{"id":"161acc6c-b626-49b7-a171-0b2302ba0cd5","arxiv_id":"2507.12995","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"An optically levitated silica nanoparticle was released, fell freely for up to 0.25 ms under gravity, was recaptured by a second optical tweezer, and showed an approximately 190-fold growth in position uncertainty.","lead":"Researchers released a glass nanoparticle trapped by a laser beam, let it fall under gravity for up to 0.25 milliseconds, and caught it again in a second laser trap placed below. The same particle can be recycled through many falls, and its position uncertainty grew about 190-fold, a step toward matter-wave experiments with large objects.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"At the headline τ=0.25 ms point, the QPD readout is acknowledged to be sub-linear; the linear-calibration-based ξq=189(18) is therefore an unquantified, likely downward-biased estimate, not a validated measurement of state expansion.","rationale":"The paper's main claim is the expansion factor at τ=0.25 ms, and the weakest link is the measurement of that point. The reader's weakest assumption concerned unbiasedness of the backward-filtered covariance matrix; that is related but broader. The more specific and more damaging issue is that the QPD transduction is nonlinear at exactly the amplitudes reached at the headline time, and the paper's own explanation for the model-data deviation is an unquantified attribution. Since q0 and σq share the same linear calibration, a global calibration error would cancel in ξq, but a nonlinearity does not: it suppresses the large-amplitude signal used for σq while leaving the small-amplitude q0 essentially linear, biasing the ratio. The likely direction of the bias is downward, so the qualitative result of a roughly two-hundred-fold expansion is probably not in jeopardy; this is why the appropriate verdict remains conditional rather than reject. The proposed test—nonlinear inversion of the raw traces, or using the independently measured Duffing frequency shift as an amplitude ruler—would settle whether the reported 189(18) is accurate or should be presented as a lower bound. This does not change the reader's CONDITIONAL verdict, but sharpens the condition that must be met for acceptance.","tokens_in":16328,"tokens_out":13507,"duration_ms":176148,"concrete_test":"Re-analyze the raw QPD-y traces for τ=0.25 ms with a nonlinear inversion of the interferometric response (e.g., V(y)∝sin(4πy/λ), or the measured transfer curve) and recompute the ensemble covariance and ξq; if ξq shifts by more than the quoted 2σ uncertainty, the reported value is not robust. As an independent cross-check, estimate the RMS amplitude at recapture from the measured Duffing frequency shift using Eq. S22 and the independently ray-traced waist, and compare with the linearly calibrated σq=110(15) nm; a disagreement beyond stated errors confirms the readout nonlinearity is biasing the headline expansion.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central quantitative result is ξq=189(18), σq=110(15) nm at τ=0.25 ms. Displacements are obtained by converting QPD voltage to meters with a linear calibration factor from small-amplitude thermal spectra (Supplement Sec. IV, Eq. S20). At τ=0.25 ms the position distribution has σq≈110 nm, so instantaneous excursions reach several hundred nm, comparable to λ/4≈266 nm, where the interferometric readout is explicitly acknowledged to become sub-linear (main text: 'This deviation can be attributed to sub-linear transduction ... significant for oscillation amplitudes comparable to a quarter of the laser wavelength'). Because ξq is a ratio of variances, the numerator is suppressed by the nonlinearity while the denominator (q0≈0.6 nm) is in the linear regime, so the reported ratio is biased. Consistent with this, the parameter-free prediction of Eq. (2), using Ωy/2π=141.2 kHz and Γ/2π=24 Hz, gives ξq≈224, about 16% above the reported 189. The paper treats the deviation as an attribution, not a calibration, and does not correct the headline value. The qualitative conclusion of a large expansion is probably safe, since the bias is likely downward, but the specific claim ξq=189(18) is not a validated measurement.