{"id":"972e30c6-ff14-4eb1-94aa-2be856b2ce83","arxiv_id":"2411.18832","paper_version":3,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For zero-mean Gaussian states, path entanglement reduces the quantum Fisher information for phase shifts, and the optimal states are unentangled, properly ordered squeezed states.","lead":"Zero-mean Gaussian states, such as squeezed light, lose phase sensitivity when their spatial modes become entangled with each other. This clarifies that path entanglement is not a resource for phase estimation with these practical states, and keeping interferometer paths unentangled is optimal.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Equation (5) has a factor-of-two error: H = ||V1||^2 - 1 should be H = (||V1||^2 - 2)/2, making the main-text derivation of Eq. (7) inconsistent as written.","rationale":"The reader's flagged weakest assumption was that every pure zero-mean Gaussian state with a given squeezing spectrum is reachable from Vin by a passive transformation. That assumption is standard and correct via the Bloch-Messiah decomposition, so it is not a genuine vulnerability. The actual load-bearing problem is an internal algebraic inconsistency in the main text: Eq. (5) has a factor-of-two error that breaks the derivation of Eq. (7), the key trade-off formula. While Eq. (7) is independently correct (it matches known single-mode QFI and the determinant-purity relation), the paper's written derivation cannot be followed to that conclusion. Because the central claim rests on Eq. (7) and the proof of Theorem 1 is relegated to the Supplementary Information, this error must be resolved before the main text is reliable. The proper verdict is to accept only after the authors correct Eq. (5) (and confirm that the SI uses the correct factor). This is a specific, verifiable issue rather than a general worry about the orbit of Gaussian states.","tokens_in":12262,"tokens_out":30685,"duration_ms":257881,"concrete_test":"Evaluate Eq. (4) and Eq. (5) for a single-mode squeezed vacuum with squeezing r. Eq. (4) gives H = (e^{-4r} + e^{4r} - 2)/2; Eq. (5) as printed gives H = e^{-4r} + e^{4r} - 1. For r = 0, these yield 0 and 1 respectively, a direct contradiction. Also substitute Eq. (6) into Eq. (5) and compare with Eq. (7): the printed version produces H = 16n1^2 + 16n1 + 3 - 2/µ^2, while the corrected form H = (||V1||_F^2 - 2)/2 reproduces H = 8n1^2 + 8n1 - (1/µ^2 - 1). This analytical check settles whether the derivation is merely misprinted or conceptually wrong.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"From the stated QFI formula Eq. (4), for a single phase-shifted mode (G = diag(1,1,0,...,0)), the trace gives H = (Tr(V1^2) - 2)/2, i.e. H = ||V1||_F^2/2 - 1. The paper instead writes H = ||V1||_F^2 - 1 in Eq. (5). The difference is not cosmetic: for vacuum (V1 = I), Eq. (4) gives H = 0, while Eq. (5) gives H = 1. Substituting the paper's own Eq. (6), ||V1||_F^2 = Tr(V1)^2 - 2 det(V1), into the printed Eq. (5) yields H = 16n1^2 + 16n1 + 3 - 2/µ^2, whereas the central result Eq. (7) is H = 8n1^2 + 8n1 - (1/µ^2 - 1). The correct factor reproduces Eq. (7) exactly. Thus the main-text derivation of the trade-off is not self-consistent; a reader following Eqs. (5)-(6) cannot reach Eq. (7). This is a concrete flaw in the argument as presented, even if the final formula happens to be correct. If the Supplementary proof of Theorem 1 relies on the same mis-scaled expression, the optimality claim could inherit the error; if not, Eq. (5) still needs correction before the main-text derivation is valid.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript studies zero-mean Gaussian states of N bosonic modes and asks how path entanglement affects the quantum Fisher information (QFI) for phase-shift estimation. Starting from the formula for the QFI of a pure zero-mean Gaussian state under a phase rotation, it derives a single-mode trade-off H = 8n_1^2 + 8n_1 - (1/µ^2 - 1), where µ is the purity of the reduced state of the phase-shifted mode; a two-mode decomposition showing that differential-phase sensitivity is degraded by entanglement while common-phase sensitivity is not; and an N-mode optimization theorem stating that optimal QFI states are exactly the decoupled, properly ordered squeezed states with H = 2 Σ_a g_a^2 sinh^2(2r_a). The