{"id":"a872c6c4-1504-4132-9c0c-5d4a88165a27","arxiv_id":"2608.03622","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"A sample-half-inserted Hong-Ou-Mandel interferometer converts the usual artifact fringe into a high-Fisher-information signal and measures optical path changes of a few nanometers with O(10^7) photons.","lead":"Researchers describe a modified Hong-Ou-Mandel interferometer in which a sample covers only half of one photon's path, producing sharp fringes called a dip-bump-dip. They report measuring nanometer-scale optical path changes with 10,000 times fewer photons than an earlier HOM experiment.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Reported 1.22 nm accuracy is a self-consistency check, not an independent validation, because the same fitted sinusoid defines both the 'true' path increment and the estimation model.","rationale":"The reader's weakest_assumption identifies exactly this circularity: the 'true' optical-path increment is derived from the same fitted interference curve used to convert coincidence counts into path difference. My read agrees, and the concern is load-bearing because the abstract explicitly claims 'average accuracy of 1.22 nm.' The Fisher-information-based precision claim (4.09 nm) is more robust, since it is supported by the theoretical model and the observed sinusoidal fringe structure, and the FI enhancement is plausible from Eq. 2. However, the accuracy validation needs an independent reference. The proposed concrete test would settle the question directly. Since the reader already conditioned the verdict on this issue, no change to the verdict is needed.","tokens_in":9743,"tokens_out":16687,"duration_ms":185301,"concrete_test":"Replace the model-derived 'true' value with an externally calibrated path increment: e.g., use a piezo-actuated mirror to introduce a known optical-path change of approximately 33.7 nm between two measurements, while keeping the sample untouched, and run the same 200-trial estimation procedure. Alternatively, measure the sample's temperature-dependent optical path with a separate calibrated interferometer (or with published thermo-optic coefficients) and compare the SHOM estimate to that external value. If the discrepancy remains ≤ 2 nm in this independent test, the circularity concern is resolved; if the discrepancy grows, the reported 1.22 nm accuracy was an artifact of using the same model for calibration and estimation.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central accuracy claim (average accuracy 1.22 nm) rests on comparing the SHOM estimate of the 38°C→40°C optical-path increment against a 'true' value that is itself derived from the same data and the same model used for the estimate. Specifically, Fig. 2(c) is fit to the sinusoid of Eq. 2; the fitted period (24.03°C for 405.0 nm) is used to convert temperature to optical path and to assert that 2°C corresponds to 33.7(3) nm. The 200-trial counts at 38°C and 40°C are then converted to an optical-path difference using that same fitted model, yielding 33.89±3.06 nm. Subtracting this from the 'true' 33.7 nm gives an accuracy of 1.22 nm. If the functional form in Eq. 2 is imperfect (e.g., visibility varies with temperature, the central-feature condition τ=T/2 is not exactly maintained, or the optical path is not strictly linear in temperature over the fitted range), both the calibration and the estimate inherit the same systematic error. The comparison therefore measures self-consistency, not agreement with an external reference. The Fisher-information and precision claims are less affected by this circularity, but the absolute accuracy headline is not independently established.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes and experimentally implements a modified Hong-Ou-Mandel interferometer in which the sample is inserted into only half of the signal-photon spatial mode (SHOM). The coincidence probability is modeled by Eq. (2), which predicts a dip-bump-dip structure with a sharp central feature whose phase depends on the sample-induced delay T. The authors calculate the classical Fisher information and find ~5×10^6 ps^-2 for SHOM versus ~30 ps^-2 for standard HOM, and they report an experimental maximum FI of 3.40(6)×10^6 ps^-2. Using 200 one-second trials (O(10^7) photons) at 38°C and 40°C, they estimate an optical-path increment and report an average precision of 4.09 nm and an average accuracy of 1.22 nm, comparing favorably with Ref. [24]. The paper also frames the previously reported QOCT 'artifact' as a metrological resource.","tokens_in":10076,"tokens_out":11717,"duration_ms":131947,"significance":"If the headline claims hold, the SHOM configuration is a simple, elegant way to increase per-photon Fisher information in HOM-type interferometry by orders of magnitude, and it could be useful for thickness and optical-path metrology. The theoretical model is explicit and falsifiable, the experimental data are presented in detail, and the authors honestly acknowledge that the dip-bump-dip structure was previously known in QOCT. The large FI enhancement is theoretically convincing and the measured SHOM FI is high even with imperfect visibility. However, the reported accuracy is not independently validated, and the empirical enhancement factor lacks a same-setup HOM baseline. These issues affect the central quantitative claims in the abstract and Table I, though they are in principle addressable.","major_comments":[{"comment":"The accuracy claim is self-calibrated. The 