REVIEW 3 major objections 5 minor 27 references
Nonlocal Metasurface Lens for Long-Wavelength Infrared Radiation
T0 review · 3 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash
Pith's one-line read A 1.45-micron germanium film focuses 10.3-micron infrared light, using a square-lattice resonance whose phase stays fixed as the pattern is tuned.
desk verdict A credible experimental demonstration of an LWIR nonlocal metalens with a genuinely new square-lattice meta-unit, but the headline geometric-phase stability claim is supported only indirectly. 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 a quasi-bound state in the continuum in a partially etched, high-index-contrast photonic-crystal slab. The meta-unit is a square lattice of unperturbed crosses with displaced crosses at interstitial sites; the displacement vector (delta_1, delta_2) controls the mode's coupling strength, while its direction delta controls the radiation polarization angle and hence the geometric phase. The paper classifies the mode as a TE q-BIC with B2 irreducible representation at the X point of the square lattice, whose p1 parent-group construction gives Phi approximately 2 delta. A fixed radius delta_0 = $\sqrt$($delta_1^{2}$ + $delta_2^{2}$) keeps Q and the resonance wavelength constant while the phase is spatially varied.
What would settle it
Measure single meta-units with varying displacement direction delta at fixed delta_0 to see whether the resonant wavelength and Q stay constant while the output polarization angle rotates by delta; alternatively, resolve the transmitted focal spot with a true circular-polarization analyzer to confirm that the focused light is left-circularly polarized rather than merely linearly polarized.
Extended reading notes
Core claim
The central claim is that a square lattice of crosses, with smaller crosses displaced by amounts delta_1 and delta_2 at interstitial sites, supports a quasi-bound state in the continuum whose resonance wavelength and Q-factor stay constant as the displacement direction delta = atan2(delta_2, delta_1) rotates, while the resonant geometric phase follows Phi approximately 2 delta. Because lambda and Q remain fixed, a hyperboloidal phase profile can be written into the device purely by patterning displacement directions, with no compensating changes to the meta-unit library. Fabricated in germanium on zinc selenide, the lens focuses the converted circular-polarization signal at 10.31 microns; the paper reports a focal spot within a ~400 nm band, a focal-spot integrated intensity of 17.5% of the transmitted signal, and an estimated RCP-to-LCP focusing efficiency of about 4%.
Load-bearing premise
The whole phase-encoding scheme rests on the fabricated partial-etch perturbation actually realizing the B2 symmetry at the X point of the square lattice, with geometric phase Phi approximately 2 delta; if the real structure's symmetry class differs from that classification, the measured focus would not follow from the design.
Editorial extensions
If this is right
- Long-wave infrared metalenses can be made about 1.45 microns thick rather than roughly 10 microns, which is compatible with optical lithography and large-area manufacturing.
- Because the design is rooted in symmetry rather than in a specific material, the same square-lattice cross platform could be transferred to visible and short-wave infrared wavelengths.
- The near-isotropic dispersion makes radial nonlocal lenses practical, avoiding the direction-dependent astigmatism seen in rectangular-lattice dimer designs.
- A device that is simultaneously spectrally selective and focusing can act as a narrowband filter and lens in one compact component, relevant to thermal imaging and chemical fingerprinting in the LWIR.
- Adding layers or engineered chirality could push conversion efficiency beyond the roughly 25% per-port limit of the single-layer geometric-phase scheme.
Reading between the lines
- An untested extension implied by the symmetry argument is that other lattice symmetries, such as hexagonal, should give even more isotropic dispersion and a distinct generalized geometric phase; the paper mentions this possibility but does not demonstrate it.
- A single-meta-unit measurement of resonant wavelength, Q, and output polarization as a function of delta would separate the symmetry classification assumption from the lens-integrated demonstration.
- The 4% efficiency estimate assumes that the unconverted background is not circularly polarized; a full Stokes or circular-polarization-resolved measurement could revise the efficiency up or down.
