{"id":"7b7627c7-36f6-46f6-942a-c4108daeeccd","arxiv_id":"2508.11023","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"The KiQQer experiment demonstrated hybrid fiber/free-space entanglement distribution between two edge nodes through a central source, reporting CHSH S = 2.63 for the full hybrid path.","lead":"A three-node quantum network on the TU Delft campus distributed polarization-entangled photon pairs over a 200-meter fiber link and a 150-meter free-space link, using commercially available hardware. The field test reports CHSH values above the classical bound, supporting the feasibility of hybrid fiber/free-space quantum networks built from off-the-shelf components.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"S=2.63 rests on a 'selective ROI' and discarded sync-error datasets; without a pre-registered selection rule, the entanglement claim is not independently checkable.","rationale":"The reader's weakest_assumption identifies the same load-bearing concern: the reported S-values depend on a selective ROI and on discarding some datasets, and these choices are not shown to be independent of the measured visibility. I independently checked the manuscript for all references to ROI and dataset exclusion. Figure 6's caption says 'selective region of interest (ROI)'; Figure 8's caption says 'narrow ROI'; Section III states that 'synchronization errors of timetaggers caused some data sets to be discarded.' These are the only places the analysis pipeline is described. The central claim is explicitly 'The results in Fig. 8 confirm successful entanglement distribution in the three-node configuration,' so the whole paper rests on S=2.63. The paper does provide supportive internal consistency: QBER 2.88%, detection rate 15 counts/s per channel combination, a 13.6 dB key-rate reduction consistent with the estimated 13.8 dB optical loss, and a stated FWHM-based integration window. That support makes the claim plausible but not fully verifiable. The lack of uncertainties is not by itself fatal; a 30 s integration at 15 counts/s gives a Poisson error on the order of 0.1, so S=2.63 would be nominally significant if the counts were independent and the selection rule were fixed. The problem is the uncontrolled selection step, which can bias S upward without showing up in Poisson statistics. Because the authors could resolve this by releasing raw timetag data and a precise pre-registered analysis rule, the appropriate verdict is CONDITIONAL, unchanged from the reader's verdict. No evidence of fabrication is present; the concern is about transparency and falsifiability of the central quantitative claim.","tokens_in":15857,"tokens_out":4713,"duration_ms":54435,"concrete_test":"Download the raw time-tag records for the 11–12 Oct. 2024 A–B run and rerun the CHSH analysis under a pre-registered rule: (1) fix the coincidence window to the independently measured FWHM (1077±106 ps) centered at zero delay, not per-point; (2) keep every dataset whose timetagger sync flag was set before data inspection; (3) compute S from raw counts with Poisson/bootstrap uncertainties. If the recomputed S is below 2, or if its lower confidence bound does not exceed 2, the central claim fails. A useful robustness check is to vary the ROI width from ±0.5 ns to ±2 ns and report whether S remains above 2 for all widths.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The demonstration's only quantitative evidence for the headline claim is the reported CHSH visibility, S=2.63 (Fig. 8). Section III says each point in Fig. 6 is taken 'over a selective region of interest (ROI)' and Fig. 8 is 'over a narrow ROI'; the text immediately after Fig. 8 adds that 'synchronization errors of timetaggers caused some data sets to be discarded.' No raw counts, no error bars, and no pre-specified ROI or exclusion rule are given. If the ROI boundaries or the accept/reject decision were chosen after looking at the measured coincidence counts, the fitted visibility and S can be inflated. This is the load-bearing assumption: retained data faithfully represent the true photon-pair correlations. Standard coincidence gating around an independently measured FWHM would be acceptable, and the paper does state that integration windows were set from measured FWHM values (995±87 ps and 1077±106 ps), but the words 'selective'/'narrow' and the unreported dataset exclusions leave an uncontrolled degree of freedom. Without a fixed rule or raw data, S=2.63 cannot be independently checked.