{"id":"bbab5e72-b144-4956-92da-9b610126a387","arxiv_id":"2608.13091","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"A predictive open-system DLCZ theory, validated in cold rubidium, shows Stokes-anti-Stokes correlation is robust to readout tuning and improves with shorter write pulses, enabling temporal slicing without feed-forward.","lead":"Quantum networks need photon pairs with controllable timing and correlations. This paper builds a propagation-inclusive quantum model of DLCZ Raman sources in rubidium and shows that shorter write pulses and classically timed readout can engineer the photon correlation.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Retrieval model's admitted omission of incoherent |2> fluorescence can shift the predicted anti-Stokes background and break the claimed flat g^(2) under retrieval tuning; needs a quantitative bound.","rationale":"The reader's weakest assumption identifies the retrieval model's omission of incoherent |2>-population fluorescence, and the paper itself flags this limitation in the sentence following Eq. (21). I consider this the most load-bearing concern because the central claim of 'robust correlations under retrieval tuning' relies on the balance between correlated coincidence and accidental background; an incoherent fluorescence background that does not rescale with the coherent retrieval rate would break that balance and shift the predicted g^(2) vs Ω_c/Δ_c curves. The write-stage theory predicts an incoherent |2> population of order 1% for the 100-ns pulses, comparable to the mean spin-wave excitation, so the omission is not obviously negligible. The paper's qualitative trends (shorter write pulses enhance correlation, temporal slicing works without feed-forward) are likely robust, and the explicit acknowledgement of the limitation is a point in the authors' favor. I therefore do not change the conditional acceptance verdict, but the concern should be resolved by the proposed control measurement before the quantitative validation claim is taken at face value. The unavailability of code and data is a secondary reproducibility concern that reinforces conditional acceptance without altering the central technical issue.","tokens_in":29239,"tokens_out":9372,"duration_ms":97999,"concrete_test":"Measure the anti-Stokes singles rate R_as with the write pulse blocked (so no spin-wave coherence and no coherent retrieval) but with the read pulse on under the same storage timing and detection settings as Fig. 6; if R_as(no-write) exceeds, say, 20% of the coherent R_as(write) at Ω_c=2Γ, then incoherent |2>-population fluorescence is non-negligible and Eqs. (19)–(21) must be extended with a Langevin fluorescence source proportional to ⟨σ22(T)⟩ before the correlation predictions can be trusted.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The retrieval model in Eqs. (19)–(21) explicitly assumes 'the atomic population remains predominantly in |1>' and 'does not include fluorescence associated with incoherent population transferred to |2>'. This is not a remote caveat: the write-stage theory itself predicts incoherent |2> population via spontaneous emission from |3> (Eq. (3) with Γ32 σ33). For the experimental parameters (Ω_d=2Γ, Δ_d=15Γ, T_d=100 ns), the excited-state population is ρ33 ≈ (Ω_d/2Δ_d)^2 ≈ 0.004, yielding a time-integrated incoherent transfer to |2> of order (Γ/2)ρ33 T ≈ 0.008. This is comparable to the mean spin-wave excitation μ ≈ 0.01 implied by the measured g^(2) ≈ 15–80 (since g^(2)−1 ~ 1/μ). The read pulse drives the |2>→|4> transition, so this incoherent population can fluoresce at the anti-Stokes frequency into the collected mode. Such a background would not rescale with the coherent retrieval rate R_as; it enters the accidental background and the normalization denominator ⟨a†_as a_as⟩ in Eq. (22) without a corresponding correlated-coincidence term. The central claim that g^(2) stays flat under tuning Ω_c and Δ_c (Figs. 5 and 6) rests on the correlated signal and accidental background shifting in parallel; an incoherent fluorescence floor breaks that parallel scaling and could introduce curvature or offsets in the predicted correlation. The paper's qualitative trends would survive, but the quantitative 'experimentally validate' claim and the extracted correlation values would shift.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript develops a propagation-inclusive open-system quantum theory for DLCZ-type spontaneous Raman scattering in atomic ensembles. The theory combines Heisenberg-Langevin equations with Maxwell-Schrödinger propagation, retains write-induced population redistribution, and is used to predict Stokes generation, spin-wave storage, anti-Stokes retrieval, and time-resolved Stokes–anti-Stokes cross-correlations. The authors benchmark the analytic large-detuning limit against the independent undepleted-medium model of Ref. [76], and they report experiments on a cold 87Rb ensemble that show: (i) a nearly unchanged integrated cross-correlation when the retrieval coupling strength and detuning are tuned, (ii) enhanced correlations for shorter write pulses, and (iii) temporal