Pith. sign in

REVIEW 4 major objections 5 minor 25 references

A Method for Fabricating CMOS Back-End-of-Line-Compatible Solid-State Nanopore Devices

T0 review · 4 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read Ion beam–deposited silicon nitride can serve as a CMOS-compatible membrane for solid-state nanopores that detect single DNA molecules.

desk verdict First IBD SiNx nanopore demonstration with real DNA sensing, but the LPCVD-comparable noise claim needs a denominator. read the letter →

arxiv 2411.17416 v1 pith:N4A7SYO4 submitted 2024-11-26 physics.app-ph cond-mat.mtrl-sciphysics.bio-ph

Mohamed Yassine Bouhamidi , Chunhui Dai , Michel Stephan , Joyeeta Nag , Justin Kinney , Lei Wan , Matthew Waugh , Kyle Briggs
show 3 more authors
Jordan Katine Vincent Tabard-Cossa Daniel Bedau
This is my paper · ORCID
classification physics.app-phcond-mat.mtrl-sciphysics.bio-ph
keywords solid-statenanoporeionbeamdepositionsiliconnitridemembranecontrolledbreakdownCMOSBEOLintegrationDNAtranslocationlow-noise
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 tries to establish that silicon nitride membranes made by ion beam deposition (IBD), a room-temperature process compatible with CMOS back-end-of-line integration, can replace the usual high-temperature LPCVD membranes in solid-state nanopore sensing. If true, this removes the main thermal obstacle to packing many nanopores onto a chip with on-chip electronics, a step needed for using nanopores as the read head for molecular data storage. The authors show that IBD membranes resist KOH etching, stay insulating, support controlled-breakdown nanopore formation, and produce low-frequency noise comparable to that of standard LPCVD nanopores. They also demonstrate translocation of 2 kbp DNA through pores in both annealed and non-annealed IBD membranes.

What carries the argument

The central object is the IBD-deposited SiNx membrane itself, a room-temperature physical vapor deposited film produced by reacting a Si target's sputtered ions with nitrogen plasma, with no organic precursors and therefore minimal hydrogen and carbon incorporation. Its high density, low wet etch rate, and compressive stress are the properties that carry the argument: they let the membrane survive KOH release, remain insulating (>1 GΩ resistance), and support controlled breakdown (CBD) nanopore fabrication. CBD is the second mechanism—an automated voltage-ramp process that creates a single nm-sized pore in the insulating membrane and then enlarges it with voltage pulses. The combination of a BEOL-compatible dense membrane and CBD yields low-noise pores that can translocate DNA.

What would settle it

A direct test would be to characterize the stoichiometry and defect density of the actual freestanding IBD membranes (e.g., by XPS or TEM) across many chips and correlate those with nanopore yield and leakage current. If hidden porosity, pinholes, or Si/N composition gradients appear in the released membranes, or if the >1 GΩ resistance and low-noise pore results are found only on a small fraction of chips, the claim that the IBD process itself produces reliable nanopore membranes would be falsified.

Watch

Extended reading notes

Core claim

The central claim is that SiNx films deposited by ion beam deposition—a purely physical, room-temperature process—are suitable for solid-state nanopore fabrication. The measured 20 nm films are dense (3.07 g/cm3, about 96% of stoichiometric LPCVD SiNx), uniform (1.2% non-uniformity), and highly resistant to hot KOH (wet etch rate 0.172 nm/hr, the lowest among the methods compared). Nanopores made by controlled breakdown in these membranes show 1/f noise in the 1 Hz–100 kHz range similar to nanopores in LPCVD SiNx, and they can detect 2 kbp DNA translocations with single-file and folded events. Annealing the IBD film changes the initial pore size distribution after breakdown (5.5±3 nm non-annealed vs 20±13 nm annealed), which the authors attribute to grain formation in the annealed film. The paper concludes that IBD SiNx is the best BEOL-compatible candidate for CMOS-integrated solid-state nanopore systems.

Load-bearing premise

The load-bearing assumption is that the properties measured on flat 20 nm IBD films (density, stress, wet etch rate) hold for the freestanding 20 nm membranes after KOH release and controlled breakdown, so that the membranes stay insulating and mechanically intact during pore formation.

