{"id":"f4fdb51a-82a8-48ff-8398-f05c476bb03e","arxiv_id":"2507.04246","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"For dispersive quantum thermometry with N00N states, the exact quantum Fisher information scales at best linearly with the number of atoms, and N00N photon number gives no Heisenberg advantage.","lead":"This paper computes exact precision bounds for measuring temperature with a dispersive atom-field probe, showing that entangled N00N light states give at best a standard quantum limit, not the Heisenberg limit claimed in an earlier work. The authors also run the scheme on IBM's Brisbane quantum processor and in Qiskit simulations, including atomic ensembles prepared at effective negative temperatures.","discovery_kind":"first_principles","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Joint-state QFI not computed: no-Heisenberg conclusion is derived only for the bright-port reduced state, so the abstract's 'entire system' claim is unsupported.","rationale":"After checking the derivation, I find the reduced-state calculation (Eqs. 17–20) algebraically consistent: tracing out mode b and qubits from Eq. A10 and evaluating QFI for the diagonal ρ_a yields Eq. 20, and the Taylor expansion in Eq. 28 is plausible for small M. The experimental part uses standard Qiskit circuits and the N=1 data match the depolarized model with pγ=0.95, though pγ is not independently calibrated. The load-bearing gap is the mismatch between the proven statement and the claimed statement. The paper proves no-Heisenberg scaling for the bright-port reduced state only. The abstract and intro claim this holds for the joint atom-field evolution, citing analysis 'both for the entire system and when measurements are restricted to the light degree of freedom only' (Sec. I), but the full-system QFI is never computed. Because QFI is monotone under partial trace, the single-port bound does not constrain the full-state QFI; the discarded mode b and qubit correlations could increase the QFI, potentially restoring Heisenberg scaling. This is not an algebraic error but an unsupported generalization. A concrete numerical check on the full-state QFI for small M would settle it. I therefore see no reason to change the reader's CONDITIONAL verdict; the concern is real but remediable.","tokens_in":22097,"tokens_out":6794,"duration_ms":72603,"concrete_test":"Compute the QFI of the full joint state ρ_abT(t) = U(t)(ρ_NOON⊗ρ_qubits)U†(t) for M=1 (or a few small M) as an exact function of N and T, using spectral decomposition of the 2^M-qubit Hilbert space. Alternatively, compute the QFI of the two-mode field state Tr_qubits[ρ_abT(t)] before the second beam splitter. Plot Q_full versus N (e.g., N=1,2,3,4) at fixed t, χ, ε, T. If Q_full grows as N^2, the central 'entire system' claim fails and the paper must be revised to claim only single-port SQL. If Q_full remains O(N) (or saturates), the joint-state assertion survives. Also verify whether the M^2 term in Eq. 28 is absent for small M by direct evaluation.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim (abstract; Sec. I) is that the joint atom-field evolution achieves at best the standard quantum limit in N. However, the derivation computes the QFI only of the reduced single-mode state ρ_a(t)=Tr_{\\a}[Uρ0U†] (Eq. 11), yielding Eq. 20 after tracing out mode b and all M qubits. No QFI for the full joint state ρ_abT(t) (Eq. A10) or for the two-mode field state is ever evaluated. Since QFI is monotone under partial trace, Q(ρ_full) ≥ Q(ρ_a), so the single-port result cannot bound the joint-state precision. Correlations retained in the full state—e.g., the off-diagonal field coherence weighted by ⟨e^{±iεNχt \\hat{M}_{qubits}}⟩—can carry additional temperature information; a joint measurement over both ports and/or the atoms could in principle yield QFI scaling differently with N (potentially N^2). Thus the abstract's 'without any assumptions or approximations' and 'joint atom-field evolution' overstate what is proven: the no-Heisenberg conclusion is established only for the scheme that discards mode b and the qubit register. Note also Sec. III.A claims QFI scales 'at best linearly with M' while Eq. 28 contains an M^2 term; the statement is only asymptotically true after the α^{2M} suppression, not a general bound.