slicing of the retrieved anti-Stokes wavepacket by classical read-pulse timing without Stokes-triggered feed-forward. The central claim is that the theory predicts and the experiments validate these correlation-engineering features.","tokens_in":29600,"tokens_out":5564,"duration_ms":56475,"significance":"If the framework withstands scrutiny, it would be a valuable unified description of DLCZ-type SRS that goes beyond undepleted-medium and adiabatic approximations, with a transparent large-detuning limit that reduces to the known result of Eq. (16). The paper also gives explicit falsifiable predictions—the approximate flatness of g^(2) under retrieval tuning and the pulse-duration dependence—and demonstrates a practical temporal-gating capability. Credit is due for the in situ SFWM-based OD calibration, the numerical convergence checks reported in Methods, and the honest statement of the retrieval model's limitations around Eqs. (19)–(21). However, the experimental validation is largely qualitative in the key comparison figures, and the omission of incoherent |2> fluorescence in the retrieval model is a load-bearing approximation for the correlation calculations. The work is a solid candidate for publication after major revision, provided the identified technical points are addressed with quantitative bounds or experimental controls.","major_comments":[{"comment":"The retrieval model explicitly assumes the atomic population remains predominantly in |1> and states that it does not include fluorescence from incoherent population transferred to |2>. This omission is load-bearing. The write-stage theory itself generates incoherent |2> population via the Γ32 σ33 term in Eq. (3); for the experimental parameters (Ωd=2Γ, Δd=15Γ, Td=100 ns), ρ33 ≈ (Ωd/2Δd)^2 ≈ 0.004 gives a time-integrated incoherent transfer to |2> of order 0.008, comparable to the mean spin-wave excitation μ ≈ 0.01 implied by the measured g^(2) ≈ 15–80. The read pulse drives the |2>→|4> transition, so this population can contribute anti-Stokes-frequency fluorescence into the collected mode. Such a background enters the denominator ⟨a†_as a_as⟩ in Eq. (22) without a corresponding correlated-coincidence term, breaking the parallel rescaling that underlies the predicted flat g^(2) under retrieval tuning in Figs. 5 and 6. The authors should provide a quantitative bound on the incoherent-fluorescence contribution—either by extending the retrieval model to include the |2> population dynamics, or by an experimental control that measures the anti-Stokes background with the read pulse applied but no write pulse.","section":"Storage and readout dynamics, Eqs. (19)–(21) and Eq. (22)"},{"comment":"The abstract claims the theory is 'experimentally validate[d]', but the reported agreement is qualitative in the two central comparison figures. In Fig. 6(a), the model yields g-bar ≈ 15 and the data are described as 'consistent with this estimate in overall magnitude' rather than matching it. In Fig. 9(b), the theory predicts approximately 100 for the first temporal section after backgrounds, whereas the measured value is approximately 80; the text concedes 'this quantitative difference'. Given that the central scientific claims are quantitative correlation values, the authors should provide a quantitative agreement metric (e.g., chi-squared or residuals with statistical and systematic error bars) for the retrieval-tuning and pulse-duration scans in Figs. 6 and 8, and should temper the abstract's language if the agreement remains order-of-magnitude only.","section":"Experimental observations, Figs. 6(a) and 9(b)"},{"comment":"The proposed mechanism for the flat DLCZ correlation—that retrieval reshapes the conditional readout signal and the accidental background approximately in parallel—is supported only by the model curves for a few parameter points (Fig. 5a–c and Fig. 6). No direct numerical or experimental breakdown of the correlated-coincidence versus accidental-background contributions is provided. Without such a decomposition, the reader cannot assess whether the predicted flatness is robust or an artifact of the specific integration windows and background model in Eq. (23). The authors should plot, or tabulate, the correlated and accidental contributions separately as functions of Ωc and Δc, and ideally compare them with a measured accidental-coincidence baseline from, e.g., anti-correlated time windows.","section":"Two-photon correlation properties, Eq. (22) and Figs. 5–6"}],"minor_comments":[{"comment":"The phrase 'remains lacking' is a strong claim that should be softened or supported with a more explicit comparison to the cited recent theories in Refs. [78,81–84], which already treat propagation and spin-wave dynamics in related settings.","section":"Introduction, paragraph 4"},{"comment":"The definition of the complex decay parameters γ13 = γ23 = Γ3 − 2iΔd is standard but unconventional; since the HLEs use γ13/2 and γ23/2, a short parenthetical explaining that these are twice the complex optical-coherence decay rate would avoid confusion.","section":"Methods, 'Undepleted large-detuning limit'"},{"comment":"The