Editorial extensions

If this is right

  • If the claim holds, solid-state nanopore arrays can be fabricated on CMOS wafers at temperatures below 400°C, enabling on-chip amplification and readout of nanopore signals.
  • The IBD film's wet-etch resistance indicates it can survive standard KOH-based membrane release, allowing BEOL-compatible fabrication to reuse mature silicon micromachining steps.
  • Since IBD allows tuning of the Si:N ratio, future deposition recipes can optimize stoichiometry, potentially reducing leakage current and improving pore formation control.
  • Annealed IBD membranes produce larger initial pores (20±13 nm), which could be exploited as a direct route to large nanopores without lengthy conditioning.
  • The demonstrated DNA translocation and noise performance suggest IBD membranes can match LPCVD membranes for single-molecule sensing, the key metric for nanopore sequencing applications.

Reading between the lines

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

  • One implicit consequence is that the annealing-induced grain structure could be engineered to control pore size: if breakdown preferentially occurs at grain boundaries, controlling grain size could make initial pore diameter a tunable parameter rather than an uncontrolled outcome.
  • The paper compares IBD to ICPCVD and PEALD only on wet-etch resistance; a direct comparison of electrical noise and DNA sensing on those alternative membranes would test whether density and wet-etch resistance are the decisive predictors of nanopore quality.
  • The authors note a ~25% discrepancy between target and effective membrane thickness inferred from DNA blockage depths; a testable extension is to correlate ellipsometry thickness with the pore-thickness extracted from translocation data across many chips to calibrate the local-thinning RIE process.
  • A further testable extension is to fabricate IBD membranes with tunable Si/N ratios and measure both the breakdown voltage distribution and leakage current, checking whether silicon-rich films raise leakage and affect pore formation statistics.
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 paper proposes ion beam deposition (IBD) of SiNx as a CMOS back-end-of-line (BEOL)-compatible route to fabricate solid-state nanopores. It characterizes 20 nm IBD SiNx films by ellipsometry, X-ray reflectometry, stress measurement, and wet-etch testing, and compares them with LPCVD, ICPCVD, and PEALD films. It then fabricates nanopores by controlled breakdown in both non-annealed and annealed IBD membranes, reports representative power spectral densities and DNA translocation data, and concludes that IBD membranes can yield low-noise nanopores with signal-to-noise ratios comparable to LPCVD, opening a path toward CMOS-integrated nanopore systems.

Significance. If fully supported, this work would address a genuine bottleneck in solid-state nanopore integration: the thermal incompatibility of LPCVD SiNx with CMOS BEOL processing. The film characterization is quantitative and useful, especially the wet-etch resistance comparison among deposition methods, and the demonstration of DNA translocations in IBD membranes is a concrete step toward the stated goal. The paper is also commendable for benchmarking directly against LPCVD and for using an automated controlled-breakdown protocol. However, the central claim that IBD membranes reliably produce low-noise nanopores is currently supported by representative traces and lacks the statistical basis that a fabrication-oriented paper needs; this makes the significance conditional on additional yield and noise data.

