Pith. sign in

REVIEW 3 major objections 5 minor 1 cited by

Entanglement-based clock syntonization for quantum key distribution networks. Demonstration over a 50 km-long link

T0 review · 3 major / 5 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read Photon pairs from a working QKD link can synchronize the users' clocks to within 12 ps.

desk verdict Solid field demonstration of a known photon-pair sync protocol, but the 12 ps clock-offset claim is really a coincidence-peak servo residual and should be reframed. read the letter →

arxiv 2501.16796 v1 pith:6Q23SUBJ submitted 2025-01-28 quant-ph

Yoann Pelet , Grégory Sauder , Sébastien Tanzilli , Olivier Alibart , Anthony Martin This is my paper · ORCID
classification quant-ph MSC 81P94
keywords quantumkeydistributionentanglement-basedQKDclocksynchronizationsyntonizationrubidiumenergy-timeentanglementcoincidencemeasurementfield-deployedfibernetwork
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 demonstrates that two remotely located rubidium clocks can be kept aligned to within 12 picoseconds at all times using only the time-correlated photon pairs that an entanglement-based quantum key distribution (QKD) system already generates. This turns clock synchronization from an extra hardware service into a byproduct of the QKD protocol itself, requiring no dedicated reference signal or additional fiber. The demonstration runs continuously for 48 hours over 48 km of deployed optical fiber and maintains an average secret key rate of 7 kbps throughout. The authors argue that this level of stability keeps the 80 picosecond coincidence peak completely inside the 120 picosecond coincidence window, which is what makes uninterrupted key generation possible.

What carries the argument

The central mechanism is a closed correction loop built around a cross-correlation histogram of photon detection times. Bob's time-tagged detections are sent to Alice, who computes a coincidence histogram with 4 ps bins over a 1 ns window at a rate of roughly twice per second; fitting the central peak gives the current clock offset. Consecutive peak positions yield the frequency difference between the two rubidium clocks, and Alice's clock frequency is adjusted accordingly. This loop carries the entire argument: it converts the quantum correlation signal into a continuous, high-precision syntonization service that needs no external time reference.

What would settle it

Feed both end stations an independent common time reference, such as common-view GNSS or a White Rabbit link, over the same 48-hour period and compare that reference to the clock offset reported by the QKD-based synchronization loop; any excursion beyond 12 ps in the difference would falsify the claim.

Watch

Extended reading notes

Core claim

The central claim is that the coincidence peak reconstructed from Alice and Bob's detection timestamps gives a precise, unbiased estimate of the relative clock offset, and that feeding this estimate into a periodic frequency-correction loop can hold the two remote rubidium clocks together indefinitely. Without correction, the clocks drift by roughly 240 ps in ten seconds and up to 27 ns over 48 hours; with correction, the residual drift over a 48-hour run has a mean of 0.08 ps and a variance of 9 ps, never exceeding 12 ps. The paper also shows that the same loop survives external perturbations, such as a magnetic disturbance that worsens one clock's stability by a factor of twenty, keeping the drift below 44 ps. This is sufficient for QKD because the coincidence peak is only 80 ps wide and must remain inside a 120 ps coincidence window.

Load-bearing premise

The load-bearing premise is that the fitted coincidence peak is a precise and unbiased measure of the clock offset, combined with the assumption that the rubidium clocks drift slowly enough for corrections every few seconds to keep the residual under 12 ps.

Editorial extensions

If this is right

  • Over a full 48-hour run, the residual clock drift stays below 12 ps, which keeps the 80 ps coincidence peak fully inside the 120 ps coincidence window and therefore preserves the optimal secret key rate.
  • The protocol tolerates an external perturbation that degrades one clock's passive stability by a factor of twenty, keeping the drift below 44 ps and the system QKD-ready.
  • Synchronization holds up to 32 dB of total transmission losses; beyond that, the loop needs more than 60 seconds to accumulate enough coincidences and the precision worsens to worse than ±80 ps.
  • Dedicating a second wavelength channel to synchronization could raise the loss limit from 32 dB to about 48 dB, since the source can generate up to 40 pairs of 100 GHz channels.
  • Because the photon pairs carry the timing information, the protocol passively corrects optical path variations induced by temperature changes in the deployed fibers, a problem that standard synchronization methods must estimate and rectify separately.

