{"id":"50506c14-d4df-4334-8f3f-7792d5e7b998","arxiv_id":"2611.16794","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"The frequency difference between two independent narrowband thermal light sources can be read from the beat oscillation in their Hong-Ou-Mandel interference coincidences.","lead":"This paper demonstrates that the beat pattern in Hong-Ou-Mandel interference between two faint thermal light sources encodes the frequency difference between them, extractable from simple coincidence count rates. It offers a practical, roughly twelve times faster calibration tool for matching the frequencies of independent narrowband single-photon sources in quantum networks.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek/deepseek-v4-flash","headline":"The model predicts g^(2)(0)=1, but the data in Fig. 5(a) show a clear peak near τ=0 reaching g^(2)≈1.8, not 1; this discrepancy is not discussed and undermines the claim of quantitative agreement.","rationale":"The reader's verdict is CONDITIONAL, with the main concern being that g^(2)(0)=1 is hard to reconcile with thermal bunching and the plotted data. I agree that this is the most serious issue. However, I think it is more load-bearing than the reader suggests: if the model cannot reproduce the zero-detuning lineshape, then the entire claim of quantitative agreement for the beat frequency is undermined, because the envelope I(τ)/I(0) is not independently validated. The reader also notes the fitting protocol and lack of error bars, but those are secondary. The paper's own sentences acknowledge 'slightly mismatch' due to experimental imperfections, but do not address the fundamental discrepancy at τ=0. The classifier labels [b79] as a limitation, but it is too vague to cover this specific issue. Therefore, I recommend REJECT unless the authors can show that the model actually does produce g^(2)(0)>1 when the full expression is used, or that the data normalization is different from what is plotted. The concrete test would settle this.","tokens_in":23920,"tokens_out":718,"duration_ms":9001,"concrete_test":"Recompute the theoretical curve for Fig. 5(a) using the full expression (Eq. 65) without setting κ1=0, or include the idler contribution that would produce g^(2)(0)>1. If the resulting curve still predicts g^(2)(0)=1, then the model is fundamentally inconsistent with the data and the beat-frequency extraction is not validated. If the full model does produce g^(2)(0)>1, then the authors should state why the simplified Eq. (14) was used and whether the beat-frequency extraction is robust to this correction.","verdict_should_be":"REJECT","load_bearing_attack":"The central closed-form result, Eq. (14)/(66), gives g^(2)(0)=1 for any detuning, including zero detuning. However, the experimental data in Fig. 5(a) (zero detuning) show a pronounced peak near τ=0 reaching g^(2)≈1.8, which is consistent with thermal bunching but inconsistent with the model's prediction of a flat g^(2)=1. The authors do not address this tension. If the model cannot reproduce the zero-detuning lineshape, then the beat-frequency extraction at non-zero detuning (Figs. 5b–d) may rely on an envelope I(τ)/I(0) that is not independently validated. The paper states that Γs and Γi are extracted from cross-correlation data and that κ1 is set to zero, but the resulting envelope (Eq. 18) is a difference of two exponentials that yields g^(2)(0)=1 regardless of the fitted rates. The data clearly show g^(2)(0)>1, so either the model is missing a contribution (e.g., from the idler or from imperfect mode matching) or the experimental g^(2) is not normalized correctly. Without resolving this, the quantitative agreement claimed for the beat frequency is on uncertain footing.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek/deepseek-v4-flash","summary":"The paper proposes and demonstrates a method to measure the frequency difference between two independent narrowband thermal light sources using Hong-Ou-Mandel-like interference. The key result is that the absolute value of the frequency detuning can be extracted from the beat signal in the second-order correlation function g^(2)(τ). The authors present a theoretical model leading to Eq. (14)/(66): g^(2)(τ) = 1 + 2|RT|^2 [I^2(τ)/I^2(0)] [1 - cos(δω τ)]. They experimentally realize this with two whispering-gallery-mode resonators (WGMRs) pumped by separate lasers, measuring g^(2)(τ) for detunings of 0, 16, 49, and 82 MHz. They claim good agreement with theory. They also compare with heralded single-photon HOM interference, showing similar beat profiles, and argue that the thermal-state measurement (two-fold coincidence) is faster (10 min) than the single-photon measurement (2 hours), thus useful for frequency matching of narrowband sources.","tokens_in":24168,"tokens_out":1995,"duration_ms":17662,"significance":"If the central claim holds, the method provides a practical calibration tool for frequency-matching independent narrowband photonic sources, which is important for quantum networks. The use of thermal states (which are naturally available from SPDC sources) instead of single photons for the beat measurement reduces measurement time