{"id":"72a5e64e-842a-403c-b846-001d341c991c","arxiv_id":"1908.11839","paper_version":2,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"A CaF transition with magnetic-field sensitivity of -4.7(2) Hz/G supports a 6.4(8) ms rotational coherence time in a magnetic trap, with simulations suggesting >1 s is feasible.","lead":"Polar molecules in a magnetic trap usually lose their rotational quantum coherence because magnetic fields shift their transition frequencies. The authors found a CaF transition with almost zero field sensitivity and kept the molecules coherent for 6.4 milliseconds, pointing a route to the second-long coherence needed for quantum technologies.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"No significant objection identified; measured 6.4 ms coherence is well-supported, and the >1 s projection is explicitly labeled as simulation-based.","rationale":"The paper's central measured result is robust: the Ramsey fringes in the magnetic trap give a coherence time of 6.4(8) ms, and a Monte-Carlo simulation with no free parameters reproduces this as 6.1 ms. The three identified decoherence mechanisms are individually characterized, and the free-space measurements with and without a spin echo support the interpretation that no additional background decoherence is present. The reader's weakest assumption, adiabatic following of the local field direction, is indeed the most uncertain element of the projected >1 s coherence in a biased trap. However, this assumption is physically standard, the non-adiabatic region in the current quadrupole trap is estimated to be very small compared with the cloud size, and the agreement between the simulation and the measured 6.4(8) ms provides indirect validation. The paper also carefully labels the >1 s result as a simulation-based suggestion rather than a demonstrated measurement, and it explicitly discloses the limitations of the electronic-structure predictions for molecules other than CaF. There is no internal inconsistency, no unsupported data analysis, and no overclaim that would warrant changing the ACCEPT verdict.","tokens_in":10532,"tokens_out":27581,"duration_ms":273675,"concrete_test":"Re-run the Monte-Carlo Ramsey simulation with the spin degree of freedom propagated by the full time-dependent Schrödinger equation, including near-zero-field Landau-Zener transitions, instead of assuming adiabatic field following, for the 45 G/cm quadrupole trap. If the predicted 1/e coherence time moves by more than the experimental 0.8 ms uncertainty, the adiabatic-based decomposition and the >1 s projection would be called into question.","verdict_should_be":"UNCHANGED","load_bearing_attack":"After reviewing the measured Ramsey data, the parameter-free Monte-Carlo simulation, and the stated limitations, I find no load-bearing flaw in the central claim. The 6.4(8) ms coherence time is a direct measurement with stated statistical uncertainty, and the simulation's 6.1 ms agreement supports the mechanistic decomposition. The >1 s projection is explicitly presented as a simulation-based suggestion rather than a demonstrated result, and the paper flags its own limitations (Yb calculations excluded, experimental gr values only for CaF). The most uncertain element is the adiabatic-following assumption behind the geometric-phase mechanism (iii), which controls the proposed bias-field suppression; however, this assumption is standard for slowly varying fields, the zero-field non-adiabatic region in the 45 G/cm trap is estimated to be sub-micron, and the matching simulation provides indirect validation. I therefore do not see a concrete soft spot that would change the verdict.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript presents a combined theoretical and experimental study of magnetically insensitive rotational transitions in 2Σ molecules and their use for long rotational coherence in magnetic traps. The authors derive an expression for the residual magnetic sensitivity of transitions between stretched rotational states, calculate the relevant molecular g-factors (g_l and g_r) by DFT and Hartree-Fock for a series of laser-coolable molecules, and identify several candidate transitions with sensitivities below 10 Hz/G. For calcium monofluoride (CaF), they measure the magnetic sensitivity of the |1⟩_str ↔ |2⟩_str transition to be −4.7(2) Hz/G and demonstrate a Ramsey coherence time of 6.4(8) ms for a magnetically trapped sample. A Monte-Carlo simulation reproduces the observed decay (6.1 ms) and is used to attribute the dephasing to three mechanisms: residual magnetic spread, Doppler motion, and a geometric phase from adiabatic following