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports a field demonstration of polarization-entanglement distribution over a three-node network at the TU Delft campus, using a central node (Charlie) with a bichromatic entangled-photon source and two edge nodes connected by a roughly 150 m free-space link (Alice) and a roughly 200 m fiber link (Bob). The central claim is that CHSH measurements on both links show Bell-inequality violations, with S = 2.55 for the C-B fiber link and S = 2.63 for the hybrid A-B path, corresponding to QBER values of about 2.8% and estimated secret-key rates of 770 bits/s and 33 bits/s. The paper also describes the synchronization architecture, a quantum time-transfer analysis, and the integration of commercial components from the Dutch quantum ecosystem.","tokens_in":16088,"tokens_out":6279,"duration_ms":74960,"significance":"If the reported S-values are statistically robust and the analysis rules are fixed, the experiment is a meaningful systems-integration milestone: it would demonstrate hybrid fiber/free-space entanglement distribution with off-the-shelf equipment, with a concrete loss budget, timing characterization, and application-oriented extrapolations. The strengths are the level of hardware detail, the explicit link-loss accounting, and the demonstration of simultaneous classical and quantum channel operation. However, the central quantitative evidence for entanglement rests on CHSH S = 2.63, and the manuscript currently does not make that evidence independently checkable: the ROI selection and dataset-retention rules are not specified, no uncertainties or raw counts are given, and the conversion from measured visibilities to S is not shown. The significance of the demonstration is therefore conditional on the authors providing a transparent, fixed analysis pipeline.","major_comments":[{"comment":"The load-bearing evidence for entanglement is not independently checkable because the analysis rules are not specified. The caption of Fig. 6 states that each point is taken over a 'selective region of interest (ROI)', the caption of Fig. 8 states that each point is over a 'narrow ROI', and the paragraph after Fig. 8 says that 'synchronization errors of timetaggers caused some data sets to be discarded.' The manuscript does not define the ROI boundaries, does not state whether the ROI was fixed before inspecting the coincidence counts, and does not give the number of discarded datasets or the discard criterion. If the ROI or acceptance rule was chosen after looking at the measured visibility, the reported S could be inflated. Please provide a precise, pre-specified analysis rule (e.g., a fixed time window derived from the independently measured FWHM of 995 +/- 87 ps and 1077 +/- 106 ps), report the raw counts for all retained and discarded runs, and show that the same S-values are obtained with a uniform ROI applied to all points.","section":"Section III, Figs. 6 and 8"},{"comment":"No uncertainties are reported for the CHSH values, visibilities, or QBER. At the quoted average rate of 15 counts/s per channel combination and 30 s integration, each correlation point contains on the order of 450 counts, so Poisson fluctuations are at the few-percent level; the resulting uncertainty in S is not negligible and must be quantified. Please report confidence intervals for S (e.g., via Poisson Monte Carlo or bootstrap) using the same data-selection rules, and verify that the lower confidence bound exceeds 2. The manuscript should also state whether accidental coincidences were subtracted and, if so, how the accidentals were estimated.","section":"Section III, Fig. 8"},{"comment":"The calculation of S from the measured data is not shown. The text reports average visibility, QBER, and S-values, and the figure captions refer to sinusoidal fits, but the four correlation functions E(theta, theta') used in the CHSH expression are not given, and the specific measurement settings are not stated. Please present the exact Bell settings, the four E-values for both the C-B and A-B links, and the formula used to combine them. The normalized arbitrary-unit plots alone do not allow a reader to reproduce S = 2.55 or S = 2.63.","section":"Section III, Figs. 6 and 8"}],"minor_comments":[{"comment":"The text says the integration window was set to 1.1 ns, 'i.e., events within a +/- 550ns' from zero delay; this should read +/- 550 ps, and the typo should be corrected in any revision.","section":"Section III, paragraph after Fig. 8"},{"comment":"The appendix states that wavelength-division multiplexing allows simultaneous quantum and classical communication over a single fiber link connecting Nodes C and A, but the main text and Fig. 1 place the fiber link between C and B and the free-space link between C and A; this contradiction should be resolved.","section":"Appendix E"},{"comment":"The text says the free-space synchronization protocol introduces 200 ps jitter, while Fig. 3 reports a measured FWHM of 541 ps; please reconcile the two numbers or