{"id":"c3e3cb7b-d0e3-414b-9973-8a5b3a3903b0","arxiv_id":"2412.02180","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"An in-situ beam-position scan of the Virgo filter cavity finds round-trip optical losses vary strongly with input-mirror hit position (42-87 ppm) but weakly with end-mirror position (53-61 ppm).","lead":"This paper measures how optical losses inside a 285-meter suspended cavity change as the laser beam moves across the mirrors, finding 42 to 87 ppm on the input mirror and 53 to 61 ppm on the end mirror. Lower losses are essential for squeezed-light quantum noise reduction in gravitational-wave detectors, so the map helps pick the best cavity axis.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The loss map itself is credible, but the 'no major loss mechanism neglected' conclusion rests on an unvalidated Eq. 5 large-angle scattering map and a global-minimum-to-range comparison; a point-by-point predicted-vs-measured map check is needed.","rationale":"The central empirical contribution, a reproducible spatially resolved in-situ loss map with <4 ppm statistical error and 1.3 ppm systematic error, is well supported and should stand. My concern is narrower: the abstract and Sec. VII draw a conclusion about completeness of the loss budget from the proximity of the measured minimum (42 ppm) to the pre-installation estimate (30.3-39.3 ppm). That inference is only as strong as the predicted map, and the predicted map's largest uncertain term is large-angle scattering, built with Eq. 5's uniformity assumption that the authors themselves describe as problematic. The paper does not show a residual map between prediction and measurement; a global-minimum-to-range comparison is a weak test because the predicted range is a spatial-variation range, not a confidence interval at a given point. This is why I agree with the reader that conditionality is appropriate. The verdict should remain conditional: accept the empirical map, but require either a point-by-point model/measurement comparison or a softened conclusion. No rejection or acceptance upgrade is warranted.","tokens_in":12796,"tokens_out":11048,"duration_ms":117043,"concrete_test":"Register and subtract the predicted total RTL map (Oscar small-angle map + Eq. 5 large-angle TIS map convolved with the 10.5-mm Gaussian + 1.4 ppm middle-angle scattering + 3.9 ppm end-mirror transmission) from the measured input-mirror RTL map in Fig. 3 (left), using the stated scanned-region coordinates. If the residual at the 42-ppm minimum exceeds the combined ~4-ppm measurement uncertainty, or if residuals show spatial structure correlated with neither map, the agreement is coincidental and the 'no missing mechanism' claim is unsupported. Additionally, re-measure ARS(theta, phi) at 3-5 points spanning the TIS map (including the minimum and a high-loss point) with an imaging scatterometer such as SPARSE and recompute each TIS using the actual angular distribution instead of Eq. 5; if the large-angle scattering budget shifts by more than about 5 ppm, the Eq.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The measured RTL map (Sec. V) is repeatable and well characterized (statistical error <4 ppm, systematic 1.3 ppm), so the central empirical finding is credible. The load-bearing weakness is the Sec. VII conclusion that no major loss mechanism has been neglected, which rests entirely on the pre-installation budget of Sec. VI. The dominant uncertain term in that budget is large-angle scattering (14-24 ppm), and it is constructed via Eq. 5: TIS(x,y) is obtained by scaling the TIS of one representative point using the ratio of ARS values at a single scattering angle (10 deg). This assumes the angular dependence of scattered light is identical at every point on the mirror, an assumption the authors themselves call 'problematic' because micro-roughness, point defects, and contaminants scatter differently. If the angular dependence varies, the large-angle scattering map is unreliable at the location of the measured 42 ppm minimum, and the predicted range 30.3-39.3 ppm is not a trustworthy benchmark. The paper also never overlays the predicted loss map on the measured RTL map; it compares only a global minimum to a broad spatial-variation range, so the closeness of 42 ppm to the predicted range could be coincidental even if the average budget is roughly correct. A point-by-point residual test would either validate the model or expose missing or misestimated loss terms.