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper analyzes dispersive thermometry of M identical non-interacting two-level atoms prepared in a Gibbs state at temperature T, probed by the field mode of a nonlinear Mach-Zehnder interferometer initially in a N00N state. The authors derive a closed-form expression for the quantum Fisher information (QFI), Eq. (20), of the reduced single-mode state at the bright port after tracing out the second field mode and all qubits. They show that the field-excitation number N enters only through the effective parameter Neff = ε N χ t / 2, so increasing N cannot produce Heisenberg scaling; they contrast this with the phase-averaging assumption of Ref. [67]. They then implement the scheme with Qiskit circuits and IBM_brisbane hardware for N = 1, 2, 3 and for positive and effective negative temperatures, finding that the simulated QFI is consistent with Eq. (20) and with the standard quantum limit.","tokens_in":22407,"tokens_out":11108,"duration_ms":121529,"significance":"If correct, the central result is a useful correction to the thermometry literature: it provides an exact, closed-form counterexample to the Heisenberg-limit claim of Ref. [67] for this specific dispersive N00N thermometry setup. The algebraic derivation of Eq. (20) is explicit and self-contained, and the binomial and de Moivre steps appear sound. The proposed digital simulation is a practical way to benchmark the theoretical prediction, and the inclusion of effective negative temperatures adds breadth. However, the paper's framing overstates what is proven: the exact no-Heisenberg conclusion is established only for the single bright-port reduced state, not for the joint atom-field state or the two-mode field state, and the M-scaling claim made around Eqs. (20)-(28) is not a general linear bound. These gaps are fixable, and a short additional calculation would close the main one.","major_comments":[{"comment":"The paper claims to analyze the QFI 'for the entire system and when measurements are restricted to the light degree of freedom only' (Sec. I) and to show that the 'joint atom-field evolution' achieves at best the standard quantum limit. The derivation, however, computes the QFI only of the reduced state ρ_a(t) = Tr_a[U ρ_0 U^†] defined in Eq. (11), after tracing out mode b and all M qubits. Since the QFI is monotone under partial trace, Q(ρ_full) ≥ Q(ρ_a), so the no-Heisenberg bound for the single-port state cannot bound the precision of arbitrary joint measurements on the full state. The conclusion is therefore not established for the joint state as written. This gap can be closed: conditioned on a fixed qubit basis string with q excitations, the field state is independent of T (the phase e^{-i ε N χ t q} depends only on the fixed q), so the full joint state is a classical-quantum state whose QFI reduces to the classical Fisher information of the thermal distribution, Eq. (9), which is linear in M and independent of N. Adding this calculation, or restricting the abstract and Sec. I claims to the single-port scheme, is necessary.","section":"Abstract; Sec. I; Sec. III, Eqs. (11) and (20)"},{"comment":"The statement that the QFI 'scales, at best, linearly' with M is not supported by Eq. (28). That expansion contains a positive M² term, and for fixed α < 1 the factor M² α^{2M} in Eq. (20) is non-monotonic: it grows initially and then decays to zero as M → ∞. The correct statement from Eq. (20) is that Q → 0 for large M and that the maximum over M is achieved at finite M; it is not a linear upper bound. This affects the wording in Sec. III.A and in the caption of Fig. 7, where 'standard quantum limit' scaling is asserted without a precise asymptotic bound.","section":"Sec. III.A, Eqs. (20) and (28)"},{"comment":"The hardware experimental data in Fig. 6 are for N = 1, M = 1 only; the higher-N and higher-M points in Fig. 7 come from digital simulation with Qiskit, not from the hardware experiment. The conclusion in Sec. VI that the standard quantum limit is 'indeed achieved in the single-photon excitation subspace' is therefore not demonstrated by a scaling curve, since a single (N, M) point cannot distinguish the SQL from any other scaling. I suggest softening the experimental claims or presenting a genuine hardware scaling scan.","section":"Secs. IV and VI"}],"minor_comments":[{"comment":"The phrase 'without any assumptions or approximations' overstates the result: the derivation is exact only within the stated model (Gibbs-state sample, dispersive Hamiltonian of Eq. (10), no losses, no qubit-qubit interactions). Consider replacing it with 'exact within the model considered here'.","section":"Abstract and Sec. I"},{"comment":"The coefficients c1(T) and c2(T) in Eq. (28) also depend on Neff through α and θ, so the notation c_i(T) is misleading; the coefficients should be