{"id":"1e1df5ed-dcbf-428b-ba9b-984c6605779c","arxiv_id":"2501.08917","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":7,"one_line_summary":"Equal internal losses during type-II SPDC make signal and idler second-order correlations differ, and this asymmetry is used to propose a loss-characterization method.","lead":"Quantum sources in waveguides lose photons while pairs are being created, and this paper computes how those losses reshape the spectrum, mode structure, and correlations of the emitted light. It also proposes using two correlation measurements to read off the waveguide's internal loss, a step toward better characterization of integrated quantum light sources.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Fig. 4 isolines are generated under idealized assumptions (Eq. 33, αp=0, frequency-independent Markovian loss); if real waveguides violate these, the inferred ᾱ and r shift, so the proposed method needs a sensitivity analysis before it can be called an experimental determination.","rationale":"The paper's central applied contribution is the loss-determination method in Sec. III C: given measured g2s and g2i, the isolines in Fig. 4 yield ᾱ and r. This is a calibration procedure, so the load-bearing condition is that the theoretical isolines are accurate for the real waveguide. The authors' own closing caveat concedes that 'higher spatial modes, frequency-dependent losses, or waveguide imperfections can significantly change... g(2) behavior,' but the three demonstration points s1–s3 are synthetic data generated with the same idealized model, so they provide no evidence about calibration error under model misspecification. The most concrete unrealized risk is pump loss: the simulations set αp=0, but in integrated sources the pump can suffer scattering/absorption comparable to or larger than signal/idler loss; an exponential pump decay changes the effective nonlinear interaction profile along z and will shift both g2s and g2i in a way that the two-parameter (ᾱ,r) inversion would misattribute to signal/idler loss. Similarly, the linear dispersion Eq. (33) omits GVD, which for sub-ps pulses over cm waveguides can alter the JSA and mode numbers. The mathematics of the Gaussian-state formalism itself appears internally consistent, and the low-gain click-detector definition of g^(2) is appropriate in the stated spontaneous regime. A quantitative sensitivity check would settle whether the method is robust; without it, the paper's applied claim is conditional, exactly the reader's verdict. I agree with the reader's weakest assumption and recommend no verdict change.","tokens_in":15126,"tokens_out":19849,"duration_ms":221217,"concrete_test":"Run the same master-equation solver with the parameters of Sec. III but add a second-order dispersion term to Eq. (33) and set αp = 1 dB/cm; then re-derive the (g_s^(2), g_i^(2)) isolines for ᾱ, r. If the point corresponding to the true (ᾱ,r) = (5 dB/cm, 0) moves by more than the experimental g^(2) uncertainty (≈0.01), the Fig. 4 calibration is not portable to realistic waveguides.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The applied claim in Sec. III C (Fig. 4) is a calibration statement: for a waveguide with 'known dispersion,' measured g_s^(2) and g_i^(2) determine ᾱ and r via the isolines. This calibration is only as good as the model used to generate the isolines. The model has three restrictive ingredients: (i) linear dispersion, Eq. (33), with no GVD or chirp; (ii) Markovian, frequency-independent α_s, α_i with αp=0; (iii) a single spatial mode. Any of these, if violated in a real device, changes the JSA and hence the mode numbers and g^(2). The authors acknowledge this in the closing paragraph of Sec. III C ('higher spatial modes, frequency-dependent losses... can significantly change...'), but no quantitative sensitivity analysis is provided; the three worked examples s1–s3 are generated with the same model, so they cannot reveal model misspecification. In particular, a pump loss αp>0 changes the effective interaction profile along z and shifts the isolines in the same direction as a change in α_s/α_i; the method as stated has no way to disentangle this. The central claim is therefore conditionally valid under the idealized model, but its status as an experimental determination of internal losses is not established until the sensitivity of Fig. 4 to these assumptions is quantified.