Pith. sign in

REVIEW 2 major objections 5 minor 34 references

Ion-Implanted Silicon Nanoregions Enable Ultra-Low-Loss Trimming of Cladded Photonic Integrated Circuits

T0 review · 2 major / 5 minor · reviewed 2026-07-10 · grok-4.5

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read Silicon ions implanted into the cladding of photonic chips form a high-index nanoregion that trims optical responses with under 0.001 dB excess loss per π phase shift, multi-month ambient stability, and automated correction of a photonic cr

desk verdict Solid experimental methods paper: Si++ FIB into HSQ cladding gives cladding-confined, ultra-low-loss non-volatile trim on cladded SiN and LN, with automation and multi-month stability that the field actually needs. read the letter →

arxiv 2607.08266 v1 pith:RC4WSTQK submitted 2026-07-09 physics.optics physics.app-ph

Zhongyu Tang , Shabnam Taheriniya , Seongmin Jo , Xinyu Ma , Akhil Varri , Anna P. Ovvyan , Vincent Spreter , Liam McRae
show 3 more authors
Xiansong Meng Wolfram H. P. Pernice Rongyang Xu
This is my paper
classification physics.opticsphysics.app-ph
keywords photonicintegratedcircuitspost-fabricationtrimmingsiliconionimplantationHSQcladdingultra-low-lossdirectionalcouplermicroringresonatorcrossbar
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

As photonic integrated circuits grow more complex, tiny fabrication errors stack up and ruin circuit performance, so reliable post-fabrication correction is essential. This paper shows that implanting silicon ions into the transparent cladding that covers the waveguides creates a thin, high-index silicon-rich pocket that sits only in the cladding and never reaches the waveguide core. The guided light’s evanescent field still feels the higher index, so the phase and coupling of devices such as directional couplers and ring resonators can be permanently adjusted. Because the waveguide itself is left untouched, the extra loss stays below 0.001 dB for every π of phase shift, and the trimmed response holds steady for months under ordinary laboratory conditions. The same process can be automated across an entire 4×4 photonic crossbar, cutting average channel-to-channel imbalance from 78.5% down to 5.9%. The result is a practical, non-volatile trimming route that works on already-cladded chips and on more than one material platform.

What carries the argument

The Si-rich amorphous nanoregion formed by 70 keV Si++ implantation into the HSQ cladding: it raises the local refractive index so that the guided-mode evanescent field experiences a controllable effective-index shift while remaining confined above the waveguide core.

What would settle it

Direct insertion-loss measurements on long waveguides or multi-device test structures that have received the same implantation doses and geometries used for the reported π-phase-shift figure, compared against unimplanted controls; if the excess loss rises above ~0.001 dB/π or if post-implantation STEM shows silicon diffusion into the waveguide core, the ultra-low-loss claim fails.

Watch

Extended reading notes

Core claim

Silicon ion implantation into the HSQ cladding of photonic integrated circuits forms a spatially confined, high-index silicon-rich nanoregion that permanently alters the effective index of the guided mode without measurable damage to the underlying waveguide, delivering non-volatile trimming with excess loss below 0.001 dB per π phase shift, multi-month ambient stability, and automated reduction of average crossbar channel variation from 78.5% to 5.9%.

Load-bearing premise

That the silicon-rich zone stays fully confined to the cladding and leaves the waveguide undamaged enough for the excess-loss figure of under 0.001 dB per π to hold in general use, a claim that rests on STEM-EDS/EELS maps that are themselves limited by detection thresholds and on loss values extracted from ring-resonator resonance shifts rather than a direct multi-device loss campaign.

Editorial extensions

If this is right

  • Cladded silicon-nitride and lithium-niobate circuits can be permanently trimmed after final packaging without continuous heater power.
  • Directional-coupler ratios and ring extinction ratios can be corrected bidirectionally by choosing implantation location (waveguide versus gap).
  • Automated sequential implantation can equalize large cascaded networks such as photonic crossbars to a few-percent channel uniformity.
  • The same cladding-confined index shift works across platforms without redesign of the waveguide core.
  • Long-term ambient stability removes the need for hermetic packaging or post-trim annealing for many applications.