Signed reviews
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper proposes and experimentally demonstrates a nonlocal metasurface lens for long-wavelength infrared radiation, operating near 10.3 µm with a 1.45 µm thick amorphous germanium film on a zinc selenide substrate. The central design is a square lattice of unperturbed crosses with perturbed crosses at interstitial sites; the displacement vector (δ1, δ2) controls the q-BIC resonance, while the angle δ is claimed to impart a geometric phase Φ ≈ 2δ on the converted circular polarization. The authors fabricate a 900 µm square metalens with a hyperboloidal phase profile and characterize it using two single-QWP configurations: one with circularly polarized input and linear output analysis, and one with linearly polarized input and circularly polarized output resolution. They report a focal spot for RCP input at 10.31 µm over a ~400 nm bandwidth, a focal spot in the LCP-resolved output channel under linear input, no corresponding RCP-resolved focus, low chromatic aberration, and an estimated RCP-to-LCP focusing efficiency of ~4%. The paper positions the device as an ultrathin, polarization- and spectrally-selective LWIR lens, enabled by a square-lattice meta-unit with isotropic dispersion and a resonance wavelength stable against the phase-encoding perturbation.
Significance. If validated, this result is significant: it extends nonlocal, q-BIC-based metasurface optics into the LWIR with a deeply subwavelength device thickness, and it introduces a square-lattice meta-unit whose resonant geometric phase does not require frequency re-adjustment, addressing a known limitation of rectangular dimer designs. The experimental work includes full-wave simulations, fabricated devices, wavelength-resolved focusing scans, and a circular-polarization-resolved measurement showing a focus in the expected LCP output but not in RCP. The authors also explicitly state the main characterization limitations, which is helpful. The main risk lies in the indirect validation of the phase-encoding law, which is central to the lensing claim.
major comments (3)
- [Results – Design; Appendix] The phase-encoding relation Φ ≈ 2δ is assumed from the symmetry classification of the q-BIC as a B2 mode at the X point of a square lattice with p1 symmetry, taken from Ref. [2], and is not directly verified for the fabricated geometry. The fabricated meta-unit has a partial etch depth (0.87 µm in a 1.45 µm film) and finite rod shapes, while the Appendix classification is derived for an idealized perturbation. Figure 2b–d shows simulated phase Φ(δ1, δ2) for the ideal geometry, but no single-meta-unit phase retrieval or equivalent simulation for the actual fabricated profile is provided. Since the focal spot is the only experimental evidence for the assumed Φ ≈ 2δ law, a deviation of the real meta-units from the assumed symmetry would scramble the encoded phase profile and lower the efficiency; the reported ~4% efficiency cannot distinguish such phase errors from ordinary losses. Please provide a direct validation, e.g., interferometric or Fourier-plane phase measurement of the fabricated meta-units, or full-wave simulation of the exact fabricated geometry demonstrating Φ ≈ 2δ across the δ and wavelength ranges used.
- [Results – Characterization] The claim of RCP-to-LCP conversion and focusing is inferred from two complementary single-QWP configurations rather than from a single measurement with circular input and circular-resolved output. The paper itself states, after the linear-polarizer analysis, that "a final proof requires to resolve the output with a QWP," and the CP-resolved configuration in Fig. 3g–k uses linearly polarized input. The observation of a focus in the LCP output under linear input is strong evidence, but it does not by itself exclude the possibility that the LCP input component also contributes to focusing if the actual device response differs from the idealized design. Given that both the focal-spot fraction (17.5%) and the efficiency estimate (~4%) depend on assigning the focused LCP signal to the RCP input component, a direct two-QWP measurement—or an equivalent polarimetric decomposition of the output for circular input—would substantiate the central conversion claim.
- [Results – Characterization] The RCP-to-LCP focusing efficiency of ~4% is presented without uncertainty and without a transparent accounting of the measurement conditions. The derivation uses ρ ≈ 12%, a focal-spot fraction of 17.5% of total intensity, and a factor of two from the wasted LCP input component; the text does not state at which wavelength and polarizer angle these quantities are evaluated, how the background is defined, or what systematic and statistical errors are associated with the integrated intensities. Since the efficiency is a headline quantitative result of the demonstration, please report uncertainties and a precise definition of each ratio, or downgrade the claim to an order-of-magnitude estimate.
minor comments (5)
- [Figure 3 caption] The caption appears to label two panels with "(i)": one as "corresponding 1D linecut" and one as "Longitudinal 2D far-field scans." Please renumber the panels consistently and correct the cross-references in the text.