argument is analytic and parameter-free, and the main results are stated as explicit formulas that can be checked independently.","tokens_in":12597,"tokens_out":10538,"duration_ms":95629,"significance":"If correct, the paper gives a clean and somewhat counterintuitive result: for zero-mean Gaussian states, path entanglement is not a resource for phase sensitivity and in fact degrades it. The two-mode formulas are directly relevant to squeezed-light interferometry, and the explicit characterization of optimal states is useful for experiment design. The strengths include the absence of fitted parameters, the use of QFI as an estimator-independent figure of merit, and the explicit discussion of the limits of the result, including displaced states and frequency-dependent settings. The main weaknesses are a normalization error in Eq. (5), a missing factor of 2 in Eq. (15), and the fact that the proofs of Proposition 1 and Theorem 1 are deferred to the Supplementary Information, which is not present in the manuscript text.","major_comments":[{"comment":"Equation (5) is off by a factor of 2. From Eq. (4) with G = diag(1,1,0,...,0), one obtains H = (Tr(V_1^2) - 2)/2 = ||V_1||_F^2/2 - 1, not ||V_1||_F^2 - 1. Combining Eq. (6) with the printed Eq. (5) gives H = 16n_1^2 + 16n_1 + 3 - 2/µ^2, which contradicts the central result Eq. (7); the corrected normalization reproduces Eq. (7) exactly. This is not a cosmetic issue, because a reader following Eqs. (5)-(6) cannot derive Eq. (7). Please correct Eq. (5), adjust Eqs. (8)-(9) accordingly, and verify that the Supplementary Information proof of Theorem 1 uses the same corrected normalization.","section":"Trade-off in the Single Phase Shift Scenario, Eq. (5)"},{"comment":"The bound in Eq. (15) is missing the factor 2 from Eq. (14). Since Theorem 1 gives H = 2 Σ_a g_a^2 sinh^2(2r_a), constraining r_a ≤ r yields H ≤ 2 ||G||^2 sinh^2(2r), not H ≤ ||G||^2 sinh^2(2r). The printed inequality is false already for N = 1, g_1 = 1, where the exact value is H = 2 sinh^2(2r). Please correct this bound and revisit the comparison with the Schatten-norm bound of ref. [54] with consistent constants.","section":"Decoupled Squeezed States Optimize the QFI, Eq. (15)"},{"comment":"The proofs of Proposition 1 and Theorem 1 are deferred entirely to the Supplementary Information, which is not available in the manuscript text. Since Theorem 1 is the central optimality claim, the submitted package should include the Supplementary Information, and the proof should be checked against the corrected QFI normalization. In addition, the main text states above Eq. (3) that every pure zero-mean Gaussian state can be generated from V_in by a passive transformation; this orbit assumption underlies the optimization in Theorem 1 and should be justified either in the main text or in the Supplementary Information.","section":"Decoupled Squeezed States Optimize the QFI, Proposition 1 and Theorem 1"}],"minor_comments":[{"comment":"The text says indices i,j,k range from 0 to 2N and indices a,b range from 0 to N, but the canonical operators are labelled q_1,...,q_2N and the modes are 1,...,N; the range should start at 1.","section":"Notation after Eq. (1)"},{"comment":"After correcting Eq. (5), the notation ||V_1||^2 should be defined consistently as the squared Frobenius norm of the 2x2 block, since the text currently introduces ||·||_2 as 'the Frobenius norm squared' but Eq. (8) still displays the uncorrected expression.","section":"Eq. (5) and Eq. (8), notation for the Frobenius norm"},{"comment":"The derivation of Eq. (11) is algebraically clear, but it would help to state explicitly that the equality of the purities obtained by tracing either mode follows from the Schmidt decomposition argument already cited in the text; this is presently only asserted in a parenthetical.","section":"Eq. (11), presentation of the two-mode formula"}],"recommendation":"major_revision","confidential_remarks":"The factor errors in Eq. (5) and Eq. (15) are likely typographical, but they occur in load-bearing formulas, so a careful revision is needed before acceptance. I do not see grounds for rejection: with the corrected normalization the main formulas are internally consistent, and the central claims are defensible if the Supplementary Information proofs are valid."