{"id":"a4fd2a93-16ae-4d02-86fb-b37abe5012f7","arxiv_id":"2508.02070","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"low","formal_verification":"none","parameter_count":1,"one_line_summary":"Multi-mode N00N states achieve Heisenberg scaling for distributed quantum sensing of a global phase average, demonstrated in a four-mode 2002 state experiment with 2.74 dB gain over the standard quantum limit.","lead":"This paper shows theoretically and experimentally that multi-mode N00N states can measure the average of several phases spread across distant nodes with Heisenberg scaling, beating the standard quantum limit. A four-mode, two-photon version measured the average of two phases with about 2.74 dB sensitivity gain, suggesting a path toward entanglement-enhanced sensor networks.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Experimental 2.74 dB gain is post-selected; theoretical CRB/QCRB saturation is sound.","rationale":"I agree with the reader that the paper merits conditional acceptance rather than full acceptance. My independent check of the theoretical core found no error: for ν = (1/d, ..., 1/d), the inverse of the QFIM in Eq. (4) yields the announced 1/N² bound, and the diagonal CFIM of the local BS+PNRD measurement in Eq. (5) reproduces the same scalar bound at the optimal phase point, so the saturation claim is not the bottleneck. The bottleneck is purely experimental: the 2.74 dB number is produced from the post-selected two-photon subensemble, and the paper's own limitation statement confirms this. Since the authors explicitly label the experiment as a proof of concept, this is an acknowledged limitation rather than an internal contradiction; however, the abstract and conclusion state the enhancement unconditionally, so the framing needs adjustment. A conditional verdict is therefore appropriate, and I would keep the reader's verdict unchanged while asking for (i) the post-selection caveat in the abstract and conclusion, and (ii) either an unconditional reanalysis or a loss-adjusted SQL comparison.","tokens_in":9867,"tokens_out":21921,"duration_ms":260824,"concrete_test":"Recompute the Fisher information from the raw detection records without restricting to two-photon events: build the empirical likelihood from all outcomes (zero, one, and two photons across the two nodes), evaluate the observed information at the same phase settings, and compare with both the ideal SQL (FI = 2) and the loss-degraded SQL based on the measured mean photon number per trial. If the unconditional FI does not exceed the ideal SQL of 2 (or, more fairly, the loss-adjusted SQL), then the 2.74 dB value should be reported explicitly as a post-selected proof-of-concept number, not as the sensitivity of the full protocol.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central experimental claim — the 2.74 dB enhancement over the SQL — rests on a post-selected likelihood. In the final paragraph of the experimental section the authors write: 'we used post-selection, which does not take into account the imperfections of the experiment; however, this does not affect the proof-of-concept of quantum-enhanced sensitivity.' That sentence correctly limits the result, but the abstract and conclusion restate the enhancement without the qualifier ('achieving a 2.74 dB sensitivity enhancement over the standard quantum limit'). Because the Fisher information of 3.76 and the MLE standard deviation are computed from the conditional two-photon probability set {P_1^{11}, P_1^{20}, P_1^{02}, P_2^{11}, P_2^{20}, P_2^{02}}, the claimed 2.74 dB is a property of the subensemble in which both photons are detected, not of the unconditional measurement process. This does not threaten the theoretical half of the paper, which I verified: Eq. (4) gives ν^T F_Q^{-1}ν = 1/N², and the local BS+PNRD CFIM in Eq. (5) saturates it at the optimal working point. The load-bearing weakness is thus confined to how the experimental result is framed: if 'achieves' is read as an unconditional sensitivity, the current data do not establish it.