- [Results – Characterization] The text refers to "dashed line in Fig. 3i" when discussing chromatic aberration, but the longitudinal scans appear to be in Fig. 3k; please fix the reference.
- [Methods] In the optical characterization paragraph, the resonance wavelength is printed as "10.31mm" and should be "10.31 µm."
- [Results – Characterization] The conversion-efficiency paragraph derives the ~2% value as 17.5% of the ~12% LCP-output ratio; this assumes that the focal-spot fraction of the total transmitted intensity equals the focal-spot fraction of the LCP component. Please state this assumption explicitly.
- [Results – Characterization] The reflectance spectrum in Fig. 3a is described as exhibiting a q-BIC with Q ≈ 100, but the measurement is unpolarized and the Fano line is "significantly smoothed"; please report an uncertainty on the extracted Q or describe the fitting procedure used.
Circularity Check
No significant circularity: the central result is an experimental metalens demonstration benchmarked against external measurements; the geometric-phase rule is a cited design input, not a fitted prediction.
full rationale
Walking the derivation chain, the metalens phase profile is designed from the geometric-phase relation Φ≈2δ, which is imported from prior selection-rule theory (Ref. [2]) rather than re-derived in this paper. This is not circular because the paper's central claim is an experimental realization tested against externally observable behavior: a focal spot at 10.31 μm for RCP incidence, no focusing for LCP incidence, a ~400-nm operating bandwidth, and a measured chromatic focal shift below 500 μm. The phase rule is not fitted to the focal spot, nor is the focal-spot outcome defined in terms of the design rule. The Q∝1/δ0^2 scaling is used as a design guide and independently reproduced by full-wave simulations (Fig. 2e), not fitted to the final device efficiency. The estimated RCP-to-LCP focusing efficiency of ~4% is arithmetic from measured integrated intensities (η_F=17.5%, ρ≈12%) with explicit corrections for the unconverted background and the wasted LCP input component; no parameter is adjusted to force a target value. The paper does rely on self-citations (Refs. [2,3,4]) for q-BIC selection rules and the nonlocal-metasurface concept, and the Appendix classifies the q-BIC as a B2 mode at the X point using Ref. [2]. However, these are peer-reviewed, parameter-free symmetry rules with stated assumptions that do not include the present lens result, so they constitute external support rather than a self-supporting chain. The paper itself acknowledges one experimental verification gap: 'a final proof requires to resolve the output with a QWP' (Results - Characterization). That is a limitation of the polarization evidence, not a definitional or fitted-input circularity. No equation in the paper makes the predicted output equivalent by construction to the design input, and no fitted parameter is renamed as a prediction. Verdict: no significant circularity.
Assumptions & free parameters
free parameters (3)
- Perturbation magnitude delta0 =
0.4 micrometers
- Unit-cell dimensions P, L, W, D =
P = 3.15 micrometers, L = 2.1 micrometers, W = 0.65 micrometers, D = 1.1 micrometers
- Film thickness and etch depth =
1.45 micrometers total film, 0.87 micrometers etch depth
assumptions (4)
- domain assumption The q-BIC selection rules of Ref. [2] apply to the B2 mode at the X point of the square lattice, so the perturbation vector (delta1, delta2) controls radiative coupling and polarization properties.
- domain assumption The geometric phase for a p1-symmetric meta-unit satisfies Phi approximately 2*delta, while the conventional p2 dimer satisfies Phi approximately 4*alpha.
- domain assumption Amorphous germanium and zinc selenide have sufficiently low absorption at 10.3 micrometers and a high enough index contrast to support the designed q-BIC.
- domain assumption The Q-factor scaling Q proportional to 1/delta0^2 holds for the perturbation range used in the fabricated device.
