{"id":"a956fb78-5bcd-4694-b0b1-439c69cd18f1","arxiv_id":"2607.29056","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"The strontium 461 nm cooling transition is measured at 650.503815(5) THz using grating-chip fluorescence and slow-beam velocity methods, a 60-fold precision gain over the 1938 value.","lead":"Researchers measured the exact frequency of strontium's key blue laser-cooling transition and obtained 650.503815(5) THz, about 60 times more precise than the 1938 value. They used a tiny grating chip in vacuum both to trap atoms and to create the laser geometry, validating the result with two independent methods.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"WLM fractional-error transfer to 461 nm is not independently validated; a wavelength-dependent calibration error would shift all four datasets by a common unknown amount and could make the 5 MHz combined uncertainty an underestimate.","rationale":"The reader identified the WLM fractional-error transfer as the weakest assumption. I agree that this is the weakest link in the absolute frequency claim. The concern is load-bearing because the quoted 5 MHz uncertainty is dominated by the WLM absolute-accuracy term, and the transfer model is validated only with synthetic calibration offsets, not against an independent absolute reference at 461 nm. A wavelength-dependent calibration error would be common to all datasets and would not appear as scatter among R1–R3. The V1 method has a larger statistical uncertainty and contributes modest weight, so its model uncertainties are secondary. The proposed optical-comb comparison is a direct, decisive test: it either confirms the transferred WLM scale at the claimed uncertainty or reveals a common systematic shift. Since the reader's verdict is already CONDITIONAL with essentially this condition, my read does not change the verdict; it reinforces the need for the direct 461-nm verification before the result is treated as closed.","tokens_in":15676,"tokens_out":11325,"duration_ms":127746,"concrete_test":"During one R1-type run, measure the 461 nm probe laser simultaneously with the WLM (using the same 778-nm calibration transfer chain) and with an optical frequency comb referenced to the KRISS hydrogen-maser/clock chain. Compare the comb-determined frequency with the WLM-corrected value. If the difference exceeds the reported 5 MHz combined uncertainty, the fractional-error transfer model is falsified at 461 nm and the central frequency must be re-derived using a direct 461-nm reference rather than the inferred correction.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim rests on the wavelength-meter calibration chain described in §2.3 and Fig. 6. Offsets measured at 556, 578, 759, and 778 nm are transferred to 461 nm using a common fractional-error model, δf/f = const. The supporting experiment applies synthetic calibration offsets at 578 nm and shows that the WLM responds multiplicatively at the other reference wavelengths. That test verifies the response of the WLM to a calibration-setting change, but it does not establish that the residual absolute calibration error of the instrument is wavelength-independent in the multiplicative sense required. All four corrected dataset values share this assumption, and the R1–R3 agreement does not bound a common error because all three spectroscopic datasets use the same transfer model. The quoted combined uncertainty of 5 MHz is dominated by the manufacturer's 3.3 MHz (1σ) WS8-10 absolute-accuracy term, which is itself assumed to be valid at 461 nm after transfer. If the true WLM error at 461 nm deviates from the fractional model by more than about 1–2 MHz, the central frequency would shift coherently across all datasets while the reported uncertainty would remain near 5 MHz. This is the most load-bearing concern because it directly affects the absolute scale of every measurement, unlike the V1 model uncertainty, which carries only a small weight in the combined mean.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports a new absolute-frequency measurement of the 88Sr 1S0→1P1 transition at 461 nm, using two complementary methods on a grating-chip platform: spatially resolved fluorescence spectroscopy of a thermal beam with a retro-reflected probe (datasets R1–R3) and velocity mapping of a slow atomic beam from a 2D grating magneto-optical trap (dataset V1). Combining the four datasets with a covariance-based uncertainty budget and a Birge-ratio correction yields f̄88 = 650.503 815(5) THz, which improves upon the 1938 solar-spectrum value by more than a factor of 50 in uncertainty. The paper includes a detailed systematic budget with correlated terms, multi-isotope hyperfine-constrained fitting, and a wavelength-meter transfer model based on reference lasers at 556, 578, 