caption does not define the dashed and solid curve styles in panel (b); the text refers to 'orange solid curves' for Stokes and the remaining curves for anti-Stokes, but adding a legend or explicit curve-type key would improve readability.","section":"Fig. 3 caption"},{"comment":"The comparison of DLCZ peak g^(2) for 20-ns and 100-ns write pulses is presented as a prediction, but the experimental data in Fig. 8 are integrated over the full anti-Stokes window. The caption should state explicitly that the 20-ns curve is not directly compared to a measurement in Fig. 7(d).","section":"Experimental observations, Fig. 7(d)"},{"comment":"The data availability statement says data are available 'upon reasonable request'; given the experimental validation is central to the abstract, the authors should consider depositing the time-tagging datasets in a public repository to strengthen reproducibility.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"The paper is likely technically sound in its theory core, and the experimental results are interesting, but the 'experimentally validate' claim in the abstract overstates the level of agreement in the data. The retrieval-model omission of incoherent |2> fluorescence is a genuine risk for the correlation predictions, and the authors should be asked to bound it quantitatively. If the authors can supply the requested decomposition and either a bound or an experimental control, the paper could become a strong contribution to the quantum-network photon-source literature. I also note that the code is only available 'upon request'; for a theory-heavy paper, a public code repository would be beneficial."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThis paper is a credible, incremental extension of the group's open-system SFWM formalism to DLCZ-type Raman sources, and the experimental support is good for the qualitative predictions, less good for the absolute numbers. It deserves a serious referee.\n\nThe genuinely new bits: the write-stage theory retains time-dependent atomic populations and optical coherence while solving the coupled Heisenberg-Langevin and Maxwell-Schrödinger equations with a Laplace transform and piecewise-constant propagator. In the undepleted large-detuning limit it reduces to the Raymer-Mostowski rate, which is a solid sanity check. The experiments test three predictions: the near-flat cross-correlation under retrieval tuning, the correlation enhancement with shorter write pulses, and the temporal slicing of the anti-Stokes wavepacket without Stokes-triggered feed-forward. Those are useful, especially the slicing result. The citations to the group's own SFWM papers are fair, since the formalism was built there.\n\nThe soft spots. First, the retrieval model explicitly omits fluorescence from incoherent population transferred to |2>. The stress-test note estimates that population is comparable to the mean spin-wave excitation. That is a real concern on paper. But the coherent retrieval is phase-matched and directional, while incoherent fluorescence is roughly isotropic, so the collected fraction should be much smaller. Also, the measured g^(2) values (20-40) are actually higher than the model's ~15, so an omitted extra background would pull the prediction the wrong way. I'd still want a quantitative bound, but I don't think it breaks the qualitative claims.\n\nSecond, the validation is semiquantitative. 'Consistent in overall magnitude' is not a match, and in Fig. 9(b) the theory gives ~100 for the first section versus ~80 measured after backgrounds. That's acceptable for trends but not a precision test. Third, no public code or data; 'available upon reasonable request' is now below the standard for a quantitative quantum-optics paper.\n\nWho this is for: groups building memory-compatible DLCZ sources or comparing DLCZ with SFWM. My recommendation: send to peer review, with requests for (1) an explicit estimate or bound on the incoherent fluorescence in the retrieved mode, (2) public code/data, and (3) a more careful comparison including error bars and the integration windows. The paper is honest about its limitations, which makes it easier to work with.","headline":"A credible, semiquantitative extension of the group's SFWM theory to DLCZ; deserves peer review, with the unquantified incoherent-fluorescence background as the main referee request.","tokens_in":30129,"tokens_out":6156,"would_cite":true,"duration_ms":61311,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["42.50.Ct","42.50.Gy"],"model":"deepseek-v4-flash","headline":"The paper proposes a propagation-inclusive open-system quantum theory for DLCZ-type spontaneous Raman scattering and validates it experimentally.","keywords":["DLCZ protocol","spontaneous Raman scattering","atomic ensemble quantum memory","Heisenberg-Langevin equations","Maxwell-Schrödinger propagation","photon-pair correlation","temporal slicing","anti-Stokes retrieval"],"falsifier":"Measure the time-resolved conditional anti-Stokes autocorrelation with a Hanbury Brown-Twiss setup: the Gaussian model predicts $g^{(2)}_{as-as|s} = (4g^{(2)}_{s-as}-2)/(g^{(2)}_{s-as})^2$, so a clean