of the local field direction. The simulation-based projection suggests that coherence times exceeding 1 s are feasible for small, cold clouds in a biased magnetic trap.","tokens_in":10671,"tokens_out":18249,"duration_ms":164137,"significance":"If the claims hold, the work establishes a practical route to long rotational coherence in magnetically trapped polar molecules, directly relevant to quantum simulation, quantum computation, and hybrid molecule–superconducting-resonator interfaces. The central measurement is direct, with quoted statistical uncertainties and a calibration of the magnetic field using a transition whose Zeeman shift is well understood. The simulation is benchmarked to the data rather than fitted to it (6.1 ms simulated versus 6.4(8) ms measured), which is a notable strength. The paper also reports open data on Zenodo and transparently discusses limitations, including the exclusion of Yb from the calculations and the fact that experimental g_r values are only available for CaF. The identification of a class of magnetically insensitive rotational transitions is a useful predictive contribution, supported by the excellent agreement between calculated and measured g-factor ratios for CaF.","major_comments":[{"comment":"The three individual decoherence rates quoted (60 s⁻¹, 50 s⁻¹, and 100 s⁻¹) sum to 210 s⁻¹, whereas the full simulation gives a 1/e time of 6.1 ms, i.e., a total decay rate of about 164 s⁻¹, and the experiment gives 6.4(8) ms. If the mechanisms were independent and each produced an exponential decay, the rates should add; the observed discrepancy indicates either correlations between mechanisms or that the single-mechanism decays are not exponential over the fitted range. The authors should state explicitly how the rates are defined and why they are not additive, since this decomposition is used to rank the mechanisms and to motivate the proposed improvements (bias-field trap and smaller cloud).","section":"Decoherence characterization, paragraph following Fig. 3(a)"},{"comment":"The Monte-Carlo simulation is described only verbally; no parameters (initial cloud size and temperature, trap field geometry, microwave k-vector, pulse durations, or the procedure for 'artificially removing' mechanisms) are given in the manuscript or in the included supplemental material. Because the >1 s coherence-time projection rests on this simulation, a reproducible description should be provided, at least in a supplementary file, so that the decomposition and the projection can be independently assessed.","section":"Monte-Carlo simulation description (main text) and the >1 s projection"}],"minor_comments":[{"comment":"The stated relation 1/α = (λ/2π)√(m/2k_BT) appears to differ by a factor of 2 from the standard Doppler-dephasing expression 1/α = (λ/2π)√(2m/k_BT). The reported temperature 37(4) μK should be re-checked; as written, the same 1/α value of 11.4 ms would correspond to T ≈ 148 μK under the standard formula.","section":"Free-space Ramsey measurement, text near Fig. 3(b)"},{"comment":"The adiabatic-following assumption underlying mechanism (iii) is asserted without a quantitative adiabaticity estimate; the statement that the zero-field non-adiabatic region in the 45 G/cm trap is sub-micron would be more convincing with a Landau-Zener parameter or an equivalent estimate, especially because the >1 s projection relies on this assumption.","section":"Adiabatic-following assumption, mechanism (iii)"},{"comment":"The main text states that the molecules are cooled to about 50 μK, while the free-space Doppler analysis yields 37(4) μK. These values should be reconciled, particularly in light of the factor-of-2 concern above.","section":"Experimental temperature consistency"},{"comment":"The hydrides have much larger g_l and g_r values that dominate the vertical scale of Fig. 1; a brief note in the text clarifying that the survey targets laser-coolable molecules with small rotational g-factors would help the reader interpret the figure.","section":"Table I and Fig. 1"}],"recommendation":"major_revision","confidential_remarks":"The core measurement and its interpretation are strong, and the issues identified here are about internal consistency and transparency of the simulation rather than a flaw in the main result. I would support acceptance after the authors clarify the non-additivity of the decoherence rates and provide a full description of the Monte-Carlo simulation."