Reading between the lines

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

  • The 12 ps bound is the loop's own measurement of its residual drift; an independent time reference, such as common-view GNSS or a distributed optical clock signal, would be needed to verify that the absolute offset is really that small.
  • If the coincidence-peak estimator remains unbiased under asymmetric dispersion, detector jitter drift, or temperature-induced path changes, the same mechanism could serve as a self-contained time-transfer service for any two nodes equipped with single-photon detectors, not only QKD users.
  • At high losses, where coincidence integration takes more than a minute, a predictive clock-drift model or Kalman filter could maintain the 12 ps bound with sparser peak measurements, since rubidium clocks drift slowly and predictably.
  • In a multi-user entanglement network, each entangled pair defines a pairwise syntonization relation, so the same photon pairs could in principle be used to align clocks across more than two nodes without dedicated synchronization hardware.
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

3 major / 5 minor

Summary. The paper reports a field implementation of a photon-pair-based time synchronization protocol on an entanglement-based QKD link deployed over 48 km of optical fiber in the Métropole Côte d'Azur. The protocol uses the strong temporal correlations of energy-time entangled photon pairs transmitted through the quantum channel to estimate and correct the relative drift between two rubidium clocks at the end stations. The authors demonstrate a sustained secret key rate of about 7 kbps and claim that the clocks never drift apart by more than 12 ps over 48-hour runs. They also argue that this approach requires less hardware than conventional synchronization methods, such as White Rabbit or dedicated reference-signal distribution, and that it passively compensates daily optical path variations.

Significance. If the central claim is fully supported, this is a practically useful result: it shows that an entanglement-based QKD link can be self-synchronizing to tens of picoseconds without dedicated synchronization hardware. The paper's strengths include a real-field deployment with a sustained key rate, a concrete complexity estimate for the histogram computation that makes frequent updates feasible, and a clear presentation of the system architecture. The protocol itself is taken from Ho et al. (ref. 17), but the new contribution is its integration into an operational QKD network and the characterization of its long-term behavior. The main weakness is that the headline 'under 12 ps at all time' claim is based on the feedback-loop residual, which does not directly measure the clock offset if optical path variations are absorbed by the loop.

major comments (3)
  1. [Section V, Abstract, and Conclusion] The central claim that 'the clocks never drift apart by more than 12 ps' is not established by the data shown. The blue curve in Fig. 4 is the residual of the coincidence-peak position after feedback correction, and this quantity is the sum of the clock offset and the differential optical-path delay between the two arms. Section IV explicitly states that the feedback is 'resilient to slow external perturbations, such as the variation of the optical path between night and day', and the Conclusion states that the protocol 'passively corrects the optical path variations'. Therefore slow path changes are absorbed into the clock-frequency corrections, and the 12 ps bound applies to the combined servo residual rather than to the clock offset alone. The authors should either quantify the optical-path stability (for example with a co-propagated classical reference or an independent time reference) or restrict the claim to the coincidence-peak position. As written, the abstract's guarantee is overstated.
  2. [Section V, Fig. 4] The statistical summary 'mean value of 0.08 ps and a variance of 9 ps' does not support the statement 'never drifting apart from more than 12 ps during the whole experiment'. A variance of 9 ps (standard deviation 3 ps) is compatible with peak excursions well above 12 ps over a 48-hour run, especially if the residual has non-Gaussian tails. The maximum absolute residual for the unperturbed run should be reported; for comparison, the perturbed run in Fig. 5 states a maximum drift of 44 ps. As written, the 'at all time' bound is not backed by the displayed metric.
  3. [Section II.B and Section IV] The assumption that δt_photons is 'mostly fixed in time' (Section II.B) is in tension with the claimed passive correction of daily optical path variations (Section IV and Conclusion). Unless the optical path variation is independently bounded below 12 ps, the feedback loop's corrections to Alice's clock frequency do not yield a measurement of the true clock offset. The paper should provide such a bound or explicitly acknowledge that the reported 12 ps stability is a property of the combined clock-plus-channel system, not of the clocks alone.
minor comments (5)
  1. [Title and Abstract] The title says a '50 km-long link' while the abstract and body state 48 km of deployed fiber; the numbers should be harmonized.
  2. [Fig. 2 caption and Section IV] The caption says the histogram is plotted twice per second, while Section IV says that 'only one measurement of δt_clock is required every few seconds' and also that histograms can be calculated 'several times per second'; the actual update rate should be stated consistently.
  3. [Section III] There is a typo in 'Ctimstamp' (should be 'C_timestamp'), and the sentence beginning 'more specifically' after a period is ungrammatical.
  4. [Fig. 1 caption] 'oranges lines' should read 'orange lines'.
  5. [Conclusion] 'pacing two rubidium atomic clocks' should be 'synchronizing' or 'pairing'.

