Pith. sign in

REVIEW 4 major objections 5 minor 42 references

Implementation of on-chip multi-channel focusing wavelength demultiplexing with regularized digital metamaterials

T0 review · 4 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash

Pith's one-line read A fabricated two-channel wavelength demultiplexer with a 2.4×10 μm² footprint separates 1520 nm and 1580 nm light with measured peak insertion loss of −1.77 dB and −2.10 dB.

desk verdict A genuinely new focusing WDM geometry with promising measured performance, but the missing grating-coupler calibration means the headline insertion-loss and crosstalk numbers aren't yet trustworthy. read the letter →

arxiv 1909.11136 v1 pith:NPPPGGKM submitted 2019-08-10 physics.app-ph physics.optics

classification physics.app-phphysics.optics
keywords siliconphotonicswavelengthdemultiplexerdigitalmetamaterialsinversedesignobjectivefirstmethodfabricationtolerancefocuseddeviceson-chipdivisionmultiplexing
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

This paper reports a two-channel on-chip wavelength demultiplexer, built from regularized digital metamaterials, that fits in 2.4 by 10 square microns and needs no long adiabatic taper. The design takes 1520 nm and 1580 nm light from a 10-micron-wide input waveguide and routes each wavelength into its own 0.48-micron-wide output waveguide. The fabricated device shows measured peak transmissions of −1.77 dB and −2.10 dB with crosstalk of −25.17 dB and −12.14 dB. The paper also simulates a three-channel version, arguing that the design method extends to other compact 'focused' devices. A sympathetic reader would care because this points to dense photonic circuits where wavelength splitting occupies a few tens of square microns rather than hundreds of microns of taper plus device area.

What carries the argument

The load-bearing mechanism is a regularized digital metamaterial: a design region discretized into silicon/air pixels whose permittivity distribution is optimized to perform mode conversion. The optimization is a two-stage pipeline: an objective-first stage with ADMM produces a continuous permittivity distribution that largely satisfies the design targets, and a modified steepest descent stage applies a projection scheme to binarize the pixels and a low-pass filter enforcing a minimum feature size (here R=3, equivalent to 120 nm). The device performance is defined by overlap integrals between input and output modes, so the optimizer directly maximizes the power routed from the 10-μm input mode to the fundamental TE modes of the two output waveguides at their target wavelengths.

What would settle it

Re-measure the fabricated chip against an on-chip reference: a straight 10-micron-wide waveguide with the same grating couplers at both ends. If the difference between demultiplexer transmission and reference transmission does not reproduce the reported −1.77 dB and −2.10 dB peaks with the stated crosstalk, then the missing grating-coupler calibration is the cause and the device-level numbers need revision.

Watch

Extended reading notes

Core claim

The central claim is that inverse-designed digital metamaterials can be made both fabricable and high-performing if the optimization enforces binarization and a minimum feature size, and that this enables a 'focused' geometry in which a wide input waveguide is demultiplexed directly into narrow output waveguides. The demonstrated two-channel demultiplexer splits 1520 nm and 1580 nm light from a 10-μm input into two 0.48-μm outputs, with measured peak transmission of −1.77 dB in the upper channel and −2.10 dB in the lower channel, and crosstalk of −25.17 dB and −12.14 dB. The paper attributes the gap between simulated and measured spectra to etching imperfections, especially under-etching of small holes, and reports fabrication-tolerance simulations that allow ±20 nm in etch depth and ±10 nm in plane size. It also presents a simulated three-channel focused demultiplexer with crosstalk below −17 dB.

Load-bearing premise

The reported insertion-loss and crosstalk values are only trustworthy as device-level figures if the transmission of the input and output grating couplers was measured and divided out, and the characterization section does not describe such a normalization.

Editorial extensions

If this is right

  • If the measured performance is accurate, a wavelength demultiplexer can occupy about 24 square microns and attach directly to a 10-micron grating coupler, eliminating the hundreds-of-microns adiabatic taper normally needed.
  • The same regularized optimization pipeline should be reusable for other compact 'focused' components, such as mode splitters, mode converters, and power splitters.
  • Enforcing a 120-nm minimum feature size makes the inverse-designed pattern compatible with standard fabrication, which matters for moving such devices from simulation to production.
  • The simulated three-channel device suggests the approach is not limited to two wavelengths, although no three-channel fabrication result is reported.

