Pith. sign in

REVIEW 2 major objections 5 minor 19 references

The paper argues that a spherical plano-convex microlens array can triple the thermal sensitivity of spintronic Poisson bolometer arrays, cutting NEDT from 30 mK to 10 mK in full-image simulations.

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 ↗

T0 review · deepseek-v4-flash

2026-08-03 16:36 UTC pith:YDGUNHIS

load-bearing objection Useful first design study for microlens arrays on spintronic Poisson bolometers, but the headline 30-to-10 mK NEDT gain is not supported by the simulation as presented. the 2 major comments →

arxiv 2512.12491 v9 pith:YDGUNHIS submitted 2025-12-13 physics.optics physics.app-ph

Design of Microlens Arrays for Thermal Imaging with Spintronic Poisson Bolometers

Ziyi Yang , Leif Bauer , Zubin Jacob This is my paper
classification physics.optics physics.app-ph
keywords microlens arraysspintronic Poisson bolometermid-wave infrarednoise-equivalent differential temperaturecollection efficiencyconcentration factorfinite-difference time-domainthermal imaging
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.

This paper aims to show that placing a spherical plano-convex microlens array over a spintronic Poisson bolometer array can substantially improve the thermal-imaging sensitivity of the detector. Sensitivity is quantified by the noise-equivalent differential temperature (NEDT), which the authors' radiometric-stochastic model estimates dropping from 30 mK to 10 mK as microlens diameter increases, assuming lossless lenses. It also derives practical design rules: collection efficiency is best for active sensor sizes above 20 microns and small lens diameters, while concentration factor is best for active sizes 5–20 microns. A fabricated aluminum-oxide sample demonstrates that the chosen lens geometries are physically attainable. The work matters because the spintronic Poisson bolometer's tiny photosensitive area limits its low-light sensitivity; a microlens array is a passive, wafer-level fix.

Core claim

The central claim is that the collection efficiency of a microlens array, computed from FDTD simulations of plane-wave illumination and then fed into a Poisson counting model, determines a direct trade-off between lens diameter and sensor active area. For the spintronic Poisson bolometer, the full-image simulation shows NEDT improving from 30 mK to 10 mK when the microlens diameter is increased from about 20 microns to 34 microns at 100% lens transmission, with correspondingly better image contrast. The authors identify two regimes: for active sizes larger than 20 microns the collection efficiency is maximized with small lens diameters, whereas for active sizes between 5 and 20 microns the c

What carries the argument

The central objects are the microlens array itself and two scalar performance metrics: the collection efficiency (CE), the fraction of optical power incident on the pixel that lands on the sensor active area, and the concentration factor (CF), the ratio of CE with the microlens to CE without it. These are computed from FDTD simulations and then coupled into a radiometric-stochastic model: a Poisson random variable whose mean is the product of system detection efficiency, incident photon rate plus dark rate, and response time, with the incident flux scaled by the main lens and the microlens CE. The microlens geometry (diameter, sag, radius of curvature) enters through CE, which is what modula

Load-bearing premise

The FDTD collection efficiencies assume collimated (plane-wave) illumination, but in the imaging configuration the MLA sits at the image plane of a fast main lens that delivers a cone of rays; if the focused cone spreads light beyond the active area, the predicted CE and NEDT gain would be too optimistic.

What would settle it

Measure the focal-plane power distribution of a fabricated MLA using a focused beam with the same numerical aperture as the main lens (f/0.5 as specified) and compare the measured CE with the plane-wave CE; or directly measure NEDT of an array with and without MLA under a blackbody scene. If measured CE under realistic cone illumination is substantially below the plane-wave value, the predicted NEDT improvement would not hold.

