{"id":"d1956f52-9a2d-4233-8457-5884a0a8c80c","arxiv_id":"1908.08722","paper_version":1,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"low","formal_verification":"none","parameter_count":3,"one_line_summary":"A measurement with ultracold KRb molecules yields (1/μ)(dμ/dt) = (0.30 ± 1.00_stat ± 0.16_syst) × 10^-14 year^-1, a fivefold improvement over prior molecular limits.","lead":"Physicists used ultracold potassium-rubidium molecules to set a new limit on whether the ratio of electron mass to proton mass changes over time. Their microwave measurement improves the best molecular test by a factor of five and demonstrates a promising technique for future precision searches.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"No significant objection identified: the Kμ sensitivity uncertainty is not load-bearing because the final limit is statistics-dominated.","rationale":"I read the paper as an experimental null measurement of a temporal drift in the electron-to-proton mass ratio. The central claim is that the observed limit, (1/μ)(dμ/dt) = (0.30 ± 1.00_stat ± 0.16_syst)×10^-14 yr^-1, improves the best previous laboratory molecular limit by a factor of five and is statistics-limited. For that claim to fail, an unaccounted systematic would have to drift with time at a level comparable to the 94 mHz per-point statistical uncertainty, or the scaling factor K would have to be wrong by a very large amount. The reader flagged the Pashov-potential-derived K as the weakest assumption. I examined that assumption and found it non-load-bearing: K enters Eq. (4) linearly, and the quoted fractional uncertainty in K is only 0.4%. A 1% error in K changes the quoted statistical uncertainty by about 1×10^-16 yr^-1, i.e., 1% of the reported 1.00×10^-14 uncertainty, which is far below the experimental resolution. Even an implausibly large 10% error in K would change the statistical uncertainty to about 1.1×10^-14 yr^-1, still a factor of five better than the SF6 result. I also reviewed the systematic error budget: the interleaved Zeeman correction handles the dominant magnetic-field drift, the BBR shift coefficient uncertainty contributes at most 10 mHz of time-varying shift, the GPS-disciplined Rb clock contributes below 1 mHz, and the residual dc Stark shift is measured and its fluctuation argued to be below 1 mHz. None of these approaches the statistical floor, and the authors explicitly identify BBR as the largest systematic, consistent with the table. The quoted 'data available upon request' is a practical limitation on reproducibility but not on the correctness of the central claim. I therefore find no load-bearing concern that should change the ACCEPT verdict.","tokens_in":8813,"tokens_out":13863,"duration_ms":149411,"concrete_test":"Independently recompute the sensitivity W = ∂(E_X,v=86 − E_a,v=16)/∂ln μ using the Pashov et al. [15] coupled-channel potential curves with an independent numerical solver, including the singlet-triplet mixing and the hyperfine wavefunction composition used for the |i⟩–|f⟩ transition, and verify that the difference equals −9.45(4) THz. If the independent W deviates from the quoted value by more than 1%, re-evaluate Eq. (4); a deviation below 1% would confirm that the sensitivity coefficient is not a hidden systematic.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The reader's weakest assumption—that Kμ = −14890(60), computed from the Pashov potential curves [15], is correct—would matter only if the final μ-drift limit were sensitive to K at the few-percent level. It is not. Equation (4) scales the measured frequency drift by 1/K, so a 1% error in K changes the central value by about 3×10^-17 yr^-1 and the quoted statistical error by about 1×10^-16 yr^-1, both far below the reported 1.00×10^-14 statistical and 0.16×10^-14 systematic uncertainties. Even a fivefold underestimate of the K uncertainty (2%) would leave the factor-of-five improvement over the Shelkovnikov limit intact. The stated K uncertainty is 60/14890 ≈ 0.4%, and the measured transition frequency ν = 634.96 MHz is determined experimentally, so the conversion is robust. The other systematic effects considered—second-order Zeeman correction, BBR fluctuation, reference clock, Stark shift, density shift—are each at or below the 10 mHz systematic budget, which is an order of magnitude below the 94 mHz statistical uncertainty per spectrum. The central claim that the result is statistically limited is internally consistent with the quoted error budget.