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The core result is worth knowing: for zero-mean Gaussian states, path entanglement degrades phase-shift QFI, and the optimal states are decoupled, properly ordered squeezed states. That is a clean, non-obvious statement, and the comparison to the earlier specific-state results (refs. [41,42]) is fair. The intuition about rotating the squeezed ellipse out of the phase-shift plane is helpful, and the common-phase exception is a nice symmetry observation. The paper is also careful about the zero-mean limitation and the frequency-independent formalism.\n\nBut there is a concrete problem in the main text. Eq. (4) gives H = Tr((VG)^2 - G^2)/2. For a single phase-shifted mode, that reduces to H = (||V1||^2 - 2)/2, not ||V1||^2 - 1 as printed in Eq. (5). The stress-test note is correct: the printed Eq. (5) gives H = 1 for vacuum, which is impossible. If you substitute Eq. (6) into the printed Eq. (5), you get a different expression than Eq. (7). The correct factor reproduces Eq. (7) exactly. So the final trade-off formula is right, but the derivation as written in the main text is inconsistent. That is a fixable typo-level error with real consequence: a reader cannot follow Eqs. (5)-(7) as printed. The authors need to correct Eq. (5) and double-check that the supplementary uses the same scaling.\n\nThe other soft spot is that the proofs of Proposition 1 and Theorem 1 live in the Supplementary, which was not available in the extracted text. The optimality claim depends on the claim that every pure zero-mean Gaussian state with the given squeezing spectrum lies on the K V_in K^{-1} orbit. That is standard for pure Gaussian states, but I could not verify it from the main text alone. The reader's concern there is legitimate.\n\nNone of this undermines the central insight, and I found no sign of fitting or circularity. The result is new, the math is mostly transparent, and the connection to gravitational-wave interferometry gives it practical relevance. This paper deserves a serious referee; the main things to check are the corrected Eq. (5) and the supplementary proofs. I would not desk-reject it, but I would insist on the fix before publication.","headline":"Solid and interesting trade-off result, but the main text has a factor-of-two error in Eq. (5) that makes the derivation as printed inconsistent; the final formula survives.","tokens_in":13100,"tokens_out":1643,"would_cite":true,"duration_ms":14380,"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":"Entangling the paths of a squeezed-light interferometer reduces its phase sensitivity, and the best sensitivity comes from unentangled squeezed states.","keywords":["quantum Fisher information","Gaussian states","squeezed states","path entanglement","phase estimation","interferometry","quantum metrology","entanglement trade-off"],"falsifier":"Directly measure the differential-phase QFI of a family of two-mode zero-mean Gaussian states with fixed per-mode squeezing but increasing entanglement (e.g., by passing two squeezed vacua through a variable beam splitter). Equation (12) predicts $H = (g_1^2 + g_2^2) H_{\\rm sqz}(n) - (g_1-g_2)^2(1/\\mu^2 - 1)$; any entangled state with differential QFI above the unentangled value $2 H_{\\rm sqz}(n)$ would refute the trade-off. Alternatively, numerical search over the symplectic-orthogonal group $K$ for an $N$-mode state with $H > 2\\sum_a g_a^2 \\sinh^2(2r_a)$ at fixed $r_a$ would falsify Theorem 1.","tokens_in":12088,"feed_emoji":"🔬","tokens_out":9191,"duration_ms":68184,"temperature":0.7,"pith_summary":"This paper tackles the common assumption that entanglement always helps quantum measurements. It shows the opposite for a broad and experimentally relevant class: squeezed-vacuum (zero-mean Gaussian) states used for phase estimation. For these states, path entanglement strictly reduces the quantum Fisher information (QFI): the single-mode phase-shift QFI is $H = H_{\\rm sqz}(n_1) - (1/\\mu^2 - 1)$, where $\\mu$ is the purity of the reduced mode, and the entanglement penalty $1/\\mu^2 - 1$ grows with entanglement. The same penalty hits two-mode differential-phase measurements hardest and vanishes for common-phase measurements. The paper concludes that optimal $N$-mode sensitivity is reached by decoupled, properly ordered