{"id":"44eb1eed-067a-4103-9e92-21ea891c5d2a","arxiv_id":"quant-ph/0603041","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"A planar-lightwave-circuit one-way QKD system demonstrated 100-km key distribution with 6% QBER, limited only by detector noise, without active phase compensation.","lead":"This 2006 paper reports a one-way quantum key distribution system built from planar lightwave circuit interferometers that distributed keys over 100 km of optical fiber with a 6% bit error rate. It matters because it shows how integrated optics can remove backscattering and active phase-stabilization problems that limited earlier two-way QKD systems, potentially extending the practical distance for quantum cryptography.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Distance claim rests on a single-photon security model applied to the actual weak-coherent-pulse source; without a WCP-compatible security analysis the 98/118 km final-key distances are not established for the demonstrated system.","rationale":"Read as an experimental demonstration, the paper's physical results are credible: 100 km QBER of 6%, stability data, and a QBER-distance model with an independently measured dark count probability. The one-way architecture also makes the dominant two-way backscattering mechanism structurally absent, so the reader's chosen weakest assumption about backscattering is a secondary concern. The more serious issue is the mismatch between the security model used for the final-key rates and the actual WCP source. The paper explicitly acknowledges the PNS vulnerability for WCP, yet still reports final key rates and distance limits derived from a single-photon proof. Those numbers are not secure key rates for the demonstrated system. Because the central conclusion claims 'possible key distribution distances are limited completely by the photon detector noise' without carrying the WCP caveat, the paper overstates the practical secure distance. This does not invalidate the physical demonstration, so the verdict remains CONDITIONAL, with the condition being a WCP-compatible security analysis or a clear statement that the distance limits are not secure distances for the WCP source.","tokens_in":8542,"tokens_out":15990,"duration_ms":160642,"concrete_test":"Recompute the secure key rate for the actual WCP source using a security proof that accounts for photon-number statistics, e.g., the GLLP bound with no decoy states, using the paper's measured distance-dependent transmittance (0.205 dB/km), mu=0.1, and QBER from Fig. 4. If the secure distance (positive key rate) is substantially below the claimed 98/118 km, the 'limited completely by detector noise' conclusion is unsupported for the demonstrated system. As a stricter pass, repeat the experiment with decoy-state intensities and recompute the finite-key rate; only if the decoy-state distance remains detector-limited is the conclusion upheld.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central quantitative claim, that possible key distribution distances are limited completely by photon detector noise, is supported by final-key rates computed with a single-photon QKD model (reference [40]) and an 11.5% QBER threshold. The demonstrated source, however, is an attenuated laser with mean photon number 0.1, i.e., weak coherent pulses. The paper itself states that this scheme 'is not unconditionally secure' and that WCP is vulnerable to the photon-number-splitting attack, in which Eve obtains information without introducing errors. Privacy amplification using a single-photon proof therefore does not bound Eve's information for the actual source. The final-key rates and the 98/118 km limits in Fig. 4 are thus not valid for the system as built and operated; at best they are upper bounds that ignore PNS. The conclusion's unqualified 'possible key distribution distances are limited completely by the photon detector noise' inherits this gap. The backscattering-free claim is better supported: a one-way geometry has no counterpropagating intense pulse, and the QBER-distance fit with independently measured dark counts is a reasonable consistency check. The security-model mismatch is the load-bearing weakness because it undermines the quantitative distance claim rather than a secondary architectural advantage.