{"id":"db91362c-5207-4587-8923-439d749fb091","arxiv_id":"2511.23416","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"A boundary time crystal used as a light source yields a quantum Fisher information rate scaling as N^4 in the number of emitters, and a replica-crystal decoder gives phase-estimation error scaling as N^{-1.222}, exceeding the Heisenberg limit.","lead":"This paper proposes using the light emitted by a boundary time crystal, a driven collection of atoms that keeps oscillating, as a probe for measuring tiny optical phase shifts. If the predicted scaling holds, it points toward a new class of light sources for precision interferometry.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The N^{-1.222} beyond-Heisenberg protocol exponent is fit from three points at fixed, non-optimized Δφ; the N-dependence of the optimal working point may be the real source of the apparent scaling.","rationale":"The reader's verdict is CONDITIONAL, with the weakest assumption identified as the three-point power-law fit and the ideal lossless assumption. I agree that the finite-N fitting is the key vulnerability, but I would sharpen it: the fixed-Δφ choice is more dangerous than the mere number of points. Fig. 4(c) explicitly shows that the optimal Δφ moves with N, so a fixed working point cannot separate the true sensitivity scaling from the drift of the optimum. The ideal-lossless assumption is less load-bearing for the theoretical claim, since the QFI bound is derived under ideal monitoring and the authors explicitly list finite detection efficiency as future work; it would not invalidate the proof-of-principle. The analytical QFI result Eq. (4) is supported by a parameter-free superspin derivation in the ω/ω_c→∞ limit and by numerical benchmarks, so the fundamental N^4 scaling is credible. The protocol's beyond-Heisenberg scaling, however, needs the larger-N/optimized-Δφ test before it can be accepted as an asymptotic resource. This leaves the verdict CONDITIONAL, unchanged from the reader.","tokens_in":24370,"tokens_out":6803,"duration_ms":72420,"concrete_test":"For N=30,50,80,120 at ω/ω_c=4, compute the large-deviation function θ_c(s,Δφ) from Eq. (S24) by sparse diagonalization of the tilted cascaded Lindbladian (dimension (N+1)^2, feasible to several hundred). For each N, sweep Δφ over, e.g., 10^{-4} to 0.1, evaluate δφ(N,Δφ)=sqrt(∂_s^2 θ_c)/|∂_φ∂_s θ_c| at s=0, locate the Δφ* that minimizes δφ, and fit δφ(N,Δφ*(N)) versus N. If the envelope exponent remains above 1, the protocol claim survives; if the exponent moves toward 1 or below, or tracks the fixed-Δφ points, the N^{-1.222} result is a preasymptotic artifact of the chosen phase offsets.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central protocol claim is δφ ∝ N^{-α} with α=1.222±0.018, inferred in Fig. 4(b) from N=6,10,20 at Δφ=0.005 (and α=1.04 at Δφ=0.01). The load-bearing premise is not merely that this power law persists; it is that fixing Δφ at these values measures the achievable sensitivity envelope. Fig. 4(c) shows that |∂_φ I_T| has a maximum whose location moves toward Δφ=0 as N increases. Therefore, for each N, the estimation error is minimized at a different Δφ*(N). The three plotted points use one fixed Δφ, so as N grows the working point drifts away from the optimum. The fitted exponent then mixes two effects: the true N-scaling of the optimal error and the N-scaling of the optimal phase offset. Without optimizing Δφ per N or providing an analytic argument for Δφ*(N), the reported α is not established as an asymptotic sensitivity exponent. A second, smaller issue is that the comparison to the QFI limit (green dotted line, ∝N^{-2}) uses the ω/ω_c→∞ value, while the numerics are at ω/ω_c=4, where the actual QFI is below f_{φ,∞}; the gap between the protocol and the fundamental bound at the working point is thus not quantified.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes using the output field of a boundary time crystal (BTC) as a resource for optical phase estimation. The authors derive the quantum Fisher information (QFI) rate for phase imprinted on the emitted light, finding f_φ ≈ κ N(N+2)[(N−1)(N+3)/135 + 2/3] ∝ N^4 in the deep time-crystal limit (Eq. 4), which exceeds the N^2 Heisenberg scaling of the stationary regime. They then analyze two measurement protocols: an average homodyne current protocol, which saturates the QCRB only in the stationary regime, and a 'perfect absorber' protocol in which the phase-shifted light is injected into a replica BTC acting as a decoder. For the perfect absorber protocol in the time-crystal regime, they report an estimation error scaling δφ ∝ N^{−1.222} at Δφ=0.005 (and