REVIEW 4 major objections 3 minor 6 references
Development of a UV-transparent Lens Array for Enlarging the Effective Area of Multichannel SiPMs
T0 review · 4 major / 3 minor · reviewed 2026-08-14 · deepseek-v4-flash
Pith's one-line read Placing a UV-transparent plano-convex lens over each pixel of a 64-channel silicon photomultiplier concentrates gap-bound photons onto active silicon, giving a measured 10–30% relative gain in photon detection efficiency at 30–60°…
desk verdict A useful, niche instrumentation result with a plausible 10-30% PDE gain, but the measurement section needs error bars and crosstalk control before the number is secure. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The load-bearing object is the individual plano-convex spherical lens assigned to each pixel. With a 3.2×3.2 mm footprint, a height of 2.0 mm, and a radius of curvature of 2.3 mm, each lens is sized to cover one 3×3 mm sensor pixel plus its surrounding 0.2 mm gap. Working together with the optical grease coupling, the lens refracts off-axis photons so they land in the sensitive area, and at steep incidence it uses total internal reflection at the lens–air boundary and reflection from the silicon surface to send photons toward neighboring pixels. The reason the numbers line up is geometric: the gap fraction is 0.121 of the pixel pitch, so increasing effective area by $\frac{1}{1-0.121}\approx 1.138$ would exactly cancel the loss, and the measured 10–30% gains bracket that needed recovery.
What would settle it
Take the same SiPM with and without the lens array at fixed overvoltage and record the waveform of each detected photon. If the 10–30% gain is geometric, the extra counts should be single-cell pulses appearing only in pixels whose simulated active-area coverage grows; if optical crosstalk is the cause, the extra counts should come as correlated multi-cell bursts immediately after a photon and should disappear at lower overvoltage or with a crosstalk-suppressed device.
Extended reading notes
Core claim
The central claim, stated in the paper's own terms, is that a lens array of 8×8 plano-convex spherical lenses—each lens 3.2×3.2 mm in footprint with a 2.0 mm height and 2.3 mm radius of curvature, made of UV-transparent material and coupled to the SiPM array with optical grease—recovers the loss of photon detection efficiency caused by the 0.2 mm pixel gaps. The reported evidence is a measured relative PDE increase of about 10–30% for incidence angles of 30–60°, consistent across 310, 375, 465, and 635 nm LED light, while simulation gives a 13.5% effective-PDE gain for a flat 30–60° distribution, nearly the 13.8% needed to offset the 12.1% fill-factor loss. The paper also reports that at large angles the mechanism is not simple focusing: photons reflect off the silicon surface and are totally reflected at the lens–air boundary, sending some light to neighboring pixels. The authors note that possible increases in optical crosstalk should still be evaluated.
Load-bearing premise
The conclusion that the lens array recovers the gap loss rests on the measured count increase being caused by geometric concentration of photons onto the sensitive silicon, not by an increase in optical crosstalk or a change in avalanche probability—and the paper explicitly says the crosstalk possibility has not yet been evaluated.
Editorial extensions
If this is right
- For incidence angles of 30–60°, the lens array raises the effective PDE of a 64-channel SiPM array to nearly the value it would have without the 0.2 mm gaps, across the tested 310–635 nm range.
- The design lets camera builders reduce silicon area—and therefore unit cost—by widening pixel gaps, because the molded lens array optically fills the missing area.
- In a Cherenkov camera whose optics deliver 30–60° rays, the lens array adds detected photons per unit focal-plane area without growing the silicon wafer.
- Because steep-angle photons are redirected to neighboring pixels, multi-channel readout and image reconstruction must treat the lens array as a light-sharing optical element, not a purely per-pixel concentrator.
Reading between the lines
- If the geometric interpretation is correct, the lens concept generalizes to any gap width: the required recovery factor is set by the ratio of lens aperture to active pixel area, so taller or higher-index lenses could recover even wider gaps.
- The wavelength-dependent gains hint that the thin layers on the silicon act as an antireflection coating; a wave-optical treatment could turn the lens array into a tool for shaping the spectral response of a focal plane.
- The measured redistribution of steep-angle photons onto neighboring pixels would effectively low-pass-filter the image, which could smooth pixel-to-pixel nonuniformities at the cost of requiring deconvolution in analysis.
- A decisive extension would be to repeat the measurement on a crosstalk-suppressed SiPM at low overvoltage; if the 10–30% gain survives, geometric concentration is confirmed.
Signed reviews
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper reports the design, simulation, and prototype measurement of a UV-transparent plano-convex lens array intended to recover the geometric fill-factor loss of a multichannel SiPM array (Hamamatsu S13361, 3x3 mm^2 pixels with 0.2 mm gaps). The lens array has one lens per pixel, with height 2.0 mm and radius of curvature 2.3 mm, chosen by ROBAST ray-tracing simulation to maximize PDE for 30-60 degree incidence, as relevant for a Schwarzschild-Couder Cherenkov camera. Measurements with four LED wavelengths (310, 375, 465, 635 nm) show a relative PDE increase of roughly 10-30% at 30-60 degrees, with smaller increases at 0-40 degrees and discrepancies from simulation at 50-70 degrees. The authors conclude that the lens array can effectively reduce the dead-area loss, while noting that optical crosstalk and image degradation need further evaluation.
