{"id":"dfe99a04-b286-4066-aed9-87f3a7bfb266","arxiv_id":"2607.28749","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"Mirrorless atom gravimetry is limited by detector resolution: optimal momentum transfer n*≈0.93 m/(σ_p k0 T) and sensitivity floor Δg≈4.4 σ_p/(mT√N).","lead":"This paper works out how the finite sharpness of a detector limits 'mirrorless' atom interferometers that measure gravity with large momentum transfer. It gives simple formulas for when removing the mirror pulse helps and for the best laser-kick number, which will guide the design of compact atom gravimeters.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Claimed resolution-limited floor for fringe-based readout is not a bound: optimizing mirror asymmetry (u→0) and n→∞ makes CFI unbounded within the same model.","rationale":"The paper's core derivation for the mirrorless configuration is mathematically sound, but the abstract and Sec. IV state the ceiling as a property of 'fringe-based readout' generally. The model's own asymmetric-timing family (Eq. 23) allows a joint optimization over mirror asymmetry and momentum transfer that makes the fringe CFI unbounded: with u→0 and n→∞ at fixed v=a u^2, the phase derivative grows as n while the contrast factor remains finite (R(e^{-v})≈0.239), so the fringe term scales as h* H(u) with H(u)~u^{-2}. This is not a boundary effect of the leading form; it persists in the exact Eq. (14) because the approximation remains valid when σ_p<<δp. Therefore the claimed resolution-limited sensitivity floor is not universal for fringe-based readouts; it applies only when the mirrorless endpoint is fixed. The paper's Sec. V considers asymmetry only at fixed n, so it does not uncover this divergence. A reader following the design rule n*≈0.93 m/(σ_p k0 T) would be misled into thinking that this is the optimal momentum transfer for all momentum-resolved readouts, whereas the model actually predicts that a nearly symmetric sequence with much larger n yields higher CFI. The verdict should be CONDITIONAL: accept the mirrorless-specific results, but require the claims to be restricted to the mirrorless configuration or augmented with the joint optimization over asymmetry and n.","tokens_in":22252,"tokens_out":32186,"duration_ms":285555,"concrete_test":"Numerically maximize the exact CFI of Eq. (14) over continuous n and u∈[0,1] for σ_p=4×10^-3 ℏk0, δp=1/(10√2)ℏk0, T=200 t0, using the Appendix G numerics. Confirm that the maximum is not at u=1 and increases as u→0 with n→∞, exceeding the mirrorless ceiling F_C=h*m^2T^2/(4σ_p^2).","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The ceiling (Eq. 19) is derived by maximizing the CFI of the mirrorless configuration (s=Tπ) over n. However, the model includes the continuous asymmetry family (Eqs. 2, 23). For the leading-form CFI (Eq. 15), the fringe term at asymmetry u=s/Tπ and momentum-transfer parameter a=4σ_p^2n^2k0^2Tπ^2/m^2 is F_fringe = (m^2T^2/(16σ_p^2)) h(a u^2) H(u), with h(v)=vR(e^{-v}) ≤ h*=0.20683 and H(u)=[2-(1-u)^2]^2/u^2. Since H(u)~1/u^2 as u→0, choosing u→0 and n→∞ while keeping a u^2 fixed at a* (so the wave-packet separation σ_pΔz/ℏ = √(a u^2) ≈0.93 stays at the resolving length) makes F_fringe diverge as 1/u^2. Thus the claimed 'resolution-limited sensitivity floor' Δg_min=4.4σ_p/(mT√N) is not a bound on fringe-based readout; it is an artifact of fixing the mirrorless endpoint. Section V optimizes asymmetry only at fixed n and misses this joint divergence.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies a solvable model of a light-pulse atom gravimeter with instantaneous lossless n-photon beam splitters, a Gaussian source, and Gaussian detection blur, continuously interpolating between Kasevich–Chu and mirrorless geometries. It derives closed-form expressions for the momentum-space output distributions and the classical Fisher information for momentum-resolved readout, including a fringe-phase-averaging approximation with exponentially small error. The central claims are: (i) mirrorless operation delivers a fourfold QFI enhancement over KC but only outperforms an equal-n KC device when σ_p < 0.91 m/(n k0 T); (ii) fringe-based mirrorless readout has an optimal momentum transfer n* ≈ 0.93 m/(σ_p k0 T) and a resolution-limited sensitivity floor Δg_min ≈ 4.4 σ_p/(mT√N); (iii) partial mirror asymmetry recovers part of the gain when this criterion is violated. The derivations are validated against direct numerical simulation of the pulse sequence at the 10^-8 level.","tokens_in":22541,"tokens_out":10957,"duration_ms":101624,"significance":"The paper is a technically thorough contribution with several strengths: closed-form propagator-level solutions, two independent derivations of the QFI, exact (up to a well-controlled average) CFI expressions, and numerical validation at high precision. The parameter-free constants (e.g., h* = 0.20683) are