{"id":"cdaca641-a863-424c-a2a2-0dfe062ea50e","arxiv_id":"2506.11881","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"low","formal_verification":"none","parameter_count":1,"one_line_summary":"A trapped atom interferometer using Floquet-Bloch bands is made first-order insensitive to lattice depth noise at special 'magic' operating points.","lead":"Researchers built an atom interferometer inside a laser lattice and engineered its energy bands so the measurement is insensitive to laser power noise. This could lead to compact, portable force sensors that do not need tall free-fall towers.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Magic condition ∂ϕ/∂V0=0 misses the δV noise channel: lattice intensity fluctuations scale both V0 and δV, so common-mode laser noise immunity is not demonstrated by the static V0 scans.","rationale":"The paper is a strong experimental demonstration of a programmable trapped interferometer: the parameter-free match in Fig. 3c, the phase-scan control in Fig. 4f, and the measured flatness of imbalance versus V0 are independent supports. My concern is not that the magic condition fails to produce ∂ϕ_Int/∂V0 = 0; the data support that. The issue is that the quantity the paper calls 'lattice intensity noise' couples to both V0 and δV. The magic condition is a derivative at fixed δV, so it is not the right condition for common-mode power fluctuations. This is an internal mismatch between the claimed immunity and the defined condition, not a disagreement with external consensus. The experiment's static V0 scans would look identical even if ∂ϕ_Int/∂δV were large, so the headline claim is currently under-supported. A numerical derivative test would settle whether the δV channel is actually a problem; if small, the claim stands. The reader's quasistatic mapping issue is related but secondary: even in the purely quasistatic limit the common-mode channel remains untested.","tokens_in":23206,"tokens_out":6931,"duration_ms":113358,"concrete_test":"Using the Floquet-Bloch solver of Eq. S47, compute the total common-mode derivative dϕ_Int/dε = V0 ∂ϕ_Int/∂V0 + δV ∂ϕ_Int/∂δV at the nominal magic point (V0 ≈ 8.85 E_R, δV ≈ 0.35 E_R, Δq = 0.4 ℏk_L, T_B = 10.7 ms). If |δV ∂ϕ_Int/∂δV| is not small compared with the V0 term (or compared with π over the loop), then the magic condition does not suppress intensity noise; an experimental check would then be to repeat the Fig. 2d force scan while adding a slow ±1% common-mode modulation of the lattice beam power and compare fringe shifts with those from V0-only offsets.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section IV defines magicness as ∂ϕ_Int/∂V0 = 0 and claims intrinsic immunity to lattice intensity noise. But the lattice potential is V(x,t) = -(V0 + δV sin ωt) cos²(k_L x); a fractional intensity fluctuation ε changes both V0 → V0(1+ε) and δV → δV(1+ε), since δV is the modulation amplitude of the same beam. To first order, the intensity-noise sensitivity is dϕ_Int/dε = V0 ∂ϕ_Int/∂V0 + δV ∂ϕ_Int/∂δV. The paper nulls only the first term. The second is generically nonzero: the gap at the avoided crossing scales linearly with δV (Eq. S12), and the Stokes phase depends on δ (Eq. S23), so both quasienergy bands and scattering phases vary with δV. The experimental verification (Fig. 2b–d) varies V0 at fixed δV = 0.35 E_R (Supplemental §4.3), so it probes ∂ϕ_Int/∂V0 but cannot detect the δV channel. Even granting the reader's quasistatic-noise assumption, the defined magic condition is not matched to common-mode laser power noise. This directly affects the headline claim of an intrinsically noise-tolerant, magic-band force sensor.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper reports a continuously trapped matter-wave interferometer built from Bloch oscillations of a non-interacting 7Li BEC in a one-dimensional amplitude-modulated optical lattice. The modulation creates Floquet-Bloch band crossings that act as Landau-Zener beamsplitters and mirrors, forming a Mach-Zehnder loop in momentum space whose output is a force-dependent population imbalance. The central new claim is the existence of 'magic' Floquet-Bloch band structures, defined by ∂ϕ_Int/∂V0 = 0, for which the interferometric phase is first-order insensitive to lattice depth. The authors verify this by measuring the imbalance versus lattice depth and force scans at