759, and 778 nm.","tokens_in":16009,"tokens_out":7545,"duration_ms":94910,"significance":"If the result stands, the paper supplies a long-needed laboratory re-measurement of a foundational Sr cooling transition, replacing a nine-decade-old value with a 60-fold reduction in uncertainty. The uncertainty treatment is a genuine strength: common-mode terms are propagated through a covariance matrix, the Birge ratio is used to inflate the weighted-mean uncertainty, and the spectroscopy statistical errors are conservatively taken as pixel distributions rather than standard errors of the mean. The two methods are independent in their physics, and their consistency is encouraging. However, the absolute accuracy of the final value rests entirely on the wavelength-meter calibration chain; the validation of the fractional-error transfer model does not directly bound wavelength-dependent calibration errors at 461 nm. That gap is load-bearing for a metrology claim of this kind. The manuscript also does not provide raw data or analysis code, which limits independent verification.","major_comments":[{"comment":"The fractional-error transfer model is the central calibration assumption. The controlled experiment applies synthetic offsets at 578 nm and observes proportional responses at 556, 578, 759, and 778 nm; this validates the response of the WLM to a change in calibration setting, but it does not establish that the residual absolute calibration error at 461 nm is exactly proportional to frequency. All four datasets share this transfer model, so a wavelength-dependent nonlinearity at 461 nm would shift every dataset by a common unknown amount, and the covariance analysis—which only includes the manufacturer's 3.3 MHz absolute-accuracy term plus 0.56–4.22 MHz transfer scatter—could underestimate the true uncertainty by more than the quoted 5 MHz. Please provide a direct 461 nm calibration against an optical frequency comb or an independently known 461 nm reference, or quantify and include a mo","section":"§2.3, Fig. 6"},{"comment":"The capture-velocity model contains an 'effective interaction length D', but D is never given a numerical value, an uncertainty, or a statement of whether it is a fitted parameter or a measured quantity. If D is free, the V1 fit has an unlisted degree of freedom; if D is fixed, its uncertainty must be propagated. The claim that the V1 result is 'scale-invariant with respect to the atomic velocity' needs to be demonstrated explicitly, for example by showing that the extracted resonance frequency is insensitive to D over a reasonable range. Although V1 carries only a 13 MHz statistical weight in the combined mean, the method is presented as an independent determination and must be fully specified.","section":"§2.2.2, V1 model"}],"minor_comments":[{"comment":"Typo: 'guarantied' should be 'guaranteed'. Also, clarify in the text or Fig. 6 that the 461 nm response is inferred, not directly measured.","section":"§2.3"},{"comment":"The statement that agreement between the two methods provides 'strong confidence in the accuracy' should be tempered: both methods share the WLM absolute-accuracy and transfer-model systematics, so the agreement validates consistency but not the absolute scale.","section":"§3.2"},{"comment":"For a metrology claim, 'data available from the corresponding author upon reasonable request' is weak. Consider depositing the processed spectra, velocity maps, and analysis code in a public repository.","section":"§4 / Data availability"},{"comment":"The claim of being 'the first reported laboratory re-determination since 1938' is strong. Please add a literature-search statement or soften the wording unless a comprehensive search has been performed.","section":"§1"}],"recommendation":"major_revision","confidential_remarks":"The WLM transfer issue is the main risk. The authors appear to have optical frequency comb infrastructure at KRISS; a direct 461 nm calibration would substantially strengthen the paper. If such a measurement is not possible within the revision timeline, the authors should add an explicit wavelength-dependent calibration model-error term and discuss the resulting impact on the combined uncertainty. The paper is otherwise well-structured and the uncertainty budget is unusually transparent."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"You should know this paper is a serious metrology effort, not a rough estimate. It reports the first re-determination of the 88Sr 1S0→1P1 transition frequency since the 1938 solar-spectrum value, coming in at 650.503815(5) THz—about 345 MHz above the old value and 60 times more precise. Two independent methods—spatially resolved fluorescence spectroscopy using a grating chip for retro-reflection, and velocity measurements of a 2D grating MOT beam—agree within uncertainties. The uncertainty budget is unusually honest: covariance terms, a Birge-ratio correction, month-long WLM stability records, and a conservative treatment of correlated systematics.