Significance. If the measured 10-30% relative PDE increase is genuine, the lens array would recover most of the 12.1% geometric fill-factor loss in a multichannel SiPM array without enlarging the silicon substrate, which is relevant for compact focal-plane Cherenkov cameras. The design uses no fitted parameters for the central comparison; the lens geometry was fixed before the measurement and the simulated and measured PDE increases are independent. The paper is honest about the simulation-measurement mismatch at large angles and about the unassessed crosstalk contribution. However, the measurement section is too thin to firmly establish the central claim as a quantitative PDE gain rather than an integrated-charge artifact.
major comments (4)
- [Section 4, Fig. 3] The measurement section does not report error bars, statistical uncertainties, or systematic checks such as LED stability, dark-count subtraction, or pulse-height calibration. The claim of a 10-30% relative PDE increase is presented as a ratio of detected counts without any uncertainty estimate, so the reader cannot judge whether the observed differences are significant relative to, for example, LED intensity drift or baseline fluctuations. This is a load-bearing omission because the central quantitative claim rests entirely on these measurements.
- [Section 4] The text states that the comparison uses 'the numbers of detected photons by one of the 64 channels' but does not specify how the channel was selected, how the lens array was aligned to the pixel, or whether the counts were obtained from integrated charge or from resolved single-photoelectron pulses. If the readout integrates charge, the measured ratio could be inflated by lens-induced optical crosstalk or by a change in the effective gain, so the attribution of the increase to geometric concentration is not uniquely established.
- [Conclusion, Fig. 3] The simulation overpredicts the relative PDE increase at 50-70 degrees and fails to reproduce the two-peak structure observed in the measurement. The paper attributes this to inadequate modeling of the thin SiO2 and Si3N4 layers and the silicon-silicone interface. This is not a fatal flaw for the measured claim, but it means the simulation cannot be used to validate the large-angle behavior, and the design optimization based on the simulation is only partially supported by the data.
- [Conclusion] The authors themselves state that 'possible increase of optical crosstalk [6] should also be evaluated.' Because the lens concentrates photons onto smaller silicon areas and redirects photons onto neighboring pixels at large angles, the measured count increase could partly arise from crosstalk-induced additional avalanches rather than from an increased probability of detecting primary photons. Without an explicit crosstalk measurement or a single-photoelectron analysis, the reported gain cannot be unambiguously interpreted as a PDE increase.
minor comments (3)
- [Section 2] The notation '3.2 × 3.2-mm2 region' is dimensionally awkward; it should read '3.2 mm × 3.2 mm region' or '3.2 x 3.2 mm^2 area.'
- [Fig. 4] The count maps in Fig. 4 are useful, but the color scale is not clearly labeled with units; stating that it is 'detected photons per bin' would improve readability.
- [References] Reference [6] is cited as 'These Proceedings (2019)' without a title or page numbers; if this is a conference proceedings, a fuller citation should be provided.
Circularity Check
No significant circularity: the measured 10-30% PDE increase is an independent comparison against the bare SiPM, with simulation used only for design and post-hoc modeling.
full rationale
The paper's central claim is an empirical ratio: the number of detected photons with the lens array coupled is compared with the number without the lens array under the same LED illumination. This measurement is not derived from the simulation or from any fitted parameter. The lens geometry was chosen by ray-tracing simulation before the prototype was built, and the measured result is an independent check of that design. The post-production simulation that adds thin-film layers is explicitly described as an attempt to reproduce the measurement ('We performed a more realistic simulation after the production and performance test of the lens array to reproduce the measurement result'), and it is disclosed as such; it does not feed back into the measured claim. The self-citations are to prior development tools (hollow Winston cones, ROBAST ray-tracing software) and are not load-bearing justifications of the measured PDE gain. The acknowledged possibility of increased optical crosstalk is an interpretation caveat and a correctness risk, not a circularity: the paper itself flags that this should be evaluated. No equation is shown to reduce to itself, no fitted input is renamed as a prediction, and no uniqueness claim is imported from the authors' prior work. Accordingly, the derivation chain is self-contained with respect to circularity.
Assumptions & free parameters
free parameters (2)
- Lens height =
2.0 mm
- Lens radius of curvature =
2.3 mm
assumptions (3)
- domain assumption Geometric ray optics accurately models light propagation through the lens, grease, and SiPM layers; thin-film interference is ignored.
- domain assumption The PDE curve of Hamamatsu S12571-050C used in the design simulation is representative of the S13361-3050AS array used in measurements.
- domain assumption A flat distribution of incidence angles between 30 and 60 degrees represents the Cherenkov camera illumination.
Cite this review
Pith. "Pith review of Development of a UV-transparent Lens Array for Enlarging the Effective Area of Multichannel SiPMs." pith.science (2026). https://pith.science/paper/HW6P4TCN
@misc{pith2026190805510,
author = {Pith},
title = {Pith review of: Development of a UV-transparent Lens Array for Enlarging the Effective Area of Multichannel SiPMs},
year = {2026},
howpublished = {\url{https://pith.science/paper/HW6P4TCN}},
note = {Machine review of arXiv:1908.05510}
}
abstract
We developed a UV-transparent lens array that can increase the photon detection efficiency of a silicon photomultiplier (SiPM) array comprising of 64 pixels ($3\times3$ mm$^2$ each) and 0.2-mm gaps. Through the plano-convex spherical lens on each $3.2\times3.2$ mm$^2$ region, we showed that the loss of photon detection efficiency due to the pixel gaps could be recovered as the incident photons get concentrated on the sensitive regions of the SiPM array. By using a prototype lens array, we achieved approximately 10%-30% relative increase in photon detection efficiency in our target angles of incidence of 30-60 deg.
Figures
Figures from the paper (2 more)
Reference graph
Works this paper leans on
Reviewed August 14, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.