Cite this review
Pith. "Pith review of Nonlocal Metasurface Lens for Long-Wavelength Infrared Radiation." pith.science (2026). https://pith.science/paper/LDJYK4LL
@misc{pith2026250504856,
author = {Pith},
title = {Pith review of: Nonlocal Metasurface Lens for Long-Wavelength Infrared Radiation},
year = {2026},
howpublished = {\url{https://pith.science/paper/LDJYK4LL}},
note = {Machine review of arXiv:2505.04856}
}
read the original abstract
Dielectric metasurfaces are structured thin films with thickness smaller than the wavelength that aim at replacing and enhancing conventional bulk optical components by structuring local resonances across an aperture. At visible and near-infrared frequencies, titania or silicon are routinely used as substrates to realize these ultrathin devices, ideally suited for conventional nanofabrication techniques. Unfortunately, directly scaling these design and material approaches to long-wave infrared frequencies is not practical, due to challenges in the required thicknesses and the presence of phonon absorption lines. Nonlocal metasurfaces based on extended resonances with a local geometric phase offer a compelling design platform that can address these challenges. They enable ultrathin metasurfaces, as they leverage lattice resonances, while they also offer multi-functionalities and frequency-selectivity, and they can be implemented in a range of low-loss material platforms. Here, we demonstrate nonlocal metalenses based on germanium thin films on a zinc-selenide substrate, operating around 10.3{\mu}m within a deeply subwavelength device thickness of 1.45{\mu}m (14% the free-space wavelength). We showcase a novel meta-unit geometry based on a square lattice with highly isotropic dispersion features, supporting a resonant geometric phase that is highly stable in frequency, simplifying the rational design of complex metasurface operations. The introduced platform promises highly multi-functional, low-profile meta-optics with enhanced meta-unit designs, compatible with the challenging thermal spectral region for imaging and sensing applications.
Reference graph
Works this paper leans on
-
[2]
Overvig, A. C., Malek, S. C., Carter, M. J., Shrestha, S., & Yu, N. (2020). Selection rules for quasibound states in the continuum. Physical Review B, 102(3), 035434
work page 2020
-
[1]
Overvig, A. and Alù, A. Diffractive nonlocal metasurfaces. Laser Photonics Rev. 16, 2100633 (2022)
work page 2022
-
[3]
Multifunctional nonlocal metasurfaces
Overvig, A. C., S. C. Malek, and N. Yu. "Multifunctional nonlocal metasurfaces." Physical Review Letters 125.1 (2020): 017402
work page 2020
-
[4]
Malek, S. C. et al. Multifunctional resonant wavefront -shaping meta-optics based on multilayer and multi-perturbation nonlocal metasurfaces. Light Sci. Appl. 11, 246 (2022)
work page 2022
-
[5]
Shastri, K. and Monticone, F. Nonlocal flat optics. Nat. Photonics 17, 36–47 (2023)
work page 2023
-
[6]
Overvig, A.C., et al. "Zone-folded quasi-bound state metasurfaces with customized, symmetry - protected energy-momentum relations." ACS Photonics 10.6 (2023): 1832-1840
work page 2023
-
[7]
Song, J. H. et al. Nonlocal metasurfaces for spectrally decoupled wavefront manipulation and eye tracking. Nat. Nanotechnol. 16, 1224–30 (2021)
work page 2021
-
[8]
Nolen, J.R., Overvig, A.C., Cotrufo, M. et al. Local control of polarization and geometric phase in thermal metasurfaces. Nat. Nanotechnol. 19, 1627 –1634 (2024). https://doi.org/10.1038/s41565- 024-01763-6
doi:10.1038/s41565- 2024
Show all 27 references
-
[9]
III et al
Barton, D. III et al. High-Q nanophotonics: sculpting wavefronts with slow light. Nanophotonics 10, 83–88 (2020)
2020
-
[10]
Lawrence, M. et al. High-quality factor phase gradient metasurfaces. Nat. Nanotechnol. 15, 956 – 61 (2020)
2020
-
[11]
