{"id":"1648e9ce-00d1-4df6-beaa-c240a966c922","arxiv_id":"2607.28184","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"low","formal_verification":"none","parameter_count":4,"one_line_summary":"Entanglement swapping between two independent integrated SiN micro-ring sources with <60 MHz bandwidth yields net swapped-state visibility 0.88±0.06, enough in principle to violate a Bell inequality.","lead":"Researchers swapped entanglement between two independent chip-scale photon sources whose light is narrow enough for atomic quantum memories. The result shows a practical path toward fibre-linked quantum repeaters built from integrated photonics rather than bulky lab optics.","discovery_kind":"new_application","skeptic_critique":{"model":"grok-4.5","headline":"Bell-violation claim rests on net visibility after subtracting dominant Raman background that leaves raw V below threshold.","rationale":"The reader correctly isolated the load-bearing hinge: the Bell-capable wording is carried entirely by the background-subtracted figure while raw lies below threshold. Internal consistency of the net numbers (HOM 0.99, ET visibilities, V_φ, product bound ≈0.874 vs measured 0.88) is solid, and the 1.6 THz fibre stabilisation plus <60 MHz integrated sources are genuine advances. No deeper inconsistency in the swapping protocol, coincidence-window arguments, or shared-Franson choice appears. The residual risk is therefore engineering/presentation honesty about what “would be sufficient to violate a Bell inequality” means once Raman is treated as real noise rather than removable accidentals—exactly the condition the CONDITIONAL verdict already imposes. No verdict shift is required.","tokens_in":12311,"tokens_out":532,"duration_ms":40405,"concrete_test":"Using the reported four-fold and background rates (or the Supplementary subtraction procedure), form the mixed state ρ=p|Ψ−⟩⟨Ψ−|+(1-p)ρ_noise with p the true-pair fraction inside the analysis windows; evaluate CHSH S (or equivalent visibility) of ρ with no post-selection subtraction. If S≤2 under measured rates, the physical state is not Bell-capable and the claim must stay explicitly net-only/cladding-contingent.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The strongest claim is that net swapped-state visibility V_net=0.88±0.06 shows the final entanglement would suffice to violate a Bell inequality. Raw visibility is only 0.54±0.04 (Results §3.2), below 1/√2. The gap is cavity-enhanced Raman scattering in the SiO2 cladding (§3.1), treated as subtractable accidentals. Background subtraction is standard for source characterisation, but for a repeater those noise photons are physical: they occupy the same spectral-temporal modes, would be stored or detected with true pairs, and degrade the actual state. Eq. (5)’s product bound and the abstract’s Bell statement both use net figures. If Raman cannot be removed at device level (or filtered without losing the <60 MHz advantage), the physical swapped state as produced does not clear the Bell threshold. That is the hinge between a high-quality integrated-source demonstration and a Bell-capable resource for repeaters.","agreement_with_reader":"agree"},"referee_report":{"model":"grok-4.5","summary":"The manuscript reports entanglement swapping between two independent continuous-wave-pumped SiN microring-resonator photon-pair sources with linewidths <60 MHz (compatible with several solid-state atomic memories). Independent pumps detuned by ~800 GHz are used; herald photons (Ch27) undergo a partial Bell-state measurement, and the non-degenerate target photons (Ch19/Ch35) are analysed in a single actively stabilised folded Franson interferometer that also serves as a 1.6 THz frequency/phase bridge. Background-subtracted HOM visibility reaches 0.99±0.01 in both frequency-resolved and temporally filtered regimes; the swapped-state interferometric contrast yields raw visibility 0.54±0.04 and net visibility 0.88±0.06. The latter exceeds 1/√2, and a product bound constructed from independently measured energy-time visibilities, HOM visibility and residual phase noise is consistent with the observed net value. Stabilisation is realised entirely with commercial fibre components and Red-Pitaya-based PDH locks.","tokens_in":12550,"tokens_out":1227,"duration_ms":37840,"significance":"The work is a concrete experimental step toward field-deployable first-generation quantum repeaters. Generating memory-compatible narrowband pairs on integrated platforms, performing entanglement swapping with fully independent pumps, and closing a 1.6 THz phase/frequency loop with only off-the-shelf fibre optics are all non-trivial and directly relevant to multi-node architectures. The high net HOM visibility and the consistency check via Eq. (5) strengthen confidence that the sources themselves are of high quality. Explicit recognition that Raman scattering in the SiO2 cladding is the dominant background, together with a concrete path (air cladding / non-degenerate visible–telecom pairs), gives the result a clear forward trajectory. These strengths make the paper suitable for a specialised quantum-optics or quantum-network journal once the net-versus-raw framing is tightened.","major_comments":[{"comment":"Abstract and Results §3.2: the claim that the swapped entanglement “would be sufficient to violate a Bell inequality” rests exclusively on the background-subtracted visibility V_net=0.88±0.06. The raw visibility is only 