Reading between the lines

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

  • Editorial inference: the paper does not describe a reference calibration for the grating couplers used in the characterization, so the reported −1.77/−2.10 dB values should be read as device-plus-coupling numbers until such a normalization is shown; the true on-chip loss could be lower or higher depending on the couplers' behavior.
  • Editorial inference: the three-channel device is promising, but its real test is fabrication; a fabricated three-channel version with the same calibration would directly show whether the objective-first initialization remains effective as the number of output ports grows.
  • Editorial inference: the etching-depth tolerance analysis assumes all holes vary uniformly; a more demanding test would vary hole sizes independently, which is closer to what real lithography and etching produce.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 5 minor

Summary. The manuscript reports the inverse design, fabrication, and characterization of a two-channel focusing wavelength demultiplexer based on regularized digital metamaterials. The design pipeline first uses the objective-first method to obtain a continuous permittivity distribution and then applies a modified steepest descent method with projection and low-pass filtering to enforce binarization and a minimum feature size of 120 nm. The resulting 2.4 μm × 10 μm device is designed to split 1520 nm and 1580 nm light from a 10-μm-wide input waveguide into two 0.48-μm-wide output waveguides, eliminating the need for an adiabatic taper. The authors report measured insertion losses of -1.77 dB and -2.10 dB and crosstalk of -25.17 dB and -12.14 dB for the upper and lower channels, respectively, and a numerical tolerance analysis claiming robustness to ±20 nm etch-depth and ±10 nm plane-size variations. A three-channel demultiplexer is also designed and simulated but not fabricated.

Significance. If the experimental results are taken at face value, the paper demonstrates that regularized inverse-designed digital metamaterials can realize an ultracompact wavelength demultiplexer that directly interfaces with a grating coupler without a taper, which would be a useful contribution to high-density silicon photonics. The optimization pipeline, including the minimum feature size constraint and the combination of objective-first initialization with steepest-descent refinement, is technically interesting and the three-channel numerical extension indicates potential scalability. However, the experimental validation as presented is incomplete: the measured spectra are not normalized to any reference, so the quoted insertion loss and crosstalk values cannot currently be separated from the grating-coupler and source response. The numerical tolerance study is also based on a restrictive uniform-perturbation assumption.