major comments (4)
  1. [§3B, Figure 5] Section 3B defines the low-noise criterion as '<10 pA^2/Hz at 1 Hz on the PSD,' yet it reports only that 19 non-annealed and 14 annealed pores were fabricated, with representative PSDs shown in Figure 5; the paper does not state how many pores met this threshold, how many were excluded, or how the displayed traces were selected. This is load-bearing for the central claim that IBD membranes 'can successfully fabricate low-noise nanopores' and for the BEOL-integration conclusion, so please add the denominator, the distribution of PSD values at 1 Hz for all fabricated pores, and the selection rule for the representative traces.
  2. [§3B, Figure 4f] Figure 4f shows that annealed IBD pores have initial sizes of 20±13 nm when 8 nm was targeted, and the text states that 'catastrophic' pore growth is 'exceedingly rare' in LPCVD without quantifying its frequency in either IBD condition. Because the controlled-size claim and the comparison with LPCVD depend on this, please report the number of pores per condition that exhibited rapid growth, the distribution of post-fabrication sizes, and the criteria used to distinguish fast from slow growth.
  3. [Conclusion, Figure 5, Figure 6] The conclusion that the signal-to-noise ratio is 'comparable' to LPCVD is supported, in the current manuscript, by one LPCVD PSD and representative IBD PSDs (Figure 5), with no statistical comparison across devices and no SNR metric for the translocation events in Figure 6. Please provide quantitative noise statistics (e.g., PSD magnitude at 1 Hz, or event SNR) for multiple pores in each membrane type and test the comparison statistically.
  4. [§2B, §3A] The density, stress, and wet-etch data are obtained on 20 nm blanket films, while the nanopore results are obtained on freestanding membranes after KOH release, ProTEK coating, dicing, and cleaning; the manuscript does not re-characterize the released membranes (e.g., stoichiometry, defect density) or report how many released membranes survived to the pore-formation stage out of the total fabricated. Please add a yield statement and, if possible, a characterization of the freestanding film, so that the blanket-film properties can be linked to the membrane behavior.
minor comments (5)
  1. [§3A, Figure 3d, Figure 4 caption] There are several typographical errors: 'BOEL' should be 'BEOL' in Section 3A, 'represnt' should be 'represent' in the Figure 4 caption, and the Figure 3d caption contains the incomplete sentence fragment 'the four different membranes.'
  2. [§2C] The low-noise threshold is defined as applying 'for most pores'; please specify which pores it applies to and how exceptions were handled, since this directly affects the interpretation of the noise results.
  3. [§3B, Figure 4f] Please define how the 'initial pore size' was measured (e.g., from ionic current, TEM, or SEM) and at what point after fabrication it was determined, so that the size distribution can be interpreted.
  4. [§3B, effective thickness equation] The equation for effective pore thickness should define all symbols and state its assumptions; given that access resistance and the finite pore length are neglected, the resulting thicknesses (25 nm and 14 nm) should be labeled as rough estimates rather than precise deviations from the 20 nm target.
  5. [§3B] The hypothesis that annealing converts amorphous SiNx into a grain structure and thereby explains larger initial pore sizes is plausible but unsupported; please either add structural characterization (e.g., TEM or XRD) or label the explanation explicitly as a hypothesis.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the central claim is an empirical process demonstration benchmarked against LPCVD, not a fitted input renamed as a prediction.

full rationale

The paper's central claim is that ion-beam-deposited SiNx membranes can yield low-noise nanopores via controlled breakdown. This is an experimental demonstration: film properties (thickness, density, stress, wet etch rate) are measured directly, nanopore fabrication is performed in situ, and noise and translocation data are reported against an LPCVD benchmark. No quantity is fitted to a subset of data and then presented as a prediction; the effective-thickness estimate from DNA blockage is a post hoc consistency check, not a derived prediction. Self-citations [19], [21], and [23] cover fabrication details, the automated controlled-breakdown protocol, and low-noise instrumentation; these are established methods used as tools, and the paper's conclusion does not depend on accepting any self-cited result as a substitute for its own measurements. The unreported denominator for the noise threshold and the use of representative traces raise yield-reporting and selection concerns, but those are correctness or statistical issues, not circularity: the claim is not equivalent to its inputs by construction. Therefore no circular step is identified.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

The central claim rests on domain assumptions about controlled breakdown, noise metrics, and wet etch resistance as proxies for device suitability, plus hand-chosen process parameters. No new entities are postulated.