'true' optical-path increment of 33.7(3) nm per 2°C is obtained by fitting the center-coincidence sinusoid of Fig. 2(c) with the model of Eq. (2). The same fitted sinusoid is then used to convert the 38°C/40°C count difference into the estimated optical-path increment (33.89±3.06 nm). Any systematic error in the model (e.g., imperfect visibility, temperature-to-path nonlinearity, or deviation from the τ=T/2 condition) enters both the calibration and the estimate. The difference of 1.22 nm therefore measures self-consistency, not agreement with an external reference. Please recalibrate the temperature-to-optical-path conversion with an independent method (e.g., a calibrated delay stage or a separate interferometric measurement), or clearly relabel the reported quantity as a consistency check and remove the accuracy claim from the abstract, Fig. 3","section":"Fig. 3 and Table I"},{"comment":"The claim of five-orders-of-magnitude FI enhancement is supported experimentally only for SHOM; the HOM baseline of ~30 ps^-2 is taken from the theoretical curve in Fig. 1(f), not from a measurement under identical conditions. To make the 'demonstrate' claim self-contained, compute the HOM FI from the measured HOM dip of Fig. 2(a) using the same fitting and FI-calculation pipeline as for the SHOM data in Fig. 2(f), or perform a direct same-setup measurement of HOM center counts as a function of sample temperature. This would rule out baseline-model dependence and quantify the enhancement under the actual experimental visibility and counting conditions.","section":"Fig. 2(a), Fig. 2(f)"}],"minor_comments":[{"comment":"The axis labels in Fig. 3 read 'Measured thickness changed nm' and 'Set thickness changed nm', but the paper measures optical-path increments, not thickness. Please correct the labels to 'Measured optical-path increment' and 'Set optical-path increment' to match the text.","section":"Fig. 3"},{"comment":"The phrase 'improvement in measurement precision by approximately two orders of magnitude' is ambiguous. The achieved absolute precision in Table I is comparable to Ref. [24] (4.09 nm vs 4.8 nm). Clarify that the two-order improvement is for a fixed number of photons or fixed number of trials, not the absolute precision reported here.","section":"Abstract and conclusion"},{"comment":"There are several typographical errors: 'the the optical path' in the paragraph beginning 'Next, we present'; 'Tempeture changed ℃' in the Fig. 3 axis; 'Set up' instead of 'Setup' in Table I; and 'detials' in Ref. [25]. Please proofread.","section":"Experimental results"},{"comment":"The definition of 'single interference event' is not explicit. A trial is a one-second coincidence-count integration containing many photon pairs. The FI in Eq. (4) is per trial in a binomial model; please state clearly that the reported experimental precision of 4.09 nm is obtained from 200 trials, each with ~10^4 coincidence counts, so that the reader does not conflate single-pair FI with the total accumulated statistics.","section":"Fisher information"},{"comment":"The sentence 'the relatively low Fisher information per trial in ordinary HOM measurements typically necessitates tens of thousands of repetitions' should be reconciled with the actual number of trials used here (200). The improvement in trial count is 175-fold, but the photon-number reduction is four orders of magnitude; the distinction is important and should be stated explicitly.","section":"Introduction"}],"recommendation":"major_revision","confidential_remarks":"The circularity of the accuracy calibration is the main concern. The authors should either provide an independent calibration of the temperature-to-optical-path conversion or explicitly downgrade the 'accuracy' claim to a consistency check. The comparison with Ref. [24] in Table I depends on this calibration, so it should be revised accordingly. The FI enhancement claim is likely sound but would be strengthened by a same-setup HOM baseline measurement. The related work by Lualdi et al. [41] is acknowledged but not discussed; the editor may wish the authors to clarify the novelty relative to that work in the revision."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's my take. The SHOM idea is a good one: taking the dip-bump-dip artifact that QOCT has known about and turning it into a metrological resource. The Fisher-information argument is clean, and the experimental demonstration is real. They show a genuinely steep slope in coincidence probability as a function of sample delay, and a single trial carries FI on the order of 10^6 ps^-2, which is why they can get nanometric precision with ~10^7 photons. The precision number (4.09 nm standard deviation) is a reasonable statistical statement.\n\nThe soft spot is the accuracy headline. The 'true' optical-path change for 2°C is derived from fitting the same sinusoid (Fig. 2c) that is then used to convert the measured coincidence difference into path change. So the claimed 1.22 nm accuracy is a self-consistency check between two uses of the same calibration curve, not an agreement with an external reference. If the temperature-to-path conversion is off, or the model in Eq. 2 is imperfect, both the 'true' and the estimated values shift together. That doesn't invalidate the Fisher-information or precision claims, but it does mean the absolute accuracy number is not established.