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a distributed quantum sensing scheme based on multi-mode N00N states, with the goal of estimating the average of d spatially distributed phases. Theoretically, it derives the quantum Fisher information matrix for these states (Eq. (4)), obtains the quantum Cramér-Rao bound 1/N^2 for the average phase, and shows that local measurements consisting of a 50/50 beam splitter and photon-number-resolving detectors achieve the same bound through the classical Fisher information matrix (Eq. (5)). It compares this with separable N00N states, which give d/N^2, and with MePe states, which also give 1/N^2. Experimentally, the authors generate a four-mode 2002 state, distribute it over two nodes, estimate the average of two phases using maximum likelihood estimation, and report a Fisher information of 3.76 versus the standard quantum limit value of 2, corresponding to a 2.74 dB enhancement. The experimental section explicitly states that the result is post-selected on two-photon detection events and is presented as a proof of concept.","tokens_in":10041,"tokens_out":9411,"duration_ms":97709,"significance":"If the results hold, the paper establishes that multi-mode N00N states are a viable discrete-variable resource for distributed quantum sensing with Heisenberg scaling, complementing continuous-variable and spin-squeezed approaches. The theoretical analysis is sound: the QFIM in Eq. (4) and the resulting QCRB 1/N^2 check out under the stated phase-encoding convention, and the CFIM in Eq. (5) is consistent with saturation of the bound at the optimal working point. The experimental demonstration, while post-selected, is a useful proof of concept and includes measured visibilities, bootstrapped error bars, and a clear comparison with the SQL and HS. The main weakness is the unqualified restatement of the post-selected 2.74 dB enhancement in the abstract and conclusion, even though the body correctly limits the claim.","major_comments":[{"comment":"The reported Fisher information of 3.76 and the 2.74 dB enhancement are computed from the post-selected two-photon probability set {P_1^{11}, P_1^{20}, P_1^{02}, P_2^{11}, P_2^{20}, P_2^{02}}, as the authors explicitly state in the final paragraph of the experimental section. The body correctly limits the claim to a proof of concept, but the Abstract and the Conclusion restate the enhancement without this qualifier, e.g., 'achieving a 2.74 dB sensitivity enhancement over the standard quantum limit.' Since the unconditional measurement includes loss and other imperfections not captured by the post-selected conditional probabilities, the headline claim overstates what the data establish. Please reword the Abstract and Conclusion to state that the enhancement is conditional on two-photon detection events, or provide an unconditional sensitivity estimate that accounts for the discarded events.","section":"Experimental results, final paragraph; Abstract; Conclusion"}],"minor_comments":[{"comment":"The text reads '∆ϕSQL = 1/√µN and ∆ϕHS = 1/√µN'; as printed the two bounds are identical. The Heisenberg-limited standard deviation should be 1/(√µ N), not 1/√(µN). Please correct this typo, otherwise the HS curve in Fig. 3(c) is mislabeled.","section":"Experimental section, Fig. 3(c) discussion"},{"comment":"The main text states that the local BS+PNRD measurement saturates the QCRB for the multi-mode N00N state, but the calculation is delegated entirely to the Supplemental Material. A brief statement in the main text of the conditional probabilities or the optimal working point would make this central claim easier to verify and more self-contained.","section":"Eq. (5) and local-measurement saturation"},{"comment":"There are several reference formatting typos: Ref. [5] should read 'AVS Quantum Sci.' rather than 'A VS Quantum Sci.', and Ref. [6] contains a stray 'J.' before 'Integrable atomtronic interferometry.' Please check the reference list.","section":"References"},{"comment":"The notation |ψ⟩_j for the j-th node is used in the caption and Eq. (2), but the normalization convention is not immediately clear until one compares with Eq. (2). Aligning the notation between the figure caption and the equation would improve readability.","section":"Fig. 1 caption"}],"recommendation":"minor_revision","confidential_remarks":null},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's my read on arXiv:2508.02070. The theoretical half is sound and the experiment is a clean proof-of-principle, but the headline 2.74 dB sensitivity gain is post-selected, and the abstract and conclusion don't carry that qualifier. That's the one thing to keep in mind.