free parameters (4)
  • Annealing temperature and duration = 800 C for 240 s
    Chosen by hand to modify film stress and structure; central to the annealed membrane results.
  • Local thinning RIE time = 4.5 to 6 min
    Chosen to reach <20 nm thickness; final thickness varies and affects pore formation and effective thickness estimates.
  • Low-noise selection threshold = <10 pA^2/Hz at 1 Hz
    Used to decide which pores proceeded to DNA translocation experiments; this selection criterion affects the reported single-molecule data.
  • Target pore size after conditioning = 8 nm and 20 nm
    Design targets for the enlargement protocol; final initial pore sizes show broad distributions, so the target is not tightly controlled.
assumptions (5)
  • domain assumption Controlled breakdown (CBD) creates a single nanopore in a dielectric membrane when a voltage ramp induces dielectric breakdown.
    The paper uses CBD to fabricate pores without independent verification of the pore formation mechanism beyond the onset of ionic current; protocol is from ref [21].
  • domain assumption Low-frequency 1/f noise magnitude serves as a proxy for nanopore stability and sensing viability.
    Section 2C uses a PSD threshold at 1 Hz; the paper cites ref [23] for this relationship but does not independently validate it.
  • domain assumption Wet etch resistance in KOH indicates chemical stability and film quality relevant to nanopore device fabrication.
    Section 3A compares wet etch rates among deposition methods to conclude that IBD is the best BEOL-compatible candidate.
  • ad hoc to paper Annealing converts amorphous SiNx to a grain structure, explaining the larger initial pore sizes in annealed membranes.
    Section 3B offers this explanation as speculation without microstructural characterization such as TEM or XRD.
  • standard math The formula L = V * sigma * pi * d_DNA^2 / (4 * delta_I) estimates effective pore thickness from DNA blockage.
    Used in Section 3B; assumes a cylindrical pore and a known DNA diameter, and is described as a first-order approximation.

how reviews work

0 comments
Cite this review

Pith. "Pith review of A Method for Fabricating CMOS Back-End-of-Line-Compatible Solid-State Nanopore Devices." pith.science (2026). https://pith.science/paper/N4A7SYO4

@misc{pith2026241117416,
  author       = {Pith},
  title        = {Pith review of: A Method for Fabricating CMOS Back-End-of-Line-Compatible Solid-State Nanopore Devices},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/N4A7SYO4}},
  note         = {Machine review of arXiv:2411.17416}
}
read the original abstract

Solid-state nanopores, nm-sized holes in thin, freestanding membranes, are powerful single-molecule sensors capable of interrogating a wide range of target analytes, from small molecules to large polymers. Interestingly, due to their high spatial resolution, nanopores can also identify tags on long polymers, making them an attractive option as the reading element for molecular information storage strategies. To fully leverage the compact and robust nature of solid-state nanopores, however, they will need to be packaged in a highly parallelized manner with on-chip electronic signal processing capabilities to rapidly and accurately handle the data generated. Additionally, the membrane itself must have specific physical, chemical, and electrical properties to ensure sufficient signal-to-noise ratios are achieved, with the traditional membrane material being SiNX . Unfortunately, the typical method of deposition, low-pressure vapour deposition, requires temperatures beyond the thermal budget of CMOS back-end-of-line integration processes, limiting the potential to generate an on-chip solution. To this end, we explore various lower-temperature deposition techniques that are BEOL-compatible to generate SiNx membranes for solid-state nanopore use, and successfully demonstrate the ability for these alternative methods to generate low-noise nanopores that are capable of performing single-molecule experiments.

Figures

Figures reproduced from arXiv: 2411.17416 by the authors.