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops a Gaussian-state formalism, based on the spatial Langevin and master equations, to model pulsed type-II spontaneous parametric down-conversion (SPDC) in lossy waveguides. It computes the joint spectral intensity, Mercer-Wolf mode numbers, normalized second-order correlation functions g(2), and Hong-Ou-Mandel (HOM) interference patterns as functions of internal propagation losses. For a frequency-degenerate type-II source operated under pump-idler group-velocity matching, the authors find that g(2)_s and g(2)_i respond differently to internal losses even when the signal and idler loss coefficients are equal, and they propose this asymmetry as the basis of a method for experimentally determining internal waveguide losses. The paper also reports that increasing losses can increase HOM dip visibility while narrowing the dip.","tokens_in":15411,"tokens_out":7746,"duration_ms":81300,"significance":"The proposed loss-determination method is potentially useful because g(2) is invariant under external frequency-independent losses, offering a way to access internal loss in integrated SPDC sources without needing absolute calibration of transmission and detection efficiencies. The theoretical framework is self-consistent, goes beyond the two-photon approximation by using Gaussian states, and provides several falsifiable qualitative predictions (e.g., g(2)_s ≠ g(2)_i under equal losses, HOM visibility increasing with loss). The central caveat is that the applied claim of 'experimental determination' is presently supported only by idealized numerical simulations, with no sensitivity analysis and no experimental validation. The manuscript is carefully written, but the practical usefulness of the method hinges on the robustness of the theoretical calibration curves, which is not yet established.","major_comments":[{"comment":"The inversion method relies on theoretical isolines generated under a restrictive set of assumptions: the linear dispersion relation of Eq. (33) with no group-velocity dispersion or chirp, Markovian frequency-independent losses with αp = 0, and a single spatial mode. The closing paragraph of Sec. III C acknowledges that higher-order effects can change g(2), but no quantitative sensitivity analysis is given. In particular, finite pump loss αp modifies the effective interaction profile along the waveguide and can shift the (ᾱ, r) isolines in the same direction as a change in the signal/idler loss ratio, so the two-parameter inversion in Fig. 4(d) would return biased estimates. Since the abstract and Sec. IV claim the method can be used for experimental determination of internal losses, the paper should either weaken that claim or provide, at least for the specific waveguide example, an analysis of how the inferred ᾱ and r respond to realistic deviations from the ideal model (e.g., αp comparable to αs, inclusion of dispersion, frequency-dependent losses, multi-spatial-mode effects).","section":"Sec. III C, Fig. 4"},{"comment":"The numerical results are presented without discretization or convergence details. The frequency-grid size N, the frequency step, the integration step for the master equations, and any convergence checks are not reported. Because the proposed method treats the theoretical isolines of Fig. 4 as exact calibration curves, the numerical convergence of those isolines is directly load-bearing: a too-coarse grid or an insufficiently converged integration could shift the isolines and invalidate the illustrative inversions s1-s3. The authors should report these numerical parameters and provide a brief convergence test (e.g., showing that g(2)_s and g(2)_i change by less than a stated tolerance when N or the step size is varied).","section":"Secs. II A and III"}],"minor_comments":[{"comment":"In the definition of I_a(t), the second equality contains ξ_b(ω_m) in the sum; since both fields are the signal field, this should be ξ_a(ω_m). The subsequent text says 'for the idler field, replace a by b', confirming this is a typographical error.","section":"Eq. (11)"},{"comment":"The text in Sec. III A refers to 'Fig. 2(e)' when discussing the signal and idler spectra, but in the figure caption the spectra are labeled as panel (d). Please correct the cross-reference.","section":"Fig. 2 and Sec. III A"},{"comment":"The statement that 'the Mercer-Wolf expansion is nothing more than a diagonalization of the matrix D with the use of a unitary matrix V' is slightly imprecise for type-II PDC; since D is block diagonal, one should specify that V = Va ⊕ Vb with