{"id":"32eb66dd-2963-48c3-b31b-a783ab2706bd","arxiv_id":"2505.12865","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Frequency-modulated trapping plus Kalman-filter feedback creates steady-state entanglement between two levitated nanoparticles with weak Coulomb coupling and low measurement efficiency.","lead":"This paper proposes combining periodic modulation of the trapping beams with optimal feedback control to entangle two levitated nanoparticles coupled by Coulomb interaction. The scheme can generate conditional and unconditional entanglement at low measurement efficiencies and modest coupling strengths.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Equation (11) for the unconditional excess-noise covariance has a sign error in the measurement-induced diffusion term; the reported unconditional entanglement may depend on this incorrect equation.","rationale":"The reader identified the independence of the two homodyne detectors as the weakest assumption. That is a legitimate experimental caveat, but it is a physical implementation assumption rather than an internal mathematical check. The sign inconsistency in Eq. (11) is more directly connected to the paper’s central numerical claim: the unconditional entanglement in Fig. 3 is obtained from V_u = V_c + V_ex, and Eq. (11) governs the excess-noise part. Since V_ex is defined as a covariance of conditional first moments, the measurement-induced Ito term must be positive; the printed negative sign would make V_ex tend toward a negative semidefinite steady state in the no-feedback limit, which is unphysical. If the numerics used the correct plus sign, the paper needs only a typographical correction and a statement of the convention used. If the numerics used the printed minus sign, the unconditional entanglement maps could be substantially altered or spurious. The absence of code makes this impossible to resolve from the manuscript alone, so the appropriate action is to require the authors to supply the corrected equation and a numerical cross-check. This is a conditional acceptance condition, not a rejection: the underlying scheme and the conditional-entanglement results may still be sound, and the error could be typographical.","tokens_in":11456,"tokens_out":17027,"duration_ms":191547,"concrete_test":"Independently re-derive Eq. (11) from Eq. (5a) by Ito’s lemma, or equivalently verify that for K = 0 the sum of Eqs. (5b) and (11) yields dV_u/dt = A V_u + V_u A^T + N. Then recompute the Fig. 3 steady-state V_u for the headline point (g/ω_m = 0.2, η = 0.5, α = 0.2, independent feedback) using both the printed '-' sign and the corrected '+' sign, checking positive semidefiniteness of V_u - V_c and the resulting unconditional logarithmic negativity. If the reported entanglement is reproduced only with '+', the manuscript must be corrected and the numerical results re-verified.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"From Eq. (5a), d⟨X⟩_c = A⟨X⟩_c dt + B u dt + 2V_c C dW. By Ito’s lemma, the excess-noise matrix V_ex = E[⟨X⟩_c⟨X⟩_c^T] receives a positive diffusion contribution +4V_c C C^T V_c. Since the conditional Riccati term in Eq. (5b) is -4V_c C C^T V_c, the V_ex equation must contain +4V_c C C^T V_c in order to reproduce the open-loop unconditional covariance dV_u/dt = A V_u + V_u A^T + N when K = 0. The printed Eq. (11) instead contains -4V_c C C^T V_c. If used literally, the steady-state V_ex can become negative semidefinite, violating V_u ≽ V_c and artificially lowering the logarithmic negativity computed from V_u. All unconditional-entanglement claims in Fig. 3 are computed from V_u, so this sign inconsistency is directly load-bearing. Because no simulation code is provided, it is impossible to tell whether the figures were produced with the correct plus sign or with Eq. (11) as written.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript proposes a continuous-measurement control scheme for entangling two optically levitated nanoparticles in free space. The trap frequencies are periodically modulated, the particles interact through a Coulomb potential, and two homodyne detectors monitor their positions. The authors use a Kalman filter to obtain conditional Gaussian states and an LQR/Bayesian feedback law to reduce excess noise, and they compute conditional and unconditional logarithmic negativity from the corresponding covariance matrices. They report that frequency modulation enhances both conditional and unconditional entanglement, allowing detection efficiencies as low as eta ~ 0.1-0.5 and Coulomb couplings |g|/omega_m ~ 0.1-0.2, and they explain the effect