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

If this is right

  • For spintronic Poisson bolometer arrays, a plano-convex Al2O3 microlens array improves NEDT by a factor of three (30 mK to 10 mK) at unit lens transmission, and improves image contrast.
  • The design rules provide a parameter window: for active sizes greater than 20 microns, small lens diameters maximize collection efficiency; for active sizes between 5 and 20 microns, the concentration factor is highest.
  • The fabrication demonstration with thermal reflow shows that sag, diameter, and spherical profile are attainable in practice, with measured transmission around 60% across the MWIR band.
  • The same microlens design rules apply to other MWIR detectors with limited active area, giving a general pathway to raise effective fill factor.
  • At the measured 60% transmission, the NEDT improvement is reduced but still present relative to the no-lens baseline.

Where Pith is reading between the lines

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

  • If the plane-wave assumption is relaxed to the cone of rays from a fast main lens, the CE values may shift; this could be tested by simulating with a focused beam or by measuring focal spots under realistic illumination.
  • The design rules suggest a scaling law: for fixed lens pitch, there is an optimal active size where CF peaks; a detector designer can tune active area versus in-pixel electronics based on this curve.
  • A testable extension is to fabricate a full MLA-integrated bolometer array and measure NEDT directly, comparing against the 30-to-10 mK prediction.
  • The Poisson counting model implies that the NEDT improvement is not simply linear in CE, because dark counts and integration time set a floor; further sensitivity gains may require reducing the dark count rate in parallel.

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

2 major / 5 minor

Summary. The paper proposes design guidelines for spherical plano-convex microlens arrays (MLAs) to improve the effective fill factor and sensitivity of spintronic Poisson bolometer (SPB) arrays in the mid-wave infrared (MWIR). The authors use FDTD simulations to compute collection efficiency (CE) and concentration factor (CF) as functions of microlens diameter and sensor active size, fabricate an Al2O3 MLA sample to demonstrate feasibility, and build a radiometric-stochastic model to predict noise-equivalent differential temperature (NEDT). The full-image simulation reports an NEDT improvement from 30 mK to 10 mK as the microlens diameter increases at 100% transmission, with correspondingly better image contrast.

Significance. If the quantitative result were supported, the paper would offer a practical, generalizable route to improving sensitivity in low-fill-factor MWIR detector arrays without modifying the detector stack. The paper is transparent about its forward-model structure: CE from FDTD and detector parameters (SDE, λ_dark) are inputs, not fitted quantities, so there is no obvious circularity. The design-rule maps in Fig. 4 and the fabricated MLA sample are useful contributions. However, the central NEDT improvement relies on a plane-wave CE applied to a fast-cone imaging geometry, and the main-lens parameters in Table 2 are internally inconsistent. The qualitative design rules may survive, but the headline sensitivity number is not yet established.

major comments (2)
  1. [Table 2] The main-lens parameters are internally inconsistent. The thin-lens equation with Zobj=30 cm and Zimg=4 cm gives f ≈ 35.3 mm; with D = 25.4 mm, this yields f/# ≈ 1.39, not 0.5. Conversely, f/0.5 with D = 25.4 mm would require f = 12.7 mm, contradicting Zimg. Since the radiometric model and the 'small opening angle' approximation depend on the actual f/#, the NEDT values in Fig. 5(b) must be recomputed for a consistent set of parameters.
  2. [Eq. (9) and Fig. 5] The transfer model in Eq. (9) applies CE as a scalar multiplier, but CE is computed in FDTD under normally incident plane-wave illumination ('Plane-wave illumination is assumed for all the simulations'). In the full-image model, the MLA sits at the image plane of a finite-f/# main lens; the incident light is a converging cone with half-angle ~21° for f/1.39 (or larger for f/0.5). The paper's claim that 'both the main lens and the microlens have small opening angles and the Abbe sine condition is satisfied' is not justified for these speeds. Oblique rays will shift the focus and introduce aberrations, so CE will differ from the plane-wave value. Because NEDT scales linearly with CE, the predicted 30→10 mK improvement is not supported. The authors should compute CE for angled illumination (e.g., focused beam or angular spectrum of plane waves) and verify its variation across the collection
minor comments (5)
  1. [Fig. 3 / Table 2] The fabricated MLA has sag ~2 µm (design 2 µm, actual 1.79 µm), while the full-image simulation uses S = 6 µm. The text states the sample demonstrates that 'the chosen geometrical parameters are realistic,' but the sample does not replicate the simulated sag. Please clarify whether 6 µm sag was fabricated elsewhere or whether the feasibility claim applies only to the spherical profile at smaller sag.
  2. [Eq. (4)] The NEDT definition is unclear: µ_pixel is described as averaging over pixels, but NEDT is normally a per-pixel temporal noise divided by responsivity. Please clarify whether σ is temporal noise, and define the averaging operation explicitly.
  3. [Fig. 5(b)] The paper quotes the 100% transmission improvement (30→10 mK), but the fabricated MLA has ~60% transmission. Please explicitly report the NEDT for T = 60% for the selected parameters, since this is the experimentally relevant case.
  4. [Introduction / Fig. 3] The text states Al2O3 has 80% transmission, but the fabricated sample transmits ~60% across MWIR. The discrepancy (Fresnel losses, fabrication quality) should be discussed to avoid confusion.
  5. [FDTD simulations] CE is computed at λ = 4 µm only, while the scene spans 3–5 µm. Please comment on the wavelength sensitivity of CE and justify using the central wavelength as representative.