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports a new laboratory limit on the temporal variation of the electron-to-proton mass ratio μ using a ~635 MHz transition between the v=86 level of the X^1Σ^+ state and the v=16 level of the a^3Σ^+ state of ultracold ⁴¹K⁸⁷Rb molecules produced by photoassociation and STIRAP. The transition is chosen for its large sensitivity coefficient K_μ = −14890(60), calculated from published potential curves. From 16 months of intermittent measurements, the authors derive (1/μ)(dμ/dt) = (0.30 ± 1.00_stat ± 0.16_syst) × 10⁻¹⁴ yr⁻¹, a factor of five improvement over the previous most stringent molecular limit (Shelkovnikov et al.). The measurement is statistics-limited, with a 94 mHz statistical uncertainty per spectrum versus a 10 mHz systematic budget.","tokens_in":9015,"tokens_out":10212,"duration_ms":100083,"significance":"If the result stands, it provides the most accurate molecular constraint on μ-variation and demonstrates the power of ultracold, near-degenerate molecular levels for precision metrology. The analysis is internally consistent: the conversion from the measured fractional frequency drift to the μ drift is arithmetically sound, the error budget in Table I is dominated by counting statistics, and the sensitivity coefficient is derived from independent potential curves rather than from the variation data, so there is no circularity. The paper also gives a concrete path toward further improvement via molecular fountains or lattice trapping. These strengths justify publication; only local presentation issues remain.","major_comments":[],"minor_comments":[{"comment":"The title contains a typo: 'ra tio' should read 'ratio'.","section":"Title"},{"comment":"The phrase 'full-width-of-half-maximum' should be 'full width at half maximum'.","section":"Fig. 3 caption"},{"comment":"The E_cal values for the F1=5/2 manifold (0.0000, 0.0009, 0.0008, 0.0004) appear to be residuals rather than absolute energies, contrary to the column header, which states that both E_exp and E_cal are measured from the |S=0,F1=3/2,F=0⟩ state; please clarify the column definition or correct the entries.","section":"Supplementary Table 1"},{"comment":"The description of how the 4 GHz uncertainty in W was estimated is too brief; please specify which vibrational levels were compared and how the uncertainties of the vibrational level intervals were propagated into the sensitivity uncertainty.","section":"Methods, Uncertainty of the sensitivity"},{"comment":"The statement that the atomic-clock results of Refs. [9] and [10] are 'essentially measuring the variation of the electron-to-proton magnetic moment ratio' is an oversimplification; a more precise characterization of what those measurements constrain would be helpful.","section":"Introduction"},{"comment":"The phrase 'Observed limit' could be replaced with 'Measurement' or 'Constraint', since the result is a measurement with error bars consistent with zero variation.","section":"Abstract"}],"recommendation":"minor_revision","confidential_remarks":null},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"You should read this one. It is the first experimental realization of DeMille's proposal to use near-degenerate vibrational levels from different electronic potentials to get a large sensitivity coefficient, and it works: the KRb microwave transition at 635 MHz has |K| ≈ 15000, and the measured limit on μ-variation is (0.30 ± 1.00 stat ± 0.16 syst) × 10^-14 yr^-1, a factor of five better than the old SF6 molecular limit. The experiment is well executed: ultracold molecules from photoassociation and STIRAP, a resolved hyperfine structure, a transition chosen to be first-order Zeeman free, and a 16-month run with the Zeeman correction done simultaneously on a companion transition. The error budget is dominated by statistics (94 mHz vs 10 mHz systematic), and the sensitivity coefficient is computed from published potential curves in a way that is not load-bearing here—even a few percent error in K would not change the conclusion. The BBR systematic is the softest spot: the shift coefficient is 0.08(20) Hz at 300 K, meaning the uncertainty is larger than the central value, and the 10 mHz fluctuation estimate comes from a temperature range of ±2 °C. That is not fatal because the statistical floor is an order of magnitude higher, but it is worth probing in any follow-up. Also, the claim that the sensitivity is determined 'almost model independently' is a bit strong—the potential curves are experimental, but the analysis still depends on the deperturbation model—still, this is a minor overstatement, not a flaw in the result. The paper is honestly written, cites the relevant theory and potential curve work properly, and the free