depths around the predicted magic depth V0 ≈ 8.85 E_R, and they demonstrate a parameter-free scaling of the fringe frequency with loop size. They also show programmable interferometer variants with pulsed beamsplitters, a second modulation tone, and controllable beamsplitter phase.","tokens_in":23454,"tokens_out":11298,"duration_ms":147421,"significance":"The Floquet band-engineering approach is inventive, and the experimental implementation is careful: the use of non-interacting 7Li at the Feshbach zero crossing, the multiple interferometer geometries, and the parameter-free analytical prediction for the force response are notable strengths. If the noise-immunity claim is fully established, the magic band structure concept could be broadly useful for trapped-atom force sensing. However, as detailed in the major comment, the reported condition and data address only the static lattice-depth channel and do not cover common-mode intensity noise that also modulates δV; the current evidence is therefore narrower than the abstract's claim of intrinsic immunity to lattice intensity noise. The authors honestly acknowledge that multi-Bloch-oscillation scaling is only simulated (Appendix E), which is a favorable sign of care.","major_comments":[{"comment":"Section IV defines the magic condition as ∂ϕInt/∂V0 = 0 and cites Figs. 2b–2d as verification. This condition is not sufficient for the claimed insensitivity to lattice intensity noise. For the lattice potential V(x,t) = -(V0 + δV sin ωt) cos²(k_L x), a fractional laser-intensity fluctuation ε changes both V0 and δV: V0 → V0(1+ε) and δV → δV(1+ε). The first-order phase response is dϕInt/dε = V0 ∂ϕInt/∂V0 + δV ∂ϕInt/∂δV. The paper nulls only the first term. The second term is generically nonzero: the avoided-crossing gap is proportional to δV (Supplemental Eq. S12), and the Stokes phase depends on δ, hence on δV, through Eqs. (S20)–(S23). The experimental verification keeps δV fixed at 0.35 E_R (Supplemental §4.3) and varies only V0, so it tests ∂ϕInt/∂V0 but cannot detect the δV channel. Even granting a quasistatic noise assumption, the central claim of an intrinsically noise-tolerant magic-band sensor is not established as stated. The authors should either show that the full directional derivative V0 ∂ϕInt/∂V0 + δV ∂ϕInt/∂δV vanishes (or is sufficiently small) at the operating point, using both theory and data that vary δV or inject intensity noise, or re-scope the noise-immunity claim to static lattice-depth changes with separately stabilized modulation depth.","section":"IV (and Supplemental §1.3, §4.3)"}],"minor_comments":[{"comment":"In Eq. (2), the upper integration limit is written as q'_r = q_r + Δq inside the integral sign; this is unconventional and should be reformatted as ∫_{q_r}^{q'_r}, with the relation q'_r - q_r = Δq stated separately.","section":"Eq. (2)"},{"comment":"The section refers interchangeably to 'lattice depth dependence' and 'lattice intensity noise'; as the major comment explains, for an amplitude-modulated lattice these are not equivalent, and the terminology should distinguish static depth changes from common-mode intensity fluctuations.","section":"Section IV"},{"comment":"The Introduction states that the interferometer is 'intrinsically insensitive to ... laser phase', but Appendix D demonstrates tolerance to initial-quasimomentum and pulse-duration variations only; no measurement or theoretical argument for laser-phase insensitivity is presented.","section":"Section I and Appendix D"},{"comment":"Appendix E explicitly states that large multi-Bloch-oscillation loops have not yet been experimentally observed; the abstract and Section I should make clear that the scaling to large loop areas is a theoretical projection rather than an experimental demonstration.","section":"Appendix E"},{"comment":"The quantitative evidence for the magic condition would be strengthened by extracting and plotting the fitted interferometer phase versus lattice depth, since the raw imbalance is a cosine of the phase and its flatness is only indirect evidence for ∂ϕInt/∂V0 = 0.","section":"Fig. 2b"}],"recommendation":"major_revision","confidential_remarks":"The δV noise-channel concern is the key technical issue. I recommend that the editor send the major comment to the authors and request either additional data or a revised, appropriately scoped claim. The paper is otherwise strong and could be suitable after this point is resolved."