derived, not fitted, and the paper is honest about the idealized nature of the model. If the central 'resolution ceiling' claim were correct, the design criteria would be directly useful for LMT gravimeter architecture. However, the ceiling as stated is not a bound on the interpolating family considered in the same paper, and this undermines the abstract-level claim. The paper's analysis of the mirrorless endpoint is sound, but the generalization to 'fringe-based readout' without qualification is not.","major_comments":[{"comment":"The claimed 'resolution-limited sensitivity floor' (Eqs. 19–20) is not a bound on the model's fringe-based readout. For the asymmetric-timing family, take u→0 and n→∞ while keeping x ≡ β u² fixed. From Eq. (23), F_fringe/(n²k0²Tπ⁴) = [2-(1-u)²]² R(e^{-x}) → R(e^{-x}) as u→0, while n² ∝ β = x/u² diverges as 1/u². Thus F_fringe ∝ x/u² → ∞ for any fixed x>0, and the exact CFI (Eq. 14) shares this divergence because ∂_g φ̃_g ≈ n k0 Tπ² → ∞ while the contrast factor remains constant. The ceiling h*m²T²/(4σ_p²) is an artifact of fixing the mirrorless endpoint u=1; the full model has no finite supremum. This directly contradicts the abstract's 'resolution-limited sensitivity floor' and the statement that Eq. (19) is a ceiling on fringe-based readout. The authors must either restrict the ceiling claim explicitly to the mirrorless configuration (and revise the abstract), or analyze the joint (n,u","section":"Secs. IV and V, Eqs. (15), (23)"},{"comment":"The abstract states 'Fringe-based readout exhibits an optimal momentum transfer n* ... and a resolution-limited sensitivity floor Δg≃4.4σ_p/(mT√N), independent of photon momentum.' This is misleading because n* and the floor are derived only for the mirrorless case (s = Tπ). For the interpolating family, the infimum of Δg over s and n is zero within the model. The paper should explicitly qualify these results as applying to the mirrorless configuration, and should discuss whether any physical constraint (e.g., finite pulse duration, velocity acceptance, or an upper bound on n) restores a finite optimum. This is not merely cosmetic: the practical conclusions about when to use conventional LMT versus mirrorless operation depend on whether the asymmetric family can circumvent the resolution penalty.","section":"Abstract and Sec. IV"}],"minor_comments":[{"comment":"The relative error bound is stated as K(C_e) e^{-2κ_e²Σ²}, but K(C_e) is not defined or bounded. A short derivation or an explicit expression for K in terms of Fourier coefficients would make the claim more quantitative.","section":"Appendix D, Eq. (D12)"},{"comment":"The figure fixes n=5 and shows the normalized CFI across u. This is useful for fixed-n comparison, but the text should caution that the joint optimization over (n,u) is not shown and, as discussed, leads to unbounded CFI in the idealized model.","section":"Fig. 4"},{"comment":"The sentence 'The resolution floor, Eq. (20), is Δg√N≈5×10^{-4} m s^{-2} for these numbers' uses the mirrorless floor without noting that the asymmetric family could, in principle, give better sensitivity in the same idealized model. The practical estimates should state this caveat explicitly.","section":"Sec. VI, first paragraph"}],"recommendation":"major_revision","confidential_remarks":"The paper is carefully executed and the closed-form derivations are a genuine contribution. The main issue is that the headline 'resolution-limited sensitivity floor' is not actually a limit of the model: the same model's asymmetric-timing family yields unbounded CFI as u→0, n→∞ with βu² fixed. This is a load-bearing interpretational error, though fixable by reframing the claims and adding the joint optimization or a physical argument for why the divergence is excluded. I do not see grounds for rejection, because the mathematical core is sound and the numerical validation is exemplary."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Let me give it to you straight: the paper is a serious, careful analysis of how detector resolution limits LMT in mirrorless atom gravimeters, but its headline result—a universal resolution-limited sensitivity floor—is not actually a bound of the model. The floor comes from fixing the mirrorless configuration (s=Tπ) and optimizing only n. The model includes a continuous mirror-asymmetry parameter u, and if you let u→0 with n→∞ while keeping σp Δz/ℏ fixed at the optimal value ≈0.93, the fringe CFI behaves as R(a*) n² k₀² Tπ⁴, which grows without bound. In other words, a slightly asymmetric interferometer with large momentum transfer evades the 'ceiling' entirely and recovers the usual n² scaling (with a reduced contrast factor 0.239 rather than 1). Section V optimizes u only at fixed n and misses this joint divergence.