Circularity Check

2 steps flagged · score 6.0 of 10

The 'never more than 12 ps' clock-offset claim is the residual of the same coincidence-peak feedback loop that corrects the clocks, so slow optical-path drift is absorbed and the true clock offset is not independently measured.

  1. self definitional [Section IV (Frequency drift correction) and Section V (Results), Fig. 4]
    "These calculations allow the extraction of the temporal evolution of δtclocks, which can then be linked to the frequency drift between the two clocks, to correct it. ... We achieve a long term synchronization for the whole duration of the QKD protocol, with the clocks never drifting apart from more than 12 ps during the whole experiment."

    δtclock is measured as the coincidence-peak position, and the same measurement drives the correction of Alice's clock frequency. Fig. 4's 'Corrected drift' is therefore the residual error of this feedback loop, not an independent measurement of the clock offset. The peak position equals the clock offset plus the differential optical-path delay; since the paper states the protocol 'passively corrects the optical path variations induced by the deployed fibers', slow path changes are absorbed into the clock correction. The reported <12 ps residual is small by construction of the loop, so the conclusion that the clocks themselves never drift by more than 12 ps does not follow from the data.

  2. other [Section II.B and Section VI (Conclusion)]
    "while the position of the coincidence peak (δtphotons) is mostly fixed in time, the coincidence window’s position (δtclock) is affected by the drift of the user’s clocks. ... this protocol also passively corrects the optical path variations induced by the deployed fibers."

    The argument that the measured peak drift is clock drift relies on δtphotons being constant. The paper simultaneously claims the protocol passively corrects optical-path variations, which means δtphotons is not constant. The loop cannot distinguish clock drift from path drift; therefore the identification of the corrected residual with clock offset is an unverified assumption, not a measured quantity. This assumption is what makes the 12 ps claim circular: the output (clock drift) is defined as the residual of a loop that absorbs both effects.

full rationale

The synchronization scheme itself is taken from an external source (ref. 17, Ho et al. 2009), so the protocol concept is not circular. The authors' self-citation (ref. 7) describes the QKD testbed and is not load-bearing for the synchronization claim; the deployed link and sustained key rate are independent experimental facts. The circularity sits in the certification of the headline stability number. The measured quantity, the coincidence-peak position, is the controlled variable of a feedback loop that adjusts Alice's clock frequency. The reported 12 ps residual is therefore the loop's own error signal, which is kept small by construction. Because the paper explicitly says the protocol passively corrects optical-path variations, the loop absorbs slow path changes into the clock correction, so the true clock offset can wander beyond 12 ps while the reported residual stays below it. The sustained 7 kbps key rate independently confirms that the coincidence peak remains inside the coincidence window, but it does not confirm that the clocks themselves are syntonized to 12 ps. The central claim thus partially reduces to a self-referential control measurement, warranting a score of 6.