written as c_i(T, Neff).","section":"Sec. III.A, Eq. (28)"},{"comment":"The binomial probability displayed after Eq. (A20) appears to be mistyped: it should be p(q) = p_1^q (1 - p_1)^{M-q}, not p_1^q (1 - p_1)^q.","section":"Appendix A, Eq. (A20)"},{"comment":"There is a typo in the final paragraph of Sec. V: 'N00N sates' should be 'N00N states'.","section":"Sec. V"},{"comment":"Consider stating explicitly that the QFI maximized over Neff is independent of N up to periodicity, because N and t are redundant in Neff. This would prevent the reader from interpreting the overlapping curves in Fig. 7 as a numerical accident rather than a structural consequence of Eq. (20).","section":"Sec. III.A and Fig. 7"}],"recommendation":"major_revision","confidential_remarks":"The main theoretical result appears to be correct for the reduced single-mode state, and the comparison with Ref. [67] is a useful contribution. The missing joint-state QFI calculation is the key gap; the authors can likely close it with a short additional calculation, since the full state is a classical-quantum mixture whose QFI should reduce to the thermal Fisher information. The hardware content is thin (N = M = 1), so I would encourage either more hardware scaling data or a clear separation between the theoretical result and the simulated proof-of-principle."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The real contribution here is replacing the approximation exp(iNχt<M_qubits>) with the exact average <exp(iNχt M_qubits)> in the dispersive thermometry model. That gives a closed-form QFI, Eq. (20), for the single-mode bright-port state, and it does undercut the Heisenberg-limit claim of Stace's 2010 paper for this specific setup. I checked the derivation in Appendix B, including the binomial sums and the α=1 limit, and the math is consistent. The digital simulation and the IBM hardware demo provide a reasonable sanity check, though the fitted depolarizing probability p_gamma=0.95 is not independently calibrated, so the experimental agreement is qualitative at best.\n\nThe soft spot is exactly what the stress-test note flags. The abstract says the results apply to the \"joint atom-field evolution\" and the intro claims QFI for \"the entire system,\" but only the QFI of the reduced single-mode state is ever computed. Since QFI is monotone under partial trace, the joint-state QFI could in principle be larger; the no-Heisenberg conclusion is established only for the scheme that discards mode b and the qubit register. This is a scope overclaim, not an algebraic error, and it is fixable by rewriting the claims to say \"for the bright-port reduced state.\" There is also a minor wording issue in Sec. III.A: \"at best linearly with M\" is asymptotic, since Eq. (28) contains an M^2 term; the α^{2M} suppression argument is correct for fixed α, but the phrasing should be clearer that this is leading-order scaling.\n\nWho is this for? People working on quantum thermometry and dispersive measurements. As a correction to a specific published claim about N00N dispersive thermometry, it is a useful and citable result. After the claims are aligned with what is actually proven, this would be a solid paper.\n\nRecommendation: send it to peer review. The referee should ask the authors to restrict the abstract and intro to the bright-port reduced-state scheme, or to actually compute the joint-state QFI if they want to keep the broader claim.","headline":"A sound exact calculation that kills Heisenberg scaling for the bright-port reduced state, but the abstract's 'joint atom-field' claim is broader than what is proven.","tokens_in":22957,"tokens_out":1956,"would_cite":true,"duration_ms":22088,"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":"An exact calculation shows that dispersive atom-field thermometry with N00N states is limited to standard quantum limit precision, contradicting earlier Heisenberg-scaling claims.","keywords":["quantum thermometry","quantum Fisher information","dispersive atom-field interaction","N00N states","standard quantum limit","Heisenberg limit","Mach-Zehnder interferometer","digital quantum simulation"],"falsifier":"A concrete decisive check is to compute the exact quantum Fisher information of the full joint state after the dispersive interaction — field modes of both interferometer ports together with the $M$-atom thermal register — without tracing anything out. If in any regime (for example, near the best working point $N_\\mathrm{eff} \\approx \\pi/2$ at low temperature) that joint QFI grows faster than linearly in $N$ or $M$, the