systematic deviation from that relation would show the Wick factorization misses relevant correlations. Alternatively, scan the read-pulse detuning beyond $|\\Delta_c|\\approx 3\\Gamma$ in a high-optical-depth ensemble, where DLCZ predicts a near-flat $g^{(2)}_{s-as}$ while SFWM predicts a strong drop.","tokens_in":29055,"feed_emoji":"⚛️","tokens_out":5687,"duration_ms":59934,"temperature":0.7,"pith_summary":"The paper presents a predictive theory of DLCZ-type spontaneous Raman scattering in atomic ensembles that includes both the time-dependent redistribution of atomic populations during the write pulse and the spatiotemporal propagation of the generated fields, then validates the theory experimentally. It claims this framework quantitatively predicts Stokes generation, spin-wave storage and retrieval, anti-Stokes wavepackets, and time-resolved Stokes-anti-Stokes correlations. Experiments with a cold rubidium-87 ensemble confirm three key predictions: the normalized cross-correlation stays nearly unchanged when the retrieval coupling strength or detuning is tuned, shorter write pulses enhance the correlation, and the retrieved anti-Stokes wavepacket can be temporally sliced by classical read-pulse timing without needing Stokes-triggered feed-forward. If correct, this gives memory-compatible photon sources a quantitative design tool for spectral, temporal, and correlation engineering.","feed_headline":"Shorter write pulses boost DLCZ photon-pair correlations","feed_subtitle":"A propagation-inclusive theory predicts robust Stokes-anti-Stokes correlations and temporal slicing without feed-forward.","key_machinery":"The load-bearing machinery is the coupled Heisenberg-Langevin and Maxwell-Schrödinger equation set for the atomic coherences, populations, and Stokes/anti-Stokes fields, with time-dependent coefficients $\\alpha(T)$ and $\\beta(T)$ that encode the write-induced population redistribution and optical coherence. These time-dependent coefficients make the evolution matrix non-commuting at different times, so the authors solve the dynamics with a piecewise-constant time-ordered propagator combined with a spatial Laplace transform, and they introduce a phenomenological replacement $\\mathrm{OD}_{\\mathrm{eff}}(T)=\\mathrm{OD}\\,\\langle\\hat{\\sigma}_{11}(T)\\rangle$ to capture the leading depletion effect. This machinery yields explicit predictions for Stokes photon rates, the spin-wave initial condition for retrieval, retrieved anti-Stokes wavepackets, and the two-photon cross-correlation function used throughout the experimental comparisons.","core_discovery":"The central claim is that a propagation-inclusive open-system quantum theory, built from coupled Heisenberg-Langevin equations and Maxwell-Schrödinger propagation while retaining the full time dependence of write-driven atomic populations, can predict the Stokes generation rate, spin-wave evolution, retrieved anti-Stokes wavepacket, and the complete time-resolved normalized cross-correlation $g^{(2)}_{s-as}(T,\\bar{T})$ in DLCZ-type SRS. The theory reproduces transient Raman buildup followed by population-transfer-induced suppression, and in the undepleted large-detuning limit it reduces to analytic Green's functions with a phenomenological effective optical depth correction. The key physical distinction from continuous spontaneous four-wave mixing is that retrieval parameters reshape the conditional readout signal and the accidental background approximately in parallel, leaving the normalized correlation nearly unchanged, while shorter write pulses increase the correlation because correlated coincidences scale roughly linearly with the mean spin-wave excitation number but accidental backgrounds scale roughly quadratically. The experiments confirm these trends and demonstrate temporal slicing of the retrieved wavepacket through the classically controlled read pulse.","pith_inferences":["The same framework could likely be extended to cavity-enhanced DLCZ sources by adding a cavity input-output relation, which would let the theory predict how cavity parameters modify the retrieved wavepacket and correlation robustness, though the paper does not treat cavities.","The model's Gaussian Wick-factorization prediction for the conditional anti-Stokes autocorrelation, $g^{(2)}_{as-as|s} = (4g^{(2)}_{s-as}-2)/(g^{(2)}_{s-as})^2$, is a sharp, quantitatively testable signature that could be checked with a Hanbury Brown-Twiss setup on the heralded field.","The near-flat correlation under retrieval tuning suggests a potentially useful design rule: one can narrow the retrieved anti-Stokes bandwidth for better spectral matching to a downstream quantum memory without sacrificing the pair correlation, a consequence the paper mentions but does not develop into an explicit optimization procedure.","The time-dependent effective optical depth correction may serve as a fast analytic surrogate for full numerical calculations in the large-detuning regime, useful for parameter searches, although the paper explicitly notes it captures only the leading depletion effect."],"forward_implications":["DLCZ-type SRS can be bandwidth- and