{"id":"d4d299c9-7899-43c9-9abf-33226895d268","arxiv_id":"2607.24318","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Resonant strong-drive selective population trapping with blue-sideband refocusing prepares metrologically useful Fock mixtures up to η≈0.5 with ≤10% error and recovers up to 9 dB Fisher information.","lead":"A resonant laser pulse sequence prepares non-thermal motional Fock mixtures in a trapped ion without ground-state cooling, running far faster than prior weak-drive methods. Routine pulse-length and detuning tweaks keep errors low at strong coupling and restore metrological gain for displacement sensing.","discovery_kind":"new_method","skeptic_critique":{"model":"moonshotai/kimi-k3","headline":"Headline numbers at η≈0.5 rest on tight motional truncations (N=14–24) with no reported convergence test; the discarded thermal tail (~1% at n̄=5, N=24) and boundary anti-JC coupling are the same size as the claimed few-percent error floors.","rationale":"The reader's weakest assumption — that the counter-rotating blue sideband is the sole leakage channel and that a calibrated (f,δ) suffices for the full sequence — is related to but distinct from my concern. The paper actually gives unusually good support for the sole-channel diagnosis (the by-hand term removal restoring machine-precision trapping, plus the recoil-free reset control), so I do not think that is where the claim is softest. Where I agree with the reader is that the (f,δ) values are fitted per operating point on the same trace-distance metric used as the figure of merit, which is fine as calibration but means the headline numbers inherit any bias in the simulated metric — and the untested motional truncation is precisely such a bias source, sitting at the same order of magnitude as the reported residuals. My read therefore sharpens rather than replaces the reader's caveat. I do not recommend changing the CONDITIONAL verdict: the concern is fully checkable with the shipped code, the mechanism diagnosis is independently corroborated, and the qualitative claim (resonant SPT survives strong driving and is correctable by routine calibrations) would survive even a modest truncation-driven shift in the quantitative edge at η≈0.5. The existing conditions (experimental validation, heating/noise) already cover the dominant risks; I would simply add \"N-convergence at the benchmark point\" to the list of things to settle before the dB figures are quoted as design targets.","tokens_in":13592,"tokens_out":5258,"duration_ms":174710,"concrete_test":"Using the authors' public code, rerun the Fig. 7 benchmark (η=0.5, n̄=5, readout n=4) and the Fig. 5A refined curve at N=32 and N=40, re-calibrating (f,δ) at each N. If the η=0.5 trace distance moves by more than ~20% (e.g. 0.10→0.12+) or the recovered Fisher-information gain shifts by more than ~1 dB, the headline \"≤10% up to η≈0.5 / 4–9 dB recovery\" is truncation-limited and must be restated with a convergence bound; if the results are stable to N=40, the central claim stands as reported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The strongest claim is quantitative: preparation error ≤10% up to η≈0.5 and 4–9 dB Fisher-information recovery (Figs. 5, 7). All of it comes from QuTiP simulations with motional truncation N=14 (default), N=16 (Fig. 5), and N=24 (the n̄=5 metrological benchmark, Fig. 7), and no convergence study in N is reported anywhere. Three concrete pressures make this load-bearing. (1) At the benchmark point n̄=5, the thermal tail above the cutoff is P(n>24)=(5/6)^25≈1% — directly comparable to the few-percent trace distances being reported. (2) The trap-state ladder n=m² places n=16 at the very edge of the N=16 space used for Fig. 5, and the n=25 trap state falls outside the N=24 benchmark space entirely; the ~4% of initial population that the ideal protocol would funnel into m=4,5 manifolds has no consistent target in the truncated space. (3) The counter-rotating coupling at the boundary, ην√(N+1)/2 = 1.25ν at η=0.5, N=24, is not small against its 2ν detuning, so the anti-JC excursion — the very leakage channel the paper diagnoses and corrects — is artificially clipped precisely where it is strongest. The truncation could therefore either inflate or mask the residual infidelity, and the (f,δ) calibration is then partially compensating a basis artifact. Separately, the single-pulse refocusing rationale assumes Ωg(n)=√((2ν)²+η²ν²(n+1)) depends weakly on n, but at η=0.5, n=4 this varies ~15% across the populated manifold, so the analytic justification is stretched at exactly the point where the headline \"9 dB recovery\" is extracted — though the by-hand term-removal test and the public code lend the empirical claim real support.","agreement_with_reader":"partial"},"referee_report":{"model":"moonshotai/kimi-k3","summary":"The manuscript proposes a resonant implementation of the authors' earlier Selective Population Trapping (SPT) protocol for preparing non-thermal Fock-state mixtures of a trapped ion's motion. Using a polaron-frame transformation that is exact in the Lamb–Dicke parameter η, the authors identify an operating point (∆=0, Ω=ν) at which a resonant Jaynes–Cummings interaction — and hence