paper's general claim that the joint evolution is standard-quantum-limited is false; if it remains linear, the single-port result is confirmed and extended. A simpler comparison experiment is to measure photon statistics at the dark port and correlations with the atomic register and check whether the total classical Fisher information exceeds the bright-port QFI of Eq. (20).","tokens_in":21947,"feed_emoji":"🌡️","tokens_out":16714,"duration_ms":157667,"temperature":0.7,"pith_summary":"This paper tries to settle what precision is actually achievable when the temperature of a small sample is read out through a dispersive atom-light interaction. The authors claim that, for $M$ identical two-level atoms in a Gibbs state probed by a single field mode prepared in a N00N state, the exact quantum Fisher information is bounded at the standard quantum limit: the precision scales at best linearly with $M$, and the photon number $N$ only rescales the interaction time through $N_\\mathrm{eff} = \\varepsilon N\\chi t/2$. This directly contradicts an earlier claim of Heisenberg ($N^2$) scaling for the same setup, which the authors trace to an illegitimate replacement of the operator phase $\\langle e^{iN\\chi t \\hat{M}_\\mathrm{qubits}}\\rangle$ by $e^{iN\\chi t \\langle \\hat{M}_\\mathrm{qubits}\\rangle}$. If the paper is right, entangled field states such as N00N states are not a useful resource for this kind of thermometry, and the practical path to better precision is more atoms, not more photons in the probe. The authors also implement the scheme as a digital quantum simulation and report agreement with the standard quantum limit for $N = 1, 2, 3$, including for atomic ensembles at effective negative temperatures.","feed_headline":"N00N states can't beat the standard limit in dispersive thermometry","feed_subtitle":"An earlier N-squared claim relied on a phase approximation; the exact bound is linear in the atom number.","key_machinery":"The load-bearing object is the exact probability $p_0(T) = \\langle \\cos^2(N\\chi t\\,\\hat M_\\mathrm{qubits}/2)\\rangle$ for the bright port, evaluated over the thermal mixture. Writing the cosine in complex exponentials and using the binomial expansion turns $p_0$ into the real part of a power of $1 + e^{-\\varepsilon/T + 2iN_\\mathrm{eff}}$, which De Moivre's formula converts into $p_0 = \\frac{1}{2}(1 + \\alpha^M\\cos(M\\theta))$ with $\\alpha \\le 1$ and an $M$-independent angle $\\theta$. The two-outcome form of $\\rho_a$ then gives the Fisher information in closed form via $Q(T) = p_0'(T)^2/[p_0(T)(1-p_0(T))]$. The factor $\\alpha^{2M}$, the modulus of the thermal characteristic function raised to the $M$-th power, is what suppresses superlinear scaling, and the identity $\\langle e^{iN\\chi t \\hat M_\\mathrm{qubits}}\\rangle \\ne e^{iN\\chi t \\langle \\hat M_\\mathrm{qubits}\\rangle}$ is what separates this calculation from the earlier Heisenberg-scaling one.","core_discovery":"On the paper's own terms, the discovery is a closed-form expression for the quantum Fisher information of the temperature, computed without any approximation of the dynamics. For the dispersive Hamiltonian $H = \\varepsilon\\chi\\,\\hat a^\\dagger \\hat a \\sum_{k=1}^M \\hat\\sigma_+^{(k)}\\hat\\sigma_-^{(k)}$ and a N00N input $|\\mathrm{N00N}\\rangle = (|N,0\\rangle + |0,N\\rangle)/\\sqrt{2}$, the reduced state of the probed mode is $\\rho_a(t) = p_0(T)|0\\rangle\\langle 0| + p_N(T)|N\\rangle\\langle N|$, and the resulting QFI is $Q(T) = M^2\\alpha^{2M} g^2 \\cos^2(M\\theta+\\Phi)\\,/\\,[1-\\alpha^{2M}\\cos^2(M\\theta)]$, where $0 < \\alpha \\le 1$ depends on temperature and on $N_\\mathrm{eff} = \\varepsilon N\\chi t/2$. Because $\\alpha^{2M}$ decays exponentially with $M$ except at isolated points where the QFI vanishes, the leading term in $M$ is linear, and the apparent $M^2$ term can never dominate; similarly, $N$ enters only inside $N_\\mathrm{eff}$, so it merely adjusts the interaction time. The paper's diagnosis of the earlier Heisenberg claim is the strict inequality $\\langle e^{iN\\chi t \\hat{M}_\\mathrm{qubits}}\\rangle \\ne e^{iN\\chi t\\langle \\hat{M}_\\mathrm{qubits}\\rangle}$: replacing the first (exact) quantity by the second converts a decaying characteristic function of the thermal excitation distribution into a sharp phase, which is what produced the spurious $N^2$ scaling.","pith_inferences":["The no-Heisenberg conclusion is proven for the reduced bright-port state; the paper's framing that it covers the entire joint atom-field system is an extrapolation beyond the shown