pith. sign in

arxiv: 2604.12889 · v1 · submitted 2026-04-14 · ⚛️ physics.optics

Building reliable 3D photonic integrated circuits and cavities at the wafer scale

Pith reviewed 2026-05-10 14:31 UTC · model grok-4.3

classification ⚛️ physics.optics
keywords 3D photonic integrated circuitswafer-scale fabricationsilicon nitride waveguideshigh-Q cavitiesinter-layer transitionschemical mechanical polishingoptical cavities
0
0 comments X

The pith

Etch-back polishing and a kappa-engineered taper produce reliable wafer-scale 3D photonic circuits with losses of 0.07 dB/cm and the first two-layer high-Q cavities.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper establishes a fabrication route that stacks two silicon nitride waveguide layers into functional 3D photonic circuits across entire wafers while maintaining low and uniform losses. An etch-back chemical mechanical polishing step flattens the spacer layer between the waveguides so that light can move between layers consistently. A new taper design shapes the coupling strength along its length to achieve high transition efficiency in a smaller footprint than conventional straight tapers, raising a reliability metric by 75 percent. The resulting circuits show typical waveguide losses of 0.077 dB/cm and 0.068 dB/cm together with inter-layer transitions of only 6 mdB. These low losses then allow optical cavities to occupy two separate device layers and still reach high quality factors.

Core claim

The central claim is that etch-back assisted chemical mechanical polishing delivers high wafer-scale uniformity of the spacer layer between two silicon nitride waveguide layers, while a kappa-engineered taper overcomes the usual efficiency-footprint trade-off in inter-layer transitions. These advances together yield 3D photonic integrated circuits with typical losses of 0.077 dB/cm and 0.068 dB/cm on the two layers and 3D transition losses as low as 6 mdB. The low transition loss in turn enables the first class of 3D high-Q optical cavities that occupy two distinct device layers.

What carries the argument

The kappa-engineered taper, which varies the coupling coefficient along the transition length to optimize efficiency within a compact footprint, paired with the E-CMP process that planarizes the inter-layer spacer for uniform coupling across the wafer.

If this is right

  • Reliable 3D PICs become practical at full wafer scale with propagation losses low enough for dense photonic systems.
  • High-Q cavities that span two device layers become available, expanding the design space for resonators.
  • The same process and design rules supply a scalable route for applications in photonic interconnects, computing networks, high-density sensors, and nonlinear photonics.

Where Pith is reading between the lines

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

  • The method could be extended to three or more layers to increase circuit density without a proportional rise in loss.
  • Full statistical yield data across many wafers would be required to confirm readiness for volume manufacturing.
  • The low-loss 3D transitions could be combined with other material platforms to address different operating wavelengths or nonlinear effects.

Load-bearing premise

The E-CMP process produces the claimed high uniformity of spacer thickness across the full wafer and the kappa-engineered taper delivers its performance gains without introducing new loss mechanisms or yield drops not captured in the reported typical values.

What would settle it

A wafer-scale map of spacer layer thickness that shows variations exceeding a few nanometers, or repeated measurements of 3D transition loss that average well above 6 mdB across many devices and multiple wafers.

read the original abstract

Three-dimensional (3D) photonic integrated circuits (PIC) are emerging as an indispensable scheme for high density and multifunctional photonic systems. However, the wafer-scale scaling of PICs towards a 3D configuration is constrained by two key factors: (i) the trade-off between inter-layer taper efficiency and footprint, and (ii) wafer-scale uniformity of inter-layer transition loss. In this work, we introduce etch-back assisted chemical mechanical polishing (E-CMP) to achieve high wafer-scale uniformity of the spacer layer. Moreover, we break the efficiency-footprint trade-off by demonstrating a novel $\kappa$-engineered taper, achieving a reliability metric that is 75\% higher than the traditional linearly tapered structure. Building on these design and fabrication developments, we enable reliable 3D PICs with typical loss of 0.077 and 0.068 dB/cm on two silicon nitride (SiN) waveguide layers and typical 3D transition loss as low as 6 mdB. Furthermore, the low 3D transition loss enables the first class of 3D high-Q optical cavities occupying two distinct device layers, providing new design space for high-Q optical cavities. The scalable fabrication process and design methodology provide routes for wafer-scale reliable 3D PICs that are promising in a series of applications ranging from photonic interconnects and computing networks to high-density photonic sensors and nonlinear photonics.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 0 minor

Summary. The manuscript introduces etch-back assisted chemical mechanical polishing (E-CMP) to achieve high wafer-scale uniformity of the spacer layer and a novel κ-engineered taper to overcome the efficiency-footprint trade-off in inter-layer transitions for 3D photonic integrated circuits on silicon nitride. Building on these, the authors report typical propagation losses of 0.077 dB/cm and 0.068 dB/cm on two waveguide layers, 3D transition losses as low as 6 mdB, a 75% improvement in a reliability metric over linear tapers, and the first 3D high-Q cavities spanning two device layers.