0.54±0.04, below 1/√2. The dominant background is cavity-enhanced Raman scattering in the SiO2 cladding (§3.1), which occupies the same spectral-temporal modes as the pairs and would be stored or detected in a real repeater link. While background subtraction is standard for source characterisation, the repeater-oriented narrative requires an explicit statement that the physical state as presently produced does not clear the Bell threshold, and that the Bell-capable figure of merit is contingent on a future cladding or spectral redesign. The present wording (“would be”) is slightly ambiguous on this point.","section":"Abstract and §3.2"},{"comment":"§2.2 and §3.2: the BSM coincidence window (±0.7 ns) and the target analysis window (±1.5 ns) are free parameters that directly enter both the reported visibility and the claim of “high spectral purity.” The Supplementary Material is said to discuss window dependence, but the main text should quantify how V_net and the four-fold rate change when the windows are varied by a factor of two, so that a reader can judge robustness of the Bell-threshold conclusion.","section":"§2.2, §3.2"}],"minor_comments":[{"comment":"Fig. 3 caption and main text: the fit models for the cross-correlations and HOM traces are deferred entirely to the Supplement. A one-sentence statement of the functional form (double-exponential convolved with Gaussian jitter; quantum-beat envelope) in the main text would aid readability.","section":"Fig. 3"},{"comment":"Table 1 lists linewidths extracted from cross-correlations; the corresponding loaded Q-factors quoted in the Introduction (>3×10^6) are not tabulated. Adding them would make the memory-compatibility claim immediately verifiable.","section":"Table 1"},{"comment":"Eq. (5) is introduced as an estimate of the expected swapped visibility. Clarify whether V_HOM is taken as the net or raw value; consistency with the product ~0.874 requires the net figure.","section":"§3.2, Eq. (5)"},{"comment":"The four-fold coincidence rate (~0.03 Hz) is stated once; a brief remark on the integration time per phase point (~3 h) and total acquisition span (several days) would help readers assess practicality.","section":"§3.2"},{"comment":"Minor typographical inconsistencies appear (e.g., “highphotonindistinguishability”, missing spaces after commas in the abstract). A careful copy-edit pass is warranted.","section":"Abstract"}],"recommendation":"minor_revision","confidential_remarks":"The net-versus-raw issue is the only load-bearing presentational weakness; once the abstract and §3.2 explicitly flag that the Bell-capable resource is contingent on Raman mitigation, the paper is ready. The experimental craftsmanship (independent pumps, 1.6 THz fibre bridge, high-Q integrated sources) is genuine and should not be undervalued. Fit for a quantum-optics / quantum-network specialist journal is good; breadth for a very high-impact general journal is more marginal because of the still-low rates and the cladding-limited background."},"author_rebuttal":null,"desk_editor":{"model":"grok-4.5","letter":"Punchline: this is the first entanglement-swap demo between two independent integrated MRR pair sources at <60 MHz, with independent pumps ~800 GHz apart and all-fibre phase/frequency lock over 1.6 THz. That systems piece is the real advance, not a new protocol.\n\nWhat they did well. Sources are memory-bandwidth compatible (37–56 MHz). HOM net visibility 0.99±0.01 in both resolved and filtered windows is clean. Swapped contrast is measured carefully (simultaneous anti-correlated ports, piecewise session offsets, product bound V_ET1·V_ET2·V_HOM·V_φ ≈ 0.874 matching V_net = 0.88±0.06). Stabilisation without high-finesse cavities is practical and well argued. Citations to Samara, prior MRR work, and memory bandwidths look fair; no circularity in the visibility product check.\n\nSoft spots, in proportion. Raman from the SiO2 cladding is severe: raw V = 0.54±0.04 sits below 1/√2, so the Bell-capable statement lives only on the background-subtracted number. Subtraction is standard for source characterisation, and they flag cladding redesign and non-degenerate visible/NIR pairs as the path forward, but a real repeater would see those noise photons. Four-fold rate ~0.03 Hz and multi-day acquisition are engineering limits, not fatal. Shared Franson is a deliberate proxy for memory interferometry; fine for proof-of-principle if not oversold as a deployed node. Data not public is a minor annoyance.\n\nWho it is for: people building memory-compatible integrated sources and repeater stabilisation. Not a theory paper. It deserves a serious referee. I would engage—cite the stabilisation and bandwidth result, keep the Bell language tied to net figures and the Raman mitigation path.","headline":"Real experimental first on narrowband integrated entanglement swapping; Bell claim is honest only if kept on the net figure after Raman subtraction.","tokens_in":13195,"tokens_out":485,"would_cite":true,"duration_ms":18418,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["03.67.Hk","42.50.Dv","42.65.Lm","03.65.Ud"],"model":"grok-4.5","headline":"Integrated photon sources under 60 MHz bandwidth can perform entanglement swapping with Bell-capable net visibility using only commercial fibre components.","keywords":["entanglement swapping","quantum repeaters","integrated photonics","narrowband photon pairs","silicon nitride micro-ring resonators","time-bin