through unequal normal-mode squeezing induced by the Coulomb coupling. The parameters are taken from recent levitated-optomechanics experiments.","tokens_in":11670,"tokens_out":13952,"duration_ms":153573,"significance":"If the reported results survive scrutiny, the paper provides a concrete route toward unconditional entanglement of levitated particles with considerably relaxed experimental requirements compared with earlier Markovian-feedback proposals. The mechanism, namely modulation-induced squeezing of the two normal modes with a Coulomb-coupling-induced imbalance, is physically coherent and offers a testable prediction for the dependence of entanglement on g/omega_m, modulation depth, and detection efficiency. The paper applies standard Gaussian filtering and LQG control consistently, uses realistic parameters, and does not fit any free parameter to the target entanglement. No code or raw data is provided, so numerical reproducibility rests on the equations as written.","major_comments":[{"comment":"Eq. (11) as printed has a sign error in the measurement-induced diffusion term. From Eq. (5a) with M = A - B K_opt, the excess-noise covariance obeys dV_ex/dt = M V_ex + V_ex M^T + 4 V_c C C^T V_c by Ito's lemma, because the stochastic term 2 V_c C dW contributes a positive diffusion. The printed equation contains -4 V_c C C^T V_c. In the open-loop limit K_opt = 0, the sum of Eq. (5b) and Eq. (11) would not reproduce dV_u/dt = A V_u + V_u A^T + N, and V_u can lose positive semidefiniteness. Since the unconditional entanglement results in Fig. 3 are computed from V_u = V_c + V_ex, this is load-bearing. Please correct Eq. (11), state explicitly which sign was used in the numerics, and re-run or confirm all unconditional results.","section":"Sec. II, Eq. (11)"},{"comment":"The protocol assumes that the two homodyne detectors measure the back-scattered light of the two particles independently, with block-diagonal C and independent Wiener increments dW_k. If the scattered fields from the two traps overlap or a single detector collects light from both particles, the photocurrents become correlated, C acquires cross terms, and the Kalman-filter estimate and the resulting feedback are no longer those simulated. The paper should state this assumption explicitly in the experimental-implementation discussion and, ideally, provide a sensitivity estimate showing that small cross-detection does not destroy the entanglement.","section":"Sec. II, Eqs. (3)-(6)"},{"comment":"The LQR controller minimizes the cooling cost in Eq. (8); it is optimal for the chosen P and Q, not for entanglement. The abstract and introduction use the terms 'optimal control scheme' and 'optimal Bayesian feedback,' which are justified only with respect to that cost. If the authors intend to claim that the protocol is optimal for generating entanglement, an optimization over P and Q or a direct derivation of an entanglement-maximizing controller is needed. As written, the term 'optimal' should be qualified.","section":"Sec. II, Eq. (8)"}],"minor_comments":[{"comment":"The term 'B u' in Eq. (5a) should read 'B u dt' to keep the equation dimensionally consistent.","section":"Sec. II, Eq. (5a)"},{"comment":"The photocurrent notation I_k(t) = sqrt(eta K_ba) <x_k>_c dt + dW_k(t) is dimensionally awkward; writing I_k dt = sqrt(eta K_ba) <x_k>_c dt + dW_k would be clearer.","section":"Eq. (3)"},{"comment":"The caption of Fig. 3 says 'Time-averaged Conditional entanglement', but the text and the panels describe unconditional entanglement; please correct the caption.","section":"Fig. 3 caption"},{"comment":"The terminal or boundary condition for the backward Riccati equation is not specified, and for a periodically modulated A(t) the method used to obtain the steady-state periodic K_opt should be described.","section":"Sec. II, Eq. (9)"},{"comment":"The effective Hamiltonian in Eq. (16) is derived under the assumptions alpha << 1 and g << omega_m, while the simulations use alpha = 0.2; the text should note that Eq. (16) is only a qualitative explanation and that the numerics solve the full time-dependent equations.","section":"Sec. III, Eq. (16)"}],"recommendation":"major_revision","confidential_remarks":"The Eq. (11) sign is very likely a typographical slip that can be fixed by replacing '-' with '+' in the last term. My recommendation is not a rejection: the physical mechanism and parameter claims are plausible, and the central conditional-entanglement results in Figs. 2, 4, and 5 do not rely on Eq. (11). However, the unconditional claims hinge on the corrected equation, so I ask the authors to confirm the numerics explicitly. I would also encourage the authors to provide the simulation code or raw data for Fig. 3 to settle the issue unambiguously."