the trapping mechanism — survives strong driving, giving a preparation speed-up of 10²–10³ over the weak-driving protocol. The residual infidelity at large η is attributed to the counter-rotating blue-sideband term discarded by the single RWA of the derivation; this diagnosis is corroborated by an ablation test in which removing that term by hand restores trapping to machine precision. Two routine calibrations — a percent-level pulse-duration refocusing factor f and a Bloch–Siegert compensating detuning δ — are shown in full Lindblad simulations (with recoil) to keep the trace distance to the ideal trapped distribution at or below ~10% up to η≈0.5 and to restore 4–9 dB of displacement-sensing Fisher information relative to the uncorrected sequence. A useful negative result shows that smooth amplitude shaping is counterproductive at the resonant point, while time-dependent detuning is the compatible shaping degree of freedom.","tokens_in":14068,"tokens_out":5469,"duration_ms":96751,"significance":"If the quantitative claims hold, this is a practical contribution: it removes both the weak-driving and Lamb–Dicke restrictions of the original SPT scheme without requiring ground-state cooling, using only calibrations (pulse length and detuning scans) that are standard in ion-trap laboratories. Particular strengths deserve emphasis: (i) the polaron derivation is non-perturbative in η, with a single, explicitly identified approximation whose leading correction is then diagnosed and suppressed — a clean logical structure; (ii) the \"sole leakage channel\" claim is supported by a controlled numerical ablation rather than asserted; (iii) the full-sequence simulations use the lab-frame Hamiltonian with a Lindblad reset including photon recoil, not the effective model being validated; (iv) the (f,δ) calibration landscape is experimentally falsifiable via a routine two-parameter scan; and (v) simulation code is publicly available on GitHub. The main vulnerability is that every headline number at strong coupling rests on tight motional truncations with no reported convergence study, at exactly the point where the discarded thermal tail and basis-boundary effects are the same size as the rep","major_comments":[{"comment":"All quantitative headline results — preparation error ≤10% up to η≈0.5 (Fig. 5A) and the 4–9 dB Fisher-information recovery (Fig. 7) — come from simulations with motional truncations N=14 (default), N=16 (Fig. 5), and N=24 (the n̄=5 metrological benchmark), and no convergence study in N is reported anywhere. Three specific pressures make this load-bearing rather than cosmetic: (a) at the benchmark point n̄=5, N=24, the discarded thermal tail is P(n>24)=(5/6)^25≈1%, directly comparable to the few-percent trace distances reported; (b) the trap-state ladder n=n0m² places n=16 at the very edge of the N=16 space used for Fig. 5, and the n=25 trap state falls entirely outside the N=24 benchmark space, so the ~1% of initial population that the ideal protocol would funnel into the m=5 manifold has no consistent target, and the ideal distribution ptr(m) of Eq. (14) is itself modified by the trunc","section":"§3.1, Figs. 5 and 7; §3, simulation parameters"},{"comment":"The mechanistic justification for the refocusing correction states that Ωg(n)=√((2ν)²+η²ν²(n+1)) 'depends only weakly on n', so that a single global factor f 'collectively refocuses the excursions of all relevant trap states'. Quantitatively this is stretched exactly where the correction is most needed: at η=0.5, Ωg(n)/ν ranges from ≈2.12 (n=1) to ≈2.55 (n=9), a ~20% spread, and with τ=2π/(ην) the refocusing phases Ωg(n)τ differ by O(π) across the populated trap manifold — so no single f can satisfy Ωg(n)fτ ∈ 2πZ simultaneously, and the excursions at large n are also the largest (matrix element ∝√(n+1)). The numerical calibration landscape (Fig. 5B) does show that optima exist and that f stays within ~5% of unity on the nearest branch, so the scheme empirically works; but the analytic narrative as written overstates why. The authors should either quantify the residual defocusing across t","section":"§3.1, paragraph following the identification of the counter-rotating term"}],"minor_comments":[{"comment":"Fig. 4B is labelled 'old protocol (Ω=0.001)' while Fig. 3 compares against Ω=ν/100 and ν/1000; please state the drive strength used for the legacy-protocol curve in Fig. 4B unambiguously, and state the motional truncation used for Figs. 2–4 (the N=16/24 values are only given for Fig. 5 onward).","section":"Figs. 3–4"},{"comment":"The phrase 'trace distance to the ideal trapped distribution' is used throughout, but it is never stated whether this is the trace distance between the full final motional density matrix and a target state built from ptr(m) of Eq. (14), or a classical distance between populations. Please define it operationally in §3.","section":"§3, below Eq. (14)"},{"comment":"The choice Γ=1000ν for the reset and ΩY=100ν for the Y pulse should be briefly justified: the Y pulse is simulated with the full Eq. (12) including D(iη), but at ΩY/ν=100 its own strong-driving corrections could inject errors not present in the effective description; a sentence confirming