squeezed states, with $H = 2\\sum_a g_a^2 \\sinh^2(2r_a)$.","feed_headline":"Path entanglement degrades squeezed-light phase sensitivity","feed_subtitle":"In zero-mean Gaussian interferometers, unentangled properly ordered squeezed states give the best phase precision.","key_machinery":"The load-bearing identity is the QFI of a pure zero-mean Gaussian state under a passive phase-shift unitary, $H = {\\rm Tr}\\big((V G)^2 - G^2\\big)/2$, with $G = {\\rm diag}(g_1,g_1,\\dots,g_N,g_N)$, derived from symplectic properties. The argument converts this into an explicit function of entanglement through the purity–determinant relation ${\\rm det}(V_1) = 1/\\mu^2$ for the reduced 2×2 covariance block of the phase-shifted mode, where $\\mu$ is the purity. The full $N$-mode optimality proof relies on the passivity structure $V = K V_{\\rm in} K^{-1}$ with symplectic-orthogonal $K$, together with Proposition 1, which says $V$ is decoupled (no cross-mode covariances) if and only if the state is a tensor product of the input squeezed states in some order and orientation.","core_discovery":"The central claim is that for any pure zero-mean Gaussian state obtained by a passive (photon-number-preserving) transformation $K$ acting on a product of squeezed states with covariance $V_{\\rm in}$, the QFI for a phase shift on a single mode is exactly $H = H_{\\rm sqz}(n_1) - (1/\\mu^2 - 1)$, where $H_{\\rm sqz}(n) = 8n^2 + 8n$ is the QFI of an unentangled squeezed state of mean photon number $n$. Because the reduced-state purity $\\mu$ decreases monotonically with entanglement entropy, this identity imposes a strict trade-off: entanglement between the phase-shifted mode and the rest of the system lowers the QFI below the unentangled optimum, and the factor-of-two bounds $H_{\\rm sqz}(n_1)/2 \\le H \\le H_{\\rm sqz}(n_1)$ bracket all pure states. In the two-mode case the paper derives $H = (g_1^2 + g_2^2)H_{\\rm sqz}(n) - (g_1 - g_2)^2(1/\\mu^2 - 1)$ for equally squeezed inputs, showing that differential phase sensitivity is destroyed by maximal entanglement (EPR states give $H = (g_1+g_2)^2 H_{\\rm sqz}(n)/2$), while common-phase QFI $H_{\\rm com} = {\\rm Tr}(V^2)/2 - 2$ is invariant under $K$. The paper's Theorem 1 then characterizes all optimal $N$-mode states: they are exactly the states reachable from decoupled, properly ordered squeezed states by passive transformations acting only within groups of modes sharing the same phase-shift coefficient $g_a$, and their QFI is $H = 2\\sum_a g_a^2 \\sinh^2(2r_a)$.","pith_inferences":["The paper's zero-mean restriction is crucial: for displaced inputs, passive networks can siphon displacement into one mode and make the QFI scale with $N$, so the per-mode squeezing bound would not apply. A natural next test is whether the same purity–QFI trade-off survives in the presence of photon loss.","The trade-off suggests a design rule for continuous-variable quantum sensors: place any entangling operation after the phase-sensing region rather than before it, so that entanglement is used at the measurement step rather than degrading the signal.","The identity $H = H_{\\rm sqz}(n_1) - (1/\\mu^2 - 1)$ is a candidate for a general resource-theoretic relation between QFI and a single-mode entanglement monotone; checking whether non-Gaussian states obey a similar bound with the same purity term would clarify its scope.","Because the formalism is frequency-independent, the paper's conclusions may need modification for continuous-wave light with optical cavities; whether the trade-off persists in a frequency-dependent treatment is an open question flagged by the authors themselves."],"forward_implications":["Interferometers using squeezed vacuum should avoid path-entangling operations before the phase shift: differential-phase QFI decreases linearly with the entanglement monotone $1/\\mu^2 - 1$, and maximally entangled EPR inputs are completely insensitive to differential phase.","The best possible $N$-mode phase sensitivity with fixed per-mode squeezing is $H = 2\\sum_a g_a^2 \\sinh^2(2r_a)$, achieved by decoupled, properly ordered squeezed states; no passive network can exceed it.","Common-phase measurements are immune to the trade-off, since the common-phase unitary commutes with every passive transformation $K$.","The trade-off extends beyond phase