Assumptions & free parameters 1 free parameters · 4 assumptions · 0 invented entities

The central claim rests on standard physics (energy-time entanglement correlations), one domain assumption about the peak position tracking the clock offset, and an operational premise about rubidium clock drift rate from manufacturer specifications. No invented entities. The only fitted choice is the 120 ps coincidence window, tuned experimentally; the 12 ps result is a measured residual, not a fit.

free parameters (1)
  • coincidence_window_width = 120 ps
    Section II.B: 'we determine experimentally that the optimal width of this window is 120 ps.' The synchronization tolerance (30 ps) and the 12 ps claim are defined relative to this window, so the window choice shapes the stated requirement, though the 12 ps value itself is a measured residual rather than a fit.
assumptions (4)
  • domain assumption The coincidence peak position equals the clock offset plus a fixed optical path delay, so tracking the peak tracks the clock offset and passively captures path variations.
    Sections II.B and III; this identification is the core of the protocol. If the peak position were affected by other slowly varying biases (e.g., detector jitter drift), the correction loop would chase a biased target.
  • domain assumption Rubidium clocks drift at about 7 ps per second, so correcting once every few seconds keeps the residual below the 30 ps tolerance.
    Section IV quotes the Spectratime LNRClock 1500 specification 'an average time drift of 7 ps/s'; no measured Allan deviation is presented in this paper.
  • domain assumption Fiber dispersion is fully compensated by a negative dispersion fiber and can be neglected in the peak width budget.
    Section II.B cites ref 20 for the non-local compensation; residual dispersion would broaden the 80 ps peak and tighten the synchronization requirement.
  • domain assumption The coincidence peak can be fitted with negligible bias from histograms with 4 ps bins computed 'twice per second' (Fig. 2 caption), so the servo can resolve well below 12 ps.
    The estimation precision of the peak fit is not analyzed explicitly; the 4 ps bin size and the count rates (3 MHz at Bob, 10 MHz at Alice) are stated, but the resulting fit uncertainty is not quantified.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Entanglement-based clock syntonization for quantum key distribution networks. Demonstration over a 50 km-long link." pith.science (2026). https://pith.science/paper/6Q23SUBJ

@misc{pith2026250116796,
  author       = {Pith},
  title        = {Pith review of: Entanglement-based clock syntonization for quantum key distribution networks. Demonstration over a 50 km-long link},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/6Q23SUBJ}},
  note         = {Machine review of arXiv:2501.16796}
}
read the original abstract

We present the implementation of a time synchronization protocol as part of an experimentally deployed entanglement-based quantum key distribution (QKD) link. The system is deployed over 48 km of optical fibers across the M\'etropole C\^ote d'Azur and enables secret cryptographic key exchange between two remote users, with an average rate of 7 kbps. We exploit the time correlation of paired photons generated by a high-quality source of energy-time entanglement implemented in the QKD link to synchronize two rubidium clocks located at the end stations. The level of stability achieved guarantees a time offset between the clocks under 12 ps at all time. We also show that this protocol requires less hardware than a typical synchronization protocol for QKD that would distribute a reference clock signal between the users.

Figures

Figures reproduced from arXiv: 2501.16796 by the authors.

Figure 1
Figure 1. FIG. 1. Bottom: general layout of the deployed fibers linking the source to each end station. oranges lines each represent two [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. FIG. 2. Coincidence histogram plotted twice per second to [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. FIG. 3. Rough synchronization histogram calculated during [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: FIG. 4. Corrected drift (blue) for a 48-hour-long measure [PITH_FULL_IMAGE:figures/full_fig_p004_4.png]
Figure 5
Figure 5. Figure 5: FIG. 5. Relative drift between the two remote clocks. An [PITH_FULL_IMAGE:figures/full_fig_p005_5.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Clock Synchronization for Drone-Based Entanglement Quantum Key Distribution

    quant-ph 2025-06 conditional novelty 5.0 of 10

    A two-stage protocol, GPS pulse alignment followed by entanglement-based temporal correction, achieves 24 ± 12 ps RMS clock drift in a lab emulation of a dynamic free-space drone channel.