Figure 1
Figure 1. Assessment of the scalability and typical process temperatures of various film generation techniques. In particular, we highlight low-pressure vapour deposition (LPCVD), 2D material exfoliation, 2D material synthesis, ion beam deposition (IBD), inductively coupled plasma chemical vapor deposition (ICPCVD), and plasma enhanced chemical vapor deposition (PECVD), specifically inductively coupled plasma enhanced ALD (IC… view at source ↗
Figure 2
Figure 2. Overview of the IBD-based nanofabrication process to generate a SiNX membrane (a-j) and associated resultant images (k-m). a) Bare double-side polished silicon wafer with crystal orientation <100>. b) Ion beam deposition (IBD) of SiNX on both sides of the wafer generates a film of ~100 nm. C) Alignment of the etched pit side of the wafer with its appropriate photomask and patterning, through UV photolithography, of … view at source ↗
Figure 3
Figure 3. Characterization of 20 nm thick, non-annealed, SiNX membranes deposited by ion beam deposition (IBD). a) Film properties including thickness, roughness, stress, density, and refractive index of a 20nm SiNx membranes deposited on a silicon wafer by ion beam deposition technique. b) Ellipsometry result showing the wafer scale thickness and uniformity of the film deposited on a six-inch wafer. c) Wet etch test are cond… view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Solid-state nanopore fabrication by controlled breakdown and enlargment by voltage￾pulse method on nominally 20nm thick annealed and non-annealed SiNX membranes formed by IBD. a, b) Representative example of the formation of a solid-state nanopore for non-annealed and …
Figure 5
Figure 5. Figure 5: Power spectral density (PSD) of the current trace at the sensing voltage (±200 mV) in [PITH_FULL_IMAGE:figures/full_fig_p019_5.png]
Figure 6
Figure 6. Figure 6: DNA translocation experiments (5 nM 2 kbp no-limit DNA in 3.6 M LiCl in pH 8 at - 200 mV, low-pass filtered at 100 kHz) using annealed and non-annealed SiNX membranes formed by IBD. a,b) Baseline current trace showing several molecular translocations using non￾annealed…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

25 extracted references · 25 canonical work pages

  1. [1]

    Western Digital Research, San Jose, California, United States

    work page

  2. [2]

    Department of Physics, University of Ottawa, Ottawa, Ontario, Canada

    work page

  3. [3]

    Northern Nanopore Instruments (Canada). # Equal contribution †Corresponding author: tcossa@uOttawa.ca *Corresponding author: daniel.bedau@wdc.com Abstract: Solid-state nanopores, nm-sized holes in thin, freestanding membranes, are powerful single- molecule sensors capable of interrogating a wide range of target analytes, from small molecules to large poly...

    work page

  4. [4]

    Dong, Y., Sun, F., Ping, Z., Ouyang, Q., & Qian, L. (2020). DNA storage: research landscape and future prospects. National Science Review, 7(6), 1092-1107

    work page 2020

  5. [5]

    Anavy, L., Vaknin, I., Atar, O., Amit, R., & Yakhini, Z. (2019). Data storage in DNA with fewer synthesis cycles using composite DNA letters. Nature biotechnology, 37(10), 1229-1236

    work page 2019

  6. [6]

    & Shen, Y

    Ping, Z., Ma, D., Huang, X., Chen, S., Liu, L., Guo, F., ... & Shen, Y. (2019). Carbon-based archiving: current progress and future prospects of DNA-based data storage. GigaScience, 8(6), giz075. [4]Grass, R. N., Heckel, R., Puddu, M., Paunescu, D., & Stark, W. J. (2015). Robust chemical preservation of digital information on DNA in silica with error‐corr...

    work page 2019

  7. [7]

    Shrivastava, S., & Badlani, R. (2014). Data storage in DNA. International Journal of Electrical Energy, 2(2), 119-124

    work page 2014

  8. [8]

    F., & Dekker, C

    Schneider, G. F., & Dekker, C. (2012). DNA sequencing with nanopores. Nature biotechnology, 30(4), 326-328

    work page 2012

Show all 25 references
  1. [9]

    M., Butler, T

    Derrington, I. M., Butler, T. Z., Collins, M. D., Manrao, E., Pavlenok, M., Niederweis, M., & Gundlach, J. H. (2010). Nanopore DNA sequencing with MspA. Proceedings of the National Academy of Sciences, 107(37), 16060-16065

    2010

  2. [10]

    Kwok, H., Briggs, K., & Tabard-Cossa, V. (2014). Nanopore fabrication by controlled dielectric breakdown. PloS one, 9(3), e92880

    2014

  3. [11]

    Goto, Y., Akahori, R., Yanagi, I., & Takeda, K. I. (2020). Solid-state nanopores towards single-molecule DNA sequencing. Journal of human genetics, 65(1), 69-77

    2020

  4. [12]

    C., Cohen-Karni, T., Rosenstein, J

    Larkin, J., Henley, R., Bell, D. C., Cohen-Karni, T., Rosenstein, J. K., & Wanunu, M. (2013). Slow DNA transport through nanopores in hafnium oxide membranes. ACS nano, 7(11), 10121-10128