shifts to any parametrized passive symplectic transformation, as shown in the paper's Supplementary Information section IV.","With at most $r$ per-mode squeezing, the QFI is bounded by $H \\le \\|G\\|^2 \\sinh^2(2r)$, a per-mode alternative to total-photon-number bounds."],"supporting_citations":[{"why":"Supplies the single-mode squeezed-state QFI $H_{\\rm sqz}(n) = 8n^2 + 8n$ used as the unentangled baseline in Eqs. (7), (11), and (16).","marker":"[38]"},{"why":"Supplies the purity–determinant identity ${\\rm det}(V_1) = 1/\\mu^2$ that converts QFI into a function of entanglement.","marker":"[44]"},{"why":"Supplies the covariance-matrix description of Gaussian states and the passive-transformation form $V = K V_{\\rm in} K^{-1}$.","marker":"[25]"},{"why":"Supplies the QFI formalism and additivity property used to compute common-phase QFI and the separable optimum.","marker":"[36]"},{"why":"Establishes that zero-mean Gaussian states achieve optimal differential-phase sensitivity among initially unentangled inputs, motivating the trade-off study.","marker":"[40]"},{"why":"Together with [46], shows every passive transformation is a sequence of beam splitters and phase shifts, supporting the reachability claim of Theorem 1.","marker":"[45]"},{"why":"Shows any passive transformation decomposes into beam splitters and phase shifts, used to justify the orbit of $K$.","marker":"[46]"}],"fun_headline_variants":["Path entanglement costs phase precision in squeezed states","Unentangled squeezed beams beat entangled for phase sensing","Trade-off: less entanglement, sharper interferometer sensitivity","Entanglement hurts phase sensitivity in Gaussian interferometers","Best phase precision comes from unentangled squeezed paths"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The optimization assumes that every pure zero-mean Gaussian state with a fixed squeezing spectrum can be produced from a product of squeezed states by some passive transformation (beam splitters and phase shifts); if any such state lies outside this orbit, the optimality theorem would miss it.","fun_headline_variants_meta":{"raw":{"variants":["Path entanglement costs phase precision in squeezed states","Unentangled squeezed beams beat entangled for phase sensing","Trade-off: less entanglement, sharper interferometer sensitivity","Entanglement hurts phase sensitivity in Gaussian interferometers","Best phase precision comes from unentangled squeezed paths"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000159,"raw_usage":{"total_tokens":1274,"prompt_tokens":1036,"completion_tokens":238,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":652,"completion_tokens_details":{"reasoning_tokens":165}},"tokens_in":652,"tokens_out":238,"duration_ms":3166,"temperature":1.0,"reasoning_tokens":165,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T10:52:07.639805+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Directly measure the differential-phase QFI of a family of two-mode zero-mean Gaussian states with fixed per-mode squeezing but increasing entanglement (e.g., by passing two squeezed vacua through a variable beam splitter). Equation (12) predicts $H = (g_1^2 + g_2^2) H_{\\rm sqz}(n) - (g_1-g_2)^2(1/\\mu^2 - 1)$; any entangled state with differential QFI above the unentangled value $2 H_{\\rm sqz}(n)$ would refute the trade-off. Alternatively, numerical search over the symplectic-orthogonal group $K$ for an $N$-mode state with $H > 2\\sum_a g_a^2 \\sinh^2(2r_a)$ at fixed $r_a$ would falsify Theorem 1.","supporting_citations":[{"cited_title":"Giovannetti, S","cited_arxiv_id":null,"evidence_quote":"Supplies the single-mode squeezed-state QFI $H_{\\rm sqz}(n) = 8n^2 + 8n$ used as the unentangled baseline in Eqs. (7), (11), and (16)."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the covariance-matrix description of Gaussian states and the passive-transformation form $V = K V_{\\rm in} K^{-1}$."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the QFI formalism and additivity property used to compute common-phase QFI and the separable optimum."},{"cited_title":"Serafini, F","cited_arxiv_id":null,"evidence_quote":"Together with [46], shows every passive transformation is a sequence of beam splitters and phase shifts, supporting the reachability claim of Theorem 1."}],"review_version":1}