N^{−1.04} at Δφ=0.01), claiming beyond-Heisenberg sensitivity. The QFI derivation is supported by analytic superspin calculations and exact numerics; the protocol claim rests on power-law fits through three small-system points.","tokens_in":24720,"tokens_out":3810,"duration_ms":39278,"significance":"If fully established, the N^4 QFI rate for a driven-dissipative light source would be a notable result: it shows that the temporal correlations of a time-crystal output can be a metrological resource beyond what N independent emitters provide. The paper is careful in benchmarking the analytic QFI expression against exact diagonalization (Figs. S1–S3), and it makes code and data publicly available. However, the advertised 'beyond Heisenberg' protocol scaling is currently supported only by a three-point numerical fit at hand-selected phase offsets, without an analytic explanation of the N-dependence. The QFI contribution itself is sound and valuable, but the protocol claim needs substantially stronger evidence before the title-level conclusion can be accepted.","major_comments":[{"comment":"The claimed beyond-Heisenberg exponent α=1.222±0.018 is obtained by fitting a power law through N=6,10,20 at a fixed working point Δφ=0.005 (α=1.04±0.04 at Δφ=0.01). This cannot support an asymptotic scaling claim. Fig. 4(c) shows that |∂_φ I_T| has a maximum whose position moves toward Δφ=0 as N increases, so the optimal operating offset Δφ*(N) depends on N. The three plotted values are not evaluated at the per-N optimum; the fitted α therefore mixes the N-scaling of the minimal error with the N-scaling of the optimal offset. To establish δφ∝N^{−α}, the authors should either optimize Δφ for each N and show the optimized errors obey a power law, or supply an analytic argument for the N-dependence at fixed Δφ. Without this, the protocol claim is not established.","section":"Perfect absorber protocol, Fig. 4(b)"},{"comment":"The green dotted line shows f_{φ,∞}^{−1/2}∝N^{−2}, the ω/ω_c→∞ limit of Eq. (4), but the numerical protocol points are computed at ω/ω_c=4. At finite ω/ω_c the exact QFI rate is below f_{φ,∞}, so the plot overstates the proximity of the protocol to the fundamental bound. The text should quantify the finite-frequency QFI bound (e.g., by diagonalizing the tilted master equation at ω/ω_c=4) and quote the ratio δφ/(1/√f_φ) at the working point.","section":"Fig. 4(b), comparison to QFI"}],"minor_comments":[{"comment":"The expression writes f_φ with a 1/T factor inside, although f_φ is defined as a long-time rate. Please clarify that the integral is understood in the T→∞ limit, or write f_φ = lim_{T→∞} (1/T) [ ... ].","section":"Eq. (3)"},{"comment":"The cascaded critical frequency ω_{c,casc} is used extensively but defined only implicitly. State its definition explicitly in the main text.","section":"Fig. 3 caption / text after Eq. (8)"},{"comment":"The text says both protocols saturate the QCRB in the stationary regime, but for the homodyne protocol this is true only at φ−β=0 and for the perfect absorber only in the limit Δφ→0. Please state these conditions explicitly in the summary of results.","section":"Introduction / homodyne protocol"},{"comment":"The numerical points in Fig. 4(a) are shown without error bars or markers indicating the finite-size/trajectory uncertainty. Adding them would help assess the reliability of the apparent saturation.","section":"Fig. 4(a)"}],"recommendation":"major_revision","confidential_remarks":"The QFI part of the paper is careful and the N^4 scaling is benchmarked, but the headline 'beyond Heisenberg' protocol claim is not yet supported: three fixed-Δφ points are too fragile, and the N-dependence of the optimal working point is visible in the authors' own Fig. 4(c). I would like to see either an analytic derivation of the finite-Δφ scaling or an optimized-Δφ study with more system sizes before this claim is published as a definitive result. The paper has the right tools to do this within its scope."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things to know up front. First, the analytical result that the QFI rate for a BTC light source scales as N^4 in the deep time-crystal regime is solid and worth having; it is derived cleanly, checked against exact numerics, and the paper is honest about its approximations. Second, the headline \"beyond Heisenberg\" protocol scaling (δφ ∝ N^{-1.222}) is not established. It comes from power-law fits through three small system sizes at fixed phase offsets, and the optimal working point moves with N. That is a real weakness in the paper's central practical claim.