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports trap-to-trap free-fall experiments with a charge-neutral, optically levitated nanoparticle. A dual optical tweezer is toggled off to release the particle for up to 0.25 ms, then reactivated at a vertically displaced focus to recapture it, allowing repeated cycles. The measured vertical acceleration ay = -10.9 ± 0.4 m/s² is consistent with gravity, and recapture energetics are described by Eq. (1) with fitted trap depth and displacement calibration. The central result is a measured state expansion ξq = 189(18), i.e., σq = 110(15) nm, at τ = 0.25 ms, compared with the parameter-free prediction ξq ≈ 224 of Eq. (2) built from independently measured Ωy, γ, T0, and q0. The paper further projects that with ground-state cooling and cryogenic UHV operation, expansions of order 1700 and coherence lengths of several nanometers become reachable.","tokens_in":16582,"tokens_out":14430,"duration_ms":149128,"significance":"If the quantitative claims are validated, this is a genuine experimental advance for levitodynamics: the first demonstration of a recyclable free-fall protocol for a charge-neutral nanoparticle, with the expansion predicted by a parameter-free formula (Eq. 2) that requires no constants fitted to the expansion dataset, and an independent gravitational-acceleration check of the free evolution. The loss-probability model of Fig. 4 provides a testable roadmap toward millisecond free-fall times and nanometer-scale coherence lengths. These strengths should be credited: the central prediction uses independently measured parameters, and the reported error bars on the acceleration and expansion are explicit. The main weakness is that the headline expansion value at the longest free-fall time is affected by a readout nonlinearity acknowledged in the text but not corrected, so the precise quantitative claim is not yet validated.","major_comments":[{"comment":"The headline quantitative claim, ξq = 189(18) at τ = 0.25 ms, coincides with the largest deviation from the parameter-free prediction of Eq. (2) (which gives ξq ≈ 224), and the text attributes this deviation to sub-linear transduction in the interferometric readout 'significant for oscillation amplitudes comparable to a quarter of the laser wavelength.' Because ξq is defined as the ratio of variances obtained through a linear volt-to-meter calibration (Supplement Sec. IV, Eq. S20), and because the recaptured position distribution (σq = 110 nm) extends to instantaneous amplitudes of several hundred nanometers, the numerator of this ratio is suppressed by the transduction nonlinearity while the denominator (q0 ≈ 0.6 nm) remains linear. The reported value is therefore a biased estimate of the true expansion, with a bias that is likely comparable to or larger than the 16% deviation between data and the parameter-free prediction, and the paper neither corrects nor bounds it. The authors should either recalibrate the readout in the nonlinear regime, provide a rigorous systematic-uncertainty band for ξq at τ = 0.25 ms, or rephrase the headline claim so that the measured lower bound and the model prediction are clearly distinguished.","section":"Results; Fig. 3(a), Eq. (2)"},{"comment":"The covariance estimator used to produce ξq and ξp is justified only by the low-efficiency Kalman limit (diagonal conditional covariance Σc, ≈0.5% efficiency) and by the statement that the redistribution into sidebands is a few percent; no evidence (synthetic-trajectory tests, filter consistency checks, or cross-validation against an independent estimator) is given that the inferred eigenvalues of the covariance matrix are unbiased at the operating point. This matters precisely because the headline point (τ = 0.25 ms) is also the point with up to 5 kHz Duffing-induced frequency shifts after recapture, a finite-sampling rotation δθ, and the readout nonlinearity of the previous comment. The two effects are entangled in the reported ξq = 189(18), and the text's attribution of the deviation to sub-linear transduction alone is an assumption, not a diagnosis. A quantitative validation of the estimator, or at least an estimate of its bias at the reported parameters, is needed before the expansion and compression factors can be treated as measured state-space quantities.","section":"Supplement Sec. V; Fig. 3"}],"minor_comments":[{"comment":"There is a duplicated article in 'we obtain U0 = 5.4(2)·10^5 and the and the calibration factor cf = 95(2) nm/MHz'; the duplicate 'the' should be deleted.","section":"Results"},{"comment":"Reference [3] contains a corrupted author name ('F. Kia/suppress lka' should be F. Kiałka), and references [16], [19], and [20] list placeholder page information (', 1 (2024)') rather than complete publication data; the bibliography should be updated.","section":"References"},{"comment":"The rendered caption of Fig. 3 contains stray subfigure labels ('(b) (b) Y Y') that should be removed.","section":"Fig. 3 caption"},{"comment":"In the concluding section, 'perquisite for matter–wave experiments' should read 'prerequisite'; a light language edit is otherwise recommended (e.g., 'Fig. 2(a-c) shows' with a plural subject).","section":"Concluding discussion"},{"comment":"The number of experimental realizations for the additional τ = 0.25 ms dataset (recorded with a different particle) is not stated, although the other datasets report 100 realizations; please specify it.","section":"Results; Fig. 3(a)"},{"comment":"The residual ay - g ≈ 1.1 m/s² is attributed to volt-to-meter calibration errors; the same residual could in principle contain an electric-field contribution, and a brief estimate