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents a one-way QKD system built from planar lightwave circuit (PLC) asymmetric Mach-Zehnder interferometers, operated with attenuated laser pulses (mean photon number 0.1) and gated InGaAs/InP avalanche photodiodes over spooled standard telecom fiber. The authors report QBERs as low as 6% at 100 km for a detector quantum efficiency of 5%, model the QBER with Eq. (1) as a sum of an optical-imperfection term and a detector-noise term using independently measured dark-count probabilities, and estimate final key rates using a single-photon QKD security model (Refs. [40,41]). They conclude that the system is stable without active phase compensation, free from backscattering, and that the achievable key distribution distance is limited entirely by photon detector noise.","tokens_in":8744,"tokens_out":8778,"duration_ms":81119,"significance":"The hardware contribution is credible and potentially useful: the 3-hour stability data, the polarization-insensitive design via birefringence balancing, and the QBER-distance agreement with measured dark counts are valuable experimental results. If the security analysis were properly matched to the actual weak coherent pulse source, the architecture would be of clear interest for practical one-way QKD. However, the central quantitative claim—that the possible key distribution distances of 98/118 km are limited completely by detector noise—rests on a single-photon security proof applied to a weak coherent source, so this benchmark is not established for the demonstrated system as built and operated.","major_comments":[{"comment":"The final key rates and the 98/118 km distance limits are computed using a single-photon QKD security model (Refs. [40,41]) and an 11.5% QBER threshold, but the experimental source is an attenuated laser with mean photon number 0.1, i.e., weak coherent pulses. As the manuscript itself acknowledges in the penultimate paragraph of §4, WCP-based QKD is vulnerable to the photon-number-splitting (PNS) attack, so a single-photon security proof does not bound Eve's information for the demonstrated system. Privacy amplification based on that model therefore overestimates the secure key rate, and the 'upper limit' distances shown in Fig. 4 are not valid for the system as operated. The conclusion that 'the possible key distribution distances are limited completely by the photon detector noise' is consequently not supported for secure key distribution. A WCP-compatible security analysis (e.g., decoy-state BB84 or SARG04 with a proper finite-key treatment) is needed to substantiate the quantitative distance claim; alternatively, the 98/118 km figures should be explicitly labeled as projections for an idealized single-photon source, not as demonstrated secure distances for this system.","section":"§4 (Key Distribution Experiment), final-key-rate paragraph and Fig. 4"},{"comment":"The conclusion that the system is 'free from backscattering' is inferred from the agreement between the measured QBER and Eq. (1), which includes only detector dark counts, afterpulsing, and a constant optical-imperfection term. No direct backscattering measurement (for example, a comparison with a two-way system or an optical time-domain reflectometry characterization) is reported. Moreover, the optical-imperfection term QBER_opt is set to 1%, which the authors state is 'slightly larger than the actual value' and attribute to afterpulsing noise; because this term is adjusted to make the model match the data, the fit is not a parameter-free test of the absence of distance-dependent error sources such as backscattering. A direct measurement or an explicit upper bound on any residual backscattering contribution would make the claim conclusive.","section":"§4, Eq. (1) and the backscattering conclusion"}],"minor_comments":[{"comment":"The phase-modulator naming is inconsistent: the text refers to 'the third phase modulator (PMB 3)' in §2 and to 'PMB 1 in Bob's site' in §3, while Fig. 1 labels the modulators PMA1, PMA2, and PMB1. Please make the nomenclature uniform throughout.","section":"§2 and Fig. 1"},{"comment":"In the submitted manuscript, Eq. (1) is not typeset correctly; the displayed formula is garbled and should be rewritten so that the QBER expression and its derivation from the rates R and D are clear.","section":"§4, Eq. (1)"},{"comment":"The statement that the bit error rate 'was determined solely by the detector noise' is stronger than the model supports, since Eq. (1) includes a nonzero optical-imperfection contribution QBER_opt and the text attributes part of the short-distance QBER to afterpulsing. Suggest rewording to 'dominated by detector noise at long distances' or similar.","section":"Abstract and §5"},{"comment":"The paper says the system operates 'with no active compensation,' but both AMZIs are temperature-stabilized with Peltier controllers to 0.01°C. Please clarify that 'no active compensation' refers specifically to phase and polarization tracking in the optical path, not to temperature stabilization of the interferometers.","section":"§3 and §5"},{"comment":"The afterpulsing probability of about 5% and the factor 1/4 used to relate it to the QBER are mentioned without supporting data or derivation; providing the separate measurement, or a citation with the relevant details, would improve reproducibility.","section":"§4, afterpulsing discussion"},{"comment":"The