convergence in ΩY (or noting that results are insensitive to it) would help. Similarly, comment on how the results depend on Γ.","section":"§3, pulse sequence steps (ii) and (v)"},{"comment":"A brief experimental-context remark would strengthen the paper: η up to 0.5 with Ω≃ν and a fast reset is a demanding combination; indicating which platforms/transitions could reach the strong-η end (e.g. microwave-dressed or Sagiv/deep-Lamb–Dicke-violating regimes) would help readers place the benchmark.","section":"§4 or §1"},{"comment":"Typographical: 'R W A' appears with spurious internal spacing throughout; 'Keywords:trapped ions' is missing a space; the floor brackets in Eq. (14) and the axis label 'detuning [ ]' in Fig. 5B render incorrectly; several figures (e.g. Fig. 2) have small shared axis labels that are hard to read.","section":"throughout"},{"comment":"The public GitHub repository is commendable; consider archiving the exact version used (e.g. a Zenodo DOI) so the record is stable, and including the scripts that generate Figs. 5B and 7 specifically, since these carry the headline claims.","section":"Appendix B, Data Availability"}],"recommendation":"major_revision","confidential_remarks":"The work is a direct extension of the group's earlier SPT paper (ref. 19), from which the trapping-rate formula and the ideal-target definition are imported; the new content (resonant operating point, leakage diagnosis, refocusing protocol, shaping analysis) is nonetheless substantial and self-contained. My one reservation is that the paper is entirely simulation-based and its most newsworthy numbers sit at η≈0.45–0.5, precisely where the unexamined basis truncation is most likely to matter; I would want the N-convergence check in hand before recommending acceptance. The request is easily within the authors' existing code base, so I expect a revision to be straightforward."},"author_rebuttal":null,"desk_editor":{"model":"grok-4.5","letter":"Punchline: this is a real, usable speed-up of the authors’ selective population trapping, not a rebrand. They put the ion–laser drive in the polaron frame, hit the resonant point Δ=0, Ω≃ν, and show SPT still works without ground-state cooling and without the old Ω≪ν bottleneck. That is the new result. The residual large-η leakage is cleanly pinned on the counter-rotating blue sideband; a percent-level pulse-length factor f plus a small Bloch–Siegert detuning δ brings the full multi-cycle sequence (with recoil) back to ~10% error at η≈0.5 and recovers several dB of displacement Fisher information that the bare resonant sequence loses. Code is public. Derivation is honest about the single RWA step.\n\nWhat works: the theory path is transparent, the by-hand removal of the anti-JC term restoring trapping is the right control, and they correctly show that ordinary amplitude shaping is counterproductive at resonance while time-dependent detuning is the natural knob. Comparison to the weak-drive SPT paper is fair, not circular—the speed-up and η window come from fresh lab-frame simulations.\n\nSoft spots, in proportion: no experiment and no heating/amplitude noise, which they flag. More load-bearing for the abstract numbers is the stress on motional truncation. Default N=14, Fig. 5 at N=16, metrology bench at N=24 with n̄=5; no convergence plot. At that point the thermal tail and the clipped boundary anti-JC coupling are the same size as the few-percent floors they quote, and the n=m² ladder puts higher trap states at or past the cutoff. So the ≤10% / 9 dB claims at η≈0.5 are plausible but not yet hardened. The weak-n dependence assumed for a global f is also stretched at η=0.5. Neither sinks the mechanism; both are fixable referee points.\n\nWho it is for: trapped-ion sensing and motional-state prep people who already care about Fock mixtures and force/displacement metrology. Worth a serious referee. I would engage—read the code, ask for an N-convergence panel—not dismiss.","headline":"Clean strong-drive extension of their own SPT protocol; headline η≈0.5 numbers need a truncation check before you trust the last few percent.","tokens_in":14964,"tokens_out":561,"would_cite":true,"duration_ms":20849,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"grok-4.5","headline":"A resonant laser pulse sequence prepares metrologically useful motional Fock mixtures in a trapped ion without ground-state cooling, and two routine calibrations keep errors low up to strong coupling.","keywords":["trapped ions","quantum sensing","Fock state mixtures","selective population trapping","displacement metrology","Lamb-Dicke parameter","Bloch-Siegert shift","atomic physics"],"falsifier":"Implement the full multi-cycle sequence on a trapped ion at η around 0.4–0.5 with the calibrated pulse-length factor and detuning; if the measured motional populations stay far from the ideal trapped mixture (trace distance well above ~0.1) or the displacement Fisher information fails to recover relative to the uncorrected protocol, the central claim is false.","tokens_in":14701,"feed_emoji":"⚛️","tokens_out":1009,"duration_ms":37398,"temperature":0.7,"pith_summary":"This paper aims to show that non-thermal mixtures of a trapped ion’s motional Fock states—states