Lin, L. et al. Universal narrowband wavefront shaping with high quality factor meta-reflect-arrays. Nano Lett. 23, 1355–62 (2023). Distribution A: Approved for public release - distribution unlimited. (AFRL/RW 2025-2029)
2023
-
[12]
U., Foley, M., Sokhoyan, R., Michaeli, L., & Atwater, H
Hail, C. U., Foley, M., Sokhoyan, R., Michaeli, L., & Atwater, H. A. (2023). High quality factor metasurfaces for two-dimensional wavefront manipulation. Nature Communications, 14(1), 8476
2023
-
[13]
F., and Alù, A
Overvig, A., Yu, N. F., and Alù, A. Chiral quasi -bound states in the continuum. Phys. Rev. Lett. 126, 073001 (2021)
2021
-
[14]
Overvig, A. C. , and Alù, A. Wavefront -selective fano resonant metasurfaces. Adv. Photonics 3, 026002 (2021)
2021
-
[15]
Active nonlocal metasurfaces
Malek, Stephanie C., Overvig, Adam C., Shrestha, Sajan and Yu, Nanfang. "Active nonlocal metasurfaces" Nanophotonics, vol. 10, no. 1, 2021, pp. 655 -665. https://doi.org/10.1515/nanoph- 2020-0375
2021 doi
-
[16]
C., Tsai, C
Malek, S. C., Tsai, C. C., and Yu, N. (2025). Thermally‐Switchable Metalenses Based on Quasi‐ Bound States in the Continuum. Laser & Photonics Reviews, 19(3), 2300618
2025
-
[17]
Improved refractive -index sensing performance in medium contrast gratings by asymmetry engineering,
Hardik Vyas and Ravi S. Hegde, "Improved refractive -index sensing performance in medium contrast gratings by asymmetry engineering," Opt. Mater. Express 10, 1616-1629 (2020)
2020
-
[18]
Jiaxin Zhou, Yuefeng Wang, Meng Xia, Yuhua Chen, Di Huang, and Xingwang Zhang, Nano Letters 2024 24 (31), 9658-9665 DOI: 10.1021/acs.nanolett.4c02439
2024 doi
-
[19]
Jiaxin Zhou, Meng Xia, Yuhua Chen, and Xingwang Zhang ACS Photonics 2024 11 (7), 2707 - 2712 DOI: 10.1021/acsphotonics.4c00553
2024 doi
-
[20]
Quasi - bound states in the continuum-based switchable light-field manipulator,
Run Chen, Tianyue Li, Qianhui Bi, Shuming Wang, Shining Zhu, and Zhenlin Wang, "Quasi - bound states in the continuum-based switchable light-field manipulator," Opt. Mater. Express 12, 1232-1241 (2022)
2022
-
[21]
Broadband transparent and high -Q resonant polarization meta-grating enabled by a non-local geometric-phase metasurface,
Di Sang, Mingfeng Xu, Qiang An, and Yunqi Fu, "Broadband transparent and high -Q resonant polarization meta-grating enabled by a non-local geometric-phase metasurface," Opt. Express 30, 26664-26675 (2022)
2022
-
[22]
Chen, R., Bi, Q., Li, T., Wang, S., Zhu, S., & Wang, Z. (2023). Dual-wavelength chiral metasurfaces based on quasi-bound states in the continuum. Journal of optics, 25(4), 045001
2023
-
[23]
& Huang, L
You, S., Zhou, M., Xu, L., Chen, D., Fan, M., Huang, J., Ma, W., Luo, S., Rahmani, M., Zhou, C., Miroshnichenko, A. & Huang, L. (2023). Quasi -bound states in the continuum with a stable resonance wavelength in dimer dielectric metasurfaces. Nanophotonics, 12(11), 2051 -2060. ...
2023 doi
-
[24]
Wu, X., Xiong, D., Liu, G., Wu, Y ., Yun, M., Chen, D., & Qi, X. (2024). Realizing quasi-bound states in the continuum with stable resonant wavelength through compensation mechanism. Results in Physics, 58, 107505
2024
-
[25]
F., Akozbek, N.,
Huang, L., Coppens, Z., Hallman, K., Han, Z., Böhringer, K. F., Akozbek, N., ... & Majumdar, A. (2021). Long wavelength infrared imaging under ambient thermal radiation via an all -silicon metalens. Optical Materials Express, 11(9), 2907-2914
2021
-
[26]
Aberration-free ultrathin flat lenses and axicons at telecom wavelengths based on plasmonic metasurfaces
Aieta, Francesco, et al. "Aberration-free ultrathin flat lenses and axicons at telecom wavelengths based on plasmonic metasurfaces." Nano letters 12.9 (2012): 4932-4936
2012
-
[27]
& Luo, X
Xie, X., Pu, M., Jin, J., Xu, M., Guo, Y ., Li, X., ... & Luo, X. (2021). Generalized Pancharatnam- Berry phase in rotationally symmetric meta-atoms. Physical Review Letters, 126(18), 183902
2021
Reviewed August 15, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.