Reading between the lines

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

  • Because the method never opens the cladding, it could be applied late in a foundry flow after full encapsulation and even after wire-bonding.
  • If dose and geometry maps can be derived from a single optical probe of an untrimmed circuit, closed-loop automated trimming of wafer-scale PICs becomes feasible.
  • The same confined high-index pocket may serve as a permanent phase shifter for non-volatile photonic memory or weight banks once write-once behavior is accepted.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

2 major / 5 minor

Summary. The manuscript demonstrates non-volatile post-fabrication trimming of cladded photonic integrated circuits by focused-beam Si++ implantation into HSQ cladding, forming a high-index Si-rich nanoregion that overlaps the guided-mode evanescent field without directly modifying the waveguide core. Structural evidence (STEM-EDS/EELS, AFM) shows the implant is confined to the cladding within detection limits and produces only a shallow surface trench. Optical results include bidirectional coupling-ratio control of directional couplers on Si3N4 and LiNbO3, independent resonance-wavelength and extinction-ratio tuning of microrings, Vernier-filter alignment, multi-month ambient stability, excess loss below 0.001 dB per π extracted from ring resonators, and automated trimming of a 4×4 photonic crossbar array that reduces average channel output variation from 78.5% to 5.9%. The authors position the method as a scalable, cross-platform route for cladded PICs.

Significance. If the reported loss, confinement, and stability hold under broader use, the work supplies a practical, single-step, non-volatile trimming tool that operates on already-cladded devices—an important gap relative to many PCM, Ge-amorphization, and UV-laser approaches that either require bare waveguides or incur higher excess loss. The automated crossbar demonstration and dual-platform validation (Si3N4 and LiNbO3) strengthen the claim of scalability beyond single-device proofs of concept. The combination of microscopy, spectral statistics, and a clear comparison table makes the contribution concrete and usable for the PIC community.

major comments (2)
  1. Methods §4.4 and Results §2.4: the headline excess-loss figure (<0.001 dB/π) is obtained solely from the change in extracted ring-resonator propagation loss before versus after a single dose (1.55 nC/μm^{2}) together with the resonance-shift-to-FSR conversion. Because the same dose is used for DC and ring tuning, the number is relevant, yet it remains an indirect, single-geometry extraction. A short multi-device or multi-geometry check (e.g., cut-back or Q-factor statistics on several rings/DCs at the doses and implant lengths actually used) would make the ultra-low-loss claim more robust; without it the figure should be explicitly scoped to the reported resonator geometry and dose.
  2. Results §2.1 / Fig. 2: confinement of the Si-rich region and the absence of waveguide damage are stated as “within the detection limits of the measurement.” This language is appropriate, but the central claim that the method is non-intrusive across platforms rests on it. A quantitative bound (detection limit for Si concentration or plasmon signal inside the Si3N4 core) or a brief note on whether any residual damage would be expected to appear in the loss metrology would tighten the argument.
minor comments (5)
  1. Table 1: the comparison is useful; adding a column or footnote for spatial resolution / minimum feature size of each technique would help readers judge relative practicality for dense circuits.
  2. Fig. 3 and Supporting Information: the two implantation configurations are clear, but a short statement of how the 0.3 µm width and successive 2 µm steps were chosen (beam spot, dose uniformity) would aid reproducibility.
  3. Fig. 7a reports CR stability over 120 days for five DCs; stating the number of devices and the exact storage conditions (temperature, humidity) in the caption or Methods would strengthen the long-term claim.
  4. Introduction and Conclusion: a brief quantitative comparison of the demonstrated tuning range (e.g., CR swing, resonance shift per dose) against the PCM and Ge-implantation entries already cited would better situate the dynamic range of the method.
  5. Methods §4.1: the nominal HSQ thickness is given as 500 nm while the STEM cross-section shows ~150 nm above the waveguide; clarifying the process variation or the thickness actually used for the measured devices would remove a minor inconsistency.