clarify which contributions are included in each.","section":"Section III and Fig. 3"},{"comment":"The secret-key-rate estimate uses a key-extraction efficiency factor of 0.56 from Ref. [19], which concerns decoy-state BB84; the applicability of this factor to the entanglement-based setup should be justified, or the rate should be presented as an illustrative estimate with stated protocol assumptions.","section":"Section IV.1"},{"comment":"The pointing-stability estimate of ~16 urad is given without an uncertainty; a short propagation of the stated voltage and power uncertainties would strengthen the claim.","section":"Appendix A, pointing stability"},{"comment":"The phrase 'off-the-shelf' is somewhat overstated because the optical heads are described as custom-designed by TNO; consider rephrasing to 'commercially available components together with custom optical heads.'","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"The stress-test concern is real and is the basis for the major comments: the selective/narrow ROI language and the unreported dataset exclusions leave an uncontrolled degree of freedom in the only quantitative evidence for entanglement. If the authors can supply raw data, a fixed analysis rule, and statistical uncertainties, the result is likely publishable as a systems-integration demonstration. I would not require loophole-free Bell tests for this type of network demonstration, but I would require unambiguous statistical evidence that S exceeds 2 under a pre-specified analysis."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"What you should know: this is an engineering demonstration, not a new-physics claim, and it is mostly credible. The genuinely new piece is the three-node campus network (Alice–Charlie–Bob) that combines a 150 m free-space link and a 200 m fiber link using off-the-shelf components, with a bichromatic 795/1324 nm source that is well matched to the hybrid channels. That combination, and the detailed system integration around it, is worth having on record. It does not push any fundamental boundary, but it is a solid step toward practical hybrid quantum networking.\n\nThe paper does several things well. The system description is unusually complete: loss budgets, synchronization architecture, detector specs, and partner contributions are laid out clearly. The C–B fiber link is convincing on its own: S = 2.55, 94.3% visibility, 2.8% QBER, and a secret key rate of 770 bit/s. The A–B free-space result is internally consistent too, with the 33 bit/s rate and the 13.6 dB drop matching the estimated optical loss. The citation pattern looks fine; references to the AFRL preprint and prior satellite work are appropriate.\n\nThe soft spot is the one the stress-test flagged, and it is real but not fatal. The headline A–B entanglement evidence is S = 2.63, with no uncertainty and no raw counts. The captions say each point is taken over a 'selective' or 'narrow ROI,' and the text says some datasets were discarded due to synchronization errors. The integration windows themselves are described as set from measured FWHMs (995 ± 87 ps, 1077 ± 106 ps), which is the right approach. But the words 'selective' and 'discarded' leave an uncontrolled degree of freedom: if the ROI boundaries or the accept/reject decision were made after looking at the measured coincidences, S could be inflated. I don't see evidence of fabrication, and the internal consistency is good, but the paper currently cannot be independently checked. The fix is straightforward: release the raw coincidence counts, state the ROI rule and dataset-exclusion rule explicitly, and report bootstrap or Poisson uncertainties on S.\n\nMinor point: the abstract's talk about a 'pathway to space' is overstated for a 150 m terrestrial demo, but that is marketing gloss, not a technical flaw. The QTT section is a side application and not load-bearing.\n\nWho this is for: experimentalists and systems engineers working on quantum networks, QKD deployment, and hybrid link integration. It deserves a serious peer review, not a desk reject. The condition for acceptance should be a data-release requirement—raw counts, selection rules, and error bars. With that, this would be a useful reference for the field.","headline":"A credible and useful integration milestone for hybrid fiber/free-space quantum networks, but the headline CHSH numbers need raw data and explicit selection rules before I'd trust them fully.","tokens_in":16767,"tokens_out":2089,"would_cite":false,"duration_ms":25806,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper reports successful distribution of polarization-entangled photon pairs across a campus-scale three-node network whose two arms are one fiber link and one free-space optical link, built