significantly. The theoretical derivation is detailed and the experimental demonstration is clear. However, the significance is somewhat incremental: the idea of using HOM interference with thermal states to measure frequency differences is not entirely new (e.g., Jin et al. 2015, Ref. [13]), and the main advance is the application to narrowband WGMR sources and the demonstration of faster calibration. The paper is well-structured and the supplementary material provides thorough derivations.","major_comments":[{"comment":"The central theoretical result, Eq. (14)/(66), predicts g^(2)(0)=1 for any detuning, including zero detuning. However, the experimental data in Fig. 5(a) (zero detuning) show a clear peak near τ=0 reaching g^(2)≈1.8, which is consistent with thermal bunching but inconsistent with the model's prediction of g^(2)=1. The authors do not address this discrepancy. If the model cannot reproduce the zero-detuning lineshape, then the beat-frequency extraction at non-zero detuning (Figs. 5b–d) may rely on an envelope I(τ)/I(0) that is not independently validated. The paper states that Γs and Γi are extracted from cross-correlation data and that κ1 is set to zero, but the resulting envelope (Eq. 18) is a difference of two exponentials that yields g^(2)(0)=1 regardless of the fitted rates. The data clearly show g^(2)(0)>1, so either the model is missing a contribution (e.g., from the idler or from imperfect mode matching) or the experimental g^(2) is not normalized correctly. Without resolving this, the quantitative agreement claimed for the beat frequency is on uncertain footing.","section":"Eq. (14)/(66) and Fig. 5(a)","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The theoretical curves in Fig. 5 are generated using Eq. (14) with Eq. (18), where Γs and Γi are extracted from the cross-correlation functions (signal-idler) of the same sources (Supplementary §IV). The authors state that κ1 is set to zero because κ1^2 is five orders of magnitude below ΓsΓi. However, the envelope I(τ)/I(0) depends on Γs and Γi, which are fitted to the cross-correlation data. The beat frequency δω is then presumably fitted to the data in Fig. 5 (the paper does not explicitly state whether δω is fixed by the pump calibration or fitted). If δω is fitted, the agreement becomes partially circular because the envelope and the beat frequency are both adjusted. The paper should clarify whether δω is fixed by the independent pump calibration (0.55 MHz/MHz relation) or left as a free parameter. If it is fixed, the agreement is more convincing; if fitted, the claim of quantitative agreement is weakened.","section":"Fig. 5 and parameter extraction","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"Fig. 6 shows the total coincidence counts integrated within a coincidence window as a function of detuning, with a minimum near zero detuning. The theoretical fit is shown, but the error bars are large and the fit quality is not quantified (e.g., R^2 or chi-squared). The paper claims this demonstrates that the zero-detuning point can be found. However, the data point at -20 MHz appears to have a higher coincidence count than at 0 MHz, but the error bars overlap. The statistical significance of the dip is not assessed. The authors should provide a quantitative measure of how well the dip location can be determined, given the uncertainties.","section":"Fig. 6 and frequency-domain HOM dip","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"Fig. 7 compares the normalized coincidence counts for heralded single-photon HOM interference and thermal-state HOM-like interference at zero detuning and 37 MHz detuning. The thermal-state data are normalized to a peak of 2 (as per g^(2) definition), while the single-photon data are normalized to a peak of 1 (as per HOM dip visibility). The comparison is qualitative: the beat period appears similar. However, the paper claims that the thermal-state measurement can be used to find the zero detuning of the single-photon sources. This claim would be stronger if the authors demonstrated that the detuning extracted from the thermal-state beat matches the detuning extracted from the single-photon HOM dip within uncertainties. Currently, only two detunings are compared (0 and 37 MHz), and the single-photon visibility is given but not the beat frequency. The paper should provide a direct comparison of the extracted δω from both methods.","section":"Fig. 7 and comparison with heralded single-photon sources","authors_acknowledge":null,"authors_acknowledge_p":null}],"minor_comments":[{"comment":"The paper uses the term 'HOM-like interference' to describe the thermal-state experiment, which is appropriate. However, the distinction between g^(2)(τ) for thermal states (which can exceed 1) and the HOM dip for single photons (which goes below 1) should be made clearer in the introduction.","section":"General","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"Eq. (14) in the main text and Eq. (66) in the supplementary are identical. The paper should note that the factor 2 in Eq. (14) comes from the assumption of a 50:50 beam splitter (|R|=|T|=1/√2). This is not explicitly stated in the main