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Let me cut to the point. The conditional-entanglement part of this paper is solid and worth reading; the unconditional part, which is the headline, likely rests on a sign error.\n\nWhat is good: the model is standard LQG for two Coulomb-coupled levitated particles with parametric modulation of the trap frequency. The physical mechanism—modulation squeezes both normal modes, the Coulomb interaction makes the squeezing unequal, and the mismatch produces entanglement—is clearly laid out and supported by the effective Hamiltonian. The parameters are anchored to recent experiments (Magrini, Tebbenjohanns, Delić), and the conditional results in Figs. 2, 4, 5 are computed from the conditional covariance alone, so they do not depend on the problematic equation. The claim that conditional entanglement survives at η≈0.1 and |g|≈0.2 ω_m with α=0.2 is plausible.\n\nThe problem: Eq. (11) for the excess-noise covariance has the wrong sign on the measurement-induced diffusion term. Ito applied to Eq. (5a) gives +4V_c C C^T V_c in the dV_ex equation. The printed equation has −4V_c C C^T V_c. With the minus sign, V_u = V_c+V_ex is not guaranteed positive semidefinite, and the inequality V_u ≽ V_c is violated. All unconditional entanglement numbers in Fig. 3 are computed from V_u. No code is provided, so I cannot tell whether the figures used the corrected plus sign or the sign as printed. If they used the printed sign, the reported unconditional entanglement is very likely an artifact. If they used the correct sign, the paper needs to say so explicitly, because the equation as written is wrong. Either way, the central claim is currently unsupported by the equations as presented.\n\nTwo more soft spots, both minor by comparison. The LQR minimizes a cooling cost (Eq. 8 with P=diag(ω_m,...)), not an entanglement measure; calling the controller 'optimal' for entanglement is an overstatement, though it may still be a good heuristic. And the note added admits that the latest version of Ref. [58] added trap-frequency modulation for this same problem, which weakens the novelty claim to 'independent treatment with more detail.'\n\nWho should read this: people active in levitated optomechanics and macroscopic entanglement. The conditional entanglement results are a useful data point, and the sign error is instructive as a cautionary tale. But I would not cite the unconditional claims in the current form. A serious referee should engage—the error is fixable and the core physics is plausible—but the paper needs a corrected Eq. (11), a rerun of the simulations, and a statement about what actually was computed.","headline":"Solid conditional-entanglement work undercut by a sign error in the excess-noise equation that the headline unconditional claims depend on.","tokens_in":12195,"tokens_out":4095,"would_cite":false,"duration_ms":41693,"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":"This paper proposes that periodically modulating the trapping beam, together with Kalman-filter Bayesian feedback on continuous homodyne measurements, generates steady entanglement between two levitated nanoparticles with weak Coulomb…","keywords":["quantum entanglement","levitated nanoparticles","optomechanics","Bayesian feedback","Kalman filter","parametric modulation","Coulomb coupling","continuous measurement"],"falsifier":"Measure the unconditional covariance matrix of two charged levitated particles with $g/\\omega_m \\approx 0.2$, trap modulation at $\\Omega \\approx 2\\omega_m$, independent homodyne detection near $\\eta \\approx 0.5$, and LQR feedback: if the logarithmic negativity is not positive in steady state, the central claim is refuted. A cheaper numerical check is to simulate the same protocol with correlated Wiener noise between the two detection channels; if the entanglement vanishes, the independence assumption is doing the work.","tokens_in":11242,"feed_emoji":"⚛️","tokens_out":6563,"duration_ms":59544,"temperature":0.7,"pith_summary":"The paper proposes a scheme to entangle two optically levitated nanoparticles in free space by combining periodic modulation of the trapping-beam amplitude with continuous homodyne detection and optimal Bayesian feedback. The central claim is that this combination yields both conditional and unconditional steady-state entanglement at Coulomb coupling strengths and detection efficiencies well below those required by earlier feedback-only