that beat the standard quantum limit in displacement sensing—can be prepared far faster than earlier weak-drive methods, straight from a thermal distribution and without ground-state cooling. The authors work in a polaron frame that is exact in the Lamb–Dicke parameter and find a resonant operating point (zero detuning, Rabi frequency equal to the trap frequency) where selective population trapping still works under strong driving. They identify the residual error at large coupling as one coherent process, the counter-rotating blue-sideband term dropped by the rotating-wave approximation, and suppress it with a percent-level pulse-duration refocus and a small Bloch–Siegert detuning. Full-sequence simulations keep preparation error at or below about 10% up to η≈0.5 and recover up to 9 dB of displacement Fisher information lost by the uncorrected protocol. A sympathetic reader cares because this would be a practical, cooling-free route to faster motional resources for force and displacement metrology.","feed_headline":"Fast ion Fock mixtures without cooling recover 9 dB","feed_subtitle":"Resonant driving plus two routine calibrations beat weak-drive limits for displacement sensing.","key_machinery":"The polaron-frame description of the ion–laser interaction, exact in η, which at the resonant point Ω=ν yields an effective Jaynes–Cummings generator for the trapping dynamics; the leading coherent error is the counter-rotating blue-sideband term, suppressed by a global pulse-length factor f and a static Bloch–Siegert detuning δ.","core_discovery":"At zero detuning and Rabi frequency matching the trap frequency, selective population trapping of motional Fock states survives strong driving. Residual infidelity at large Lamb–Dicke parameter η traces to the counter-rotating blue-sideband term neglected in the rotating-wave approximation; percent-level refocusing of the pulse duration plus a small compensating Bloch–Siegert detuning keep preparation error at or below the 10% level up to η≈0.5 and restore displacement-sensing Fisher information, recovering up to 9 dB relative to the nominal sequence.","pith_inferences":["The speed-up could make Fock-mixture probes competitive with squeezed-state methods in settings where phase control relative to the signal is difficult.","If motional heating and laser-amplitude noise do not spoil the refocusing, the protocol may apply directly in surface-electrode traps used for electric-field noise sensing.","Optimized multi-Fock or adaptive readout could close more of the remaining gap between refined preparation and the ideal-trap Fisher information at η=0.5."],"forward_implications":["Motional Fock mixtures for displacement metrology can be prepared orders of magnitude faster by driving at Ω=ν instead of in the weak-drive limit.","Ground-state cooling is not required before preparing these metrologically useful states from a thermal distribution.","Conventional amplitude pulse shaping is counterproductive at resonance; time-dependent detuning is the appropriate shaping degree of freedom.","Routine ion-trap calibrations of pulse duration and detuning extend usable Lamb–Dicke parameters up to about η≈0.5 while restoring sensing gain."],"fun_headline_variants":["Resonant drive preps ion Fock mixtures without cooling","Strong-drive Fock states restore 9 dB sensing Fisher info","Zero-detuning pulses trap Fock mixtures up to η≈0.5","Two calibrations fix blue-sideband error in fast Fock prep","Trap-frequency Rabi drive beats weak-drive Fock limits"],"cache_read_input_tokens":128,"weakest_assumption_plain":"That one counter-rotating blue-sideband process is the only coherent leakage channel at large coupling, so a single global pulse-length tweak and a small static detuning restore trapping for the full multi-cycle sequence with realistic reset and recoil.","fun_headline_variants_meta":{"raw":{"variants":["Resonant drive preps ion Fock mixtures without cooling","Strong-drive Fock states restore 9 dB sensing Fisher info","Zero-detuning pulses trap Fock mixtures up to η≈0.5","Two calibrations fix blue-sideband error in fast Fock prep","Trap-frequency Rabi drive beats weak-drive Fock limits"]},"model":"grok-4.5","effort":"low","cost_usd":0.005322,"raw_usage":{"total_tokens":1452,"prompt_tokens":793,"num_sources_used":0,"completion_tokens":79,"cost_in_usd_ticks":53224000,"prompt_tokens_details":{"text_tokens":793,"audio_tokens":0,"image_tokens":0,"cached_tokens":128},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":580,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":793,"tokens_out":79,"duration_ms":10550,"temperature":1.0,"reasoning_tokens":580,"cache_read_input_tokens":128,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-31T18:16:15.261855+00:00","model_set":{"reader":"grok-4.5"},"falsifier":"Implement the full multi-cycle sequence on a trapped ion at η around 0.4–0.5 with the calibrated pulse-length factor and detuning; if the measured motional populations stay far from the ideal trapped mixture (trace distance well above ~0.1) or the displacement Fisher information fails to recover relative to the uncorrected protocol, the central claim is false.","supporting_citations":[],"review_version":1}