\n\nWhat's actually new: the experimental demonstration of distributed quantum sensing with a four-mode 2002 state estimating the average of two phases. That specific experiment hasn't appeared before. The theoretical result — Heisenberg scaling 1/N² for multi-mode N00N states with ν=(1/d,...,1/d) — is a direct corollary of the QFIM already computed in the group's earlier papers [32,33]. That's not a flaw, but it means the novelty is modest. The paper does a good job comparing with separable N00N states and MePe states, and it verifies that local beam-splitter plus PNRD measurements saturate the QCRB. I checked the algebra: Eq. (4) with the stated phase encoding gives ν^T F_Q^{-1} ν = 1/N², and the CFIM in Eq. (5) gives the same at the optimal working point. So the central theoretical claim holds up.\n\nThe soft spots are about framing. The experimental Fisher information of 3.76 and the 2.74 dB gain are computed from the conditional two-photon probability set. Losses and other imperfections are discarded. The authors do say this in the final paragraph of the experimental section, but the abstract and conclusion restate the gain without the qualifier. If \"achieves\" is read as an unconditional sensitivity, the data don't support it. This is a framing issue, not a mathematical flaw. Also, the SQL and HS formulas in the text are both printed as 1/√µN, which is obviously a typo; the HS should be 1/(√µ N). That needs a fix.\n\nThe audience is quantum metrology researchers, particularly those working on distributed sensing or multi-mode N00N states. The proof of local-measurement saturation lives in the Supplemental Material; I didn't verify it line-by-line but it's consistent. The scalability discussion is speculative but clearly flagged.\n\nOverall: this is a solid incremental paper. The theory is a corollary, but the experiment is a genuine first for distributed sensing with multi-mode N00N states. A serious referee should engage with it. I'd accept it after a minor revision that tones down the unconditional language and fixes the typo. I wouldn't cite it in my own work in the next year — the post-selection and N=2 limit make it more of a proof-of-concept — but it's worth a reading group discussion.\n\nRecommendation: send to peer review.","headline":"Sound theory, clean but post-selected N=2 proof-of-principle; abstract and conclusion oversell the unconditional gain.","tokens_in":10673,"tokens_out":4593,"would_cite":false,"duration_ms":41539,"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":"Multi-mode N00N states can estimate distributed phase averages at the Heisenberg 1/N² limit, with a two-phase photonic demonstration beating the SQL by 2.74 dB.","keywords":["distributed quantum sensing","multi-mode N00N states","Heisenberg scaling","quantum Cramér-Rao bound","multiple phase estimation","photon-number-resolving detection","quantum metrology","sensor networks"],"falsifier":"To test the unconditional claim, repeat the two-node experiment without post-selection, including losses, vacuum, and all other detection outcomes when computing the Fisher information of the estimated average phase; if the variance then fails to beat the standard quantum limit (1/N for N=2) or fails to approach 1/$N^{2}$ scaling as N grows, the central claim as stated is refuted.","tokens_in":9580,"feed_emoji":"🎯","tokens_out":10050,"duration_ms":99155,"temperature":0.7,"pith_summary":"This paper claims that a single multi-mode $N00N$ state -- one entangled probe in which all $N$ photons are coherently superposed across the interferometric modes of $d$ spatially separated nodes -- can estimate the average of $d$ unknown phases with Heisenberg scaling, meaning the estimation variance shrinks as $1/N^2$ with total photon number $N$. The authors establish this at the level of both the quantum Cramér-Rao bound, the fundamental limit set by quantum mechanics, and the classical Cramér-Rao bound for a realistic local measurement consisting of a 50/50 beam splitter and photon-number-resolving detection at each node. They also show this sensitivity matches the best known bound for mode-and-particle-entangled states and beats separable $N00N$ states, whose variance degrades to $d/N^2$. To support the theory, they generate a four-mode $2002$ state, distribute it over two nodes, and estimate the average of two phases with a 2.74 dB sensitivity enhancement over the standard quantum limit, using post-selected two-photon events as a