Circularity Check

0 steps flagged

No circularity: the NEDT gain is a forward simulation from FDTD-derived CE and independently sourced detector parameters; the main weakness is an external-validity approximation, not a definitional/fitting loop.

full rationale

The derivation chain is a forward model. CE and CF are computed directly from FDTD simulations (Eqs. 1–3), with plane-wave illumination and no fitting to the final NEDT numbers. The NEDT model uses the radiometric transfer relation Φ_SPB = Φ_m · T_m · CE (Eq. 9) and the stochastic detector model μ = (SDE·λ_in + λ_dark)·t_r (Eq. 5). The parameters λ_dark = 3.9341 Mcps and SDE = 2.5×10^-8 are taken from prior SPB work (Refs. [4], [17]); they are inputs, not quantities predicted here, and they do not depend on the microlens geometry or on the claimed 30 mK → 10 mK improvement. No equation in the paper reduces to its own input by construction: the microlens design rules in Fig. 4 are direct simulation outputs, and Fig. 5 is obtained by substituting those outputs into the NEDT definition. The self-citations in Refs. [4] and [17] supply detector constants and the Poisson counting model, but the microlens result is not a restatement of those works. The plane-wave assumption and the small-opening-angle / Abbe-sine-condition approximation are potential sources of quantitative error for an f/0.5-class main lens, but that is a modeling-validity concern, not circular reasoning. Therefore no specific circular step can be quoted, and the appropriate score is 0.

Axiom & Free-Parameter Ledger

3 free parameters · 5 axioms · 0 invented entities

The central NEDT claim depends on plane-wave FDTD CE, the Poisson bolometer model and detector constants from prior same-group papers, and radiative-transfer approximations. Choices of sag, active size, and main-lens f/# affect the quantitative results. No new physical entities are postulated.