major comments (4)
  1. [Methods, Characterization; Figure 4] The measured spectra in Figure 4(b) appear to be raw fiber-to-fiber transmission through the input grating coupler, the device, and the output grating couplers. The Characterization section states only that photonic crystal grating couplers were fabricated at each port for chip-fiber coupling; it does not describe any reference waveguide, cut-back set, or normalization to the supercontinuum source spectrum. Because both the source and the grating couplers are strongly wavelength-dependent, the quoted insertion losses (-1.77 dB and -2.10 dB) and crosstalk values (-25.17 dB and -12.14 dB) cannot be attributed to the demultiplexer alone. The authors should provide reference-normalized measurements (e.g., a straight waveguide with identical grating couplers) or clearly restate the reported values as fiber-to-fiber quantities. Without this, the central experimental claim is not supported.
  2. [Experimental Results and Analyses; Figure 4] The measured lower-channel crosstalk of -12.14 dB is more than 11 dB worse than the simulated value of -23.50 dB, whereas the upper channel changes by only about 1.5 dB. The paper attributes this discrepancy to RIE lag under-etching and to the fact that the 1565 nm field is spread throughout the design region, but this explanation is qualitative and no simulation of the estimated under-etched geometry is shown to reproduce the measured spectrum. To make the fabrication-imperfection argument load-bearing, the authors should include a simulated S-parameter response for an under-etched device (e.g., the 200 nm etch-depth case from Figure 5(a)) and directly compare the resulting lower-channel crosstalk with the measurement.
  3. [Optimization Model Construction; Figure 4(a)] The design targets in the optimization specify output powers in the range [0.9,1] and [0,0.01] for the two channels, but the final regularized device with R=3 (Figure 4(a)) exhibits simulated peak transmissions of -1.39 dB (≈72.6%) and -1.45 dB (≈71.6%), which are outside the stated [0.9,1] interval. The manuscript does not explain this discrepancy or quantify how much of the performance loss is introduced by the binarization and minimum-feature-size constraints. The authors should report the performance of the unregularized continuous design alongside the regularized design and clarify whether the constraints are soft targets.
  4. [Tolerance to Fabrication Errors; Figure 5] The robustness analysis assumes that the etch depth and the plane size of all air holes vary uniformly. In practical RIE lag, the etch depth depends on the local feature size and pattern density, so a uniform shift is not necessarily the worst case and may overstate the claimed ±20 nm/±10 nm tolerances. The paper should justify the uniform-variation model or simulate a representative non-uniform perturbation, such as smaller holes being under-etched more than larger ones, to support the conclusion that the device is robust to fabrication errors.
minor comments (5)
  1. [Equation (2)] The expression for the mode conversion efficiency appears garbled: the term '†c' is not defined and the integral notation is incomplete. Please rewrite this equation with clear definitions of all symbols.
  2. [Abstract] In the final sentence, 'Benefit From' should be 'Benefiting from' for grammatical correctness.
  3. [Experimental Results and Analyses] The sentence 'The image show that all of the designed air holes were precisely reproduced' contains a subject-verb agreement error; it should be 'The image shows'. More importantly, a quantitative overlay of the SEM image with the designed layout would substantiate this claim of precise reproduction.
  4. [References] Reference [41] is cited as 'Under Review'; if this work is unpublished, the authors should provide a preprint identifier or remove the reference and rely on the discussion in the text.
  5. [Figure 4] The measured spectra in Figure 4(b) appear to cover a limited wavelength range; the authors should specify the scan range, resolution, and the number of devices measured, and include error bars or at least state the measurement repeatability.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the fabricated device performance is measured independently of the optimization targets that generated the design.

full rationale

The paper's derivation chain is: define mode-conversion targets (Eq. 2), optimize the permittivity distribution by objective-first and modified steepest descent methods with binarization and regularization, simulate the resulting pattern with FDTD, fabricate the device, and independently measure its S-parameters. The measured insertion losses and crosstalk are not obtained by fitting any parameter that was used in the optimization; the simulation and measurement are separate evidence streams. The optimization targets—power in [0.9,1] for the desired channel and [0,0.01] for the undesired channel—are design constraints, not predictions extracted from data. Citations to the authors' prior work (e.g., ref. 31) are used as algorithmic background or comparison ('Different from our previous work'), not as the justification for the reported device performance. The 'focused' terminology denotes a structural property (direct connection to a 10-μm-wide grating coupler without an adiabatic taper) rather than a renamed known result. The fabrication-tolerance analysis is a forward simulation study and does not reduce to any fitted constant. The only notable weakness is in the Characterization section, which states that grating couplers were fabricated for chip-fiber coupling but does not describe a reference normalization, so the reported dB values may include coupler or source response. That is a measurement-verification concern, not a circularity in the derivation chain, and therefore does not raise the circularity score.

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

The central claim rests on the optimization and measurement being reliable. The listed free parameters are hand-chosen optimization hyperparameters that influence the final device structure but are not fitted to the measured data. The axioms are standard electromagnetic theory and domain assumptions about single-mode operation and the uniform-error model used in the tolerance study. No new physical entities are introduced.