Reference graph

Works this paper leans on

23 extracted references · 19 canonical work pages · cited by 1 Pith paper

  1. [1]

    Labont \'e , author O

    author author L. Labont \'e , author O. Alibart , author V. D’Auria , author F. Doutre , author J. Etesse , author G. Sauder , author A. Martin , author \'E . Picholle , \ and\ author S. Tanzilli ,\ @noop journal journal PRX Quantum \ volume 5 ,\ pages 010101 ( year 2024 ) NoStop

    work page 2024

  2. [2]

    Xu , author X

    author author F. Xu , author X. Ma , author Q. Zhang , author H.-K. \ Lo , \ and\ author J.-W. \ Pan ,\ @noop journal journal Reviews of modern physics \ volume 92 ,\ pages 025002 ( year 2020 ) NoStop

    work page 2020

  3. [3]

    Pirandola , author U

    author author S. Pirandola , author U. L. \ Andersen , author L. Banchi , author M. Berta , author D. Bunandar , author R. Colbeck , author D. Englund , author T. Gehring , author C. Lupo , author C. Ottaviani , et al. ,\ @noop journal journal Advances in optics and photonics \ volume 12 ,\ pages 1012 ( year 2020 ) NoStop

    work page 2020

  4. [4]

    Rusca \ and\ author N

    author author D. Rusca \ and\ author N. Gisin ,\ @noop journal journal arXiv preprint arXiv:2411.04044 \ ( year 2024 ) NoStop

    arXiv 2024

  5. [5]

    Gr \"u nenfelder , author A

    author author F. Gr \"u nenfelder , author A. Boaron , author G. V. \ Resta , author M. Perrenoud , author D. Rusca , author C. Barreiro , author R. Houlmann , author R. Sax , author L. Stasi , author S. El-Khoury , et al. ,\ @noop journal journal Nature Photonics \ volume 17 ,\ pages 422 ( year 2023 ) NoStop

    work page 2023

  6. [6]

    Liu , author W.-J

    author author Y. Liu , author W.-J. \ Zhang , author C. Jiang , author J.-P. \ Chen , author C. Zhang , author W.-X. \ Pan , author D. Ma , author H. Dong , author J.-M. \ Xiong , author C.-J. \ Zhang , et al. ,\ @noop journal journal Physical Review Letters \ volume 130 ,\ pages 210801 ( year 2023 ) NoStop

    work page 2023

  7. [7]

    Pelet , author G

    author author Y. Pelet , author G. Sauder , author M. Cohen , author L. Labont \'e , author O. Alibart , author A. Martin , \ and\ author S. Tanzilli ,\ @noop journal journal arXiv preprint arXiv:2207.14707 \ ( year 2022 ) NoStop

    work page arXiv 2022

  8. [8]

    Wu , author L

    author author J. Wu , author L. You , author S. Chen , author H. Li , author Y. He , author C. Lv , author Z. Wang , \ and\ author X. Xie ,\ 10.1364/AO.56.002195 journal journal Appl. Opt. \ volume 56 ,\ pages 2195 ( year 2017 ) NoStop

    work page doi:10.1364/ao.56.002195

Show all 23 references
  1. [9]

    Chen , author D

    author author S. Chen , author D. Liu , author W. Zhang , author L. You , author Y. He , author W. Zhang , author X. Yang , author G. Wu , author M. Ren , author H. Zeng , author Z. Wang , author X. Xie , \ and\ author M. Jiang ,\ 10.1364/AO.52.003241 journal journal Appl. Opt...

    doi

  2. [10]

    Schmitt-Manderbach , author H

    author author T. Schmitt-Manderbach , author H. Weier , author M. F\"urst , author R. Ursin , author F. Tiefenbacher , author T. Scheidl , author J. Perdigues , author Z. Sodnik , author C. Kurtsiefer , author J. G. \ Rarity , author A. Zeilinger , \ and\ author H. Weinfurter ...

    doi

  3. [11]

    Ribezzo , author M

    author author D. Ribezzo , author M. Zahidy , author I. Vagniluca , author N. Biagi , author S. Francesconi , author T. Occhipinti , author L. K. \ Oxenl we , author M. Lon c ari \'c , author I. Cviti \'c , author M. Stip c evi \'c , et al. ,\ @noop journal journal Advanced Qu...