\n\nThe spatial mapping is the real strength. Each pixel in the CCD image gets its own complete spectrum, and the multi-isotope pseudo-Voigt fit with hyperfine constraints lets the Doppler shifts vary spatially. That is a clean way to control geometric Doppler uncertainties, and the consistency between different wavelength references (778 nm Rb two-photon for R1, 759 nm Yb lattice laser for R2/R3, 578 nm for V1) is evidence the result is not riding on a single reference.\n\nThe load-bearing assumption is the WLM calibration transfer. Offsets measured at 556–778 nm are transferred to 461 nm with a common fractional-error model, δf/f = const. The paper shows the WLM responds multiplicatively when you change the calibration setting, which is necessary but not sufficient to establish that its residual absolute error at 461 nm follows the fractional model. All four datasets share this transfer chain, so their agreement does not bound a common error of that kind. The quoted 5 MHz uncertainty is dominated by the manufacturer's 3.3 MHz WS8-10 absolute-accuracy term, assumed valid at 461 nm after transfer. If the true error at 461 nm deviates from the fractional model by more than about 1–2 MHz, every dataset shifts coherently. That is a real soft spot, not a manufactured one.\n\nThe V1 velocity model is the weaker of the two methods: it uses an unspecified effective interaction length D and a v ∝ √Γ_scatt scaling, and it carries a 13 MHz statistical uncertainty. Since V1 has little weight in the combined mean, it does not threaten the central value, but it is the least rigorous part of the paper. Also, no raw data or code are released; for a claim that corrects a 90-year-old number by 345 MHz, the authors should put the spectra and fitting code in a repository.\n\nWho is this for? Atomic physicists working on Sr laser cooling, optical lattice clocks, isotope-shift analyses, and compact cold-atom platforms. It deserves a serious referee. With the WLM point addressed—ideally by a direct measurement at 461 nm against an optical frequency comb, or at least by releasing the raw data—it would be a solid reference. I would send it to review.","headline":"First modern lab measurement of the Sr 461 nm transition frequency, with a careful covariance-based uncertainty budget; the central value is probably right, but the wavelength-meter calibration transfer to 461 nm is the one assumption to stress-test before treating the result as closed.","tokens_in":16542,"tokens_out":2107,"would_cite":true,"duration_ms":22475,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["32.30.-r","06.30.Ft"],"model":"deepseek-v4-flash","headline":"Two independent methods, one photonic chip, pin strontium's 461 nm transition at 650.503815(5) THz.","keywords":["strontium","1S0→1P1 transition","absolute frequency measurement","photonic grating chip","grating magneto-optical trap","fluorescence spectroscopy","wavelength meter calibration","461 nm"],"falsifier":"Perform a direct beat-note comparison of a 461 nm laser, frequency-doubled or comb-referenced, against the authors' stabilized probe under identical conditions; a difference larger than ~1 MHz between the comb-based reading and the wavelength-meter-derived value would disprove the fractional-error transfer model and shift the combined frequency accordingly.","tokens_in":15614,"feed_emoji":"⚛️","tokens_out":5871,"duration_ms":57439,"temperature":0.7,"pith_summary":"The paper sets out to re-measure the absolute frequency of strontium's 1S0 → 1P1 transition, the 461 nm line that underpins laser cooling of Sr atoms, whose accepted value has rested on a 1938 solar-spectrum analysis with a 310 MHz uncertainty. It reports 650.503815(5) THz from four datasets: three fluorescence-spectroscopy runs using an in-vacuum diffraction grating chip as a retro-reflector, and one independent determination from the velocity dynamics of a 2D grating magneto-optical trap. The result is consistent within its error bars with the 1938 value but shrinks the uncertainty to 5 MHz, a more than sixty-fold improvement. The authors argue that the demonstrated chip-based apparatus, with its rigid beam geometry and wavelength-meter calibration transfer, is a practical metrological platform for Sr-based clocks and quantum devices.","feed_headline":"Sr 461 nm line re-measured: 650.503815(5) THz","feed_subtitle":"Two chip-based methods cut the uncertainty from 310 MHz to 5 MHz, replacing the 1938 value.","key_machinery":"The central object is a CMOS-fabricated photonic grating chip (652 nm period, Al-coated) placed below the atoms, which simultaneously