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear Colleague,\n\nThis one is worth your time. The genuinely new thing is Eq. (2): a simple cancellation condition for stretched-state rotational transitions in 2Sigma molecules. When gl/gr sits near (2N+4)^2 - 1, the electronic and rotational Zeeman pieces nearly cancel, leaving a transition that is magnetically insensitive while both states keep moments of about one Bohr magneton. That turns a practical nuisance into a design principle.\n\nThe paper earns its central claim. The measured sensitivity of -4.7(2) Hz/G for the CaF |1>str <-> |2>str transition is a direct Ramsey measurement, calibrated against a transition whose Zeeman shift is known to high precision. The extracted gl and gr agree with the DFT/HF calculations within a few percent, so the twelve-molecule survey is not unanchored. The 6.4(8) ms trapped coherence time is a direct measurement, and the Monte-Carlo simulation reproduces it (6.1 ms) without fitting, which gives me confidence in the mechanistic decomposition into residual magnetic spread, motion-induced Doppler phase, and geometric phase. The free-space spin-echo control is a good sanity check: it shows no intrinsic decoherence above 20/s on the timescale studied.\n\nThe soft spots are modest and mostly self-flagged. The >1 s projection is explicitly a simulation-based extrapolation, not a demonstration. It leans on adiabatic following of the local field direction; that is a standard assumption for slowly varying fields, and the matching simulation gives indirect support, but the bias-field trap remains untested. Also, only CaF has experimental gr available, so the survey's predictive reach rests partly on Hartree-Fock rotational g factors. The authors say they excluded Yb because their calculations were not reliable enough, and they give the T1 diagnostic. That seems honest.\n\nCitation practice is fine. The paper cites their own earlier magnetic-trapping work where it is directly relevant, and compares fairly with the NaK magic trap result.\n\nOverall: solid, clearly written, and the kind of paper that gives people a new toolkit rather than just a new number. A serious referee should engage with it.","headline":"A genuine design principle for magnetically insensitive rotational transitions in 2Sigma molecules, backed by a clean CaF measurement and an honest Monte-Carlo simulation.","tokens_in":11213,"tokens_out":1611,"would_cite":true,"duration_ms":14508,"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":"Magnetically trapped CaF molecules in a rotational superposition keep their phase for 6.4(8) ms, and the same field-insensitive transition should support coherence beyond 1 s in a smaller, colder, biased trap.","keywords":["polar molecules","rotational coherence","magnetic trapping","CaF","Ramsey spectroscopy","stretched states","magnetic insensitivity"],"falsifier":"Load a roughly 5 µK cloud of CaF into a quadrupole trap with a large bias field so that the field-direction variation across the cloud is small, run the Ramsey sequence on the $|1\\rangle_{\\mathrm{str}} \\leftrightarrow |2\\rangle_{\\mathrm{str}}$ transition, and fit the fringe decay; if the 1/e time does not approach about 1.4 s, or if population leaks out of the stretched states, the central projection is wrong.","tokens_in":10352,"feed_emoji":"🧲","tokens_out":12133,"duration_ms":100844,"temperature":0.7,"pith_summary":"Rotating molecules can talk to each other through long-range dipolar forces, but only if their rotational superpositions stay phase-locked while trapped. This paper establishes a way to keep that lock in a magnetic trap: it identifies rotational transitions that are almost completely insensitive to magnetic fields, verifies one in CaF with sensitivity $-4.7(2)$ Hz/G, and uses it to measure a 6.4(8) ms coherence time for trapped molecules in a superposition of the $N=1$ and $N=2$ stretched (fully spin-aligned) states. The authors decompose the observed dephasing into three mechanisms and show that all three can be suppressed, by smaller, colder clouds and by a biased trap that keeps the field direction nearly fixed. If their simulation of the suppressed mechanisms is right, rotational coherence beyond 1 s is within reach, making magnetically trapped polar molecules usable for interaction-based quantum simulation and computation.","feed_headline":"Molecule rotations stay coherent 6.4 ms in a magnetic trap","feed_subtitle":"Nearly field-insensitive CaF transition is the route to >1 s coherence for quantum simulators.","key_machinery":"The load-bearing object is the stretched-state transition $|N\\rangle_{\\mathrm{str}} \\leftrightarrow |N+1\\rangle_{\\mathrm{str}}$ in a $^2\\Sigma$ molecule, i.e. states in which electron spin, rotation, and nuclear spin are all maximally projected along the magnetic field. For