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The first thing you should know: this paper reports the first experimental realization of a trapped matter-wave interferometer built from magic Floquet-Bloch bands, and the core demonstration is solid. The P-D magic loop is new, the force-response scaling matches a parameter-free analytical theory, and the programmable variants (pulsed beamsplitters, multi-frequency drives, phase control) show real flexibility. The magic condition is verified by the flatness of the imbalance-versus-depth curve and by the near-overlap of fringes at three lattice depths. That is genuine, reproducible-looking evidence, and the central construction does not reduce to the authors' own prior transport work. Credit where due: the extension of magic-depth ideas from static excited bands to Floquet-Bloch bands is a useful step, and the experimental data are internally consistent.\n\nThe soft spots are real but mostly addressable. The largest is that the paper defines magicness as ∂ϕ_Int/∂V0 = 0 and then claims immunity to lattice intensity noise, but the lattice potential is V0 + δV sinωt, so a fractional intensity fluctuation changes both V0 and δV. The experiments vary V0 at fixed δV and therefore probe only one of the two first-order directions in parameter space. The stress-test note is right that the gap at the avoided crossing and the Stokes phase both depend on δV, so the common-mode noise sensitivity is V0 ∂V0 + δV ∂δV, not just the first term. That does not break the interferometer result, but it does mean the headline 'intrinsically noise-tolerant' claim is not actually demonstrated by the static scans. Either measure dynamic amplitude noise or compute/show that the δV channel is small under the same conditions; as written, the claim outruns the evidence.\n\nTwo smaller issues. The beamsplitter modulation depth δV is empirically calibrated rather than predicted; fine for an experimental paper, but worth stating. The large-loop contrast decay is attributed to transverse gradients without a direct measurement; the supplemental rules out axial curvature and mean-field effects, so the inference is plausible, but it is still an inference. Also, no data or code is included, which would help for a paper that leans on parameter-free agreement.\n\nWho is this for? Atomic physicists working on trapped interferometry and quantum sensing. It deserves a serious referee; the experimental advance and the theory-data match justify the time. I would recommend acceptance conditional on addressing the δV noise channel and softening the noise-immunity language if it is not directly demonstrated. I would cite it if I worked in this area.","headline":"A genuinely new Floquet-Bloch interferometer with a convincing magic-depth demonstration, but the noise-immunity claim is narrower than the title suggests.","tokens_in":24038,"tokens_out":2247,"would_cite":true,"duration_ms":32853,"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":"Tuning a trapped atom interferometer to a magic Floquet-Bloch band structure makes its phase first-order insensitive to lattice-depth noise, verified near 8.85 recoil energies.","keywords":["atom interferometry","Floquet-Bloch bands","optical lattices","Landau-Zener transitions","magic band structures","Bloch oscillations","force sensing","Bose-Einstein condensate"],"falsifier":"Add a known sinusoidal modulation to the lattice laser power at a frequency comparable to the Bloch frequency or the inverse loop time and record fringe scatter at and away from the magic depth; if the magic depth does not suppress the resulting scatter, the quasistatic assumption is wrong.","tokens_in":23000,"feed_emoji":"⚛️","tokens_out":12946,"duration_ms":129209,"temperature":0.7,"pith_summary":"Continuously trapped atom interferometers can measure forces in a compact device, but noise in the trapping lattice itself adds phase errors that free-fall interferometers do not have. This paper reports a way to design the trap so that this error cancels: a 'magic' Floquet-Bloch band structure in which the interferometer phase has zero first-order dependence on the lattice depth. The authors realize it with a non-interacting lithium condensate that Bloch-oscillates through an amplitude-modulated optical lattice, reaching