free parameters (3)
  • Microlens sag S = 4 µm (Fig. 2), 6 µm (Table 2), 2 µm design/1.79 µm measured (Fig. 3)
    Chosen by hand, not optimized; affects focal length, spot size, CE, and NEDT. Values are inconsistent across figures.
  • Sensor active size = 10 µm for NEDT; 0–30 µm scan in Fig. 4
    Fixed to SPB active area from prior work; determines CE/CF and NEDT results.
  • Main lens f/# = 0.5 (Table 2); implied ≈1.39 from Zobj=30 cm, Zimg=4 cm
    Chosen for simulation; affects Ω_m and absolute NEDT. Internal inconsistency with thin-lens equation.
axioms (5)
  • domain assumption Plane-wave illumination is assumed for all FDTD simulations of microlens CE and focal spot.
    Entered in Fig. 2 and used in Eq. (9) via CE; real imaging uses a converging cone from the main lens, so CE may differ.
  • domain assumption Radiative transfer model: blackbody is Lambertian, main lens is lossless, small opening angles, Abbe sine condition.
    Stated before Eqs. (8)–(9); f/0.5 in Table 2 is not a small opening angle, making this approximation questionable.
  • domain assumption Poisson bolometer response µ = (SDE·λ_in + λ_dark)·t_r with Poissonian counts from [17].
    Borrowed SPB stochastic model; SDE and λ_dark come from prior same-group papers [4,17].
  • standard math Main-lens imaging obeys the thin-lens equation with Zobj=30 cm and Zimg=4 cm.
    Used to place the MLA at the image plane; inconsistent with the f/0.5 listed in Table 2.
  • domain assumption Al2O3 transmission is ~80% (material) and fabricated MLA ~60% as measured by FTIR.
    Used in the NEDT model for T=100% and T=60% cases.

pith-pipeline@v1.3.0-alltime-deepseek · 5120 in / 13898 out tokens · 130201 ms · 2026-08-03T16:36:31.251740+00:00 · methodology

0 comments
read the original abstract

Infrared (IR) detectors are widely used due to their ability to sense thermal radiation. Recently, a room-temperature infrared detector known as the spintronic Poisson bolometer was introduced. While offering fast digital readout, its sensitivity is currently limited by its small photosensitive area and array fill factor. Beyond this specific detector, many emerging detector architectures also require substantial in-pixel electronics or engineering tradeoffs, which can reduce the fill factor and degrade optical coupling efficiency. In this work, we present design guidelines for spherical plano-convex microlens arrays that enhance light collection in spintronic Poisson bolometer arrays in the mid-wave infrared (MWIR). Guided by the simulations, we fabricate a microlens array sample to demonstrate that the chosen geometrical parameters are experimentally attainable and compatible with the fabrication process. We then systematically explore the scaling laws between sensor active size, microlens geometries, and the resulting light collection metrics, yielding practical design rules that are broadly relevant to MWIR detector arrays with limited active area. A radiometric-stochastic model is used to quantify the resulting sensitivity improvements for the spintronic Poisson bolometer. Our work is the first systematic simulation of a microlens design with spintronic Poisson bolometer arrays, bridging microphotonics, spintronics, and thermal imaging.

Figures

Figures reproduced from arXiv: 2512.12491 by Leif Bauer, Ziyi Yang, Zubin Jacob.

Figure 1
Figure 1. Figure 1: Schematic of the imaging system and spintronic Pois￾son bolometer device structure. research aims to increase the effective fill factor of sensors. Usu￾ally, plane-wave illumination with a collimated laser is used for experiments, and there is only one pixel behind each microlens [7, 10]. Imaging-level MLA research is mostly in the visible spec￾trum, which includes MLAs for high-resolution imaging without … view at source ↗
Figure 2
Figure 2. Figure 2: (a) Plano-convex spherical microlens geometry. (b) E-field distribution. The white dashed line marks the focal length. (c) Profile at the focus. (d) Focal length versus mi￾crolens diameter. (e) Spot size versus microlens diameter. the simulated geometries using the thermal reflow method [15] [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 4
Figure 4. Figure 4: (a) Collection efficiency and (b) concentration factor of Al2O3 spherical plano-convex microlenses in MWIR, as a function of microlens diameter and sensor active size. µ = (SDE ∗ λin + λdark) ∗ tr (5) where SDE = λph/λin is the system detection efficiency, λph is the detected count rate, λin is the incident photon rate, λdark is the dark count rate, and tr is the detector response time. The total detected … view at source ↗
Figure 5
Figure 5. Figure 5: Full-image simulation. (a) Imaging settings. (b) NEDT versus microlens diameter. T is the transmission of the mi￾crolenses. (c) A better contrast is shown with a smaller NEDT. We simulate when the MLA is placed at the image plane of the main lens. Therefore, the object distance Zobj between the blackbody and the main lens, and the image distance Zimg between the main lens and the MLA are determined by the … view at source ↗