these states the two large Zeeman contributions are independent of $N$, leaving a residual magnetic sensitivity $\\Delta\\mu(N) = (g_l/((2N+4)^2-1) - g_r)\\mu_B$, where $g_l$ is the anisotropy of the electronic $g$ tensor and $g_r$ is the rotational $g$ factor. The scheme works when $g_l/g_r$ is close to $(2N+4)^2 - 1$; in CaF this ratio is about 36, so the two terms cancel at $N=1$ to 3.3(1) parts per million. Experimentally, the machinery is Ramsey spectroscopy with two microwave $\\pi/2$ pulses, confinement in a quadrupole magnetic trap, and recapture into a magneto-optical trap for state-selective detection. A Monte-Carlo simulation of the trapped Ramsey sequence identifies the three dephasing channels and is used to project the beyond-1-s outcome for a smaller cloud in a biased trap.","core_discovery":"The central discovery is that the stretched rotational states $|N\\rangle_{\\mathrm{str}} = |N, m_N=N\\rangle |S, m_S=S\\rangle |I, m_I=I\\rangle$ of a $^2\\Sigma$ molecule have Zeeman shifts that are almost independent of $N$, so the microwave transition between neighboring stretched states can have an exceptionally small residual magnetic sensitivity. In CaF, the $|1\\rangle_{\\mathrm{str}} \\leftrightarrow |2\\rangle_{\\mathrm{str}}$ transition has measured sensitivity $-4.7(2)$ Hz/G, meaning the two states' magnetic moments agree to 3.3(1) parts per million. In a quadrupole magnetic trap, Ramsey spectroscopy on this transition gives a coherence time of 6.4(8) ms. Monte-Carlo simulation reproduces this result and attributes the dephasing to the residual field sensitivity (60 s$^{-1}$), to motion through the microwave field (50 s$^{-1}$), and to a geometric phase acquired as molecules adiabatically follow the local field direction (100 s$^{-1}$). Free-space measurements with a spin echo show no intrinsic decoherence (below 20 s$^{-1}$ at 95% confidence), and simulations indicate that shrinking the cloud and using a biased trap can push the coherence time past 1 s.","pith_inferences":["The authors leave implicit a sharp quantitative test: at 5 µK in a bias-field trap the only modeled dephasing that cannot be designed away is the residual-sensitivity term, so the measured Ramsey decay time directly tests their decoherence budget.","Because a free-space spin echo removed all observable dephasing, a trapped spin-echo sequence might extend coherence beyond 1 s even before the bias-field upgrade, at the cost of reduced contrast from imperfect $\\pi$ pulses.","The same cancellation condition could be deliberately detuned, for example by choosing a different vibrational state or isotopologue, to create a small controlled field sensitivity where the magnetic field is wanted as a tuning knob rather than only as a noise source.","If non-adiabatic transitions occur in a real biased trap, they should appear as population loss from the stretched states on the longer-timescale runs; searching for that loss is a clean experimental check of the adiabatic assumption."],"forward_implications":["Rotational qubits encoded in $|1\\rangle_{\\mathrm{str}}$ and $|2\\rangle_{\\mathrm{str}}$ can stay coherent across the millisecond timescales needed for dipolar quantum gates.","A biased magnetic trap with a small cloud already cooled to 5 µK should reach coherence times beyond 1 s, since the projected decoherence rate from residual sensitivity is only 0.14 µK$^{-1}$ s$^{-1}$.","The measured $|1\\rangle_{\\mathrm{str}} \\leftrightarrow |2\\rangle_{\\mathrm{str}}$ transition also acts as a magnetic-field-insensitive frequency reference, useful for precision measurements where field inhomogeneity is a systematic error.","Magnetic chip traps near superconducting microwave resonators become viable platforms for strong molecule-photon coupling, because the transition is essentially field-insensitive while the trap itself is magnetic.","Electronic-structure calculations indicate that other laser-coolable alkaline-earth fluorides and hydroxides have transitions with $|\\Delta\\mu/h|$ below 30 Hz/G, so the approach should transfer to other molecular species."],"supporting_citations":[{"why":"Supplies the magneto-optical trap of CaF and the recapture-and-image detection used for Ramsey signals.","marker":"[17]"},{"why":"Demonstrates magnetic trapping and coherent control of laser-cooled CaF, providing the trap and state-preparation tools the experiment builds on.","marker":"[21]"},{"why":"Gives the Zeeman and fine/hyperfine Hamiltonians from which the stretched-state sensitivity formula is derived.","marker":"[23]"},{"why":"Supplies the sub-Doppler cooling that prepares the roughly 50 µK sample used in the Ramsey