entirely from commercially available…","keywords":["entanglement distribution","hybrid quantum network","free-space optical link","polarization-entangled photons","quantum key distribution","quantum time transfer","CHSH inequality","single-photon detectors"],"falsifier":"Recompute the CHSH value for the A–B path using all recorded time tags with no region-of-interest trimming, or with a region of interest and an acceptance rule fixed before unmasking the visibility curve; if the resulting S falls at or below 2, the entanglement claim fails. A reader could also rerun the campaign with the same hardware and a pre-registered analysis pipeline to see whether S = 2.63 reproduces.","tokens_in":15661,"feed_emoji":"🔗","tokens_out":7058,"duration_ms":76799,"temperature":0.7,"pith_summary":"This paper reports a field demonstration in which polarization-entangled photon pairs were distributed across a three-node network on a university campus, with one arm running over roughly 200 m of fiber and the other over a roughly 150 m free-space atmospheric link. A central node generated the pairs and sent one photon to each edge node; local polarization measurements gave correlation visibilities that violate the classical CHSH bound, with S = 2.63 for the combined fiber/free-space path. The whole system was assembled from off-the-shelf components supplied by commercial partners. A sympathetic reading is that this establishes the practical maturity of hybrid fiber/free-space entanglement distribution and shows a concrete route toward metropolitan quantum networks, entanglement-based quantum key distribution, and quantum time transfer. The authors report an estimated secret-key rate of 33 bits/s over the full A–B path and 770 bits/s over the fiber-only C–B path.","feed_headline":"Three-node quantum link spans fiber and free space with S = 2.63","feed_subtitle":"Off-the-shelf parts share polarization-entangled pairs between two edge nodes at 33 bits/s.","key_machinery":"The load-bearing component is a bichromatic entangled-photon-pair source based on spontaneous four-wave mixing in warm rubidium vapor, which generates the Bell state $|\\Phi^+\\rangle$ with the signal photon at 795 nm (resonant with rubidium’s D1 line and suited to free space) and the idler at 1324 nm (telecom O-band and suited to fiber). Around this source, the experiment combines three synchronization and multiplexing mechanisms: a White Rabbit clock distributed over fiber for the C–B arm, a free-space pilot tone with active beam steering for the C–A arm, and coarse wavelength-division multiplexing that lets classical Gigabit Ethernet share the same fiber as the quantum idler. The GHz-class linewidth and high brightness of the source let the receivers isolate true coincidence peaks with narrow regions of interest, while superconducting nanowire single-photon detectors provide picosecond timing and very low dark counts.","core_discovery":"The central discovery is that a three-node network built from commercially available hardware can distribute polarization entanglement over a hybrid channel in which one photon of each pair travels through standard optical fiber and the other through open air, with both arms synchronized to a common clock. Using a bichromatic source emitting 795 nm signal and 1324 nm idler photons in the maximally entangled Bell state $|\\Phi^+\\rangle = (|HH\\rangle + |VV\\rangle)/\\sqrt{2}$, the authors measured polarization correlations at the two edge nodes whose sinusoidal visibility curves exceed the classical bound, yielding S = 2.55 for the fiber-connected node and S = 2.63 for the full path between the two edge nodes. They report an average quantum bit error rate of 2.88%, an estimated secret-key rate of 33 bits/s, and a quantum time-transfer clock-offset residual of 594 ps after 30 s of integration. The authors state that these results confirm successful entanglement distribution in the three-node configuration.","pith_inferences":["A stronger test of the entanglement claim, not performed in the paper, would fix the region of interest and the dataset-acceptance rules before any visibility scan and then compute S from all retained events; the deployed setup makes this straightforward.","The one-way quantum time-transfer analysis suggests that a reciprocal two-way link would cancel common-mode clock drift and push clock-offset residuals below the reported 594 ps, a test the same hardware could support.","Because the free-space arm already uses active beam steering and clock recovery from a pilot tone, the deployment pattern is directly transferable to moving platforms such as drones, ships, or ground-to-satellite uplinks, though atmospheric and Doppler effects would need separate characterization.","The loss budget