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports an in-situ measurement of round-trip optical losses in the 285-m Virgo filter cavity as a function of the intra-cavity beam position on the two mirrors. Losses are inferred from the ratio of reflected power when the IR sub-carrier is on and off resonance, and the beam position is reconstructed using a calibrated dithering technique. The authors map a roughly 2-cm region on each mirror and find that the round-trip loss varies strongly with beam position on the input mirror (42-87 ppm) while remaining more uniform on the end mirror (53-61 ppm). Measurements repeated over ten days show repeatability with a statistical error below 4 ppm and an estimated systematic error of 1.3 ppm. The lowest measured loss is compared with a pre-installation budget built from small-angle scattering (Oscar simulation of a flatness map), large-angle scattering (CASI scatterometer with a single-angle scaling assumption), middle-angle scattering (PSD extrapolation), and end-mirror transmission, giving a predicted range of 30.3-39.3 ppm. The authors conclude that no major loss mechanism has been neglected and attribute the higher losses at some positions to contamination. Appendices provide the cavity-geometry relations and dithering calibration details.","tokens_in":13091,"tokens_out":7911,"duration_ms":78426,"significance":"If the measured loss map is accepted, this is a valuable experimental contribution: it provides the first precise loss-vs-position mapping of a 100-m-scale suspended Fabry-Perot cavity, with a carefully characterized measurement chain (repeatability over ten days, statistical error below 4 ppm, systematic error 1.3 ppm, beam-position uncertainty of 0.3 mm). The method is directly relevant to optimizing filter-cavity performance for frequency-dependent squeezing in current and future gravitational-wave detectors. However, the secondary claim that the pre-installation budget proves no major loss mechanism has been neglected is not supported at the same evidentiary level: the large-angle scattering term rests on an explicitly acknowledged scaling assumption in Eq. (5), and the comparison is made between a single global minimum and a broad predicted range rather than as a spatially resolved validation.","major_comments":[{"comment":"The large-angle scattering contribution to the predicted loss budget (15-24 ppm) is constructed by assuming that the angular dependence of ARS measured at one representative point (x0,y0) applies to every point on the mirror, so that TIS(x,y) is obtained by scaling with ARS(x,y;θ0,φ0). The authors themselves acknowledge in the same section that this assumption is problematic because micro-roughness, point defects, and contaminants scatter differently. Since the predicted range 30.3-39.3 ppm includes this term and the abstract and conclusions use the proximity of the 42 ppm minimum to this range to rule out missing loss mechanisms, Eq. (5) is load-bearing. The authors should either validate the assumption with angular-resolved measurements at several positions (for example with SPARSE or an integrating-sphere scanner) or explicitly refrain from the \"no major loss mechanism neglected\" conclusion.","section":"Sec. VI, Eq. (5)"},{"comment":"The comparison between measurement and prediction is made between a single global minimum (42 ppm) and a broad pre-installation range (30.3-39.3 ppm). The paper never overlays the predicted loss map on the measured RTL map: the small-angle map from Oscar (Fig. 4b), the large-angle map after convolution (Fig. 4d), and the middle-angle plus transmission terms are not combined into a spatial prediction that can be compared with Fig. 3. A point-by-point comparison, such as a residual map or a predicted-vs-measured scatter plot over the scanned region, would test whether the model explains the spatial structure of the measured losses rather than merely matching by chance at a single location. The authors should add such a comparison before using the closeness of the minimum to support the claim that no major loss mechanism is missing.","section":"Secs. V-VII"},{"comment":"The expected-loss budget is reported as a range (30.3-39.3 ppm) with no uncertainty on the small-angle scattering (10 ppm), middle-angle scattering (1.4 ppm), or end-mirror transmission (3.9 ppm). The difference between the measured minimum (42 ppm) and the upper end of the predicted range (39.3 ppm) is only 2.7 ppm, which is likely within the combined uncertainty of the budget terms. The authors should state the uncertainty on each term and on the total budget before interpreting the agreement as evidence that no significant loss mechanism has been neglected.","section":"Table II"}],"minor_comments":[{"comment":"The abstract reports a statistical error smaller than 4 ppm, while Sec. IV states that each individual measurement has a statistical error always below 1.4 ppm; the relationship between these two numbers (for example, day-to-day repeatability of the map) should be clarified.","section":"Abstract and Sec. IV"},{"comment":"The horizontal spread in the reconstructed end-mirror positions is