Significance. If the wafer-scale uniformity and low-loss performance are confirmed with comprehensive statistics, this work would meaningfully advance scalable 3D PIC fabrication by addressing key barriers to multilayer integration, enabling higher-density systems for photonic interconnects, computing, sensors, and nonlinear optics while also expanding high-Q cavity design options.

major comments (2)
  1. Abstract: The central claims of 'reliable' wafer-scale 3D PICs with the stated typical loss values and 75% reliability improvement rest on 'typical' numbers without reported wafer-scale statistics, standard deviations, die-to-die histograms, yield percentages, or number of devices measured. This directly weakens the ability to substantiate uniformity of the E-CMP spacer and the absence of hidden yield penalties in the κ-taper.
  2. Results (loss and cavity sections): The propagation and transition loss figures are presented only as typical values; without error bars, measurement counts, or wafer maps, it is impossible to distinguish consistent performance from post-selection of best devices, which is load-bearing for the 'reliable' and 'wafer-scale' qualifiers.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback emphasizing the need for statistical rigor to support the wafer-scale and reliability claims. We address each major comment below and have revised the manuscript accordingly.

read point-by-point responses
  1. Referee: Abstract: The central claims of 'reliable' wafer-scale 3D PICs with the stated typical loss values and 75% reliability improvement rest on 'typical' numbers without reported wafer-scale statistics, standard deviations, die-to-die histograms, yield percentages, or number of devices measured. This directly weakens the ability to substantiate uniformity of the E-CMP spacer and the absence of hidden yield penalties in the κ-taper.

    Authors: We agree that the presentation of 'typical' values would be strengthened by explicit statistical details. In the revised manuscript, we have updated the abstract and main text to include the number of devices measured for each metric, standard deviations on the reported loss values, die-to-die histograms, and yield percentages. These additions, drawn from our full characterization dataset, substantiate the uniformity of the E-CMP spacer layer and confirm that the κ-engineered tapers do not introduce hidden yield penalties. revision: yes

  2. Referee: Results (loss and cavity sections): The propagation and transition loss figures are presented only as typical values; without error bars, measurement counts, or wafer maps, it is impossible to distinguish consistent performance from post-selection of best devices, which is load-bearing for the 'reliable' and 'wafer-scale' qualifiers.

    Authors: We concur that error bars, measurement counts, and wafer maps are necessary to demonstrate consistency. The revised manuscript now includes error bars on all loss and Q-factor plots, specifies the number of devices and wafers measured for each result, and adds wafer maps plus histograms to the supplementary information. These changes show that the reported performance is representative rather than post-selected and support the wafer-scale reliability claims for both the waveguides and the 3D cavities. revision: yes

Circularity Check

0 steps flagged

No circularity: experimental results with no derivation chain

full rationale

The paper reports measured fabrication outcomes (E-CMP spacer uniformity and κ-engineered taper performance) and resulting typical propagation/transition losses. No equations, first-principles derivations, or predictions are presented that reduce to fitted inputs, self-citations, or ansatzes. The central claims rest on direct experimental data rather than any load-bearing theoretical step, so the work is self-contained.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The central claims rest on standard semiconductor fabrication assumptions (uniform deposition, etch selectivity, CMP removal rates) plus the unverified premise that the new taper shape improves reliability without hidden yield penalties. No free parameters or invented entities are introduced in the abstract.

axioms (2)
  • domain assumption E-CMP produces wafer-scale spacer uniformity sufficient for low inter-layer loss
    Invoked to justify the 3D transition loss of 6 mdB
  • domain assumption The κ-engineered taper breaks the efficiency-footprint trade-off without new loss channels
    Central to the 75% reliability improvement claim

pith-pipeline@v0.9.0 · 5574 in / 1417 out tokens · 31139 ms · 2026-05-10T14:31:56.156188+00:00 · methodology

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

35 extracted references · 35 canonical work pages

  1. [1]