\n\nThe new material: applying BTC output to an external optical phase (rather than estimating BTC parameters), the N^4 QFI rate, and the perfect absorber/replica decoder as a way to retrieve temporal-correlation information. The stationary-regime results, including the homodyne protocol and the HP calculation for the cascaded system, are consistent with numerics. The code/data are on GitHub. That is the credit.\n\nWhere it wobbles: the exponent α=1.222±0.018 is fit from N=6, 10, 20 at Δφ=0.005 (and α=1.04 at 0.01). Fig. 4(c) shows |∂_φ I_T| peaks at a Δφ that decreases with N. So the three points are taken at a working point that is not optimal for each N, and the fitted exponent conflates the true N-scaling with the drift of the optimum. Without either optimizing Δφ per N or an analytic argument for Δφ*(N), I would not take N^{-1.222} as an asymptotic sensitivity exponent. Also, the plot comparing the protocol to the QFI limit uses the ω/ω_c→∞ bound, while the data are at ω/ω_c=4; the actual gap is unquantified. And the abstract's phrase \"saturate the fundamental bound\" goes beyond what the paper shows: saturation is demonstrated in the stationary regime, not in the time-crystal regime.\n\nProportionate verdict: a solid core with an over-assertive extrapolation. The QFI result is a genuine contribution; the beyond-Heisenberg protocol claim needs per-N optimization, larger sizes, and a resource count in terms of emitted photons. I'd send it to a serious referee and, if it survives, cite it for the QFI derivation.","headline":"The N^4 QFI result is solid and worth citing; the beyond-Heisenberg protocol exponent is an extrapolation from three small-N points at fixed phase offsets, so treat it as provisional.","tokens_in":25203,"tokens_out":3674,"would_cite":true,"duration_ms":37679,"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":"This paper proposes using a boundary time crystal — a driven, dissipative collection of N two-level emitters — as a light source for optical phase estimation, and argues that in its time-crystal phase the temporal correlations of the emitte","keywords":["boundary time crystal","optical phase estimation","quantum Fisher information","Heisenberg limit","temporal correlations","perfect absorber","driven-dissipative systems","quantum metrology"],"falsifier":"Compute the perfect-absorber estimation error for N = 30, 40, and 50 at Δφ = 0.005 with ω/ω_c = 4 by exact diagonalization of the tilted cascaded master equation: if the fitted exponent α drops to 1 or below, the claimed N^{-1.222} scaling is a small-system artifact. Alternatively, add a finite photon-detection efficiency η < 1 and check whether the exponent α falls below 1.","tokens_in":24259,"feed_emoji":"💡","tokens_out":9094,"duration_ms":78186,"temperature":0.7,"pith_summary":"This paper tries to establish that the light emitted by a boundary time crystal is metrologically structured: in the time-crystal regime its two-time intensity correlations grow with the number of emitters N, and those correlations are a usable resource for estimating an optical phase. The fundamental limit is set by the quantum Fisher information rate, which the paper computes analytically as f_φ,∞ = κ N(N+2)[(N-1)(N+3)/135 + 2/3] ∝ N^4 — a stronger-than-Heisenberg scaling in system size while remaining standard-quantum-limited in measurement time (δφ ∝ 1/√T). To tap that resource, the paper constructs a detector that is a replica of the source — a perfect absorber for the phase-shifted light — and shows it can achieve phase-estimation errors scaling as N^{-1.222} at small phase offsets, surpassing the Heisenberg limit. A homodyne scheme saturates the bound in the stationary regime but fails to exploit the time-crystal correlations. A careful reader would care because this turns a driven many-body non-equilibrium state into a proposed light source for precision interferometry.","feed_headline":"Time-crystal light beats the Heisenberg limit","feed_subtitle":"Correlated emission yields N^4 Fisher info; a replica detector reaches N^-1.222 error scaling.","key_machinery":"The key object is the boundary time crystal (BTC): N two-level emitters with collective resonant driving at Rabi frequency ω and collective dissipation at rate κ, described by a master equation; for ω > Nκ/2 it enters a time-crystal phase with persistent oscillations. The resource is the two-time intensity correlation C(τ) of its output field, which the quantum Fisher information rate is essentially the time integral of. The matching mechanism is the perfect absorber protocol: a replica BTC with a tunable phase φ′ coupled unidirectionally to the source, whose dark state at Δφ=φ−φ′=0 makes the decoder an ideal detector for temporal correlations; the