Circularity Check

0 steps flagged · score 1.0 of 10

Experimental methods paper with measured spectra, microscopy, and device statistics; no derivation that reduces to its inputs by construction.

full rationale

This is an experimental demonstration paper, not a theoretical derivation. Load-bearing claims (Si-rich nanoregion confined to HSQ cladding with no observable waveguide damage; excess loss <0.001 dB/π extracted from ring-resonator propagation-loss differences before/after implantation at fixed dose; multi-month ambient stability of CR and resonance shift; automated PCA trimming reducing channel variation 78.5%→5.9%) rest on direct STEM-EDS/EELS maps (Fig. 2), AFM, transmission spectra (Figs. 3–6), and repeated measurements (Fig. 7), plus standard mode/implantation simulations that are not inverted to force the result. Excess-loss calculation (Methods §4.4) is a straightforward ratio of measured Δloss to measured phase shift (Δλ/FSR); it does not redefine the measured quantities. Table 1 comparisons (including one overlapping-author FIB-carbon entry) are external baselines, not premises that force the present numbers. No self-definitional loop, no fitted parameter re-labeled as prediction, no uniqueness theorem imported from prior author work, and no ansatz smuggled via citation. Minor self-citations of related group results exist but are non-load-bearing. Circularity is therefore negligible.

Assumptions & free parameters 3 free parameters · 5 assumptions · 0 invented entities

Experimental methods paper. Load-bearing content is process parameters and domain physics assumptions (evanescent overlap, Si vs cladding index contrast, loss extraction from rings), not invented particles or multi-parameter theory fits. Free parameters are experimental knobs (dose, geometry, energy) chosen to achieve target phase/coupling shifts. No new fundamental entities are postulated; the Si-rich nanoregion is directly imaged.

free parameters (3)
  • Si++ implantation dose (e.g. 1–1.55 nC/μm² for main tuning; range 0.001–5 nC/μm² in stability study)
    Dose is chosen experimentally to set resonance shift and coupling change; the ultra-low-loss and tuning-range claims depend on operating in this dose window.
  • Implantation geometry (width ~0.1–0.3 μm, length 0–25 μm, position: bus / gap / ring)
    Length, width, and placement are free experimental controls that map to CR change and phase shift; reported slopes (−3.6% and +1.5% CR per μm) are process-specific.
  • Acceleration voltage / beam energy (35 kV Si++, ~70 keV)
    Sets implant depth (~100 nm center below surface in simulation); suitability claim for cladding <1 μm rests on this choice.
assumptions (5)
  • domain assumption Evanescent field of the guided mode overlaps a cladding-confined high-index Si-rich region enough to raise neff without requiring core modification.
    Invoked in §2.1 and Fig. 1c mode simulations; central to the non-intrusive cladding-only mechanism.
  • domain assumption Within STEM-EDS/EELS detection limits, implanted Si does not measurably enter or damage the waveguide core.
    Stated in §2.1 / Fig. 2; required for ultra-low-loss and cross-platform claims.
  • ad hoc to paper Excess loss per π can be extracted from ring resonance shift (normalized to FSR) and the change in extracted propagation loss before vs after implant (Methods §4.4).
    Defines the headline <0.001 dB/π figure of merit; different metrology could change the number.
  • domain assumption Cured HSQ optical properties are close enough to SiO2 that it serves as a representative low-index cladding.
    Stated in §2.1; underpins generality to cladded PICs.
  • ad hoc to paper Surface trench (~45 nm) from partial milling has negligible modal overlap and thus negligible performance impact.
    AFM result in §2.1 / Supporting Fig. S1; used to dismiss milling as a loss channel.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Ion-Implanted Silicon Nanoregions Enable Ultra-Low-Loss Trimming of Cladded Photonic Integrated Circuits." pith.science (2026). https://pith.science/paper/RC4WSTQK