\n\nThat said, the paper earns credit. It is, as far as I know, the first to combine LMT, mirrorless operation, and finite detector resolution in one Fisher-information framework. The closed-form blurred distributions and CFI are derived in detail, cross-checked by two independent routes for the QFI and by numerical simulation for the CFI, with the fringe-phase averaging error exponentially bounded. The authors are unusually honest about their idealizations in Sec. VI. The specific results for the fully mirrorless configuration—the optimal n* ≈ 0.93 m/(σp k₀ T), the crossover σp < 0.91 m/(n k₀ T), and the envelope/fringe competition—are correct as statements about that subfamily and will be useful for compact-sensor design.\n\nThe soft spot is the interpretation. Calling Eq. (19) a 'resolution-limited ceiling on fringe-based readout' is misleading within the paper's own model. The model also includes Kasevich–Chu (u=0), whose CFI grows as n² with no resolution penalty, so a 'floor' that excludes the KC limit is not a floor on the model. The authors need to either prove a global bound (impossible, given this counterexample) or clearly scope the result to the mirrorless endpoint and discuss the asymmetric family's n² growth.\n\nBottom line: worth serious refereeing, but the central claim needs to be reframed. I would send it to review, with a request to address the joint u–n optimization. The derivations are sound; the conclusion overreaches.","headline":"Solid Fisher-information analysis of mirrorless LMT, but the claimed resolution floor is an artifact of fixing the mirrorless endpoint—asymmetry plus large n makes the fringe CFI unbounded within the same model.","tokens_in":23000,"tokens_out":17958,"would_cite":true,"duration_ms":159170,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Detector resolution, not photon momentum, sets the ultimate sensitivity of fringe-readout atom gravimeters.","keywords":["atom interferometry","large momentum transfer","gravimetry","Fisher information","detector resolution","momentum-resolved readout","mirrorless interferometer","quantum sensing"],"falsifier":"Measure the output momentum distribution of a mirrorless atom gravimeter at fixed interrogation time and detector resolution, and estimate the classical Fisher information as a function of n from repeated estimates of g. If the information does not peak near n* ≈ 0.93 m/(σ_p k0 T) and then fall off with the predicted contrast factor exp(−σ_p² n² k₀² T²/2m²), the claimed ceiling is wrong. A simpler check: if a velocity-selective diffraction pulse with acceptance narrower than the source width δp changes the sensitivity by more than the model's envelope term allows, the Gaussian-envelope assumpt","tokens_in":22143,"feed_emoji":"⚛️","tokens_out":12491,"duration_ms":105690,"temperature":0.7,"pith_summary":"Large momentum transfer (LMT) multiplies the gravitational phase of a light-pulse atom interferometer by n, and removing the mirror pulse in favor of momentum-resolved readout can quadruple the phase-carrying quantum Fisher information — the standard measure of how much a quantum state can tell about a parameter. This paper shows the two gains cannot be combined indefinitely: the fringes that carry the phase shrink to a period 1/n in momentum space while the detector's blur stays fixed, so the contrast decays and the extractable information is capped. In a solvable model with a Gaussian atomic source and Gaussian detection blur, the classical Fisher information is maximized at the momentum transfer n* ≈ 0.93 m/(σ_p k0 T), giving a resolution-limited sensitivity floor Δg ≈ 4.4 σ_p/(mT√N) that is independent of photon momentum. The resulting design rule is concrete: a mirrorless sequence beats a conventional equal-momentum-transfer sequence only when σ_p < 0.91 m/(n k0 T), which for rubidium sensors confines the mirrorless advantage to short-baseline instruments.","feed_headline":"Detector blur caps atom-gravimeter sensitivity at 4.4 σ_p/(mT√N)","feed_subtitle":"Past the optimal momentum transfer, extra recoils shrink the fringes; mirrorless gains require sharp readout.","key_machinery":"The analytical engine is the exact momentum-space composition of the π/2–T1–π–T2–π/2 pulse family (mirror asymmetry s=(T1−T2)/2). Each output port is a Gaussian envelope times a cosine fringe of wave number κ=Δz/ℏ; a Gaussian convolution lemma gives the blurred distribution: contrast multiplied by C_e = exp(−κ²σ_p²δp²/2Σ²), wave number stretched to κ_e = κδp²/Σ², and phase shifted. Two-port cross terms cancel exactly, leaving F_C = (∂_g φ̃_g)² R(C_e²) + m²T²/Σ² with R(C²)=1−√(1−C²). Maximizing the fringe term over n reduces to a fixed-point equation whose unique positive solution a*=0.864 gives h*=0.2068, hence n* and the ceiling. This converts the trade-off into closed-form design rules.","core_discovery":"The central claim: fringe-based momentum readout of a mirrorless atom gravimeter is a competition between the phase gain of large momentum transfer