entanglement","Hong-Ou-Mandel interference","phase stabilisation"],"falsifier":"Repeat the swapped-state contrast measurement without background subtraction, or after cladding redesign that removes Raman noise, and check whether the raw visibility still exceeds 1/√2 under the same coincidence windows and independent-pump conditions.","tokens_in":13138,"feed_emoji":"🔗","tokens_out":923,"duration_ms":23048,"temperature":0.7,"pith_summary":"First-generation quantum repeaters need narrowband photons that match atomic quantum memories, yet integrated chips usually make photons that are too broadband, and narrowband light is hard to keep phase-stable across separate nodes. This experiment shows that two independent silicon-nitride micro-ring sources, each producing photon pairs narrower than 60 MHz, can swap entanglement at a central station. Independent pumps differing by hundreds of gigahertz were locked together across 1.6 THz with ordinary fibre optics, yielding near-perfect Hong-Ou-Mandel interference and a background-subtracted swapped-state visibility high enough that the final entanglement could violate a Bell inequality. The result is a concrete path from laboratory chips toward field-deployed repeater nodes that do not require bulky high-finesse cavities.","feed_headline":"Narrowband chip photons swap entanglement past the Bell line","feed_subtitle":"Two independent micro-ring sources under 60 MHz lock across 1.6 THz with ordinary fibre parts","key_machinery":"A fibre 50/50 beam-splitter Bell-state measurement on the Ch27 heralds, combined with a single folded Franson interferometer that both analyses the swapped time-bin state and serves as a frequency/phase bridge locking the two independent pumps across 1.6 THz.","core_discovery":"Two independent integrated micro-ring resonators generating energy-time entangled pairs with linewidths below 60 MHz can perform entanglement swapping: a partial Bell-state measurement on the herald photons projects the remaining target photons into a time-bin entangled state whose background-subtracted visibility is 0.88 ± 0.06, sufficient to violate a Bell inequality, while phase and frequency are stabilised over a 1.6 THz span using only all-fibre commercial components and independent pumps of deliberately different frequencies.","pith_inferences":["If air cladding or alternative materials suppress cavity-enhanced Raman scattering, the raw rate and visibility may rise enough for real-time repeater operation without post-selected background subtraction.","The demonstrated 1.6 THz bridge suggests that a single distributed optical frequency comb could synchronise an entire metropolitan quantum network of heterogeneous memory platforms.","Combining these sources with on-chip absorptive memories would close the loop from integrated generation to integrated storage, the missing piece for compact field nodes."],"forward_implications":["Integrated SiN micro-rings with Q > 3×10^6 become usable entanglement resources for solid-state atomic memories that demand <60 MHz photons.","Repeater nodes operating at widely separated optical frequencies can share a common phase reference via a single fibre interferometer or an SI-traceable network tone rather than high-finesse cavities.","The same architecture can be extended to highly non-degenerate pairs that interface telecom heralds with visible/NIR memory wavelengths.","Systems-level frequency stabilisation spanning terahertz becomes a solved engineering block for multi-node quantum networks built from commercial fibre parts."],"fun_headline_variants":["Chip micro-rings swap entanglement at 0.88 visibility past Bell","Independent narrowband sources swap entanglement across 1.6 THz","Integrated sub-60 MHz photons achieve Bell-ready entanglement swap","All-fibre lock lets chip pairs swap entanglement with unlike pumps","Narrowband chip photons yield 0.88-visibility swapped entangled state"],"cache_read_input_tokens":128,"weakest_assumption_plain":"The large Raman-scattering background that cuts the raw visibility to 0.54 can be cleanly subtracted so that the remaining net visibility fairly represents the entanglement a real repeater would obtain.","fun_headline_variants_meta":{"raw":{"variants":["Chip micro-rings swap entanglement at 0.88 visibility past Bell","Independent narrowband sources swap entanglement across 1.6 THz","Integrated sub-60 MHz photons achieve Bell-ready entanglement swap","All-fibre lock lets chip pairs swap entanglement with unlike pumps","Narrowband chip photons yield 0.88-visibility swapped entangled state"]},"model":"grok-4.5","effort":"low","cost_usd":0.005631,"raw_usage":{"total_tokens":1579,"prompt_tokens":852,"num_sources_used":0,"completion_tokens":79,"cost_in_usd_ticks":56308000,"prompt_tokens_details":{"text_tokens":852,"audio_tokens":0,"image_tokens":0,"cached_tokens":256},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":648,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":852,"tokens_out":79,"duration_ms":12603,"temperature":1.0,"reasoning_tokens":648,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-31T15:19:52.805345+00:00","model_set":{"reader":"grok-4.5"},"falsifier":"Repeat the swapped-state contrast measurement without background subtraction, or after cladding redesign that removes Raman noise, and check whether the raw visibility still exceeds 1/√2 under the same coincidence windows and independent-pump conditions.","supporting_citations":[],"review_version":1}