free parameters (4)
  • Projection midpoint eta = 0.6
    Chosen by hand to control the binarization projection; affects the final device structure.
  • Projection strength beta schedule = increased from 5 to 100 for R=3 case
    Chosen by hand to gradually binarize the design; the schedule influences the optimized layout.
  • Minimum feature size R = 3 pixels (120 nm)
    Chosen as a trade-off between fabricability and simulated performance; R=4 gives worse transmission.
  • Initial permittivity parameter = 0.75 in design area
    Initial guess for optimization; influences the local optimum reached.
assumptions (4)
  • standard math Maxwell's equations in the frequency domain describe the device response (Eq. 1).
    The optimization and FDTD simulation rely on classical electromagnetism.
  • domain assumption A linear optical device can be fully characterized as a mode converter between input and output modes (Eq. 2).
    This enables the objective-first formulation; it is a standard, valid assumption for passive dielectric devices in the linear regime.
  • domain assumption The input to the design is the fundamental TE mode of the 10-um-wide waveguide and of the 0.48-um output waveguides.
    The focusing demux assumes single-mode operation of the wide input; higher-order modes would degrade the expected performance.
  • ad hoc to paper Fabrication errors can be modeled by uniform variation of all air holes in depth and plane size (Figure 5).
    The tolerance analysis applies the same etch-depth or size shift to every hole, ignoring non-uniform, position-dependent process variations like RIE lag.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Implementation of on-chip multi-channel focusing wavelength demultiplexing with regularized digital metamaterials." pith.science (2026). https://pith.science/paper/NPPPGGKM

@misc{pith2026190911136,
  author       = {Pith},
  title        = {Pith review of: Implementation of on-chip multi-channel focusing wavelength demultiplexing with regularized digital metamaterials},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/NPPPGGKM}},
  note         = {Machine review of arXiv:1909.11136}
}
read the original abstract

Adiabatic waveguide taper and on-chip wavelength demultiplexer are the key components of photonic integrated circuits. However, these two kinds of devices which designed by traditional semi-analytic methods or brute-force search methods usually have large size. Here, based on regularized digital metamaterials, we have designed, fabricated and characterized a two-channel focused wavelength demultiplexer with a footprint of 2.4 x 10 um2. The designed demultiplexer can directly connect to a grating coupler under the absence of an adiabatic waveguide taper. The objective first method and modified steepest descent method are used to design the demultiplexer which splits 1520 nm and 1580 nm light from a 10-um-wide input waveguide into two 0.48-um-wide output waveguides. Experimental results show that the insertion loss of the upper (lower) channel of the demultiplexer is -1.77 dB (-2.10 dB) and the crosstalk is -25.17 dB (-12.14 dB). Besides, the simulation results indicate that the fabrication tolerance of our devices can reach 20 nm in etching depth and 10 nm in plane size changing. Benefit From the extensibility of our design method, we can design other types of ultra-compact 'focused' devices, like mode splitters, mode converters and power splitters, and we can also design devices with more complicated functionalities, for example, we have designed a three-channel focused wavelength demultiplexer.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

42 extracted references · 42 canonical work pages

  1. [1]

    S.; Watts, M

    Sun, J.; Timurdogan, E.; Yaacobi, A.; Hosseini, E. S.; Watts, M. R. Large-Scale Nanophotonic Phased Array. Nature 2013, 493, 195-199

  2. [2]

    Skrzypczyk, P.; Salavrakos, A.; Tura, J.; Augusiak, R.; Mančinska, L.; Bacco, D.; et al

    Wang, J.; Paesani, S.; Ding, Y.; Santagati, R. ; Skrzypczyk, P.; Salavrakos, A.; Tura, J.; Augusiak, R.; Mančinska, L.; Bacco, D.; et al. Multidimensional Quantum Entanglement with Large-Scale Integrated Optics. Science 2018, 360, 285-291

  3. [3]

    J.; Kwon, K.; Henriksson, J.; Luo, J.; Wu, M

    Seok, T. J.; Kwon, K.; Henriksson, J.; Luo, J.; Wu, M. C. Wafer-Scale Silicon Photonic Switches Beyond Die Size Limit. Optica 2019, 6, 490-494

  4. [4]

    Nanophotonic Computational Design

    Lu, J.; Vučković, J. Nanophotonic Computational Design. Opt. Express 2013, 21, 13351 - 13367

  5. [5]