and the contrast loss from detector blur, and this competition has a sharp optimum. For Gaussian source and detector, each port's output is a Gaussian envelope times a cosine fringe of wave number growing with n; blur reduces contrast by exp(−σ_p²Δz²/2ℏ²). The Fisher information separates into envelope and fringe terms; maximizing the fringe term gives n* ≈ 0.93 m/(σ_p k0 T), sensitivity floor Δg ≈ 4.4 σ_p/(mT√N), independent of photon momentum, and crossover σ_p < 0.91 m/(n k0 T) for the mirrorless scheme to beat a conventional equal-n sequence.","pith_inferences":["The same fringe-compression mechanism should appear in any readout that extracts phase from momentum-space or spatial fringes, such as phase-shear and point-source imaging; the numerical constants will shift, but the qualitative picture of a resolution-limited optimum in momentum transfer is likely general.","A practical consequence the authors leave implicit: for compact gravimeters, improving detector resolution is as valuable as increasing momentum transfer, because the ceiling scales linearly in σ_p while the n² gain saturates at the optimum.","Because the controlling parameter is the dimensionless combination σ_p k0 T/m, the same design curves should transfer to other atomic species with equal values of this parameter — a testable scaling beyond the rubidium numbers in the paper."],"forward_implications":["A fixed detector resolution makes the extractable Fisher information grow with momentum transfer only up to n* ≈ 0.93 m/(σ_p k0 T); beyond that, extra recoils compress the fringes faster than they enlarge the phase, and the extractable information falls off.","The best per-shot sensitivity with fringe readout is Δg ≈ 4.4 σ_p/(mT√N); improving the detector's momentum resolution improves this directly, while adding photon momentum beyond the optimum does not.","A mirrorless sequence beats a conventional equal-n sequence only when σ_p < 0.91 m/(n k0 T). When that fails, partial mirror asymmetry (s between 0 and T/2) recovers part of the enhancement, with the optimized separation kept below the detector's resolving length.","Population-based readout, because its signal is a global phase rather than a momentum-space fringe, is insensitive to detector blur and grows as n²; at large n it favors the conventional sequence over the mirrorless one.","For rubidium, the long-baseline case (T=260 ms) requires σ_p below about 2×10⁻⁵ ℏk0 even at n=1, beyond current capability; the short-baseline case (T=10 ms) yields n* ≈ 5 and a mirrorless advantage up to n≈4."],"fun_headline_variants":["Atom gravimetry: blur sets optimal n at 0.93 m/(σ_p k0 T)","Mirrorless LMT beats conventional only if σ_p < 0.91 m/(n k0 T)","Sensitivity floor Δg ≈ 4.4 σ_p/(mT√N) regardless of photon momentum","Trade-off: LMT gain vs detector blur gives n* ≈ 0.93 m/(σ_p k0 T)","Atom gravimeter resolution limit: n* = 0.93 m/(σ_p k0 T), Δg = 4.4 σ_p/(mT√N)"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The calculation rests on the idealization that each beam splitter instantaneously transfers exactly n photon momenta with no velocity dependence and that the atomic source and detection blur are Gaussian; if real high-order diffraction pulses filter the momentum distribution or leave the internal state unlabeled, the fringe contrast and the envelope width change, and the quoted numerical thresholds will shift.","fun_headline_variants_meta":{"raw":{"variants":["Atom gravimetry: blur sets optimal n at 0.93 m/(σ_p k0 T)","Mirrorless LMT beats conventional only if σ_p < 0.91 m/(n k0 T)","Sensitivity floor Δg ≈ 4.4 σ_p/(mT√N) regardless of photon momentum","Trade-off: LMT gain vs detector blur gives n* ≈ 0.93 m/(σ_p k0 T)","Atom gravimeter resolution limit: n* = 0.93 m/(σ_p k0 T), Δg = 4.4 σ_p/(mT√N)"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000864,"raw_usage":{"total_tokens":3636,"prompt_tokens":850,"completion_tokens":2786,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":594,"completion_tokens_details":{"reasoning_tokens":2637}},"tokens_in":594,"tokens_out":2786,"duration_ms":19106,"temperature":1.0,"reasoning_tokens":2637,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-03T00:29:27.615579+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the output momentum distribution of a mirrorless atom gravimeter at fixed interrogation time and detector resolution, and estimate the classical Fisher information as a function of n from repeated estimates of g. If the information does not peak near n* ≈ 0.93 m/(σ_p k0 T) and then fall off with the predicted contrast factor exp(−σ_p² n² k₀² T²/2m²), the claimed ceiling is wrong. A simpler check: if a velocity-selective diffraction pulse with acceptance narrower than the source width δp changes the sensitivity by more than the model's envelope term allows, the Gaussian-envelope assumpt","supporting_citations":[],"review_version":1}