\n\nAlso, the comparison to the 2018 Lyons result is not a controlled baseline. The FI enhancement is real enough on paper, but claiming 'five orders of magnitude' from a different setup and a different model is a bit much. A same-setup HOM measurement would make the comparison fair.\n\nThe paper is honest about the artifact being known, and they cite the relevant QOCT work. The theory is not new physics, but the application is new. I'd want to see it through review with a request for an independent calibration method, and ideally a baseline HOM curve at similar photon number. The central result likely holds; the accuracy claim needs revalidation.\n\nI'd engage with this. It deserves peer review.","headline":"The SHOM idea is real and the Fisher-info enhancement holds up, but the headline accuracy number is a self-consistency check, not an external calibration.","tokens_in":10597,"tokens_out":3473,"would_cite":false,"duration_ms":36460,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Half-inserting a sample turns a quantum-interference artifact into a nanometer-precision ruler.","keywords":["Hong-Ou-Mandel interferometry","quantum metrology","Fisher information","optical path difference","two-photon interference","dip-bump-dip structure","spontaneous parametric down-conversion","transparent material characterization"],"falsifier":"Measure an optical-path increment created by an independent, calibrated mechanism—for example a piezo-actuated mirror or a reference etalon with known thickness—and compare the SHOM estimate against that value without using the same sinusoidal fit to define ground truth. Also test whether the central dip/bump position tracks τ = T/2 as Eq. (2) predicts and whether the measured period is exactly 2λ_p.","tokens_in":9653,"feed_emoji":"⚛️","tokens_out":12024,"duration_ms":112933,"temperature":0.7,"pith_summary":"The paper proposes and demonstrates a modified Hong-Ou-Mandel (HOM) interferometer in which the sample is inserted into only one half of the signal photon beam, so half the signal photons pass through the sample and half travel through air. This asymmetry generates a dip-bump-dip coincidence pattern whose central feature oscillates rapidly with the sample-induced optical delay, replacing the nearly flat response of a fully inserted sample. The authors show that this converts what quantum-optical-coherence-tomography studies called the central 'artifact' into the measurement signal, raising the Fisher information per event by about five orders of magnitude. With roughly 10^7 photons, they measure optical path differences with an average precision of 4.09 nm (13.63 as) and average accuracy of 1.22 nm (4.07 as), a large reduction in photon budget compared with an earlier HOM result that needed about 10^11 photons. If this holds, sample-half-inserted operation is a phase-insensitive route to fast nanometric thickness or surface characterization of transparent materials.","feed_headline":"Half-inserted sample lifts quantum-interference metrology 100,000x","feed_subtitle":"Half-inserting the sample creates a dip-bump-dip signal that needs 10^7 photons, not 10^11.","key_machinery":"The machine is the sample-half-inserted configuration itself, summarized by the SHOM coincidence probability P_SHOM(τ,T) in Eq. (2). The key term is the cosine factor cos(ω_p T/2), with ω_p the pump angular frequency, inside a Gaussian envelope; it appears because the two-photon amplitude has two components—one pair where the signal photon traversed the sample and one where it did not—so the half-inserted beam splitter creates a which-path superposition. Scanning the sample temperature T at fixed τ converts this cosine into a steep, nearly full-contrast oscillation of the coincidence counts, which is what the Fisher-information formula F=(∂_T P)^2/[P(1-P)] turns into the five-order enhanceme","core_discovery":"The central claim is that a half-inserted sample—letting only half of the signal photons pass through the material while the other half propagates in air—creates a coherent superposition of the two paths and produces the SHOM coincidence probability of Eq. (2), with the extra term cos(ω_p T/2) exp[-(σ_+^2+σ_-^2)T^2/8] modulating the usual two HOM dips. When scanned in sample temperature T at fixed delay τ, this term makes the central coincidence probability swing between 0 and 1 with a period corresponding to the pump wavelength (405 nm in optical path), instead of the nearly flat response of a fully inserted sample. The steep slope of this central dip/bump raises the classical Fisher inform","pith_inferences":["Because the method measures differential changes in optical path (here driven by temperature), a direct thickness measurement of a static sample requires separate knowledge of the thermo-optic and thermal-expansion coefficients; the paper notes this but does not demonstrate it.","The same half-inserted mechanism could be sharpened by spectral engineering—narrower pump bandwidth or frequency-resolved detection would steepen the central feature and likely lower the photon budget below 10^7; this is a testable extension not explored here.","The Fisher-information gain is computed for the ideal lossless case; under realistic loss and multi-pair emission the practical advantage remains large but likely smaller than the ideal 10^5.","Porting the half-inserted idea to N00N-state or Franson interferometers could transfer the gain to phase-sensitive measurements, though the paper only suggests this possibility."],"forward_implications":["Optical path differences in transparent materials can be estimated to about 4 nm