    2013

  5. [13]

    F., Kowalczyk, S

    Schneider, G. F., Kowalczyk, S. W., Calado, V. E., Pandraud, G., Zandbergen, H. W., Vandersypen, L. M., & Dekker, C. (2010). DNA translocation through graphene nanopores. Nano letters, 10(8), 3163-3167

    2010

  6. [14]

    Yang, C., & Pham, J. (2018). Characteristic study of silicon nitride films deposited by LPCVD and PECVD. Silicon, 10(6), 2561-2567

    2018

  7. [15]

    Olson, J. M. (2002). Analysis of LPCVD process conditions for the deposition of low stress silicon nitride. Part I: preliminary LPCVD experiments. Materials Science in Semiconductor Processing, 5(1), 51-60

    2002

  8. [16]

    W., Mackenzie, K

    Lee, J. W., Mackenzie, K. D., Johnson, D., Sasserath, J. N., Pearton, S. J., & Ren, F. (2000). Low temperature silicon nitride and silicon dioxide film processing by inductively coupled plasma chemical vapor deposition. Journal of The Electrochemical Society, 147(4), 1481

    2000

  9. [17]

    C., Mallikarjunan, A., Buchanan, I., Hausmann, D

    Faraz, T., van Drunen, M., Knoops, H. C., Mallikarjunan, A., Buchanan, I., Hausmann, D. M., ... & Kessels, W. M. (2017). Atomic layer deposition of wet-etch resistant silicon nitride using di (sec-butylamino) silane and N2 plasma on planar and 3D substrate topographies. ACS Ap...

    2017

  10. [18]

    Dergez, D., Schalko, J., Bittner, A., & Schmid, U. (2013). Fundamental properties of a- SiNx: H thin films deposited by ICP-PECVD for MEMS applications. Applied surface science, 284, 348-353

    2013

  11. [19]

    W., Dickinson, J

    Huang, L., Hipps, K. W., Dickinson, J. T., Mazur, U., & Wang, X. D. (1997). Structure and composition studies for silicon nitride thin films deposited by single ion bean sputter deposition. Thin Solid Films, 299(1-2), 104-109

    1997

  12. [20]

    Bundesmann, C., & Neumann, H. (2018). Tutorial: The systematics of ion beam sputtering for deposition of thin films with tailored properties. Journal of Applied Physics, 124(23)

    2018

  13. [21]

    Mu, G. (2022). SiN Drum Resonator Fabrication and Integrated Actuation Using Substrate Capacitors (Doctoral dissertation, Université d'Ottawa/University of Ottawa)

    2022

  14. [22]

    Seidel, H., Csepregi, L., Heuberger, A., & Baumgärtel, H. (1990). Anisotropic etching of crystalline silicon in alkaline solutions: I. Orientation dependence and behavior of passivation layers. Journal of the electrochemical society, 137(11), 3612

    1990

  15. [23]

    & Tabard-Cossa, V

    Waugh, M., Briggs, K., Gunn, D., Gibeault, M., King, S., Ingram, Q., ... & Tabard-Cossa, V. (2020). Solid-state nanopore fabrication by automated controlled breakdown. Nature Protocols, 15(1), 122-143

    2020

  16. [24]

    K., White, R

    Yen, B. K., White, R. L., Waltman, R. J., Dai, Q., Miller, D. C., Kellock, A. J., ... & Raman, V. (2003). Microstructure and properties of ultrathin amorphous silicon nitride protective coating. Journal of Vacuum Science & Technology A: Vacuum, Surfaces, and Films, 21(6), 1895-1904

    2003

  17. [25]

    Chapter 3 - Instrumentation for Low-Noise High-Bandwidth Nanopore Recording

    Tabard-Cossa, Vincent. 2013. “Chapter 3 - Instrumentation for Low-Noise High-Bandwidth Nanopore Recording.” In Engineered Nanopores for Bioanalytical Applications, edited by Joshua B. Edel and Tim Albrecht, 59–93. Micro and Nano Technologies. Oxford: William Andrew Publishing....

    2013 doi

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.