    Y.; Lu, J.; Lagoudakis, K

    Piggott, A. Y.; Lu, J.; Lagoudakis, K. G.; Petykiewicz, J.; Babinec, T. M.; Vučković, J. Inverse Design and Demonstration of a Compact and Broadband On-Chip Wavelength Demultiplexer. Nat. Photonics 2015, 9, 374-377

  6. [6]

    An Integrated-Nanophotonics Polarization Beamsplitter with 2.4 × 2.4 μm2 Footprint

    Shen, B.; Wang, P.; Polson, R.; Menon, R. An Integrated-Nanophotonics Polarization Beamsplitter with 2.4 × 2.4 μm2 Footprint. Nat. Photonics 2015, 9, 378-382

  7. [7]

    Increasing the Density of Passive Photonic-Integrated Circuits Via Nanophotonic Cloaking

    Shen, B.; Polson, R.; Menon, R. Increasing the Density of Passive Photonic-Integrated Circuits Via Nanophotonic Cloaking. Nat. Commun. 2016, 7, 13126

  8. [8]

    F.; Ding, Y.; Sigmund, O.; Frandsen, L

    Frellsen, L. F.; Ding, Y.; Sigmund, O.; Frandsen, L. H. Topology Optimized Mode Multiplexing in Silicon-On-Insulator Photonic Wire Waveguides. Opt. Express 2016, 24, 16866- 16873

Show all 42 references
  1. [9]

    M.; Bhargava, S.; Miller, O

    Lalau-Keraly, C. M.; Bhargava, S.; Miller, O. D.; Yablonovitch, E. Adjoint Shape Optimization Applied to Electromagnetic Design. Opt. Express 2013, 21, 21693-21701

  2. [10]

    Adjoint-Based Optimization of Active Nanophotonic Devices

    Wang, J.; Shi, Y.; Hughes, T.; Zhao, Z.; Fan, S. Adjoint-Based Optimization of Active Nanophotonic Devices. Opt. Express 2018, 26, 3236-3248

  3. [11]

    W.; Minkov, M.; Williamson, I

    Hughes, T. W.; Minkov, M.; Williamson, I. A. D.; Fan, S. Adjoint Method and Inverse Design for Nonlinear Nanophotonic Devices. ACS Photonics 2018, 5, 4781-4787

  4. [12]

    A.; Lin, L.; Yun, S.; Liu, L.; Werner, D

    Bossard, J. A.; Lin, L.; Yun, S.; Liu, L.; Werner, D. H.; Mayer, T. S. Near-Ideal Optical Metamaterial Absorbers with Super-Octave Bandwidth. ACS Nano 2014, 8, 1517-1524

  5. [13]

    Genetically Optimized On-Chip Wideband Ultracompact Reflectors and Fabry–Perot Cavities

    Yu, Z.; Cui, H.; Sun, X. Genetically Optimized On-Chip Wideband Ultracompact Reflectors and Fabry–Perot Cavities. Photonics Res. 2017, 5, B15-B19

  6. [14]

    C.; Sideris, C.; Jeong, J.; Hajimiri, A.; Poon, J

    Mak, J. C.; Sideris, C.; Jeong, J.; Hajimiri, A.; Poon, J. K. Binary Particle Swarm Optimized 2 x 2 Power Splitters in a Standard Foundry Silicon Photonic Platform. Opt. Lett. 2016, 41, 3868- 3871

  7. [15]

    Deep-Learning-Enabled On -Demand Design of Chiral Metamaterials

    Ma, W.; Cheng, F.; Liu, Y. Deep-Learning-Enabled On -Demand Design of Chiral Metamaterials. ACS Nano 2018, 12, 6326-6334

  8. [16]

    H.; Kojima, K.; Koike-Akino, T.; Jha, D.; Wang, B.; Lin, C.; Parsons, K

    Tahersima, M. H.; Kojima, K.; Koike-Akino, T.; Jha, D.; Wang, B.; Lin, C.; Parsons, K. Deep Neural Network Inverse Design of Integrated Photonic Power Splitters. Sci. Rep. 2019, 9, 1368

  9. [17]

    All-Dielectric Metamaterials

    Jahani, S.; Jacob, Z. All-Dielectric Metamaterials. Nat. Nanotechnol. 2016, 11, 23-36