implies roughly 13.8 dB of headroom between the free-space arm and the fiber-only arm, so increasing source brightness or receiver aperture could lift the A–B secret-key rate from 33 bits/s toward the 770 bits/s achieved on the fiber-only path."],"forward_implications":["A metropolitan quantum key distribution link can be built from off-the-shelf components today, with a demonstrated secret-key rate of 33 bits/s over the combined fiber/free-space path and 770 bits/s over the fiber-only arm.","The same entanglement can serve as a timing resource: with 30 s of data, the quantum time-transfer algorithm pinned the clock offset between two edge nodes to 594 ps, consistent with the detector jitter budget.","Because the 795 nm photon is resonant with rubidium’s D1 transition, the network can in principle connect directly to atomic quantum memories and clocks without frequency conversion.","Classical synchronization signals and authenticated encrypted control traffic can share the physical channel with photonic qubits, since the source wavelengths sit outside the telecom C-band; the demonstration ran an encrypted link alongside the free-space qubit path and multiplexed classical data with the fiber idler arm."],"supporting_citations":[{"why":"Supplies the warm-vapor bichromatic entanglement source class whose high brightness and narrow linewidth the experiment’s source builds on.","marker":"[8]"},{"why":"Establishes the metropolitan free-space quantum network context and baseline that this three-node hybrid deployment extends.","marker":"[10]"},{"why":"Provides the key-extraction efficiency factor 0.56 used to convert measured QBER and visibility into the reported secret-key rates.","marker":"[19]"},{"why":"Documents the bichromatic source specifications—brightness, g^(2), and heralding efficiency—that the experiment relies on for pair generation.","marker":"[21]"},{"why":"Supports the high system detection efficiency attributed to the superconducting nanowire single-photon detectors.","marker":"[23]"},{"why":"Supports the picosecond timing resolution attributed to the superconducting detectors, which sets the correlation-window floor.","marker":"[24]"},{"why":"Establishes the time-correlated-photon-pair method used for the quantum time-transfer analysis.","marker":"[29]"}],"fun_headline_variants":["Off-the-shelf quantum devices entangle across fiber and air","Three-node hybrid net shares entanglement at 33 bits/s","Campus-scale link uses off-the-shelf parts, S=2.63","Entanglement beats classical limit on mixed fiber-air path","Store-bought kit builds quantum network across campus"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The result stands on the assumption that the selective regions of interest and the discarding of datasets with synchronization errors were decided without reference to the measured visibilities; if those choices were data-dependent, the reported S values could be inflated.","fun_headline_variants_meta":{"raw":{"variants":["Off-the-shelf quantum devices entangle across fiber and air","Three-node hybrid net shares entanglement at 33 bits/s","Campus-scale link uses off-the-shelf parts, S=2.63","Entanglement beats classical limit on mixed fiber-air path","Store-bought kit builds quantum network across campus"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000787,"raw_usage":{"total_tokens":3430,"prompt_tokens":859,"completion_tokens":2571,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":475,"completion_tokens_details":{"reasoning_tokens":2487}},"tokens_in":475,"tokens_out":2571,"duration_ms":20452,"temperature":1.0,"reasoning_tokens":2487,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T17:28:55.183747+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Recompute the CHSH value for the A–B path using all recorded time tags with no region-of-interest trimming, or with a region of interest and an acceptance rule fixed before unmasking the visibility curve; if the resulting S falls at or below 2, the entanglement claim fails. A reader could also rerun the campaign with the same hardware and a pre-registered analysis pipeline to see whether S = 2.63 reproduces.","supporting_citations":[{"cited_title":"Merolla, B","cited_arxiv_id":null,"evidence_quote":"Establishes the metropolitan free-space quantum network context and baseline that this three-node hybrid deployment extends."},{"cited_title":"ˇSkori´ c and M","cited_arxiv_id":null,"evidence_quote":"Supports the high system detection efficiency attributed to the superconducting nanowire single-photon detectors."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supports the picosecond timing resolution attributed to the superconducting detectors, which sets the correlation-window floor."}],"review_version":2}