Va and Vb diagonalizing the signal and idler blocks separately, as is done in the subsequent sentence.","section":"Sec. II C"},{"comment":"The sentence introducing the synthetic examples uses inconsistent quotation styling for the 'measured' values (e.g., 'g(2)_s = 1.6 and g(2)_i = 1.86' with and without quotes). Please unify the notation for clarity.","section":"Sec. III C, around Eq. (34)"}],"recommendation":"major_revision","confidential_remarks":"The paper is a solid theoretical contribution from a group with relevant expertise, and the Gaussian-state framework is appropriate. The main gap is the lack of any sensitivity analysis for the proposed inversion method; because the method is the paper's main claimed novelty, this is a load-bearing issue rather than a cosmetic one. If the authors add a robustness study (or explicitly restrict the claim to the idealized model), the paper could be suitable for publication. The self-citation to Ref. [18] for the master-equation framework is appropriate and does not raise novelty concerns."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The genuinely useful part of this paper is the demonstration that equal internal losses in type-II SPDC do not just attenuate the fields—they change the mode structure and the normalized g(2) of signal and idler in different ways, and they can make the HOM dip more visible. That is a non-obvious, physically explained result, and it is correctly derived from the Gaussian/Langevin framework the group has been developing. The math in sections II and III is standard and consistent; the JSI, HOM click probabilities, and g(2) expressions all follow from the correlation matrices, and the spectral broadening and temporal skew are plausibly explained by the effective shortening of the nonlinear medium. Appendix B’s argument that external losses drop out of g(2) is correct and is the right basis for the loss-characterization idea.\n\nThe new applied claim is the Sec III C proposal: from measured g_s^(2), g_i^(2), and optionally RN, one can read off the mean loss alpha-bar and asymmetry r from isolines like Fig. 4. That is a neat idea and could be useful for integrated SPDC sources. But it is still a calibration in an idealized model. The isolines are computed with linear dispersion (no GVD/chirp), frequency-independent Markovian losses with alpha_p=0, and a single spatial mode. The authors explicitly list these limitations in the last paragraph of Sec III C, but they do not quantify them. The three examples s1-s3 are inversions of the same model that generated the isolines, so they cannot reveal model misspecification. The stress-test note is right: if pump loss is nonzero, the effective z-dependence of the interaction changes and the isolines shift in a way that could mimic a different alpha_s/alpha_i; the method as stated has no handle on that. So the claim should be read as “this works if your waveguide matches the assumed model,” not as a validated experimental determination. That is a real but proportionate limitation—it does not undermine the core physics results.\n\nThe paper would benefit from a sensitivity analysis, even a rough one, over pump loss, GVD, and frequency-dependent absorption, and from some convergence details on the frequency grid. No code or data are included, which is normal for a theory paper but would help for reproducibility. The citation pattern is fine; the self-reference to [18] is a genuine prior formalism, not an attempt to bury criticism.\n\nWho is this for? Groups working on integrated SPDC characterization and on theory of lossy nonlinear waveguides. It deserves a serious referee: the physics is coherent, the derivations check out, and the applied idea is worth testing. A referee should ask for the sensitivity analysis before publication.","headline":"Solid lossy-SPDC theory with a useful g(2)-asymmetry finding and an honest but unquantified calibration method; worth refereeing with a request for sensitivity analysis.","tokens_in":15992,"tokens_out":3030,"would_cite":true,"duration_ms":30515,"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 proposes that the gap between signal and idler $g^{(2)}$ values, caused by internal losses during type-II SPDC, can be used to measure those losses even when external losses are unknown.","keywords":["spontaneous parametric down-conversion","waveguide internal losses","second-order correlation function","Gaussian states","Langevin equation","type-II phase matching","Hong-Ou-Mandel