attributed to piezo hysteresis, but the text does not state how this systematic effect propagates into the RTL map uncertainty; please quantify or explicitly bound its impact.","section":"Fig. 2(b)"},{"comment":"The integration lower limit θ_min is not specified; please define it (presumably the CASI lower-angle limit near 3 degrees) and state the numerical value used in the TIS calculation.","section":"Sec. VI, Eq. (4)"},{"comment":"The sentence \"we use extrapolate the mirror surface power spectral density\" contains a grammatical error and should be rewritten.","section":"Sec. VI, middle-angle paragraph"},{"comment":"The footnote qualifying the statement that the GR and IR beams share the same cavity axis should be expanded or moved into the main text, as it bears on the assumed beam-size scaling between the two wavelengths.","section":"Footnote 28"},{"comment":"A table listing the numerical RTL values and uncertainties for the three repeated points would be easier to read than the plotted data alone, which is given without error bars in the figure.","section":"Fig. 2(c)"}],"recommendation":"major_revision","confidential_remarks":"The experimental methodology and the measured loss map are credible and well documented, and the paper is a useful contribution to the filter-cavity literature. The main weakness is that the strong conclusion about the pre-installation budget rests on an unvalidated scaling assumption in Eq. (5) and on a global-minimum-to-range comparison. I believe the paper can be made suitable for publication after a major revision that either validates the large-angle scattering map or substantially softens the corresponding claim."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nRead the Zhao et al. paper on mapping optical losses in the 285 m Virgo filter cavity. The measurement result is the real thing: first position-resolved round-trip loss map at this scale, with a clear input-mirror dependence (42-87 ppm) and a flatter end mirror (53-61 ppm). The team repeats the measurement over ten days, quotes sub-4 ppm statistical error and 1.3 ppm systematic, and gives a credible account of the dithering-based position readout. That part is carefully done and I would trust the map.\n\nThe soft spot is the comparison to the pre-installation budget. The abstract says the lowest measured 42 ppm is close to predicted 30.3-39.3 ppm, implying no major loss mechanism is neglected. That inference is weaker than it looks. The budget's dominant uncertain term, large-angle scattering (15-24 ppm), is constructed through Eq. 5 scaling a single representative point's TIS by the ARS ratio at 10 degrees. The authors themselves call the uniform-angular-dependence assumption problematic. So the predicted range has a wide unvalidated band, and the comparison is only global-minimum-to-range. They never overlay the predicted loss map on the measured one. A point-by-point residual test would be the obvious check. The missing raw data/code also makes it hard for a referee to re-verify the map independently, though the method description is detailed enough for reproduction by a skilled group.\n\nNone of this sinks the paper. The empirical loss map stands on its own and is a genuinely new result for the 100-m-scale filter cavity community. The 'no major loss mechanism neglected' claim should be softened or backed by a proper residual map. The citation pattern looks reasonable, mostly prior cavity-loss work and public characterizations, and the self-citations are legitimate method context.\n\nMy take: send to peer review, but the referee should demand the predicted-vs-measured map comparison and a clearer caveat on Eq. 5. A good revision, not a desk reject. I would cite the measured loss map in relevant work.","headline":"The first credible position-resolved loss map of a 285-m filter cavity, with an overclaimed budget comparison that needs a residual-map check.","tokens_in":13616,"tokens_out":1618,"would_cite":true,"duration_ms":15783,"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":"Round-trip losses in a 285-m filter cavity vary from 42 to 87 ppm with beam position on the input mirror.","keywords":["round-trip losses","filter cavity","squeezed vacuum","gravitational-wave detectors","beam position scan","scattering losses","Fabry-Perot cavity","in-situ measurement"],"falsifier":"Measure the angle-resolved scattering (ARS) at several positions across the coated input mirror's scanned region using a spatially and angularly resolving scatterometer or an integrating-sphere scanner; if the angular dependence of ARS differs between points, the total-integrated-scattering map built by scaling a single ARS curve is wrong, and recomputing the predicted loss budget with the true map would show whether the 42 ppm measured minimum genuinely matches the 30.3–39.3 