functions as the diffractive element of a 2D grating magneto-optical trap and as an end mirror whose first-order diffracted beam retro-reflects the 461 nm probe, defining a counter-propagating geometry with a 20.7° angle fixed by the grating period to better than 0.1 nm. The second load-bearing mechanism is the wavelength-meter calibration transfer model: controlled offsets at 556, 578, 759, and 778 nm map to 461 nm by the constant-fractional-error relation δf/f = constant, with a fitted transfer error of 0.048%. The data pipeline combines pixel-wise hyperfine","core_discovery":"The central claim is a new value for the 88Sr 1S0 → 1P1 transition frequency: 650.503815(5) THz. It comes from four datasets — three fluorescence-spectroscopy runs in which an in-vacuum grating chip retro-reflects the probe beam (Doppler-free midpoints extracted pixel-by-pixel), and one velocity-mapping run of a 2D grating MOT — which agree and combine through a covariance-weighted mean. The value lies 345 MHz above the 1938 determination but within its expanded uncertainty, and tightens the uncertainty from 310 MHz to 5 MHz.","pith_inferences":["If this value holds, previously reported Sr isotope shifts anchored to the 1938 center would shift by ~345 MHz in absolute terms, though relative isotope shifts are unaffected; absolute-energy comparisons with theory should be updated.","The constant-δf/f wavelength-meter model could be tested directly by measuring a comb-referenced 461 nm laser against the wavelength meter, which would either validate the 5 MHz budget or reveal a missed wavelength-dependent term.","The grating-chip retro-reflection geometry could be transferred to other species (e.g., Ca, Yb) whose cooling transitions lack a precise modern measurement, yielding similar factor-of-ten uncertainty gains on compact platforms.","The velocity-based V1 method, being immune to absolute velocity calibration, suggests a route to frequency measurements that bypass wavelength meters entirely if the MOT velocity can be tied to a frequency comb via the cooling detuning."],"forward_implications":["Sr laser-cooling and clock experiments can adopt 650.503815(5) THz as the working reference for the 461 nm cooling/detection transition, replacing a value inherited from 1938.","The agreement of two independent methods within error supports the chip-based architecture's suitability for compact cold-atom metrology.","The 345 MHz offset from the 1938 center may reconcile discrepancies between observed and Ritz values in Sr spectroscopic compilations.","The quoted 5 MHz uncertainty, dominated by the wavelength meter's 3.3 MHz absolute accuracy, sets the current practical limit of this apparatus; further gains require a comb-referenced 461 nm source.","The measured frequency provides a fixed anchor for isotope-shift and hyperfine-structure analyses of 86Sr and 87Sr that have long been referenced to the 1938 value."],"fun_headline_variants":["Sr 461 nm transition: chip-based measurement cuts error to 5 MHz","Strontium's 461 nm line: 1938 value revised with grating chip","Chip reflects light to pin Sr transition at 650.503815 THz","New Sr frequency: 650.503815 THz, uncertainty drops to 5 MHz"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The measurement rests on the assumption that the wavelength meter's error scales purely proportionally with frequency (δf/f constant) between 461 nm and the reference wavelengths; if the meter has wavelength-dependent nonlinearities that the transfer model misses, every dataset inherits the same hidden frequency shift and the quoted 5 MHz uncertainty is an underestimate.","fun_headline_variants_meta":{"raw":{"variants":["Sr 461 nm transition: chip-based measurement cuts error to 5 MHz","Strontium's 461 nm line: 1938 value revised with grating chip","Chip reflects light to pin Sr transition at 650.503815 THz","New Sr frequency: 650.503815 THz, uncertainty drops to 5 MHz"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000222,"raw_usage":{"total_tokens":1344,"prompt_tokens":854,"completion_tokens":490,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":598,"completion_tokens_details":{"reasoning_tokens":413}},"tokens_in":598,"tokens_out":490,"duration_ms":5886,"temperature":1.0,"reasoning_tokens":413,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-03T14:25:55.180027+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Perform a direct beat-note comparison of a 461 nm laser, frequency-doubled or comb-referenced, against the authors' stabilized probe under identical conditions; a difference larger than ~1 MHz between the comb-based reading and the wavelength-meter-derived value would disprove the fractional-error transfer model and shift the combined frequency accordingly.","supporting_citations":[],"review_version":1}