distinction between open and solid symbols is difficult to discern in the reproduced figure; consider using different marker shapes and labeling the QE=5% and QE=10% data directly on the plot.","section":"Fig. 4"}],"recommendation":"major_revision","confidential_remarks":"The experimental work appears sound and the paper is readable, but the security-model mismatch is a substantive gap that affects the headline distance claims. If the authors can either add a concise WCP-compatible security analysis (e.g., citing GLLP or decoy-state results) or explicitly reframe the 98/118 km figures as single-photon-source projections rather than demonstrated secure distances, the paper would be acceptable. The backscattering-free claim would also be strengthened by a direct measurement or a clear bound on residual backscattering."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a competent 2006 hardware paper that shows a one-way PLC interferometer can run BB84 over 100 km of spooled fiber with QBER around 6%, dominated by detector dark counts and afterpulsing. The stability and absence of backscattering are real achievements. The distance upper limits (98/118 km) in the final-key figure are not established for the demonstrated source, because they come from a single-photon security model applied to weak coherent pulses. The authors admit the scheme is not unconditionally secure, but the conclusion repeats the detector-noise-limited framing without that caveat.\n\nWhat's new: the same group's earlier PLC paper (ref 29) showed single-photon interference over 150 km. This one adds the full key exchange: sifting, error correction, privacy amplification, and a QBER model that fits the distance dependence using independently measured dark counts and a constant optical visibility term. That model is decent evidence the only distance-dependent error source is detector noise. The one-way design also genuinely avoids the counterpropagating pulses that cause backscattering in plug-and-play systems; the inference is indirect, but the architecture makes it plausible.\n\nSoft spots: the security-model mismatch is the main one. An attenuated laser with mean photon number 0.1 is not a single-photon source, and the PNS attack invalidates the single-photon final-key rates. The paper acknowledges this in the text but presents the 98 and 118 km numbers as upper limits without spelling out that they are only upper bounds under a model that doesn't apply to the source. That undercuts the unqualified conclusion. Also, the backscattering-free claim is not directly measured—it follows from the QBER model's agreement, which assumes no backscattering term. That's a weaker inference than the paper's tone suggests. Minor: error bars are absent, but that's typical for this era.\n\nVerdict: the experimental core is credible and the QBER model is a genuine check. The security gap is real but has to be read in context—this is a hardware demonstration, not a security proof. The paper deserves a serious referee; a revision should qualify the distance claim with a WCP-compatible security analysis or label the figures clearly as single-photon bounds.","headline":"Solid 2006 hardware demo of one-way PLC QKD over 100 km, with a security-model mismatch that overstates the distance limits.","tokens_in":9330,"tokens_out":2041,"would_cite":false,"duration_ms":18990,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"A planar-lightwave-circuit one-way interferometer performs stable, backscattering-free QKD beyond 100 km, with the distance set by detector noise.","keywords":["quantum key distribution","BB84 protocol","planar lightwave circuit","asymmetric Mach-Zehnder interferometer","one-way QKD","backscattering","InGaAs avalanche photodiode","SARG04 protocol"],"falsifier":"Measure QBER versus fiber length at fixed mean photon number while independently measuring the detector dark-count probability per gate, and test whether the data follow $QBER = QBER_{\\mathrm{opt}} + (1/2 - QBER_{\\mathrm{opt}})d/R$ with no extra length-dependent term. A systematic excess at long fibers, or a dependence on launched pulse intensity, would reveal backscattering or another unmodeled noise source; equivalently, a time-resolved backscatter measurement on the link during operation, looking for photon counts coincident with the detection window beyond the dark-count level, would settle the question directly.","tokens_in":8301,"feed_emoji":"🔑","tokens_out":7888,"duration_ms":70181,"temperature":0.7,"pith_summary":"The paper sets out to show that one-way quantum key distribution can shed the two main weaknesses of commercial two-way (plug-and-play) systems by building the interferometer on a planar lightwave circuit. It reports a BB84 key-distribution experiment over spools of standard telecom fiber out to 100 km, using an attenuated laser pulse with a