@misc{pith2026260708266,
  author       = {Pith},
  title        = {Pith review of: Ion-Implanted Silicon Nanoregions Enable Ultra-Low-Loss Trimming of Cladded Photonic Integrated Circuits},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/RC4WSTQK}},
  note         = {Machine review of arXiv:2607.08266}
}
read the original abstract

Photonic integrated circuits (PICs) have emerged as a key platform for information processing, including optical communication and computing. As PIC complexity increases, fabrication-induced response deviations accumulate, making post-fabrication trimming critical for unlocking their full potential. Here, we demonstrate that silicon ion implantation enables scalable, ultra-low-loss trimming of cladded PICs by locally forming a high-index silicon-rich region. Structural characterization confirms that the silicon-implanted nanoregion is confined to the cladding without observable damage to the underlying waveguide. The implantation-induced excess loss is below 0.001 dB per {\pi} phase shift, while the optical response remains stable over a four-month observation period. Automated trimming is demonstrated on a photonic crossbar array, reducing the average channel output variation from 78.5% to 5.9%. These results establish a practical route towards automated, ultra-low-loss post-fabrication trimming for large-scale cladded PICs.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

34 extracted references · 34 canonical work pages

  1. [1]

    Brückerhoff-Plückelmann, A

    F. Brückerhoff-Plückelmann, A. P. Ovvyan, A. Varri, H. Borras, B. Klein, L. Meyer, C. D. Wright, H. Bhaskaran, G. S. Syed, A. Sebastian, H. Fröning, W. Pernice, Nat Comput Sci 2025, 5, 377

    work page 2025

  2. [2]

    S. Hua, E. Divita, S. Yu, B. Peng, C. Roques-Carmes, Z. Su, Z. Chen, Y. Bai, J. Zou, Y. Zhu, Y. Xu, C. Lu, Y. Di, H. Chen, L. Jiang, L. Wang, L. Ou, C. Zhang, J. Chen, W. Zhang, H. Zhu, W. Kuang, L. Wang, H. Meng, M. Steinman, Y. Shen, Nature 2025, 640, 361

    work page 2025

  3. [3]

    McNab, N

    S. McNab, N. Moll, Y. Vlasov, Opt. Express 2003, 11, 2927

    work page 2003

  4. [4]

    Krasnokutska, J.-L

    I. Krasnokutska, J.-L. J. Tambasco, X. Li, A. Peruzzo, Opt. Express 2018, 26, 897

    work page 2018

  5. [5]

    Feldmann, N

    J. Feldmann, N. Youngblood, M. Karpov, H. Gehring, X. Li, M. Stappers, M. Le Gallo, X. Fu, A. Lukashchuk, A. S. Raja, J. Liu, C. D. Wright, A. Sebastian, T. J. Kippenberg, W. H. P. Pernice, H. Bhaskaran, Nature 2021, 589, 52

    work page 2021

  6. [6]

    Chetrit, J

    Ansheng Liu, Ling Liao, Y. Chetrit, J. Basak, H. Nguyen, D. Rubin, M. Paniccia, IEEE J. Select. Topics Quantum Electron. 2010, 16, 23

    work page 2010

  7. [7]

    R. Xu, S. Taheriniya, A. Varri, M. Ulanov, I. Konyshev, L. Krämer, L. McRae, F. L. Ebert, J. R. Bankwitz, X. Ma, S. Ferrari, H. Bhaskaran, W. H. P. Pernice, Nano Lett. 2024, 24, 10813

    work page 2024

  8. [8]