parameters in the hyperfine fit are standard. It is not competitive with atomic clock limits, but it is a genuinely complementary probe of electron-to-proton mass variation. I would bring it to our group meeting and would cite it if we do any cold-molecule precision work. It deserves a serious referee and should be accepted after the usual minor revisions.","headline":"A clean, statistically limited molecular measurement of μ-variation that delivers the first ultracold realization of DeMille's enhanced-sensitivity idea and improves the best molecular limit by a factor of five.","tokens_in":9582,"tokens_out":1293,"would_cite":true,"duration_ms":15143,"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":"Ultracold KRb molecules provide the most precise molecular laboratory bound on temporal variation of the electron-to-proton mass ratio, five times tighter than before.","keywords":["electron-to-proton mass ratio","ultracold molecules","KRb","variation of fundamental constants","precision spectroscopy","sensitivity coefficient","microwave transition","STIRAP"],"falsifier":"Measure the $v=86$ to $v=16$ energy interval, or the same microwave transition, in another KRb isotopologue and compare with the prediction from the same potential curves; a deviation beyond the quoted sensitivity uncertainty would invalidate $K_\\mu$. Alternatively, an independent two-photon spectroscopic determination of the interval at sub-kHz accuracy would test the potential-curve-derived sensitivity.","tokens_in":8592,"feed_emoji":"⚛️","tokens_out":7202,"duration_ms":64508,"temperature":0.7,"pith_summary":"This paper reports the most precise laboratory molecular test of whether the electron-to-proton mass ratio $\\mu = m_e/M_p$ changes over time. The authors use ultracold $^{41}$K$^{87}$Rb molecules and a microwave transition between a nearly degenerate pair of vibrational levels, one in the $X^1\\Sigma^+$ ground potential and one in the $a^3\\Sigma^+$ potential, whose energy difference is unusually sensitive to $\\mu$. From frequency measurements spread over about 16 months, they obtain $(1/\\mu)(d\\mu/dt) = (0.30 \\pm 1.00_{\\mathrm{stat}} \\pm 0.16_{\\mathrm{syst}}) \\times 10^{-14}\\,\\mathrm{year}^{-1}$, a factor of five tighter than the previous best molecular limit. Because the measurement is limited by counting statistics rather than known systematics, the result points toward substantial further improvement.","feed_headline":"KRb molecules tighten mass-ratio drift limit fivefold","feed_subtitle":"A 635 MHz KRb transition with 10^4 sensitivity gives a drift bound of (0.30 ± 1.0) × 10^-14 per year.","key_machinery":"The load-bearing object is the near-degenerate pair of vibrational levels, $v=86$ in the deep $X^1\\Sigma^+$ potential and $v=16$ in the shallow $a^3\\Sigma^+$ potential of $^{41}$K$^{87}$Rb. Their energy separation is only about 635 MHz, while the individual energies have large and opposite dependence on $\\ln \\mu$, producing $W \\approx -9.45$ THz and $|K_\\mu| \\approx 15000$. The sensitivity is computed from experimentally determined potential energy curves, so it is not tied to a free-parameter model. The protocol uses stimulated Raman adiabatic passage (STIRAP) to prepare a rovibrationally pure molecular sample, a 16 ms microwave $\\pi$-pulse to drive the transition, and a simultaneous Zeeman-sensitive transition to subtract the second-order magnetic shift.","core_discovery":"The central finding is that a microwave transition in ultracold KRb can serve as a nearly model-independent probe of the stability of $\\mu$. The transition connects $|S=0,F_1=3/2,F=0,m_F=0\\rangle$ in $v=86$ of $X^1\\Sigma^+$ with $|S=1,F_1=1/2,F=1,m_F=0\\rangle$ in $v=16$ of $a^3\\Sigma^+$, at frequency $\\nu = 634.96$ MHz. Its sensitivity coefficient is $K_\\mu = W/\\nu = -14890(60)$, with $W = \\partial \\nu/\\partial(\\ln \\mu) \\approx -9.45$ THz, so a fractional frequency measurement is magnified by roughly $1.5 \\times 10^4$ in $\\mu$. Over sixteen months of alternating measurements with a Zeeman-sensitive companion transition, the fractional frequency drift was $(-0.44 \\pm 1.47_{\\mathrm{stat}} \\pm 0.24_{\\mathrm{syst}}) \\times 10^{-10}\\,\\mathrm{year}^{-1}$, which translates to $(1/\\mu)(d\\mu/dt) = (0.30 \\pm 1.00_{\\mathrm{stat}} \\pm 0.16_{\\mathrm{syst}}) \\times 10^{-14}\\,\\mathrm{year}^{-1}$, a factor of five better than the most stringent previous molecular limit.","pith_inferences":["A direct isotope-substitution test of $K_\\mu$ is an obvious next experiment: the ratio of transition frequencies between isotopologues is a predicted function of the potential curves, so any