  10. [18]

    Y.; Sapra, N

    Dory, C.; Vercruysse, D.; Yang, K. Y.; Sapra, N. V.; Rugar, A. E.; Sun, S.; Lukin, D. M.; Piggott, A. Y.; Zhang, J. L.; Radulaski, M. et al. Optimized Diamond Quantum Photonics. arXiv:1812.02287 2019

  11. [19]

    Liang, H.; Lin, Q.; Xie, X.; Sun, Q.; Wang, Y.; Zhou, L.; Liu, L.; Yu, X.; Zhou, J.; Krauss, T. F. et al. Ultrahigh Numerical Aperture Metalens at Visible Wavelengths. Nano Lett. 2018, 18, 4460-4466

  12. [20]

    Chung, H.; Miller, O. D. High-NA Achromatic Metalenses by Inverse Design. arXiv:1905.09213 2019

  13. [21]

    Integrated Metamaterials for Efficient and Compact Free-Space-To-Waveguide Coupling

    Shen, B.; Wang, P.; Polson, R.; Menon, R. Integrated Metamaterials for Efficient and Compact Free-Space-To-Waveguide Coupling. Opt. Express 2014, 22, 27175-27182

  14. [22]

    Ultra-Compact Polarization Rotation in Integrated Silicon Photonics Using Digital Metamaterials

    Majumder, A.; Shen, B.; Polson, R.; Menon, R. Ultra-Compact Polarization Rotation in Integrated Silicon Photonics Using Digital Metamaterials. Opt. Express 2017, 25, 19721-19731

  15. [23]

    Y.; Sapra, N

    Su, L.; Piggott, A. Y.; Sapra, N. V.; Petykiewicz, J.; Vučković, J. Inverse Design and Demonstration of a Compact On-Chip Narrowband Three-Channel Wavelength Demultiplexer. ACS Photonics 2017, 5, 301-305

  16. [24]

    Adiabatic and Ultracompact Waveguide Tapers Based on Digital Metamaterials

    Liu, Y.; Sun, W.; Xie, H.; Zhang, N.; Xu, K.; Yao, Y.; Xiao, S.; Song, Q. Adiabatic and Ultracompact Waveguide Tapers Based on Digital Metamaterials. IEEE J. Set. Top. Quant. 2019, 25, 1-6

  17. [25]

    Ultracompact Dual-Mode Waveguide Crossing Based on Subwavelength Multimode-Interference Couplers

    Chang, W.; Lu, L.; Ren, X.; Li, D.; Pan, Z.; Cheng, M.; Liu, D.; Zhang, M. Ultracompact Dual-Mode Waveguide Crossing Based on Subwavelength Multimode-Interference Couplers. Photonics Res. 2018, 6, 660-665

  18. [26]

    Inverse Design and Demonstration of an Ultracompact Broadband Dual-Mode 3 dB Power Splitter

    Chang, W.; Ren, X.; Ao, Y.; Lu, L.; Cheng, M.; Deng, L.; Liu, D.; Zhang, M. Inverse Design and Demonstration of an Ultracompact Broadband Dual-Mode 3 dB Power Splitter. Opt. Express 2018, 26, 24135-24144

  19. [27]

    Y.; Lu, J.; Babinec, T

    Piggott, A. Y.; Lu, J.; Babinec, T. M.; Lagoudakis, K. G.; Petykiewicz, J.; Vučković, J. Inverse Design and Implementation of a Wavelength Demultiplexing Grating Coupler. Sci. Rep. 2014, 4, 7210

  20. [28]

    V.; Roelkens, G.; Ayre, M.; Schrauwen, J.; Taillaert, D.; Van Thourhout, D.; Krauss, T

    Laere, F. V.; Roelkens, G.; Ayre, M.; Schrauwen, J.; Taillaert, D.; Van Thourhout, D.; Krauss, T. F.; Baets, R. Compact and Highly Efficient Grating Couplers Between Optical Fiber and Nanophotonic Waveguides. J. Lightwave Technol. 2007, 25, 151-156