interference","quantum source characterization"],"falsifier":"Compare the isoline prediction against an independent loss measurement, such as a cut-back transmission calibration: measure $g_s^{(2)}$ and $g_i^{(2)}$ on the same waveguide and check whether the intersection in the $(\\bar\\alpha, r)$ plane reproduces the known loss. If the inferred losses disagree systematically, or if the two $g^{(2)}$ values remain equal under known equal internal loss, the central claim is not supported.","tokens_in":14899,"feed_emoji":"⚛️","tokens_out":8580,"duration_ms":82960,"temperature":0.7,"pith_summary":"This paper tries to establish that internal losses during waveguide-based spontaneous parametric down-conversion leave a measurable fingerprint in second-order photon correlations. Even when the signal and idler channels lose photons at the same rate, the normalized correlation functions $g_s^{(2)}$ and $g_i^{(2)}$ respond differently to loss, because the signal field experiences temporal walk-off relative to the pump while the idler does not. Since frequency-independent external losses leave $g^{(2)}$ unchanged, the gap between the two measured values isolates the internal loss. The paper shows how to turn this into a practical characterization method for nonlinear waveguides using isolines of $g_s^{(2)}$, $g_i^{(2)}$, and the photon-number ratio.","feed_headline":"Two correlation measurements size up internal loss in quantum waveguides","feed_subtitle":"Signal and idler correlation functions respond differently to internal loss, giving a loss meter immune to external losses.","key_machinery":"The computation is carried by the spatial Langevin and master equations for the second-order correlation matrices $D(z)$ and $C(z)$, which contain all information about the multimode Gaussian state under Markovian losses; fidelities of these Gaussian states with vacuum give the click and coincidence probabilities needed for $g^{(2)}$. The loss-determination method itself is an isoline-intersection procedure: for a known waveguide dispersion, $g_s^{(2)}\\left(\\bar\\alpha,r\\right)$, $g_i^{(2)}\\left(\\bar\\alpha,r\\right)$, and the relative photon number $R_N$ are computed over the plane of mean loss $\\bar\\alpha$ and loss asymmetry $r$, and the measured values pick out an intersection that estimates the losses. The Mercer-Wolf expansion, a diagonalization of the correlation matrix into broadband modes, provides the mode counts $\\mu_a$, $\\mu_b$, and $\\mu_{ab}$ used to interpret the $g^{(2)}$ values.","core_discovery":"Under pump-idler group-velocity matching in type-II SPDC, the idler travels with the pump while the signal lags behind, so photons generated early in the waveguide are more likely to be lost before the end. This makes the signal's spectral and temporal structure, and hence its measured $g^{(2)}$, progressively more sensitive to internal loss than the idler's, even when the loss coefficients are equal. External frequency-independent losses cancel out of $g^{(2)}$, so a measured difference between $g_s^{(2)}$ and $g_i^{(2)}$ indicates internal loss and, for a waveguide with known dispersion, the intersection of the theoretical isolines in the mean-loss and loss-asymmetry plane determines those loss parameters. The paper demonstrates the effect in simulations for a 1 cm waveguide with 0.5 ps pump pulses and uses it to propose an experimental method of internal-loss determination.","pith_inferences":["The asymmetry mechanism should appear in any phase-matching scheme where the two daughter fields have different group velocities; the authors only analyse pump-idler group-velocity matching, but the same walk-off argument makes $g^{(2)}$-based loss metering more widely applicable.","A practical extension would be to measure $g^{(2)}$ at several spectral filters or time delays, which could separate frequency-dependent losses from the flat-loss model considered in the paper.","The $g^{(2)}$ gap could act as a continuous in-situ health monitor for integrated quantum sources, flagging degradation of internal loss during an experiment without adding absolute-efficiency calibration."],"forward_implications":["Internal loss can be estimated from $g_s^{(2)}$ and $g_i^{(2)}$ without calibrating transmission or detection efficiency, because those external losses do not change either correlation function.","Equal $g_s^{(2)}$ and $g_i^{(2)}$ do not certify the absence of internal loss, so they should not be used as a lossless-source check.","Higher Hong-Ou-Mandel