measurements.","marker":"[48]"},{"why":"Provides the precisely known hyperfine and spin-rotation parameters used to calibrate the magnetic field in the sensitivity measurements.","marker":"[50]"},{"why":"Underlies the geometric-phase decoherence mechanism attributed to molecules adiabatically following the local field direction.","marker":"[51]"},{"why":"Demonstrates the 5 µK CaF temperature on which the projected greater-than-1-s coherence time is based.","marker":"[52]"},{"why":"Describes the biased magnetic-trap geometry proposed to suppress field-direction variation and hence the geometric-phase decoherence.","marker":"[54]"}],"fun_headline_variants":["Rotational coherence in magnetic trap hits 6.4 ms","CaF transition nearly field-free, coherence 6.4 ms","Magnetic trap yields 6.4 ms rotational coherence, could reach 1 s","Trapped molecules: 6.4 ms rotational coherence, path to 1 s","Near-insensitive CaF rotation transition gives 6.4 ms coherence"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The projection to coherence times beyond 1 s assumes that molecules in the trap adiabatically follow the local magnetic-field direction, so that the geometric phase is the only field-direction effect and a biased trap makes that effect negligible; if non-adiabatic transitions occur, or if the biased trap does not suppress field-direction variation as modeled, the projection collapses.","fun_headline_variants_meta":{"raw":{"variants":["Rotational coherence in magnetic trap hits 6.4 ms","CaF transition nearly field-free, coherence 6.4 ms","Magnetic trap yields 6.4 ms rotational coherence, could reach 1 s","Trapped molecules: 6.4 ms rotational coherence, path to 1 s","Near-insensitive CaF rotation transition gives 6.4 ms coherence"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000891,"raw_usage":{"total_tokens":3823,"prompt_tokens":907,"completion_tokens":2916,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":523,"completion_tokens_details":{"reasoning_tokens":2817}},"tokens_in":523,"tokens_out":2916,"duration_ms":20086,"temperature":1.0,"reasoning_tokens":2817,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T10:05:59.429337+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Load a roughly 5 µK cloud of CaF into a quadrupole trap with a large bias field so that the field-direction variation across the cloud is small, run the Ramsey sequence on the $|1\\rangle_{\\mathrm{str}} \\leftrightarrow |2\\rangle_{\\mathrm{str}}$ transition, and fit the fringe decay; if the 1/e time does not approach about 1.4 s, or if population leaks out of the stretched states, the central projection is wrong.","supporting_citations":[{"cited_title":"Characteristics of a magneto-optical trap of molecules,","cited_arxiv_id":null,"evidence_quote":"Supplies the magneto-optical trap of CaF and the recapture-and-image detection used for Ramsey signals."},{"cited_title":"Magnetic trapping and coherent control of laser- cooled molecules,","cited_arxiv_id":null,"evidence_quote":"Demonstrates magnetic trapping and coherent control of laser-cooled CaF, providing the trap and state-preparation tools the experiment builds on."},{"cited_title":"Brown and A","cited_arxiv_id":null,"evidence_quote":"Gives the Zeeman and fine/hyperfine Hamiltonians from which the stretched-state sensitivity formula is derived."},{"cited_title":"Molecules cooled below the doppler limit,","cited_arxiv_id":null,"evidence_quote":"Supplies the sub-Doppler cooling that prepares the roughly 50 µK sample used in the Ramsey measurements."},{"cited_title":"Precise determination of the ν and N dependence of the spin-rotation and hyperfine interactions in the CaF X2Σ1/2 ground state,","cited_arxiv_id":null,"evidence_quote":"Provides the precisely known hyperfine and spin-rotation parameters used to calibrate the magnetic field in the sensitivity measurements."},{"cited_title":"Prospects for mea- suring the electric dipole moment of the electron using electrically trapped polar molecules,","cited_arxiv_id":null,"evidence_quote":"Underlies the geometric-phase decoherence mechanism attributed to molecules adiabatically following the local field direction."},{"cited_title":"Deep Laser Cooling and Efficient Magnetic Compression of Molecules,","cited_arxiv_id":null,"evidence_quote":"Demonstrates the 5 µK CaF temperature on which the projected greater-than-1-s coherence time is based."},{"cited_title":"Magneto- static trapping fields for neutral atoms,","cited_arxiv_id":null,"evidence_quote":"Describes the biased magnetic-trap geometry proposed to suppress field-direction variation and hence the geometric-phase decoherence."}],"review_version":1}