the predicted magic depth near $V_0 \\approx 8.85\\,E_R$ for a P-D band loop, and they show that force fringes are unchanged when the depth is varied close to that point. They also demonstrate that the force response grows with the momentum-space loop area and that pulsed or multifrequency modulation gives programmable control over loop size and fringe phase. If the approach scales, it points toward compact, stable, high-sensitivity local force sensors.","feed_headline":"Atom interferometer loses sensitivity to lattice noise at magic depth","feed_subtitle":"Fringes no longer shift with lattice-depth drifts at the magic depth, stabilizing compact force sensors.","key_machinery":"The central object is the Floquet-Bloch band structure of an amplitude-modulated optical lattice: the periodic drive hybridizes static Bloch bands (here the P and D bands) into quasienergy bands, and avoided crossings between them act as tunable Landau-Zener beamsplitters. The magic condition is the identity $\\partial \\phi_{\\mathrm{Int}}/\\partial V_0 = 0$, achieved by choosing a lattice depth and loop geometry where the differential dynamical phase is stationary in $V_0$; the S band cannot satisfy it because it is repelled only from above, so the lowest usable loop uses P and D bands. Because the group velocity obeys $dx/dt = d\\tilde{E}/dq$, synthesizing a quasienergy dispersion is equivalent to synthesizing a real-space interferometer trajectory.","core_discovery":"The central claim is that amplitude modulation of an optical lattice can synthesize Floquet-Bloch bands whose avoided crossings act as beamsplitters and mirrors for a continuously trapped matter-wave interferometer, and that a subset of these band structures is 'magic': the total interferometer phase satisfies $\\partial \\phi_{\\mathrm{Int}}/\\partial V_0 = 0$ and is therefore first-order insensitive to lattice intensity. The experiment verifies this with lithium atoms in a non-interacting BEC: force scans at three lattice depths centered at $V_0 \\approx 8.85\\,E_R$ produce nearly identical fringes, while equal offsets at non-magic depths produce strongly different fringes. The paper also shows that the measured interference fringe frequency scales with loop size in line with a fit-parameter-free analytical theory, and that pulsed beamsplitters, an additional modulation tone, and a variable modulation phase can enlarge the loop, increase sensitivity, and shift the fringe phase.","pith_inferences":["Because only the first derivative is nulled, a natural extension is to search for higher-order magic loops in which $\\partial^2\\phi_{\\mathrm{Int}}/\\partial V_0^2$ also vanishes, extending protection to faster, larger amplitude noise.","The same stationary-phase design principle could be applied to other experimental parameters, such as magnetic field gradient or modulation depth, yielding multi-dimensional magic surfaces for simultaneous noise rejection.","The $1/F$ force response suggests a differential configuration: two traps operated at different effective forces would react oppositely to a common external force while sharing common-mode lattice noise, a useful geometry for weak-force searches.","Combining multifrequency Floquet synthesis with optimal control could allow loops designed to be simultaneously magic in several parameters, trading the demonstrated programmability for stability in deployed sensors."],"forward_implications":["Magic band structures exist for essentially any quasimomentum range and any set of excited bands, so the intensity-noise cancellation is not tied to one loop shape.","Force response scales with the quasimomentum loop area: measured fringes become finer as $\\Delta q$ grows, matching the analytical prediction, and simulations show continued growth when the loop spans multiple Bloch oscillations.","Landau-Zener beamsplitters make the fringe phase nearly immune to initial momentum spread and pulse-duration errors, with contrast changing only slightly for variations of $\\pm 0.1\\hbar k_L$ and $\\pm 15\\%$.","Pulsed beamsplitters avoid unwanted higher-band resonances; a second modulation tone hybridizing the P and F bands increases loop area, and shifting the phase of one beamsplitter pulse shifts the fringe phase by up to $2\\pi$.","Sensitivity in the weak-force regime scales as $1/F$, suggesting compact sensors for small forces; an