proof of concept. If correct, this offers a practical route to entanglement-enhanced sensor networks because the required measurements are local.","feed_headline":"Multi-mode N00N states reach Heisenberg limit in distributed sensing","feed_subtitle":"A four-mode 2002 state measured with local photon counting beats the standard quantum limit by 2.74 dB.","key_machinery":"The load-bearing object is the multi-mode $N00N$ state, a coherent superposition in which all $N$ photons sit entirely in one mode of one node, in either arm of the local interferometer, with every other mode empty and equal amplitude across all $d$ nodes; the four-mode $2002$ state is its $N=2, d=2$ instance. Its quantum Fisher information matrix has a uniform negative off-diagonal structure, $-N^2/d^2$, which cancels when contracted with the equal-weight vector $\\nu = (1/d,\\dots,1/d)$, yielding the Heisenberg bound $1/N^2$. The second mechanism is the local measurement of a $2\\times2$ beam splitter plus photon-number-resolving detectors, which diagonalizes the classical Fisher information matrix into entries $N^2/d$, so the classical Cramér-Rao bound reproduces the quantum bound; the saturation proof is given in the Supplemental Material. Experimentally, the state is generated from a Bell state via Hong-Ou-Mandel interference, phase-encoded with wave plates at each node, and read out with fiber beam splitters and superconducting nanowire single-photon detectors.","core_discovery":"The central discovery is that the multi-mode $N00N$ state $|\\Psi_{MN}\\rangle = (1/\\sqrt{d})\\sum_{j=1}^{d} (1/\\sqrt{2})(|N0\\rangle_j + |0N\\rangle_j)\\otimes_{k\\neq j}|00\\rangle_k$, when used to estimate the equal-weight average $\\phi = (1/d)\\sum_j \\phi_j$ of $d$ distributed phases, yields a quantum Fisher information matrix whose contraction with the weight vector gives exactly $\\nu^T F_Q^{-1} \\nu = 1/N^2$. The QFIM has diagonal entries $(2d-1)N^2/d^2$ and off-diagonal entries $-N^2/d^2$, and their cancellation along the equal-weight direction produces the Heisenberg bound. The paper further shows that a local measurement -- a $2\\times2$ beam splitter and photon-number-resolving detection at each node -- gives a diagonal classical Fisher information matrix with entries $N^2/d$, so the classical Cramér-Rao bound also saturates at $1/N^2$, meaning no joint measurement across nodes is required to reach the quantum limit. Experimentally, the authors use the four-mode $2002$ state (two photons, four modes) to estimate $(\\phi_1+\\phi_2)/2$ and report a Fisher information of 3.76, beyond the SQL value of 2, and a 2.74 dB improvement in the standard deviation over the SQL, with the estimated phase variance tracking the Heisenberg limit.","pith_inferences":["A quantitative loss-threshold analysis, which the post-selected experiment leaves open, would determine whether the $1/N^2$ scaling survives realistic channel losses; loss models for multi-mode $N00N$ states suggest the advantage persists only below a per-mode loss rate that shrinks as $N$ grows.","Because the off-diagonal Fisher information cancels only along the equal-weight direction, estimating weighted averages with non-uniform coefficients will generally yield variance larger than $1/N^2$; a test with unequal weights would map the boundary of the scheme.","The comparison with separable states implies a design rule: concentrating all $N$ photons in one multi-mode entangled state is strictly better than splitting them into $d$ local $N/d$-photon $N00N$ states, which matters for photon-budget-limited sensor networks.","Since the Heisenberg bound does not depend on $d$, an experiment with three or four nodes would directly test whether the per-photon precision survives in larger arrays, a scaling the paper motivates but does not demonstrate."],"forward_implications":["Distributed sensor networks can estimate the average of $d$ unknown phases at the Heisenberg limit using only local measurements, because the classical Cramér-Rao bound reaches $1/N^2$ with beam splitters and photon-number-resolving detection.","The multi-mode $N00N$ probe matches the known best sensitivity of mode-and-particle-entangled states and strictly outperforms separable $N00N$ states, whose variance $d/N^2$ falls below the SQL when $d>N$.","The demonstrated four-mode $2002$ state provides a proof-of-concept 2.74 dB enhancement over the SQL for the average