proposals. The physical mechanism is that the modulation squeezes both the common and differential normal modes of the two-particle system, while the Coulomb interaction makes the two modes squeeze by different amounts; that squeezing imbalance is the resource that entangles the two particles. The paper supports the claim with simulations of logarithmic negativity and an effective-Hamiltonian analysis showing two parametric resonances.","feed_headline":"Trap modulation entangles two levitated particles in free space","feed_subtitle":"Trap shaking plus Bayesian feedback entangles two levitated particles at weak coupling and ~50% detection efficiency.","key_machinery":"The central objects are the two normal modes defined by the beam-splitter transformation, plus the Kalman-filter/LQR feedback loop. After a rotating-wave approximation, the effective Hamiltonian reads $H_{\\rm eff} = \\Delta_+ \\hat{c}_+^\\dagger \\hat{c}_+ + \\frac{\\omega_m\\alpha}{4}(\\hat{c}_+\\hat{c}_+ + \\hat{c}_+^\\dagger\\hat{c}_+^\\dagger) + \\Delta_- \\hat{c}_-^\\dagger \\hat{c}_- + \\frac{\\omega_m\\alpha}{4}(\\hat{c}_-\\hat{c}_- + \\hat{c}_-^\\dagger \\hat{c}_-^\\dagger)$, with detunings $\\Delta_+ = \\omega_m + \\omega_m\\alpha^2/4 - \\Omega/2$ and $\\Delta_- = \\Delta_+ + 2g$. This Hamiltonian has two parametric resonances, at $\\Omega \\simeq 2\\omega_m$ and $\\Omega \\simeq 2\\omega_m + 4g$, where the two normal modes are strongly squeezed. The Coulomb interaction shifts one resonance relative to the other, producing a difference in the magnitude or direction of the two modes' squeezing; the homodyne measurement stabilizes the parametric dynamics, and the feedback suppresses the excess noise that would otherwise destroy the entanglement.","core_discovery":"On the paper's own terms, the discovery is that a periodically modulated optical trap, combined with continuous homodyne measurement and a Kalman-filter/linear-quadratic-regulator feedback loop, can entangle two Coulomb-coupled levitated nanoparticles in steady state without strong coupling or near-unit detection efficiency. The unconditional entanglement survives for coupling strengths as low as $|g|/\\omega_m \\approx 0.1$ and detection efficiencies around $\\eta \\approx 0.5$, while conditional entanglement appears already near $\\eta \\approx 0.1$, in both repulsive and attractive interaction regimes. The modulation induces strong squeezing in both normal modes, and the Coulomb coupling shifts the differential-mode resonance so that the two modes' squeezing ellipses differ; this difference, not the Coulomb interaction itself, is what generates the entanglement. Parameters in the simulations match recent levitated-particle experiments, including trap frequency 29.6 kHz, room temperature and atmospheric pressure.","pith_inferences":["Inference: A natural testable extension is to feed correlated noise into the two homodyne channels; if entanglement survives cross-talk between the scattered fields, the scheme is robust, whereas if it disappears, the independence of the two readouts is the truly load-bearing requirement.","Inference: The same squeezing-imbalance mechanism could in principle entangle more than two particles in a modulated common trap, as long as individual position readout remains possible, though the paper does not address multi-particle arrays.","Inference: The predicted double-‘V' structure of conditional entanglement versus modulation frequency and coupling strength provides a sharp experimental fingerprint: measuring that map would directly confirm the mechanism before attempting the full feedback protocol."],"forward_implications":["Unconditional entanglement becomes compatible with coupling strengths an order of magnitude below the mechanical frequency and detection efficiencies near 50%, both within reach of current levitated-particle platforms.","Strong conditional entanglement ($\\langle E_N^c\\rangle > \\ln 2$) can be reached by increasing the modulation depth, which the paper argues is difficult in conventional optomechanical schemes because of stability constraints.","The feedback can be implemented either with identical forces (requiring unequal charges on the two particles) or with independent forces (allowing equal charges), so the scheme adapts to different trap geometries.","Because the entanglement oscillates at the modulation period, any protocol that uses the generated entanglement must be synchronized with the modulation cycle."],"supporting_citations":[{"why":"Real-time optimal quantum control and cooling of a levitated nanoparticle via measurement and Bayesian feedback; supplies