mismatch would expose errors in the sensitivity coefficient.","If the statistical limit is pushed, the blackbody-radiation shift, currently the largest systematic, will need to be measured directly with infrared fields; the paper identifies this path but does not carry it out.","The scheme effectively turns a microwave transition into a $\\mu$-variation detector with built-in gain near $10^4$; applying the same idea to additional near-degenerate molecular pairs could create a set of independent probes with different systematic signatures."],"forward_implications":["Molecular spectroscopy now yields an independent laboratory bound on $\\mu$ drift five times tighter than the previous molecular result, complementing atomic-clock limits that probe a different combination of constants.","Because the dominant error is statistical, longer integration or larger molecular samples should improve the bound roughly as $\\sqrt{N}$ in the number of detected molecules.","The near-degeneracy enhancement can be sought in other alkali dimers or other isotopologues, where different potential curves may give even larger $|K_\\mu|$.","With proposed narrow-line laser cooling and molecular fountains or magic-wavelength lattices, linewidths near 1 Hz would convert the $\\sim 10^4$ sensitivity into a proportionally stronger $\\mu$-drift constraint."],"supporting_citations":[{"why":"Proposed the near-degenerate vibrational-level enhancement that gives $|K_\\mu| \\sim 10^4$.","marker":"[11]"},{"why":"Supplies the experimentally determined $X$ and $a$ potential curves used to compute $W$ and $K_\\mu$.","marker":"[15]"},{"why":"Describes the STIRAP transfer that produces the rovibrationally pure ultracold KRb sample.","marker":"[12]"},{"why":"Provides the previous best molecular limit on $\\mu$ variation, the baseline this measurement improves by a factor of five.","marker":"[8]"},{"why":"Provides the ab initio transition dipole moments used to estimate the blackbody-radiation shift uncertainty.","marker":"[16]"},{"why":"Provides the dipole moment used to estimate the dc-Stark shift.","marker":"[17]"}],"fun_headline_variants":["Ultracold KRb tightens mass-ratio variation limit fivefold","Fivefold better proton-electron mass ratio drift bound with KRb","KRb molecules sharpen electron-proton mass ratio drift limit by 5x","Fivefold tighter limit on electron-proton mass ratio drift from ultracold KRb"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The conversion from the measured frequency drift to a drift in $\\mu$ assumes that the sensitivity coefficient $K_\\mu = -14890(60)$, computed from the experimentally fitted potential energy curves of $^{41}$K$^{87}$Rb, is accurate to its quoted uncertainty.","fun_headline_variants_meta":{"raw":{"variants":["Ultracold KRb tightens mass-ratio variation limit fivefold","Fivefold better proton-electron mass ratio drift bound with KRb","KRb molecules sharpen electron-proton mass ratio drift limit by 5x","Fivefold tighter limit on electron-proton mass ratio drift from ultracold KRb"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000884,"raw_usage":{"total_tokens":3863,"prompt_tokens":1035,"completion_tokens":2828,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":651,"completion_tokens_details":{"reasoning_tokens":2747}},"tokens_in":651,"tokens_out":2828,"duration_ms":19893,"temperature":1.0,"reasoning_tokens":2747,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T11:30:42.624596+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the $v=86$ to $v=16$ energy interval, or the same microwave transition, in another KRb isotopologue and compare with the prediction from the same potential curves; a deviation beyond the quoted sensitivity uncertainty would invalidate $K_\\mu$. Alternatively, an independent two-photon spectroscopic determination of the interval at sub-kHz accuracy would test the potential-curve-derived sensitivity.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Proposed the near-degenerate vibrational-level enhancement that gives $|K_\\mu| \\sim 10^4$."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the experimentally determined $X$ and $a$ potential curves used to compute $W$ and $K_\\mu$."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Describes the STIRAP transfer that produces the rovibrationally pure ultracold KRb sample."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the previous best molecular limit on $\\mu$ variation, the baseline this measurement improves by a factor of five."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the dipole moment used to estimate the dc-Stark shift."}],"review_version":1}