  21. [29]

    V.; Claes, T.; Schrauwen, J.; Scheerlinck, S.; Bogaerts, W.; Taillaert, D.; O'Faolain, L.; Thourhout, D

    Laere, F. V.; Claes, T.; Schrauwen, J.; Scheerlinck, S.; Bogaerts, W.; Taillaert, D.; O'Faolain, L.; Thourhout, D. V.; Baets, R. Compact Focusing Grating Couplers for Silicon -On-Insulator Integrated Circuits. IEEE Photonic. Tech. L. 2007, 19, 1919-1921

  22. [30]

    Objective-First Design of High-Efficiency, Small-Footprint Couplers Between Arbitrary Nanophotonic Waveguide Modes

    Lu, J.; Vučković, J. Objective-First Design of High-Efficiency, Small-Footprint Couplers Between Arbitrary Nanophotonic Waveguide Modes. Opt. Express 2012, 20, 7221-7236

  23. [31]

    Ultra-Compact Broadband Polarization Beam Splitter with Strong Expansibility

    Huang, J.; Yang, J.; Chen, D.; He, X.; Han, Y.; Zhang, J.; Zhang, Z. Ultra-Compact Broadband Polarization Beam Splitter with Strong Expansibility. Photonics Res. 2018, 6, 574- 578

  24. [32]

    S.; Wang, F.; Sigmund, O

    Zhou, M.; Lazarov, B. S.; Wang, F.; Sigmund, O. Minimum Length Scale in Topology Optimization by Geometric Constraints. Comput. Method. Appl. M. 2015, 293, 266-282

  25. [33]

    Miller, D. A. B. All Linear Optical Devices are Mode Converters. Opt. Express 2012, 20, 23985-23993

  26. [34]

    Active Nanophotonics: Inverse Design and Strained Germanium Light Emitters

    Petykiewicz, J. Active Nanophotonics: Inverse Design and Strained Germanium Light Emitters. Ph.D dissertation (Stanford University. Department of Electrical Engineering) 2016

  27. [35]

    B.; Pierce, N

    Giles, M. B.; Pierce, N. A. An Introduction to the Adjoint Approach to Design. Flow Turbul. Combust. 2000, 65, 393-415

  28. [36]

    Y.; Petykiewicz, J.; Su, L.; Vučković, J

    Piggott, A. Y.; Petykiewicz, J.; Su, L.; Vučković, J. Fabrication-Constrained Nanophotonic Inverse Design. Sci. Rep. 2017, 7, 1786

  29. [37]

    V.; Piggott, A

    Vercruysse, D.; Su, L.; Trivedi, R.; Sapra, N. V.; Piggott, A. Y.; Vučković, J. Level-Set Fabrication Constraints for Gradient -Based Optimization of Optical Devices . CLEO: QELS_Fundamental Science. Optical Society of America 2018, pp.JTu2A-84

  30. [38]

    V.; Su, L.; Trivedi, R.; Vučković, J

    Vercruysse, D.; Sapra, N. V.; Su, L.; Trivedi, R.; Vučković, J. Analytical Level Set Fabrication Constraints for Inverse Design. Sci. Rep. 2019, 9, 8999

  31. [39]

    https://www.lumerical.com/

  32. [40]

    Abrokwah, K. O. Characterization and Modeling of Plasma Etch Pattern Dependencies in Integrated Circuits. Ph.D dissertation(Massachusetts Institute of Technology) 2006

  33. [41]

    Inverse Designed Tunable Four-Channel Wavelength Demultiplexer

    Han, J.; Huang, J.; Wu, J.; Yang, J. Inverse Designed Tunable Four-Channel Wavelength Demultiplexer. Under Review 2019. SUPPORTING INFORMATION

  34. [42]

    Constraints on minimum feature size. As illustrated in Figure S1, when the value of the minimum feature size R was set to 3 (4), the simulated peaking transmission was -1.39 dB (- 2.25 dB) at upper channel, -1.45 dB (-2.15 dB) at lower channel, the corresponding crosstalk was ...

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.