/arXiv arXiv 2017

  7. [7]

    Hamamatsu Photonics,MPPC (Multi-Pixel Photon Counter) Technical Note, Cat. No. KAPD9008E

  8. [8]

    Kuraray Co., Ltd.,Wavelength Shifting Fibers, product catalog (2023), http://kuraraypsf.jp/psf/ws.html

    2023

  9. [9]

    S.Kodamaetal.,PerformanceofnewKuraraywavelength-shiftingfiberswithshortdecaytime,Prog. Theor. Exp. Phys.2024(2024) 053H01 [arXiv:2311.07297]

    Pith/arXiv arXiv 2024

  10. [10]

    Eljen Technology,Plastic Scintillators, product catalog, https://eljentechnology.com/products/plastic-scintillators

  11. [11]

    Saint-Gobain Crystals,Scintillating Optical Fibers, product catalog (2023), https://www.crystals.saint-gobain.com/products/scintillating-fiber

    2023

  12. [12]

    David and H.N

    H.A. David and H.N. Nagaraja,Order Statistics, 3rd ed., Wiley-Interscience, Hoboken, NJ (2003), [DOI:10.1002/0471722162]

    doi:10.1002/0471722162 2003

  13. [13]

    Post and L.I

    R.F. Post and L.I. Schiff,Statistical Limitations on the Resolving Time of a Scintillation Counter, Phys. Rev.80(1950) 1113 [DOI:10.1103/PhysRev.80.1113]

    work page doi:10.1103/physrev.80.1113 1950

  14. [14]

    Agostinelli et al

    S. Agostinelli et al. [GEANT4 Collaboration],GEANT4 — a simulation toolkit,Nucl. Instrum. Meth. A506(2003) 250 [DOI:10.1016/S0168-9002(03)01368-8]

    doi:10.1016/s0168-9002(03)01368-8 2003

  15. [15]

    Allison et al.,Recent developments in Geant4,Nucl

    J. Allison et al.,Recent developments in Geant4,Nucl. Instrum. Meth. A835(2016) 186 [DOI:10.1016/j.nima.2016.06.125]

    doi:10.1016/j.nima.2016.06.125 2016

  16. [16]

    Abe et al.,Characterization of the optical model of the T2K 3D segmented plastic scintillator detector,Nucl

    S. Abe et al.,Characterization of the optical model of the T2K 3D segmented plastic scintillator detector,Nucl. Instrum. Meth. A1080(2025) 170757 [DOI:10.1016/j.nima.2025.170757]

    arXiv 2025

  17. [17]

    Yamazumi et al.,Performance evaluation of Luxium Solutions BCF-XL wavelength-shifting fibers, Prog

    T. Yamazumi et al.,Performance evaluation of Luxium Solutions BCF-XL wavelength-shifting fibers, Prog. Theor. Exp. Phys.2026(2026) 053H01 [arXiv:2601.06784]

    Pith/arXiv arXiv 2026

  18. [18]

    Gundacker et al.,Experimental time resolution limits of modern SiPMs and TOF-PET detectors exploring different scintillators and Cherenkov emission,Phys

    S. Gundacker et al.,Experimental time resolution limits of modern SiPMs and TOF-PET detectors exploring different scintillators and Cherenkov emission,Phys. Med. Biol.65(2020) 025001 [DOI:10.1088/1361-6560/ab63b4]

    doi:10.1088/1361-6560/ab63b4 2020

  19. [19]

    Lang et al.,Characterization of 1600 Hamamatsu 16-anode photomultipliers for the MINOS Far detector,Nucl

    K. Lang et al.,Characterization of 1600 Hamamatsu 16-anode photomultipliers for the MINOS Far detector,Nucl. Instrum. Meth. A545(2005) 852 [DOI:10.1016/j.nima.2005.02.041]. – 39 –

    work page doi:10.1016/j.nima.2005.02.041 2005

  20. [20]

    Barker et al.,Testing of Hamamatsu R5900-00-M64 Multi-Pixel PMTs for MINOS, [arXiv:hep-ex/0203001]

    M.A. Barker et al.,Testing of Hamamatsu R5900-00-M64 Multi-Pixel PMTs for MINOS, [arXiv:hep-ex/0203001]. – 40 –

    Pith/arXiv arXiv

This paper was first reviewed by deepseek-v4-flash on August 2, 2026.