    2023

  4. [12]

    Du , author L

    author author D. Du , author L. Castillo-Veneros , author G. Cui , author D. Cottrill , author J. Mart\' i nez-Rinc\' o n , author P. Stankus , author D. Katramatos , \ and\ author E. Figueroa ,\ in\ 10.1364/QUANTUM.2022.QM2B.3 booktitle Quantum 2.0 Conference and Exhibition \...

    2022 doi

  5. [13]

    Dynes , author A

    author author J. Dynes , author A. Wonfor , author W.-S. \ Tam , author A. Sharpe , author R. Takahashi , author M. Lucamarini , author A. Plews , author Z. Yuan , author A. Dixon , author J. Cho , et al. ,\ @noop journal journal npj Quantum Information \ volume 5 ,\ pages 1 (...

    2019

  6. [14]

    Calderaro , author A

    author author L. Calderaro , author A. Stanco , author C. Agnesi , author M. Avesani , author D. Dequal , author P. Villoresi , \ and\ author G. Vallone ,\ 10.1103/PhysRevApplied.13.054041 journal journal Phys. Rev. Appl. \ volume 13 ,\ pages 054041 ( year 2020 ) NoStop

    doi

  7. [15]

    author author A. V. \ Miller ,\ 10.1088/1742-6596/2701/1/012017 journal journal Journal of Physics: Conference Series \ volume 2701 ,\ pages 012017 ( year 2024 ) NoStop

    2024 doi

  8. [16]

    Wang , author Y

    author author C. Wang , author Y. Li , author W. Cai , author M. Yang , author W. Liu , author S. Liao , \ and\ author C. Peng ,\ 10.1364/AO.425085 journal journal Appl. Opt. \ volume 60 ,\ pages 4787 ( year 2021 ) NoStop

    2021 doi

  9. [17]

    Ho , author A

    author author C. Ho , author A. Lamas-Linares , \ and\ author C. Kurtsiefer ,\ @noop journal journal New Journal of Physics \ volume 11 ,\ pages 045011 ( year 2009 ) NoStop

    2009

  10. [18]

    author author C. H. \ Bennett , author G. Brassard , \ and\ author N. D. \ Mermin ,\ 10.1103/PhysRevLett.68.557 journal journal Phys. Rev. Lett. \ volume 68 ,\ pages 557 ( year 1992 ) NoStop

    1992 doi

  11. [19]

    Martinez-Mateo , author C

    author author J. Martinez-Mateo , author C. Pacher , author M. Peev , author A. Ciurana , \ and\ author V. Martin ,\ @noop journal journal arXiv preprint arXiv:1407.3257 \ ( year 2014 ) NoStop

    2014 arXiv

  12. [20]

    Steinberg , author P

    author author A. Steinberg , author P. Kwiat , \ and\ author R. Chiao ,\ @noop journal journal Physical review letters \ volume 68 ,\ pages 2421 ( year 1992 ) NoStop

    1992

  13. [21]

    Lipi \'n ski , author E

    author author M. Lipi \'n ski , author E. van der Bij , author J. Serrano , author T. W ostowski , author G. Daniluk , author A. Wujek , author M. Rizzi , \ and\ author D. Lampridis ,\ in\ @noop booktitle 2018 IEEE International Symposium on Precision Clock Synchronization for...

    2018

  14. [22]

    Valivarthi , author L

    author author R. Valivarthi , author L. Narv \'a ez , author S. I. \ Davis , author N. Lauk , author C. Pe \ n a , author S. Xie , author J. P. \ Allmaras , author A. D. \ Beyer , author B. Korzh , author A. Mueller , et al. ,\ @noop journal journal Journal of Lightwave Techno...

    2022

  15. [23]

    author author J. M. \ Merlo , author S. R. \ Mghabghab , \ and\ author J. A. \ Nanzer ,\ @noop journal journal IEEE Transactions on Microwave Theory and Techniques \ volume 71 ,\ pages 1720 ( year 2022 ) NoStop

    2022

Pith tools

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