dip visibility can accompany higher internal loss, meaning visibility alone is not a reliable proxy for biphoton indistinguishability in lossy waveguides.","Strong internal losses effectively shorten the waveguide seen by the photons: signal spectral oscillations wash out, spectra broaden, and the effective number of occupied modes grows."],"supporting_citations":[{"why":"Provides the Gaussian-state Langevin/master-equation scheme for multimode squeezed light in lossy media that the paper uses to solve the PDC dynamics.","marker":"[18]"},{"why":"Supplies the Gaussian-state and covariance-matrix formalism underlying all computed observables.","marker":"[19]"},{"why":"Gives the expression for the joint spectral intensity in terms of second-order correlation matrices, used for spectral and mode analysis.","marker":"[25]"},{"why":"Relates the normalized second-order correlation function to the number of occupied modes and provides the standard g(2)-based characterization context.","marker":"[13]"},{"why":"Gives the $g^{(2)} = 1 + 1/\\mu$ relation used to connect measured correlation functions to mode numbers.","marker":"[40]"},{"why":"Reports an experimental frequency-degenerate type-II SPDC source under pump-idler group-velocity matching, the parameter regime simulated here.","marker":"[16]"},{"why":"Provides the Gaussian-state fidelity formula used to compute vacuum probabilities for click detection in the $g^{(2)}$ and Hong-Ou-Mandel calculations.","marker":"[35]"}],"fun_headline_variants":["Internal loss found by signal-idler g^(2) gap","New loss meter: compare g^(2) of signal and idler","Signal vs idler g^(2) probes internal waveguide loss","Measure internal loss via SPDC g^(2) difference"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The method assumes the waveguide's dispersion and group velocities are known exactly enough from a linear model, that losses during propagation are frequency-independent and equal for signal and idler, and that the pump is not scattered; if any of these fail, the theoretical $g^{(2)}$ curves shift and the inferred losses are biased.","fun_headline_variants_meta":{"raw":{"variants":["Internal loss found by signal-idler g^(2) gap","New loss meter: compare g^(2) of signal and idler","Signal vs idler g^(2) probes internal waveguide loss","Measure internal loss via SPDC g^(2) difference"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000147,"raw_usage":{"total_tokens":1142,"prompt_tokens":858,"completion_tokens":284,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":474,"completion_tokens_details":{"reasoning_tokens":212}},"tokens_in":474,"tokens_out":284,"duration_ms":3318,"temperature":1.0,"reasoning_tokens":212,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T20:15:00.119528+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compare the isoline prediction against an independent loss measurement, such as a cut-back transmission calibration: measure $g_s^{(2)}$ and $g_i^{(2)}$ on the same waveguide and check whether the intersection in the $(\\bar\\alpha, r)$ plane reproduces the known loss. If the inferred losses disagree systematically, or if the two $g^{(2)}$ values remain equal under known equal internal loss, the central claim is not supported.","supporting_citations":[{"cited_title":"Christ, K","cited_arxiv_id":null,"evidence_quote":"Provides the Gaussian-state Langevin/master-equation scheme for multimode squeezed light in lossy media that the paper uses to solve the PDC dynamics."},{"cited_title":"Zielnicki, K","cited_arxiv_id":null,"evidence_quote":"Supplies the Gaussian-state and covariance-matrix formalism underlying all computed observables."},{"cited_title":"Huttner, S","cited_arxiv_id":null,"evidence_quote":"Gives the expression for the joint spectral intensity in terms of second-order correlation matrices, used for spectral and mode analysis."},{"cited_title":"Hammer, S","cited_arxiv_id":null,"evidence_quote":"Relates the normalized second-order correlation function to the number of occupied modes and provides the standard g(2)-based characterization context."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Reports an experimental frequency-degenerate type-II SPDC source under pump-idler group-velocity matching, the parameter regime simulated here."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the Gaussian-state fidelity formula used to compute vacuum probabilities for click detection in the $g^{(2)}$ and Hong-Ou-Mandel calculations."}],"review_version":1}