of the implied upper bound on the particle's residual charge would strengthen the charge-neutral claim.","section":"Results; Fig. 2(b)"}],"recommendation":"major_revision","confidential_remarks":"The manuscript fits the scope of a general physics journal in the quant-ph area. The stress-test concern about the τ = 0.25 ms point is confirmed by the manuscript itself, which explicitly attributes the deviation at that point to sub-linear transduction without correcting the headline number. The revision is feasible within the manuscript's scope: it requires re-analysis and possibly a short calibration study, not a new apparatus. I recommend major revision rather than rejection because the qualitative result (large, gravity-dominated free expansion with recycling) is credible and the parameter-free prediction of Eq. (2) is a strong asset."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The genuinely new thing here is not free fall of a levitated particle—that has been shown before—but the dual-tweezer recapture-and-recycle protocol with a charge-neutral particle, and it works. The authors release a cooled 120 nm silica sphere from an upper tweezer, let it fall up to 0.25 ms, catch it in a vertically displaced lower tweezer, and bring it back for repeated trials. The paper is honest about what limits the free-fall time and what would extend it.\n\nWhat is done well: the state-expansion prediction in Eq. (2) is parameter-free, built from independently measured trap frequency, damping rate, and initial temperature, and it tracks the data over most of the range. The energy-at-recapture model (Eq. 1) with the U0 and calibration fit parameters is consistent with independent estimates within 10%. The acceleration measurement along y, −10.9±0.4 m/s², is close enough to g that the motion is dominated by gravity, and the x and z accelerations vanish as expected. The phase-space reconstruction and the explicit discussion of Duffing shifts at recapture show care.\n\nSoft spots, in proportion: the headline number at τ=0.25 ms, ξq=189(18), is exactly the point where the authors invoke sub-linear QPD transduction to excuse the model-data deviation. The parameter-free prediction gives ξq≈224, about 16% higher. Since the nonlinearity compresses the measured variance while leaving the initial variance in the linear regime, the reported expansion ratio is likely a downward-biased estimate, not a validated measurement. The paper says as much in passing, but it does not correct or re-derive the headline value. That is a real weakness, but a localized one: the qualitative claim of a roughly two-hundred-fold expansion and the recapture/recycle capability do not depend on the exact number. The y-acceleration offset is attributed to volt-to-meter calibration without quantification; that is minor and plausibly right, but a sentence of support would help. No data or code are released, which makes independent verification harder, though the methods are detailed enough to reproduce in principle.\n\nOverall: a serious experimental paper with a new capability and mostly solid analysis. The 0.25 ms expansion point should be re-examined or re-labelled as a lower bound, and that should be fixable in revision.\n\nRecommendation: send it to a good referee. I would bring it to the reading group and would cite it once the expansion value is stated honestly.","headline":"Solid experimental demonstration of recapture-and-recycle free fall for a levitated nanoparticle; the headline expansion value is likely a bit low, but the capability is real.","tokens_in":17172,"tokens_out":1923,"would_cite":true,"duration_ms":20699,"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":"A charge-neutral optically levitated nanoparticle can be released, fall under gravity alone, and be recaptured by a displaced optical tweezer with its position uncertainty growing nearly 200-fold.","keywords":["optically levitated nanoparticle","free-fall experiment","state expansion","optical tweezer","charge neutrality","recapture","levitodynamics","matter-wave interferometry"],"falsifier":"Re-analyze the $\\tau = 0.25$ ms single-shot traces with bandpass-filter bandwidths of 1, 2, and 4 kHz and with the filter carrier shifted by $\\pm 5$ kHz, and check whether the recovered position spread stays at 110(15) nm; a systematic drift with filter settings would mean the expansion is an estimation artifact rather than real free-evolution growth.","tokens_in":16058,"feed_emoji":"⚛️","tokens_out":9478,"duration_ms":99433,"temperature":0.7,"pith_summary":"Using an optical tweezer that can be switched off, moved vertically, and switched back on, this paper shows that a charge-neutral levitated nanoparticle can be released, fall freely under gravity, and be recaptured without being lost, so the same particle can be recycled through thousands of experiments. During the free fall the particle is subject only to gravity: the measured vertical acceleration is $a_y = -10.9 \\pm 0.4$ m/s$^2$, and the horizontal accelerations vanish within error. At the longest demonstrated fall time of 0.25 ms, the vertical position uncertainty grows to $\\sigma_q = 110(15)$ nm, a state expansion factor $\\xi_q = 189(18)$ relative to the initial cooled state, in agreement with a free-evolution model that includes only the initial spread, ballistic expansion, and background-gas reheating. The authors further show that recapture is optimized by displacing the trap by the gravitational drop $g\\tau^2/2$, which minimizes the particle's energy in the recapturing trap and suppresses loss. If the projection holds, ground-state cooling and cryogenic ultra-high vacuum would extend free falls to millisecond durations, producing nanometer-scale coherence lengths that matter-wave interferometry with nanoparticles would need.","feed_headline":"Freed nanoparticle falls 0.25 ms, returns with 190-fold spread","feed_subtitle":"Charge-neutral levitated bead, recaptured after gravity-only fall, shows 190-fold position uncertainty growth.","key_machinery":"The load-bearing mechanism is the switchable, vertically displaceable optical tweezer: an acousto-optic modulator toggles the trap power on and off and shifts the focus vertically on a 300 ns timescale, so the same beam releases the particle and recaptures it after a variable free-fall time $\\tau$. The quantitative core is the covariance model for free evolution of a damped oscillator, in which the vertical position variance grows as $\\xi_q^2 \\approx 1 + \\Omega_y^2\\tau^2 + \\frac{2}{3}\\Gamma\\Omega_y^2\\tau^3$, with $\\Omega_y$ the initial trap frequency and $\\Gamma$ the background-gas reheating rate, and the recapture energy is minimized at trap displacement $d = g\\tau^2/2$. State reconstruction uses backward bandpass filtering of single-shot quadrant-photodetector traces in the low-efficiency linear-filter limit; the maximum eigenvalue of the reconstructed covariance matrix defines the expansion factor $\\xi_q$.","core_discovery":"The central discovery is that a charge-neutral optically levitated nanoparticle can be taken through a full release-free-fall-recapture cycle without destroying it, so that state expansion is obtained from free evolution instead of from applied potentials or photon recoil. The measured expansion after 0.25 ms is $\\xi_q = 189(18)$, i.e., $\\sigma_q = 110(15)$ nm along the vertical axis, following the predicted growth $\\xi_q^2 \\approx 1 + \\Omega_y^2\\tau^2 + \\frac{2}{3}\\Gamma\\Omega_y^2\\tau^3$; the acceleration data $a_y = -10.9 \\pm 0.4$ m/s$^2$ confirm that only gravity acts during the fall. The authors also establish a quantitative recapture-loss criterion: the recaptured energy is minimized when the tweezer displacement equals $g\\tau^2/2$, and their fit yields a trap depth $U_0 = 5.4(2) \\times 10^5\\,k_B T_0$. With ground-state cooling and a cryogenic ultra-high-vacuum environment, they calculate that the same scheme could reach $\\tau \\approx 1.9$ ms, an expansion of roughly 1680, and a coherence length $\\ell \\approx 4.7$ nm, more than two orders of magnitude beyond previous demonstrations.","pith_inferences":["The recycled-particle capability suggests using the same protocol as a repeated, single-mass gravity probe: the fitted relation between trap displacement and fall time gives a per-particle measurement of $g$, a use the paper does not develop.","The $\\tau = 0.25$ ms deviation from the model is attributed to sub-linear transduction; measuring at lower detection power or at a different wavelength would isolate that readout nonlinearity from a genuine expansion anomaly.","Extending the sequence to a second release after recapture would turn the setup into a pulsed free-fall sequence with a matter-wave-like geometry; the paper does not propose this, but the 300 ns switching time makes it plausible."],"forward_implications":["Each free-fall realization recycles the same nanoparticle, so thousands of state-expansion cycles can be accumulated without preparing indistinguishable copies.","Free evolution under gravity is genuinely free of electric-field and photon-recoil perturbations; the dominant remaining decoherence channel in the current setup is background gas.","With ground-state cooling and cryogenic ultra-high vacuum, the protocol projects to $\\tau \\approx 1.9$ ms, $\\xi_q \\approx 1680$, and a coherence length of 4.7 nm.","The condition $d = g\\tau^2/2$ for minimizing recapture energy gives a practical recipe for extending free-fall duration while keeping particle loss low.","Because the particle is charge-neutral, no electrode shielding or ion-trap infrastructure is needed in the free-fall region, simplifying the path to larger delocalizations."],"supporting_citations":[{"why":"Supplies the benchmark: an earlier demonstration of a twofold coherence-length increase that this protocol projects to exceed by two orders of magnitude.","marker":"[19]"},{"why":"Supplies