of two phases, with the estimated standard deviation tracking the Heisenberg limit.","The scheme extends to more than two nodes and to cases where the number of photons is smaller than the number of phases, because the Heisenberg bound $1/N^2$ is independent of $d$."],"supporting_citations":[{"why":"introduced generalized multi-mode N00N states for multiple-parameter estimation, supplying the probe-state family this paper adopts.","marker":"[31]"},{"why":"demonstrated experimental generation of multi-mode N00N states via Bell states and Hong-Ou-Mandel interference, the method adopted for the four-mode 2002 state.","marker":"[32]"},{"why":"supplies the quantum Fisher information matrix and sensitivity-bound formalism, including the MePe-state best bound 1/N^2 used for comparison.","marker":"[30]"},{"why":"established distributed quantum phase estimation with entangled photons and the 1/N^2 sensitivity benchmark this paper compares against.","marker":"[25]"},{"why":"formulated multiparameter estimation in networked quantum sensors, defining the linear global function nu^T phi and the distributed sensing scenario.","marker":"[18]"},{"why":"extended distributed quantum sensing to multiple phases with fewer photons, the scaling target for higher-mode N00N states.","marker":"[27]"}],"fun_headline_variants":["Four-mode N00N state beats quantum limit in distributed sensing","Distributed quantum sensing hits Heisenberg limit with multi-mode N00N","Heisenberg scaling in distributed sensing via multi-mode N00N states","2.74 dB gain: multi-mode N00N in distributed phase sensing"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The experimental demonstration is post-selected, since only two-photon detection events enter the 2.74 dB gain, so the claimed advantage is not established unconditionally under loss; a second load-bearing premise is that the local beam-splitter plus photon-counting measurement saturates the quantum Cramér-Rao bound, a proof relegated to the Supplemental Material.","fun_headline_variants_meta":{"raw":{"variants":["Four-mode N00N state beats quantum limit in distributed sensing","Distributed quantum sensing hits Heisenberg limit with multi-mode N00N","Heisenberg scaling in distributed sensing via multi-mode N00N states","2.74 dB gain: multi-mode N00N in distributed phase sensing"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000709,"raw_usage":{"total_tokens":3264,"prompt_tokens":1086,"completion_tokens":2178,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":702,"completion_tokens_details":{"reasoning_tokens":2101}},"tokens_in":702,"tokens_out":2178,"duration_ms":16937,"temperature":1.0,"reasoning_tokens":2101,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T05:12:22.517985+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"To test the unconditional claim, repeat the two-node experiment without post-selection, including losses, vacuum, and all other detection outcomes when computing the Fisher information of the estimated average phase; if the variance then fails to beat the standard quantum limit (1/N for N=2) or fails to approach 1/$N^{2}$ scaling as N grows, the central claim as stated is refuted.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"introduced generalized multi-mode N00N states for multiple-parameter estimation, supplying the probe-state family this paper adopts."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"demonstrated experimental generation of multi-mode N00N states via Bell states and Hong-Ou-Mandel interference, the method adopted for the four-mode 2002 state."},{"cited_title":"Gessner, L","cited_arxiv_id":null,"evidence_quote":"supplies the quantum Fisher information matrix and sensitivity-bound formalism, including the MePe-state best bound 1/N^2 used for comparison."},{"cited_title":"Liu, Y.-Z","cited_arxiv_id":null,"evidence_quote":"established distributed quantum phase estimation with entangled photons and the 1/N^2 sensitivity benchmark this paper compares against."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"formulated multiparameter estimation in networked quantum sensors, defining the linear global function nu^T phi and the distributed sensing scenario."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"extended distributed quantum sensing to multiple phases with fewer photons, the scaling target for higher-mode N00N states."}],"review_version":1}