accelerated-lattice frame transformation could cancel part of the applied force."],"supporting_citations":[{"why":"Establishes Bloch oscillations of atoms in an optical potential, the basic transport mechanism.","marker":"[21]"},{"why":"Demonstrates position-space Bloch oscillations in an ultracold gas, connecting band dispersion to real-space trajectories.","marker":"[22]"},{"why":"Supplies the Floquet engineering framework for quasienergy bands of periodically driven optical lattices.","marker":"[23]"},{"why":"Shows transport in Floquet-Bloch bands on the same apparatus, providing the experimental platform and trajectory-control method.","marker":"[24]"},{"why":"Introduces magic lattice depths for excited-band Bloch oscillations, the concept generalized here to magic band structures.","marker":"[27]"},{"why":"Provides the calibration and tuning of the Landau-Zener beamsplitter probability via modulation depth.","marker":"[35]"},{"why":"Gives the Landau-Zener-Stückelberg interferometer phase formalism used for the phase model.","marker":"[39]"},{"why":"Reviews Landau-Zener-Stückelberg-Majorana dynamics and the Stokes phase used in the interferometer phase calculation.","marker":"[40]"}],"fun_headline_variants":["Magic Floquet bands erase trap noise in atom interferometer","Continuous atom interferometry shrugs off lattice noise at magic depth","Floquet-engineered magic bands stabilize trapped atom interferometer","Noise-proof atom interferometer runs continuously in optical lattice","Lithium BEC interferometer goes magic, kills intensity noise"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The noise-tolerance claim rests on the assumption that lattice-power fluctuations are slow compared with the interferometer's duration, so each shot effectively sees a shifted static lattice depth; fast power fluctuations are not directly tested.","fun_headline_variants_meta":{"raw":{"variants":["Magic Floquet bands erase trap noise in atom interferometer","Continuous atom interferometry shrugs off lattice noise at magic depth","Floquet-engineered magic bands stabilize trapped atom interferometer","Noise-proof atom interferometer runs continuously in optical lattice","Lithium BEC interferometer goes magic, kills intensity noise"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000452,"raw_usage":{"total_tokens":2259,"prompt_tokens":913,"completion_tokens":1346,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":529,"completion_tokens_details":{"reasoning_tokens":1261}},"tokens_in":529,"tokens_out":1346,"duration_ms":12034,"temperature":1.0,"reasoning_tokens":1261,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T01:01:50.329402+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Add a known sinusoidal modulation to the lattice laser power at a frequency comparable to the Bloch frequency or the inverse loop time and record fringe scatter at and away from the magic depth; if the magic depth does not suppress the resulting scatter, the quasistatic assumption is wrong.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Establishes Bloch oscillations of atoms in an optical potential, the basic transport mechanism."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Demonstrates position-space Bloch oscillations in an ultracold gas, connecting band dispersion to real-space trajectories."},{"cited_title":"Holthaus, Floquet engineering with quasienergy bands of periodically driven optical lattices , Journal of Physics B: Atomic, Molecular and Optical Physics 49, 013001 (2016)","cited_arxiv_id":null,"evidence_quote":"Supplies the Floquet engineering framework for quasienergy bands of periodically driven optical lattices."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Shows transport in Floquet-Bloch bands on the same apparatus, providing the experimental platform and trajectory-control method."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Introduces magic lattice depths for excited-band Bloch oscillations, the concept generalized here to magic band structures."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the calibration and tuning of the Landau-Zener beamsplitter probability via modulation depth."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Reviews Landau-Zener-Stückelberg-Majorana dynamics and the Stokes phase used in the interferometer phase calculation."}],"review_version":1}