ppm prediction.","tokens_in":1812,"feed_emoji":"🔭","tokens_out":2301,"duration_ms":72976,"temperature":0.7,"pith_summary":"This paper establishes that round-trip optical losses in the 285-m Virgo filter cavity depend on where the infrared beam strikes the mirrors, not just on the mirrors' intrinsic quality. By scanning the beam over a two-centimeter region, the authors found losses varying from 42 ppm to 87 ppm on the input mirror while the end mirror stayed nearly uniform at 53–61 ppm, with a statistical error below 4 ppm across repeated measurements. The lowest measured value, 42 ppm, is close to the 30.3–39.3 ppm budget predicted from pre-installation mirror characterisation, which the authors interpret as evidence that no major loss mechanism has been overlooked. This matters because filter cavities with finesse near 10,000 are extremely sensitive to losses, which degrade squeezed vacuum and limit quantum noise reduction in gravitational-wave detectors. The methodology also provides a practical way to find the optimal cavity axis for minimal loss.","feed_headline":"Beam spot decides filter-cavity loss: 42 to 87 ppm","feed_subtitle":"In-situ scan of Virgo's filter cavity finds where losses are lowest and confirms the loss budget is nearly closed.","key_machinery":"The argument is carried by an on-off resonance reflectivity measurement: with the infrared sub-carrier beam locked to the cavity, the ratio of reflected power on resonance to off resonance gives the round-trip loss via $L \\simeq (T_1/2)(1-R_{\\mathrm{cav}})/(1+R_{\\mathrm{cav}})$, where $T_1 = 562 \\pm 1$ ppm is the input mirror transmissivity. Around this sits an automated beam-scanning system: the green alignment beam is steered with two mirrors, wavefront-sensing loops re-point the cavity axis onto the beam, and the dithering technique converts the longitudinal control error into a calibrated readout of beam position on each mirror, letting the beam move on one mirror while staying fixed on the other. The predicted-loss comparison is assembled from three scattering regimes: small-angle scattering simulated with the Oscar FFT code from a Zygo flatness map, large-angle scattering from a CASI scatterometer raster scan scaled by a single representative angle-resolved scattering curve via Eq. (5), and middle-angle scattering from a power-law extrapolation of the surface power spectral density, with end-mirror transmission (3.9 ppm) added to reach the 30.3–39.3 ppm budget.","core_discovery":"The central claim is that round-trip losses in the Virgo filter cavity are a function of beam position: scanning the input mirror changes the measured loss from 42 ppm to 87 ppm, whereas the equivalent scan of the end mirror spans only 53 ppm to 61 ppm. The measurement, repeated ten times over ten days at three fixed points, is reproducible to better than ±4 ppm. The lowest measured losses, 42 ppm, sit slightly above the 30.3–39.3 ppm budget assembled from pre-installation mirror characterisation, which the authors take to mean that no major loss mechanism has been neglected; the higher losses at other positions are attributed to contamination introduced during cavity integration. The same automated scan locates a cavity axis position whose round-trip losses are among the lowest ever measured for a cavity of this scale.","pith_inferences":["The single-point scaling assumption behind the large-angle scattering map (Eq. 5) is the weakest link in the loss-budget closure; a spatially and angularly resolved scatterometer scan of the same input mirror would directly test it, and if the angular scattering varies across the surface the predicted 30.3–39.3 ppm could shift materially.","The same beam-scanning method could be applied to the main arm cavities of current and future gravitational-wave detectors to re-optimise the cavity axis after installation or after contamination events, turning loss mapping into an operational tool rather than a one-time characterisation.","The asymmetry between input and end mirror maps hints that coating or contamination state, rather than bulk substrate properties, drives the spatially varying losses; a mirror swap or a spatially resolved scatterometer check could confirm this.","A natural extension is to monitor the loss map over months to quantify contamination growth rates and correlate loss spikes with vacuum or integration events."],"forward_implications":["Operating the Virgo filter cavity at the beam position that yields the minimum measured loss (42 ppm) gives round-trip losses among the lowest ever measured for a cavity of this scale.","The closeness of the measured minimum to the pre-installation prediction indicates that the mirror characterisation pipeline (flatness maps, scatterometry, PSD extrapolation) captures the dominant loss mechanisms.","The strong