mean photon number of 0.1 and a measured bit error rate of 6%. The authors argue that the error rate is fully accounted for by interferometer visibility plus detector dark counts and afterpulsing, meaning no backscattering or phase drift is limiting the system. A sympathetic reader would care because the result suggests a simpler, stabler route to long-distance QKD than the two-way designs currently used in commercial systems.","feed_headline":"PLC one-way QKD sends keys 100 km with no active stabilization","feed_subtitle":"QBER stays at 6% and matches a detector-noise-only model, clearing backscattering from the distance limit.","key_machinery":"The load-bearing device is a planar lightwave circuit (PLC) asymmetric Mach-Zehnder interferometer, 4.9 by 8.9 cm, with a 5-ns delay arm and a coupler made asymmetric to equalize arm losses. Two such AMZIs, one at Alice and one at Bob, form a time-division interferometer: Alice encodes phase with two modulators, Bob decodes with a third, and temperature control at 0.01 degrees Celsius both stabilizes the phase and balances birefringence, setting $\\theta_L - \\theta_S$ to a multiple of $2\\pi$ so the fringe visibility is insensitive to polarization drift in the fiber. The argument that distance is detector-limited is carried by the QBER model $QBER = QBER_{\\mathrm{opt}} + (1/2 - QBER_{\\mathrm{opt}})d/R$, whose two terms are the only error sources the authors allow; agreement of this formula with the distance-dependent data is what licenses the backscattering-free conclusion.","core_discovery":"Using a one-way, time-division interferometer built from two silica planar-lightwave-circuit asymmetric Mach-Zehnder interferometers, the authors claim to have demonstrated stable, backscattering-free quantum key distribution with no active compensation. The key observation is that the measured QBER for fiber lengths up to 100 km tracks $QBER = QBER_{\\mathrm{opt}} + (1/2 - QBER_{\\mathrm{opt}})d/R$, where $QBER_{\\mathrm{opt}} = e/(1+e)=(1-V)/2$ is set by the interferometer's extinction ratio $e$ or visibility $V$ and the second term is the detector dark-count probability per gate $d$ relative to the sifted rate $R$; agreement with this model means no additional distance-dependent error source such as backscattering is contributing. The authors obtain a QBER of about 6% at 100 km and estimate the maximum BB84 distance at 98 km with a 10%-quantum-efficiency detector or 118 km with a 5%-QE detector, limited by detector dark counts and afterpulsing rather than by the interferometer.","pith_inferences":["The paper infers the absence of backscattering from curve fits to the QBER model rather than from a direct scattering measurement; a time-resolved backscatter measurement at the operating repetition rate would be the cleanest way to confirm that the one-way PLC design fully removes the two-way systems' distance-versus-rate tradeoff.","If the detector-limited ceiling is correct, a straightforward extension is to pair the PLC interferometer with lower-dark-count or superconducting detectors; the predicted 147 km at tenfold lower dark counts is a testable number.","The same PLC platform could plausibly raise the clock rate well above 1 MHz because one-way propagation removes the pulse-collision constraint that forces long dead times in two-way systems; whether afterpulsing can be managed at higher rates is an open engineering question the paper does not settle.","The polarization-insensitive phase modulator in Bob's station costs about 3.2 dB of loss; a lower-loss polarization-diversity design would directly improve the sifted key rate and shift the distance ceiling further out without changing the architecture."],"forward_implications":["At the demonstrated performance, BB84 key distillation at 100 km works with QBER 6%; the estimated upper distance limits are 98 km with a 10%-efficiency detector and 118 km with a 5%-efficiency detector, both set by detector noise.","Improving the detector signal-to-noise ratio converts directly into distance: the paper estimates that cutting the dark-count probability by one order of magnitude at 10% quantum efficiency would push the upper limit to about 147 km.","Because the interferometer is stable without active compensation, field links do not need a feedback loop for phase tracking, simplifying deployment.","The same apparatus implements SARG04 as well as BB84 simply by changing which bits are revealed, making the one-way design compatible with a protocol more robust against photon-number-splitting attacks.","The distance limit is not fundamental: replacing weak coherent pulses with true single-photon sources removes the photon-number-splitting vulnerability, and the system is compatible with such sources."],"supporting_citations":[{"why":"Supplies the double asymmetric Mach-Zehnder time-division interferometer design that