the experimental platform and feedback technique the scheme is based on.","marker":"[33]"},{"why":"Steady-state entanglement of two levitated nanoparticles under strong Coulomb coupling and Markovian feedback; the baseline whose high coupling requirement the scheme relaxes.","marker":"[35]"},{"why":"Optimal feedback control achieving steady-state entanglement of interacting masses in free space; the precursor this paper extends by adding trap modulation.","marker":"[36]"},{"why":"Genuine EPR steering in a continuously monitored cavity magnomechanical system; provides the stochastic-master-equation formalism for conditional covariance.","marker":"[41]"},{"why":"Feedback control of quantum systems using continuous state estimation; underpins the linear-quadratic-Gaussian control derivation.","marker":"[44]"},{"why":"Opto- and electromechanical entanglement improved by modulation; source of the parametric-modulation resonance idea used here.","marker":"[54]"},{"why":"Theory for cavity cooling of levitated nanoparticles via coherent scattering; supplies the back-action decoherence rate and trap parameters used in the simulations.","marker":"[57]"}],"fun_headline_variants":["Trap modulation plus Bayesian feedback entangles levitated nanoparticles","Shaken trap and feedback entangle nanoparticles at weak coupling","Entangling levitated particles: modulation and feedback beat strong coupling","Squeezing mismatch yields entanglement via trap modulation and feedback","Bayesian feedback entangles levitated particles even with weak Coulomb coupling"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The scheme assumes the two homodyne detectors read out the two particles' positions independently, with independent noise; if the back-scattered fields from the two particles interfere or cannot be separated, the Kalman filter's state estimate is corrupted and the simulated entanglement is not produced.","fun_headline_variants_meta":{"raw":{"variants":["Trap modulation plus Bayesian feedback entangles levitated nanoparticles","Shaken trap and feedback entangle nanoparticles at weak coupling","Entangling levitated particles: modulation and feedback beat strong coupling","Squeezing mismatch yields entanglement via trap modulation and feedback","Bayesian feedback entangles levitated particles even with weak Coulomb coupling"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000503,"raw_usage":{"total_tokens":2448,"prompt_tokens":925,"completion_tokens":1523,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":541,"completion_tokens_details":{"reasoning_tokens":1439}},"tokens_in":541,"tokens_out":1523,"duration_ms":12077,"temperature":1.0,"reasoning_tokens":1439,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T20:25:28.228794+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the unconditional covariance matrix of two charged levitated particles with $g/\\omega_m \\approx 0.2$, trap modulation at $\\Omega \\approx 2\\omega_m$, independent homodyne detection near $\\eta \\approx 0.5$, and LQR feedback: if the logarithmic negativity is not positive in steady state, the central claim is refuted. A cheaper numerical check is to simulate the same protocol with correlated Wiener noise between the two detection channels; if the entanglement vanishes, the independence assumption is doing the work.","supporting_citations":[{"cited_title":"Rudolph, U","cited_arxiv_id":null,"evidence_quote":"Steady-state entanglement of two levitated nanoparticles under strong Coulomb coupling and Markovian feedback; the baseline whose high coupling requirement the scheme relaxes."},{"cited_title":"Tan, Genuine photon-magnon-phonon einstein- podolsky-rosen steerable nonlocality in a continuously- monitored cavity magnomechanical system, Phys","cited_arxiv_id":null,"evidence_quote":"Genuine EPR steering in a continuously monitored cavity magnomechanical system; provides the stochastic-master-equation formalism for conditional covariance."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Feedback control of quantum systems using continuous state estimation; underpins the linear-quadratic-Gaussian control derivation."},{"cited_title":"Mari and J","cited_arxiv_id":null,"evidence_quote":"Opto- and electromechanical entanglement improved by modulation; source of the parametric-modulation resonance idea used here."},{"cited_title":"Gonzalez-Ballestero, P","cited_arxiv_id":null,"evidence_quote":"Theory for cavity cooling of levitated nanoparticles via coherent scattering; supplies the back-action decoherence rate and trap parameters used in the simulations."}],"review_version":1}