    & Yoo, S

    Zhang, Y., Samanta, A., Shang, K. & Yoo, S. B. Scal- able 3d silicon photonic electronic integrated circuits and their applications.IEEE Journal of Selected Topics in Quantum Electronics26, 1–10 (2020)

  2. [2]

    Zhu, X.et al.Low-loss and polarization insensitive 32× 4 optical switch for roadm applications.Light: Science & Applications13, 94 (2024)

  3. [3]

    In2025 Eu- ropean Conference on Optical Communications (ECOC), 1–4 (2025)

    Xia, Y.et al.3d silicon nitride waveguide interposers for high-density scale-up chiplet interconnects. In2025 Eu- ropean Conference on Optical Communications (ECOC), 1–4 (2025)

  4. [4]

    Kaur, P.et al.Hybrid and heterogeneous photonic inte- gration.APL photonics6(2021)

  5. [5]

    M.et al.Ultra-low crosstalk, cmos compatible waveguide crossings for densely integrated photonic in- terconnection networks.Optics express21, 12002–12013 (2013)

    Jones, A. M.et al.Ultra-low crosstalk, cmos compatible waveguide crossings for densely integrated photonic in- terconnection networks.Optics express21, 12002–12013 (2013)

  6. [6]

    D.et al.Monolithically integrated multilayer silicon nitride-on-silicon waveguide platforms for 3-d pho- tonic circuits and devices.Proceedings of the IEEE106, 2232–2245 (2018)

    Sacher, W. D.et al.Monolithically integrated multilayer silicon nitride-on-silicon waveguide platforms for 3-d pho- tonic circuits and devices.Proceedings of the IEEE106, 2232–2245 (2018)

  7. [7]

    Margalit, N.et al.Perspective on the future of silicon photonics and electronics.Applied Physics Letters118, 220501 (2021)

  8. [8]

    A.et al.Extending the spectrum of fully inte- grated photonics to submicrometre wavelengths.Nature 610, 54–60 (2022)

    Tran, M. A.et al.Extending the spectrum of fully inte- grated photonics to submicrometre wavelengths.Nature 610, 54–60 (2022)

  9. [9]

    Churaev, M.et al.A heterogeneously integrated lithium niobate-on-silicon nitride photonic platform.Nature com- munications14, 3499 (2023)

  10. [10]

    Xiang, C.et al.3d integration enables ultralow-noise isolator-free lasers in silicon photonics.Nature620, 78– 85 (2023)

  11. [11]

    & Yoo, S

    Zhang, Y., Ling, Y.-C., Zhang, Y., Shang, K. & Yoo, S. B. High-density wafer-scale 3-d silicon-photonic in- tegrated circuits.IEEE Journal of Selected Topics in Quantum Electronics24, 1–10 (2018)

  12. [12]

    IEEE Journal of Selected Topics in Quantum Electronics 20, 317–322 (2014)

    Kang, J.et al.Amorphous-silicon inter-layer grating couplers with metal mirrors toward 3-d interconnection. IEEE Journal of Selected Topics in Quantum Electronics 20, 317–322 (2014)

  13. [13]

    Sodagar, M., Pourabolghasem, R., Eftekhar, A. A. & Adibi, A. High-efficiency and wideband interlayer grat- ing couplers in multilayer si/sio 2/sin platform for 3d integration of optical functionalities.Optics express22, 16767–16777 (2014)

  14. [14]

    Wan, C., Gaylord, T. K. & Bakir, M. S. Grating design for interlayer optical interconnection of in-plane waveg- uides.Applied Optics55, 2601–2610 (2016)

  15. [15]

    Xu, P., Zhang, Y., Zhang, S., Chen, Y. & Yu, S. Sin x–si interlayer coupler using a gradient index metamaterial. Optics letters44, 1230–1233 (2019)

  16. [16]

    & Yoo, S

    Shang, K., Pathak, S., Guan, B., Liu, G. & Yoo, S. Low- loss compact multilayer silicon nitride platform for 3d photonic integrated circuits.Optics Express23, 21334– 21342 (2015)

  17. [17]

    D.et al.Tri-layer silicon nitride-on-silicon photonic platform for ultra-low-loss crossings and inter- layer transitions.Optics express25, 30862–30875 (2017)

    Sacher, W. D.et al.Tri-layer silicon nitride-on-silicon photonic platform for ultra-low-loss crossings and inter- layer transitions.Optics express25, 30862–30875 (2017)