    K. Zou, K. Pang, H. Song, J. Fan, Z. Zhao, H. Song, R. Zhang, H. Zhou, A. Minoofar, C. Liu, X. Su, N. Hu, A. McClung, M. Torfeh, A. Arbabi, M. Tur, A. E. Willner, Nat Commun 2022, 13, 7662

    work page 2022

Show all 34 references
  1. [9]

    Zhang, N

    Y. Zhang, N. Farmakidis, I. Roumpos, M. Moralis-Pegios, A. Tsakyridis, J. S. Lee, B. Dong, Y. He, S. Aggarwal, N. Pleros, H. Bhaskaran, Nat Commun 2025, 17, 989

    2025

  2. [10]

    Ekici, Y

    Ç. Ekici, Y. Yu, J. C. Adcock, A. L. Muthali, M. Zahidy, H. Tan, Z. Lin, H. Li, L. K. Oxenløwe, X. Cai, Y. Ding, npj Quantum Inf 2025, 11, 21

    2025

  3. [11]

    G. J. Mendoza, R. Santagati, J. Munns, E. Hemsley, M. Piekarek, E. Martín-López, G. D. Marshall, D. Bonneau, M. G. Thompson, J. L. O’Brien, Optica 2016, 3, 127

    2016

  4. [12]

    El-Fiky, A

    E. El-Fiky, A. Samani, D. Patel, M. Jacques, M. Sowailem, D. V. Plant, Opt. Express 2019, 27, 10258

    2019

  5. [13]

    Zhang, M

    H. Zhang, M. Li, Y. Zhang, D. Zhang, Q. Liao, J. He, S. Hu, B. Zhang, L. Wang, X. Xiao, N. Qi, S. Yu, Photon. Res. 2020, 8, 1776

    2020

  6. [14]

    Zheng, H

    Y. Zheng, H. Wang, X. Jia, J. Huang, H. Yuan, C. Zhai, J. Dai, J. Shi, L. Zhang, X. Zhang, M. Zhuang, J. Liu, J. Mao, T. Dai, Z. Fu, Y. Jiao, Y. Shi, D. Dai, X. Wang, Y. Li, Q. Gong, Z. Yuan, L. Chang, J. Wang, Nature 2026, DOI 10.1038/s41586-026-10152-z

    2026 doi

  7. [15]

    Bernabé, Q

    S. Bernabé, Q. Wilmart, K. Hasharoni, K. Hassan, Y. Thonnart, P. Tissier, Y. Désières, S. Olivier, T. Tekin, B. Szelag, Solid-State Electronics 2021, 179, 107928. 14

    2021

  8. [16]

    Carolan, C

    J. Carolan, C. Harrold, C. Sparrow, E. Martín-López, N. J. Russell, J. W. Silverstone, P. J. Shadbolt, N. Matsuda, M. Oguma, M. Itoh, G. D. Marshall, M. G. Thompson, J. C. F. Matthews, T. Hashimoto, J. L. O’Brien, A. Laing, Science 2015, 349, 711

    2015

  9. [17]

    J. M. Arrazola, V. Bergholm, K. Brádler, T. R. Bromley, M. J. Collins, I. Dhand, A. Fumagalli, T. Gerrits, A. Goussev, L. G. Helt, J. Hundal, T. Isacsson, R. B. Israel, J. Izaac, S. Jahangiri, R. Janik, N. Killoran, S. P. Kumar, J. Lavoie, A. E. Lita, D. H. Mahler, M. Menotti,...