text.","section":"Eq. (14) and Eq. (66)","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"Fig. 4 shows autocorrelation functions g^(2)_ss(τ). The caption should specify that these are measured with a Hanbury Brown-Twiss setup (Det1 and Det2) and that the theoretical fit is based on a single-mode thermal state model.","section":"Fig. 4 caption","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"Fig. 5 caption should state that the theoretical curves are generated using Eq. (14) with Eq. (18) and parameters from the cross-correlation fits. It should also note that κ1 is set to zero.","section":"Fig. 5 caption","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The cross-correlation functions in Fig. 4 of the supplementary show g^(2)_is(τ) with peak values around 25-30. The paper should comment on whether these high values are consistent with the thermal state assumption and the low κ1 values.","section":"Supplementary §IV","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"Typo: 'the second WGMRs' should be 'the second WGMR'.","section":"Page 5, line 2","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"Typo: 'in interplay' should be 'an interplay'.","section":"Page 6, line 5","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"Reference [16] is incomplete: 'A manufacturable platform for photonic quantum computing, Nature , 1 (2025).' The authors should provide the full author list and DOI.","section":"References","authors_acknowledge":null,"authors_acknowledge_p":null}],"recommendation":"major_revision"},"author_rebuttal":null,"desk_editor":{"model":"deepseek/deepseek-v4-flash","letter":"Hi,\n\nThis paper proposes using the HOM-like interference between two independent narrowband thermal states to extract the absolute frequency difference between them, with a claimed twelve-fold speed advantage over using heralded single photons. The idea is sensible and the experimental demonstration of the beat signal at three non-zero detunings is the main new result. The theory derivation in the supplement is thorough, and the approximation κ1→0 is well justified (five orders of magnitude below ΓsΓi). The comparison to the heralded single-photon HOM beat (ref 25) is a nice practical touch.\n\nNow the soft spots. The stress-test concern is real: Eq. (14)/(66) gives g^(2)(0)=1 for any detuning, including zero. But the zero-detuning data in Fig. 5(a) show a clear peak near τ=0 reaching g^(2)≈1.8, consistent with thermal bunching. The authors do not discuss this discrepancy. If the model cannot reproduce the zero-detuning lineshape, the envelope I(τ)/I(0) used for the non-zero detuning fits is not independently validated. The paper states that Γs and Γi are extracted from the cross-correlation data, and the resulting envelope (Eq. 18) is a difference of two exponentials that yields g^(2)(0)=1 regardless of the fitted rates. The data clearly show g^(2)(0)>1, so either the model is missing a contribution (e.g., from the idler or imperfect mode matching) or the experimental g^(2) is not normalized correctly. Without resolving this, the quantitative agreement claimed for the beat frequency is on uncertain footing.\n\nSecond, the reader's concern about whether the δω in each theoretical curve in Fig. 5 is fitted or independently set by the pump calibration is not addressed in the main text. The supplement gives the 0.55 MHz/MHz calibration, but it is not stated explicitly that the δω values in the theory curves are fixed by that calibration rather than fitted to the beat. This matters for the strength of the agreement.\n\nThird, the correlation data in Fig. 5 lack error bars, which is unusual for a quantitative claim.\n\nNone of these issues breaks the underlying beat-frequency mechanism, which is plausible. But they need to be addressed before the claim of clean agreement is taken at face value.\n\nWho this is for: experimentalists working on frequency calibration of narrowband sources for quantum networks. The paper is worth engaging with, but the referee should insist on clarification of the g^(2)(0) discrepancy and the fitting protocol.\n\nRecommendation: send to peer review, but flag the zero-detuning model-data tension as a major concern.","headline":"The thermal-state HOM beat is a plausible calibration tool, but the model-data tension at zero detuning (g^(2)(0)=1 predicted vs ~1.8 observed) is not addressed and needs to be resolved before the quantitative agreement claim holds.","tokens_in":24888,"tokens_out":676,"duration_ms":8636},"pith_extraction":null,"created_at":"2026-11-13T12:39:29.086890+00:00","model_set":{"reader":"deepseek/deepseek-v4-flash"},"pipeline_version":"pith-pipeline@v1.4.1-daily-deepseek-or-medium","input_kind":null,"input_chars":null,"input_total_chars":null,"passes_missing":null,"published_standing":null,"published_confidence":null,"objection_audit":null,"run_record":null,"basis":{"kind":"not_recorded","source":null,"chars_given":null,"chars_total":null},"receipt":null,"falsifier":null,"supporting_citations":[],"lane_accuracy":{"status":"not yet measured","lane":{"model":"deepseek/deepseek-v4-flash","input_kind":null},"url":"https://pith.science/accuracy"},"review_version":1}