estimation error is extracted from the time","core_discovery":"The central claim is that, in the boundary time crystal's time-crystal phase, the quantum Fisher information rate for phase estimation is f_φ,∞ ≈ κ N(N+2)[(N-1)(N+3)/135 + 2/3], scaling as N^4 for large N and linear in measurement time T, computed via the superspin method in the strong-driving limit and benchmarked numerically. This exceeds the Heisenberg scaling f_φ ∝ N^2 found in the stationary regime by a bosonic large-displacement expansion. The information is carried by the emitted light's temporal correlations, specifically the two-time intensity correlation C(τ). The paper further claims that a detection protocol in which the phase-shifted output is unidirectionally guided into an ide","pith_inferences":["The asymptotic exponent α≈1.222 is inferred from exact points at N=6, 10, 20; extending the numerics to larger N or finding an analytic large-N expansion would test whether this is the true asymptotic scaling or a finite-size crossover.","The ideal lossless unidirectional cascade and unit-efficiency detection are load-bearing idealizations; the paper itself lists finite detection efficiency and local decay as open issues, and either will likely degrade the demonstrated exponent.","The replica-as-perfect-absorber idea suggests a general recipe for extracting multi-time correlation resources from other driven-dissipative light sources, and the paper's outlook proposes testing other time-crystal platforms with non-collective dissipation.","If the scaling survives practical losses, the boundary time crystal would supply quantum-enhanced phase sensitivity without needing externally prepared squeezed or entangled probe states."],"forward_implications":["In the time-crystal regime, the quantum Fisher information rate for phase estimation scales as N^4 and linearly with measurement time T, so the ultimate precision per unit time grows much faster with emitter number than the N^2 scaling of the stationary regime.","The perfect absorber protocol (a replica BTC as decoder) reaches phase-estimation errors δφ ∝ N^{-1.222} at Δφ=0.005 and ω/ω_c=4, a scaling that exceeds the Heisenberg limit.","The average-homodyne-current protocol saturates the quantum Fisher information bound in the stationary regime, giving Heisenberg scaling, but does not exploit the time-crystal correlations.","Both measurement protocols obey the standard quantum limit in time, δφ ∝ 1/√T, so the system-size enhancement is independent of integration time.","The QFI rate expression f_φ,∞ = κ N(N+2)[(N-1)(N+3)/135 + 2/3] is exact in the limit ω/ω_c → ∞ and is benchmarked numerically at finite ratios, so the N^4 scaling is a property of the model, not of a specific measurement."],"fun_headline_variants":["Time-crystal light hits N^4 sensing precision","Correlated light from time crystals beats Heisenberg limit","N^4 precision from time-crystal light sources","Time-crystal light: quantum sensing beyond Heisenberg","Collective time-crystal emission boosts phase estimation"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The beyond-Heisenberg scaling rests on fitting a power law through exact results at only N = 6, 10, and 20 at hand-chosen phase offsets, alongside the ideal assumption of a lossless unidirectional cascade and perfect detection efficiency; if finite-size corrections bend the curve or losses enter, the protocol may not exceed the Heisenberg limit in the asymptotic sense.","fun_headline_variants_meta":{"raw":{"variants":["Time-crystal light hits N^4 sensing precision","Correlated light from time crystals beats Heisenberg limit","N^4 precision from time-crystal light sources","Time-crystal light: quantum sensing beyond Heisenberg","Collective time-crystal emission boosts phase estimation"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000509,"raw_usage":{"total_tokens":2309,"prompt_tokens":734,"completion_tokens":1575,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":478,"completion_tokens_details":{"reasoning_tokens":1514}},"tokens_in":478,"tokens_out":1575,"duration_ms":10941,"temperature":1.0,"reasoning_tokens":1514,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-03T19:32:47.246104+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the perfect-absorber estimation error for N = 30, 40, and 50 at Δφ = 0.005 with ω/ω_c = 4 by exact diagonalization of the tilted cascaded master equation: if the fitted exponent α drops to 1 or below, the claimed N^{-1.222} scaling is a small-system artifact. Alternatively, add a finite photon-detection efficiency η < 1 and check whether the exponent α falls below 1.","supporting_citations":[],"review_version":1}