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

19 extracted references

  1. [1]

    F . Bao, X. Wang, S. H. Sureshbabu,et al., Nature619(2023)

    2023

  2. [2]

    Medical applications of infrared thermography: A review,

    B. B. Lahiri, S. Bagavathiappan, T. Jayakumar, and J. Philip, “Medical applications of infrared thermography: A review,” (2012)

    2012

  3. [3]

    Optical communication in space: Chal- lenges and mitigation techniques,

    H. Kaushal and G. Kaddoum, “Optical communication in space: Chal- lenges and mitigation techniques,” (2017)

    2017

  4. [4]

    Bauer, A

    L. Bauer, A. Deka, M. A. Mousa,et al., Nano Lett.25, 5599 (2025)

    2025

  5. [5]

    M. F . Rashman,T errestrial and astronomical applications of uncooled infrared technology(Liverpool John Moores University (United King- dom), 2020)

    2020

  6. [6]

    High-speed 3d shape measure- ment of transparent objects by sequential thermal fringe projection and image acquisition in the long-wave infrared,

    M. Landmann, H. Speck, Z. Gao,et al., “High-speed 3d shape measure- ment of transparent objects by sequential thermal fringe projection and image acquisition in the long-wave infrared,” inThermosense: Thermal Infrared Applications XL V ,, vol. 12536 (SPIE, 2023), pp. 162–172

    2023

  7. [7]

    Bruschini, I

    C. Bruschini, I. M. Antolovic, F . Zanella,et al., Opt. Express31, 21935 (2023)

    2023

  8. [8]

    B. G. Oripov, D. S. Rampini, J. Allmaras,et al., Nature622(2023)

    2023

  9. [9]

    E. E. Wollman, V. B. Verma, A. E. Lita,et al., Opt. Express27(2019)

    2019

  10. [10]

    C. M. Kang, S. Bianconi, T. Hamilton,et al., ACS Appl. Electron. Mater. 4(2022)

    2022

  11. [11]

    Z. Liu, G. Hu, H. Y e,et al., Light. Sci. Appl.12(2023)

    2023

  12. [12]

    Light field photography with a hand- held plenoptic camera,

    R. Ng, M. Levoy, M. Brédif,et al., “Light field photography with a hand- held plenoptic camera,” Ph.D. thesis, Stanford University (2005)

    2005

  13. [13]

    The focused plenoptic camera,

    A. Lumsdaine and T. Georgiev, “The focused plenoptic camera,” in 2009 IEEE International Conference on Computational Photography (ICCP),(2009), pp. 1–8

    2009

  14. [14]

    Microlenses arrays: Fabrication, materi- als, and applications,

    S. Cai, Y . Sun, H. Chu,et al., “Microlenses arrays: Fabrication, materi- als, and applications,” (2021)

    2021

  15. [15]

    Nussbaum, R

    P . Nussbaum, R. Völkel, H. P . Herzig,et al., Pure Appl. Opt. J. Eur. Opt. Soc. Part A6, 617 (1997)

    1997

  16. [16]

    Rogalski,Infrared and terahertz detectors(CRC press, 2019)

    A. Rogalski,Infrared and terahertz detectors(CRC press, 2019)

    2019

  17. [17]

    Infrared digital-mode poissonian bolometer,

    L. Bauer, “Infrared digital-mode poissonian bolometer,” Ph.D. thesis, Purdue University (2024)

    2024

  18. [18]

    A. C. Parr, R. Datla, and J. Gardner,Optical radiometry(Elsevier, 2005)

    2005

  19. [19]

    M. A. Mousa, U. Singh, L. Bauer,et al., Phys. Rev. Appl.24, 064044 (2025)

    2025