frequency-engineered at the retrieval stage without substantially lowering the Stokes-anti-Stokes correlation, which is useful for matching to quantum memories and frequency conversion modules.","Shorter write pulses, down to 10-20 ns, are predicted to produce much stronger two-photon correlations and improved single-photon character of the heralded anti-Stokes field, with a predicted integrated correlation near 88 at 10 ns in the modeled regime.","The classically controlled read pulse enables temporal gating and slicing of the retrieved anti-Stokes wavepacket without real-time feed-forward from Stokes detection, directly supporting time-bin and high-dimensional temporal encoding.","The theory distinguishes DLCZ-type SRS from continuous SFWM: the separated write and read stages allow spin-wave accumulation and transient Raman enhancement, whereas in SFWM the spin-wave coherence is continuously converted and does not build up the same way.","A quantitative framework now connects write-pulse duration, driving detuning, retrieval coupling strength and detuning, optical depth, and storage time to the resulting photon-pair correlations, allowing systematic optimization of memory-compatible sources."],"supporting_citations":[{"why":"Defines the DLCZ protocol as the write-store-read sequence that this paper models and extends.","marker":"[16]"},{"why":"Provides the foundational undepleted-medium stimulated Raman scattering model that the new theory goes beyond and compares against.","marker":"[76]"},{"why":"Supplies the open-system quantum Langevin description of paired photon generation on which the present Heisenberg-Langevin approach builds.","marker":"[90]"},{"why":"Establishes the authors' earlier theoretical treatment of asymmetric biphoton generation with ground-state decoherence and phase mismatch, providing the basis for the present propagation-inclusive formalism.","marker":"[96]"},{"why":"Presents the validated open-system SFWM model used for comparison and for the in situ optical-depth calibration in the experiments.","marker":"[97]"},{"why":"Supplies the previous experimental SFWM setup and detuning-asymmetry analysis that the paper's experimental configuration and retrieval-detuning comparisons rely on.","marker":"[101]"}],"fun_headline_variants":["Shorter write pulses strengthen DLCZ photon-pair correlations","DLCZ theory predicts time-resolved correlation control","Write pulse duration gates DLCZ anti-Stokes retrieval","Retrieval gating slices DLCZ anti-Stokes wavepacket"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The retrieval calculation assumes atoms stay almost entirely in the lowest ground state during readout and ignores fluorescence from atoms that were incoherently transferred to the other ground state during the write pulse; if that incoherent population is not negligible, the predicted anti-Stokes signal and accidental background would shift.","fun_headline_variants_meta":{"raw":{"variants":["Shorter write pulses strengthen DLCZ photon-pair correlations","DLCZ theory predicts time-resolved correlation control","Write pulse duration gates DLCZ anti-Stokes retrieval","Retrieval gating slices DLCZ anti-Stokes wavepacket"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001339,"raw_usage":{"total_tokens":5434,"prompt_tokens":930,"completion_tokens":4504,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":546,"completion_tokens_details":{"reasoning_tokens":4433}},"tokens_in":546,"tokens_out":4504,"duration_ms":34146,"temperature":1.0,"reasoning_tokens":4433,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T16:56:28.033726+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the time-resolved conditional anti-Stokes autocorrelation with a Hanbury Brown-Twiss setup: the Gaussian model predicts $g^{(2)}_{as-as|s} = (4g^{(2)}_{s-as}-2)/(g^{(2)}_{s-as})^2$, so a clean systematic deviation from that relation would show the Wick factorization misses relevant correlations. Alternatively, scan the read-pulse detuning beyond $|\\Delta_c|\\approx 3\\Gamma$ in a high-optical-depth ensemble, where DLCZ predicts a near-flat $g^{(2)}_{s-as}$ while SFWM predicts a strong drop.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the foundational undepleted-medium stimulated Raman scattering model that the new theory goes beyond and compares against."},{"cited_title":"Electromagnetically-induced-transparency-based paired photon generation.Phys","cited_arxiv_id":null,"evidence_quote":"Supplies the open-system quantum Langevin description of paired photon generation on which the present Heisenberg-Langevin approach builds."},{"cited_title":"& Chen, Y .-F","cited_arxiv_id":null,"evidence_quote":"Establishes the authors' earlier theoretical treatment of asymmetric biphoton generation with ground-state decoherence and phase mismatch, providing the basis for the present propagation-inclusive formalism."},{"cited_title":"& Chen, Y .-F","cited_arxiv_id":null,"evidence_quote":"Supplies the previous experimental SFWM setup and detuning-asymmetry analysis that the paper's experimental configuration and retrieval-detuning comparisons rely on."}],"review_version":1}