position dependence on the input mirror means that a single loss number is insufficient: loss budgets and quantum-noise projections for filter cavities must specify the beam position.","The scan's repeatability below 4 ppm over ten days makes the method a practical diagnostic for localising contamination spots and for checking the effectiveness of in-situ cleaning techniques.","For high-finesse filter cavities used in frequency-dependent squeezing, choosing the low-loss axis position directly improves the achievable squeezing for a given input power."],"supporting_citations":[{"why":"Provides the on-off resonance reflectivity method used to extract round-trip losses and the treatment of uncoupled power.","marker":"[19]"},{"why":"Describes the Virgo filter cavity, its control scheme, and the squeezing setup that this measurement characterises.","marker":"[21]"},{"why":"Supplies the measured input mirror transmissivity $T_1 = 562 \\pm 1$ ppm used in the loss formula and the loss budget.","marker":"[29]"},{"why":"Provides the pre-installation framework for estimating filter cavity losses from mirror characterisation.","marker":"[31]"},{"why":"Justifies treating the post-polishing flatness map as valid after coating, since ion-beam sputtering preserves surface flatness.","marker":"[32]"},{"why":"The Oscar simulation code used to compute small-angle scattering losses from the flatness map.","marker":"[34]"},{"why":"Supports the self-affine surface model used to extrapolate the power spectral density into the middle-angle scattering region.","marker":"[38]"}],"fun_headline_variants":["Beam spot on input mirror swings Virgo filter-cavity loss 42-87 ppm","Virgo filter-cavity loss depends on beam spot: up to 87 ppm","Input mirror position sets Virgo filter-cavity loss from 42 to 87 ppm","Mirror-beam map: 42-87 ppm loss in Virgo filter cavity","Input mirror spot determines loss in Virgo filter cavity: 42-87 ppm"],"cache_read_input_tokens":15744,"weakest_assumption_plain":"The predicted loss budget assumes that the angle-resolved scattering pattern measured at one representative point on the coated input mirror can be scaled to every point, even though micro-roughness, point defects, and contaminants scatter light differently.","fun_headline_variants_meta":{"raw":{"variants":["Beam spot on input mirror swings Virgo filter-cavity loss 42-87 ppm","Virgo filter-cavity loss depends on beam spot: up to 87 ppm","Input mirror position sets Virgo filter-cavity loss from 42 to 87 ppm","Mirror-beam map: 42-87 ppm loss in Virgo filter cavity","Input mirror spot determines loss in Virgo filter cavity: 42-87 ppm"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001917,"raw_usage":{"total_tokens":7556,"prompt_tokens":1041,"completion_tokens":6515,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":657,"completion_tokens_details":{"reasoning_tokens":6403}},"tokens_in":657,"tokens_out":6515,"duration_ms":45444,"temperature":1.0,"reasoning_tokens":6403,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T23:44:59.679763+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the angle-resolved scattering (ARS) at several positions across the coated input mirror's scanned region using a spatially and angularly resolving scatterometer or an integrating-sphere scanner; if the angular dependence of ARS differs between points, the total-integrated-scattering map built by scaling a single ARS curve is wrong, and recomputing the predicted loss budget with the true map would show whether the 42 ppm measured minimum genuinely matches the 30.3–39.3 ppm prediction.","supporting_citations":[{"cited_title":"Capocasa, Y","cited_arxiv_id":null,"evidence_quote":"Provides the on-off resonance reflectivity method used to extract round-trip losses and the treatment of uncoupled power."},{"cited_title":"Acernese, M","cited_arxiv_id":null,"evidence_quote":"Describes the Virgo filter cavity, its control scheme, and the squeezing setup that this measurement characterises."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the measured input mirror transmissivity $T_1 = 562 \\pm 1$ ppm used in the loss formula and the loss budget."},{"cited_title":"Straniero, J","cited_arxiv_id":null,"evidence_quote":"Justifies treating the post-polishing flatness map as valid after coating, since ion-beam sputtering preserves surface flatness."},{"cited_title":"Degallaix, Oscar: A matlab based package to simulate 12 realistic optical cavities, SoftwareX 12, 100587 (2020)","cited_arxiv_id":null,"evidence_quote":"The Oscar simulation code used to compute small-angle scattering losses from the flatness map."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supports the self-affine surface model used to extrapolate the power spectral density into the middle-angle scattering region."}],"review_version":1}