    2021

  10. [18]

    Zelaya, A

    K. Zelaya, A. Perez‐Leija, A. Shetewy, J. Friedman, M. He, K. Jamshidi, M. Miri, Adv Quantum Tech 2026, 9, e00812

    2026

  11. [19]

    D. Wang, Y. Nie, G. Hu, H. K. Tsang, C. Huang, Nat Commun 2024, 15, 10841

    2024

  12. [20]

    Chrostowski, X

    L. Chrostowski, X. Wang, J. Flueckiger, Y. Wu, Y. Wang, S. T. Fard, in Optical Fiber Communication Conference, OSA, San Francisco, California, 2014, p. Th2A.37

    2014

  13. [21]

    S. K. Selvaraja, E. Rosseel, L. Fernandez, M. Tabat, W. Bogaerts, J. Hautala, P. Absil, in 8th IEEE International Conference on Group IV Photonics, IEEE, London, United Kingdom, 2011, pp. 71–73

    2011

  14. [22]

    N. Gao, M. Chen, H. Xu, Z. Xue, N. Zhang, L. Fei, X. Wei, Materials Science in Semiconductor Processing 2020, 117, 105159

    2020

  15. [23]

    Spector, J

    S. Spector, J. M. Knecht, P. W. Juodawlkis, Opt. Express 2016, 24, 5996

    2016

  16. [24]

    Y. Xie, H. C. Frankis, J. D. B. Bradley, A. P. Knights, Opt. Mater. Express 2021, 11, 2401

    2021

  17. [25]

    Jayatilleka, H

    H. Jayatilleka, H. Frish, R. Kumar, J. Heck, C. Ma, M. Sakib, D. Huang, H. Rong, J. Lightwave Technol. 2021, 39, 5083

    2021

  18. [26]

    C. Ríos, M. Stegmaier, P. Hosseini, D. Wang, T. Scherer, C. D. Wright, H. Bhaskaran, W. H. P. Pernice, Nature Photon 2015, 9, 725

    2015

  19. [27]

    M. M. Milosevic, X. Chen, X. Yu, N. J. Dinsdale, O. Aktas, S. Z. Oo, A. Z. Khokhar, D. J. Thomson, O. L. Muskens, H. M. H. Chong, A. C. Peacock, S. Saito, G. T. Reed, J. Lightwave Technol. 2020, 38, 1865

    2020

  20. [28]

    De Paoli, S

    G. De Paoli, S. L. Jantzen, T. D. Bucio, I. Skandalos, C. Holmes, P. G. R. Smith, M. M. Milosevic, F. Y. Gardes, Photon. Res. 2020, 8, 677

    2020

  21. [29]

    Van Steenberge, F

    G. Van Steenberge, F. Ferraro, Chip Scale Review 2024, 28

    2024

  22. [30]

    Zhang, Y

    X. Zhang, Y. Liang, Y. Chen, B. Shen, J. Huang, C. Lao, Y. Wu, Z. Li, D. Liu, H. Shu, W. Hu, X. Wang, N. Cui, L. Chang, Applied Physics Letters 2024, 125, 121111

    2024

  23. [31]

    T. Lin, F. Lu, P. Zhou, Y. Chen, J. Jiang, G. Hu, B. Yun, Y. Cui, J. Opt. 2024, 26, 025802

    2024

  24. [32]

    R. Chen, Z. Fang, C. Perez, F. Miller, K. Kumari, A. Saxena, J. Zheng, S. J. Geiger, K. E. Goodson, A. Majumdar, Nat Commun 2023, 14, 3465

    2023

  25. [33]

    Faneca, I

    J. Faneca, I. Zeimpekis, S. T. Ilie, T. D. Bucio, K. Grabska, D. W. Hewak, F. Y. Gardes, Neuromorph. Comput. Eng. 2021, 1, 014004

    2021

  26. [34]

    R. Xu, Z. Tang, L. McRae, A. Varri, F. Brückerhoff-Plückelmann, X. Ma, J. R. Bankwitz, J. Römer, R. Pradip, Q. Zhang, L. Meyer, Z. Zhao, J. Dijkstra, H. Bhaskaran, R. R. Schröder, W. H. P. Pernice, S. Taheriniya, Nat Commun 2026, 17, 4562

    2026

Pith tools

Reviewed July 10, 2026 · model on record in the stance chip above.