the PLC system implements.","marker":"[7]"},{"why":"Presents the two-way Faraday-mirror design whose backscattering problems motivate the one-way approach.","marker":"[11]"},{"why":"Documents that Rayleigh backscattering limits the performance of plug-and-play two-way QKD systems.","marker":"[14]"},{"why":"Establishes that long distance and high key rate conflict in two-way systems because of backscattering, the baseline the one-way design must beat.","marker":"[25]"},{"why":"Prior demonstration of single-photon interference over 150 km with the PLC AMZI, providing the stability evidence this system builds on.","marker":"[29]"},{"why":"Provides the InGaAs APD dark-count and afterpulsing characteristics used to model the QBER.","marker":"[33]"},{"why":"Introduces the SARG04 protocol, which the paper says the same setup can implement for robustness against photon-number-splitting attacks.","marker":"[31]"},{"why":"Supplies the model relating QBER to leaked information, used to estimate the final key rates after privacy amplification.","marker":"[40]"},{"why":"Provides the security limit for QBER below 11.5% used to estimate the maximum key distribution distance.","marker":"[41]"}],"fun_headline_variants":["One-way PLC QKD hits 100 km without active stabilization","Backscatter-free QKD on PLC spans 100 km at 6% QBER","PLC one-way QKD clears long-haul barrier at 100 km","Stable QKD over 100 km using PLC, no phase correction needed"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The conclusion that the system is backscattering-free depends on the assumption that the QBER model with only optical visibility and detector dark counts/afterpulsing correctly describes the data; no direct measurement of backscattered light is made, so an unmodeled distance-dependent error source with the same signature would invalidate the claim.","fun_headline_variants_meta":{"raw":{"variants":["One-way PLC QKD hits 100 km without active stabilization","Backscatter-free QKD on PLC spans 100 km at 6% QBER","PLC one-way QKD clears long-haul barrier at 100 km","Stable QKD over 100 km using PLC, no phase correction needed"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000254,"raw_usage":{"total_tokens":1541,"prompt_tokens":889,"completion_tokens":652,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":505,"completion_tokens_details":{"reasoning_tokens":570}},"tokens_in":505,"tokens_out":652,"duration_ms":6120,"temperature":1.0,"reasoning_tokens":570,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-28T19:24:34.796804+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure QBER versus fiber length at fixed mean photon number while independently measuring the detector dark-count probability per gate, and test whether the data follow $QBER = QBER_{\\mathrm{opt}} + (1/2 - QBER_{\\mathrm{opt}})d/R$ with no extra length-dependent term. A systematic excess at long fibers, or a dependence on launched pulse intensity, would reveal backscattering or another unmodeled noise source; equivalently, a time-resolved backscatter measurement on the link during operation, looking for photon counts coincident with the detection window beyond the dark-count level, would settle the question directly.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the double asymmetric Mach-Zehnder time-division interferometer design that the PLC system implements."},{"cited_title":"Muller, H","cited_arxiv_id":null,"evidence_quote":"Presents the two-way Faraday-mirror design whose backscattering problems motivate the one-way approach."},{"cited_title":"Ribordy, J","cited_arxiv_id":null,"evidence_quote":"Documents that Rayleigh backscattering limits the performance of plug-and-play two-way QKD systems."},{"cited_title":"Subacius, A","cited_arxiv_id":null,"evidence_quote":"Establishes that long distance and high key rate conflict in two-way systems because of backscattering, the baseline the one-way design must beat."},{"cited_title":"Kimura, Y","cited_arxiv_id":null,"evidence_quote":"Prior demonstration of single-photon interference over 150 km with the PLC AMZI, providing the stability evidence this system builds on."},{"cited_title":"Stucki, G","cited_arxiv_id":null,"evidence_quote":"Provides the InGaAs APD dark-count and afterpulsing characteristics used to model the QBER."},{"cited_title":"Scarani, A","cited_arxiv_id":null,"evidence_quote":"Introduces the SARG04 protocol, which the paper says the same setup can implement for robustness against photon-number-splitting attacks."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the model relating QBER to leaked information, used to estimate the final key rates after privacy amplification."},{"cited_title":"Lütkenhaus: Phys","cited_arxiv_id":null,"evidence_quote":"Provides the security limit for QBER below 11.5% used to estimate the maximum key distribution distance."}],"review_version":1}