  18. [18]

    & Yan, J

    Zhang, L., Guo, Y., Wang, J., Li, J. & Yan, J. 3d het- erogeneous integration of silicon nitride and aluminum nitride on sapphire toward ultrawideband photonics in- tegrated circuits.ACS Photonics12, 2538–2547 (2025)

  19. [19]

    Niels, M.et al.A high-speed heterogeneous lithium tan- talate silicon photonics platform.Nature Photonics1–7 (2026). 9

  20. [20]

    F.et al.Ultra-low-loss single-mode si3n4 waveguides with 0.7 db/m propagation loss

    Bauters, J. F.et al.Ultra-low-loss single-mode si3n4 waveguides with 0.7 db/m propagation loss. InEuro- pean conference and exposition on optical communica- tions, Th–12 (Optica Publishing Group, 2011)

  21. [21]

    Jin, W.et al.Hertz-linewidth semiconductor lasers using cmos-ready ultra-high-q microresonators.Nature Pho- tonics15, 346–353 (2021)

  22. [22]

    Communications Physics5, 84 (2022)

    Ji, X.et al.Compact, spatial-mode-interaction-free, ultralow-loss, nonlinear photonic integrated circuits. Communications Physics5, 84 (2022)

  23. [23]

    Ji, X.et al.Efficient mass manufacturing of high-density, ultra-low-loss si3n4 photonic integrated circuits.Optica 11, 1397–1407 (2024)

  24. [24]

    Liang, T.-L.et al.A fully numerical method for designing efficient adiabatic mode evolution structures (adiabatic taper, coupler, splitter, mode converter) applicable to complex geometries.Journal of Lightwave Technology 39, 5531–5547 (2021)

  25. [25]

    K.et al.Shortcuts to adiabaticity in waveg- uide couplers–theory and implementation.Advances in Physics: X6, 1894978 (2021)

    Taras, A. K.et al.Shortcuts to adiabaticity in waveg- uide couplers–theory and implementation.Advances in Physics: X6, 1894978 (2021)

  26. [26]

    & Penty, R

    Yao, C., Cheng, Q., Roelkens, G. & Penty, R. Bridging the gap between resonance and adiabaticity: a compact and highly tolerant vertical coupling structure.Photonics Research10, 2081–2090 (2022)

  27. [27]

    Van Asch, J.et al.A methodical approach to design adi- abatic waveguide couplers for heterogeneous integrated photonics.Journal of Physics: Photonics6, 045013 (2024)

  28. [28]

    Pfeiffer, M. H. P.et al.Photonic damascene process for integrated high-q microresonator based nonlinear pho- tonics.Optica3, 20–25 (2016)

  29. [29]

    Liu, J.et al.High-yield, wafer-scale fabrication of ultralow-loss, dispersion-engineered silicon nitride pho- tonic circuits.Nature communications12, 2236 (2021)

  30. [30]

    Pfeiffer, M. H. P.et al.Photonic damascene process for low-loss, high-confinement silicon nitride waveguides. IEEE Journal of selected topics in quantum electronics 24, 1–11 (2018)

  31. [31]

    & Poon, A

    Wu, K. & Poon, A. W. Stress-released si3n4 fabrication process for dispersion-engineered integrated silicon pho- tonics.Optics Express28, 17708–17722 (2020)

  32. [32]

    & Zhang, Z

    Liu, S., Zhang, Y., Hariri, A., Al-Hallak, A.-R. & Zhang, Z. Fabrication of ultra-low-loss, dispersion-engineered sil- icon nitride photonic integrated circuits via silicon hard- mask etching.ACS Photonics12, 1039–1046 (2025)

  33. [33]

    Zhang, S.et al.Low-temperature sputtered ultralow-loss silicon nitride for hybrid photonic integration.Laser & Photonics Reviews18, 2300642 (2024)

  34. [34]

    Optics Letters43, 1527–1530 (2018)

    Chiles, J.et al.Deuterated silicon nitride photonic de- vices for broadband optical frequency comb generation. Optics Letters43, 1527–1530 (2018)

  35. [35]

    Bose, D.et al.Anneal-free ultra-low loss silicon nitride integrated photonics.Light: Science & Applications13, 156 (2024). METHODS Comparison of SiN deposition method and fabrication process flow To release stress and reduce the wafer bow, process flow based on low-pressure chemical vapor deposition (LPCVD) SiN usually demands complicated processes, such ...