the second recent two-fold expansion result, together forming the baseline the paper compares against.","marker":"[20]"},{"why":"Provides the parametric feedback cooling method used to initialize the nanoparticle's cold thermal state.","marker":"[25]"},{"why":"Provides the discharge protocol that makes the nanoparticle charge-neutral, eliminating electric-field coupling during free fall.","marker":"[27]"},{"why":"Provides the optimal linear-filter formalism used to justify backward bandpass filtering for state estimation.","marker":"[29]"},{"why":"Demonstrates linear-filter state estimation on levitated nanoparticles, supporting the filtering procedure used here.","marker":"[30]"},{"why":"Provides the calibration procedure that converts quadrant-photodetector voltages into displacement units and extracts particle mass and radius.","marker":"[31]"},{"why":"Supplies the optical cold-damping technique the paper proposes for improving initial cooling and thereby extending free-fall time.","marker":"[33]"},{"why":"Supplies the cryogenic ultra-high-vacuum decoherence analysis used to project millisecond free-fall operation.","marker":"[34]"}],"fun_headline_variants":["Neutral nanoparticle free falls, returns with 190x spread","Gravity-only fall widens levitated bead's spread 190x","0.25 ms free fall, 190-fold position spread for levitated particle","Charge-neutral bead free-falls, uncertainty jumps 190x at recapture"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The headline expansion number assumes the backward-filtered detector traces faithfully reconstruct the particle's true position and momentum spread, even though the readout is only about 0.5% efficient and the trap frequency shifts by up to 5 kHz after recapture; if that reconstruction is biased, the 190-fold growth is not real.","fun_headline_variants_meta":{"raw":{"variants":["Neutral nanoparticle free falls, returns with 190x spread","Gravity-only fall widens levitated bead's spread 190x","0.25 ms free fall, 190-fold position spread for levitated particle","Charge-neutral bead free-falls, uncertainty jumps 190x at recapture"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000597,"raw_usage":{"total_tokens":2810,"prompt_tokens":977,"completion_tokens":1833,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":593,"completion_tokens_details":{"reasoning_tokens":1753}},"tokens_in":593,"tokens_out":1833,"duration_ms":14164,"temperature":1.0,"reasoning_tokens":1753,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T16:34:28.827545+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Re-analyze the $\\tau = 0.25$ ms single-shot traces with bandpass-filter bandwidths of 1, 2, and 4 kHz and with the filter carrier shifted by $\\pm 5$ kHz, and check whether the recovered position spread stays at 110(15) nm; a systematic drift with filter settings would mean the expansion is an estimation artifact rather than real free-evolution growth.","supporting_citations":[{"cited_title":"Hebestreit, M","cited_arxiv_id":null,"evidence_quote":"Supplies the cryogenic ultra-high-vacuum decoherence analysis used to project millisecond free-fall operation."},{"cited_title":"Nanomechanical State Amplifier Based on Optical Inverted Pendulum","cited_arxiv_id":"2403.04302","evidence_quote":"Supplies the benchmark: an earlier demonstration of a twofold coherence-length increase that this protocol projects to exceed by two orders of magnitude."},{"cited_title":"State Expansion of a Levitated Nanoparticle in a Dark Harmonic Potential","cited_arxiv_id":"2312.13111","evidence_quote":"Supplies the second recent two-fold expansion result, together forming the baseline the paper compares against."},{"cited_title":"Kamba and K","cited_arxiv_id":null,"evidence_quote":"Provides the parametric feedback cooling method used to initialize the nanoparticle's cold thermal state."},{"cited_title":"Millen, T","cited_arxiv_id":null,"evidence_quote":"Provides the discharge protocol that makes the nanoparticle charge-neutral, eliminating electric-field coupling during free fall."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the optimal linear-filter formalism used to justify backward bandpass filtering for state estimation."},{"cited_title":"Frimmer, K","cited_arxiv_id":null,"evidence_quote":"Demonstrates linear-filter state estimation on levitated nanoparticles, supporting the filtering procedure used here."},{"cited_title":"See supplemental material for the derivation of the model and of the recapture probability, for details on the data analysis and for an extended dataset","cited_arxiv_id":null,"evidence_quote":"Provides the calibration procedure that converts quadrant-photodetector voltages into displacement units and extracts particle mass and radius."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the optical cold-damping technique the paper proposes for improving initial cooling and thereby extending free-fall time."}],"review_version":1}