with roughly 10^7 photons, about four orders of magnitude fewer than the 10^11-photon HOM benchmark, making fast quantum thickness or surface metrology practical.","The central dip/bump previously treated as an artifact in quantum optical coherence tomography becomes the information carrier, so artifact-suppression strategies such as frequency dithering, broadband pumps, or machine-learning post-processing are unnecessary for this measurement mode.","Because the working observable is the coincidence rate at a fixed delay, the method is phase-insensitive: no stabilization of the interferometric phase is needed, only control of the sample temperature that tunes T.","Within the Cramér-Rao bound, precision scales as 1/√N in the number of trials, so better temperature control, lower detector jitter, and longer integration directly push the demonstrated 4.09 nm toward sub-nanometer values.","The half-inserted operation can be generalized to other two-photon interferometers, including N00N-state or Franson interferometers, transferring the Fisher-information gain beyond HOM."],"supporting_citations":[{"why":"Supplies the baseline attosecond-resolution HOM measurement with O(10^11) photons that this work improves upon in photon efficiency.","marker":"[24]"},{"why":"Earlier HOM-based thickness measurement with 33 nm (0.1 fs) resolution, defining the resolution ceiling and birefringent-sample restriction this method addresses.","marker":"[13]"},{"why":"Foundational QOCT theory establishing the biphoton interference formalism and dispersion cancellation used to derive the SHOM coincidence probability.","marker":"[9]"},{"why":"Demonstration of QOCT where the central interference structure is reported and later labeled an artifact; provides the interpretive contrast for SHOM.","marker":"[10]"},{"why":"Derives quantum metrology timing limits for HOM, justifying the Fisher-information-per-trial as the figure of merit.","marker":"[8]"},{"why":"Gives the definition and calculation of Fisher information and quantum limits in optical interferometry used in Eq. (4) and the Cramér-Rao bound.","marker":"[28]"}],"fun_headline_variants":["Half-inserted HOM turns artifact into 100,000x sharper metrology","Dip-bump-dip: half-inserted sample gives 10^5 more info per trial","Half-inserted HOM measures nm and as with 10^7 photons, not 10^11","SHOM: artifact dip becomes a 100,000x sharper ruler for thin films","One-shot half-inserted HOM: 13.63 as precision from 10^7 photons"],"cache_read_input_tokens":2688,"weakest_assumption_plain":"The accuracy claim rests on the assumption that the 'true' optical-path increment is known independently; in practice it is derived by fitting the same sinusoidal center-coincidence data with the same model that converts counts into path difference, so if that model or the assumed linearity of optical path with temperature is imperfect, the reported accuracy is a self-consistency check rather than an external validation.","fun_headline_variants_meta":{"raw":{"variants":["Half-inserted HOM turns artifact into 100,000x sharper metrology","Dip-bump-dip: half-inserted sample gives 10^5 more info per trial","Half-inserted HOM measures nm and as with 10^7 photons, not 10^11","SHOM: artifact dip becomes a 100,000x sharper ruler for thin films","One-shot half-inserted HOM: 13.63 as precision from 10^7 photons"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000971,"raw_usage":{"total_tokens":3979,"prompt_tokens":771,"completion_tokens":3208,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":515,"completion_tokens_details":{"reasoning_tokens":3085}},"tokens_in":515,"tokens_out":3208,"duration_ms":23886,"temperature":1.0,"reasoning_tokens":3085,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-05T15:48:09.857802+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure an optical-path increment created by an independent, calibrated mechanism—for example a piezo-actuated mirror or a reference etalon with known thickness—and compare the SHOM estimate against that value without using the same sinusoidal fit to define ground truth. Also test whether the central dip/bump position tracks τ = T/2 as Eq. (2) predicts and whether the measured period is exactly 2λ_p.","supporting_citations":[{"cited_title":"Lyons, G","cited_arxiv_id":null,"evidence_quote":"Supplies the baseline attosecond-resolution HOM measurement with O(10^11) photons that this work improves upon in photon efficiency."},{"cited_title":"Branning, A","cited_arxiv_id":null,"evidence_quote":"Earlier HOM-based thickness measurement with 33 nm (0.1 fs) resolution, defining the resolution ceiling and birefringent-sample restriction this method addresses."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Foundational QOCT theory establishing the biphoton interference formalism and dispersion cancellation used to derive the SHOM coincidence probability."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Demonstration of QOCT where the central interference structure is reported and later labeled an artifact; provides the interpretive contrast for SHOM."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Derives quantum metrology timing limits for HOM, justifying the Fisher-information-per-trial as the figure of merit."},{"cited_title":"Demkowicz-Dobrzański, M","cited_arxiv_id":null,"evidence_quote":"Gives the definition and calculation of Fisher information and quantum limits in optical interferometry used in Eq. (4) and the Cramér-Rao bound."}],"review_version":1}