derivation, and computing the QFI of the full joint state (both interferometer modes plus the atomic register) is the most direct open test of that stronger statement.","The suppression mechanism is general: $p_0$ is the characteristic function of the thermal excitation-number distribution, so any probe whose coupling operator has a discrete spectrum and sits in a classical mixture will exhibit the same $\\alpha^{2M}$ decay; coherent, squeezed, or otherwise non-classical atomic preparations would evade this mechanism and are the natural place to look for genuine He","The same characteristic-function technique applies to estimating any parameter encoded through an operator with discrete spectrum — for instance the coupling strength $\\chi$, the energy splitting $\\varepsilon$, or a detuning — not just temperature; each would inherit the same standard-quantum-limit structure.","The hardware results suggest that depolarizing noise pulls the observed Fisher information below the theoretical curve; an error-mitigated version of the same circuit would test whether the standard quantum limit can be approached more closely on current devices."],"forward_implications":["If the central claim is right, N00N states and higher photon numbers offer no thermometric advantage in this setting: precision scales at best linearly with the number of atoms $M$, and the photon number $N$ merely rescales the interaction time through $N_\\mathrm{eff} = \\varepsilon N\\chi t/2$.","The earlier Heisenberg-limit prediction for this setup cannot be reached, because it relied on replacing the exact operator phase $\\langle e^{iN\\chi t\\hat M_\\mathrm{qubits}}\\rangle$ with $e^{iN\\chi t\\langle\\hat M_\\mathrm{qubits}\\rangle}$; any experimental scheme built on that approximation will cap at the standard quantum limit.","Because increasing $N$ is equivalent to tuning $t$ or $\\chi$, single-photon probes ($N=1$) already reach the optimal precision at the correct interaction time, which the authors demonstrate on hardware data.","The digital implementation also works for effective negative temperatures, where the optimal $N_\\mathrm{eff}$ stays nearly constant over a wide temperature range, so one experimental setting can serve many inverted-population temperatures."],"supporting_citations":[{"why":"The earlier quantum-thermometry paper that claimed Heisenberg-limited ($N^2$) scaling for this dispersive setup; its phase-averaging approximation is the specific target the exact calculation contradicts.","marker":"[67]"},{"why":"A review chapter also asserting that quantum thermometry can reach the Heisenberg limit under strict conditions, cited alongside [67] as the claim being revisited.","marker":"[12]"},{"why":"Supplies the equilibrium thermometry framework and the QFI $= (\\Delta H)^2/T^4$ formula that the paper generalizes to the dynamical dispersive regime.","marker":"[11]"},{"why":"Provides the spectral-decomposition formula for the quantum Fisher information and the Cramér-Rao bound formalism used throughout the exact calculation.","marker":"[48]"},{"why":"The standard source for the claim that N00N and other entangled states give Heisenberg scaling in phase estimation, which the paper argues does not transfer to dispersive thermometry.","marker":"[58]"},{"why":"The quantum-circuit framework used to build and simulate the digital implementation of the nonlinear Mach-Zehnder thermometer.","marker":"[72]"},{"why":"The quantum platform on which the experimental runs were executed, supplying the hardware data reported for the single-photon case.","marker":"[73]"}],"fun_headline_variants":["Exact thermometry bound stays linear, not Heisenberg","Nonlinear MZI simulator probes thermometry's true limit","Dispersive thermometry: exact QFI kills the N-squared claim","Quantum simulator shows atom-number linearity in temperature sensing","Approximation exposed: thermometry precision is standard-limited"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the thermometer is exactly this system — $M$ non-interacting identical two-level atoms in a Gibbs state at temperature $T$, coupled only through the dispersive Hamiltonian $H = \\varepsilon\\chi\\,\\hat a^\\dagger \\hat a \\sum_k \\hat\\sigma_+^{(k)}\\hat\\sigma_-^{(k)}$ with no losses — and that the bound is demonstrated for the reduced state of the single probed field mode, leaving open whether joint measurements on the second interferometer port and on the atoms could extract more information.","fun_headline_variants_meta":{"raw":{"variants":["Exact thermometry bound stays linear, not Heisenberg","Nonlinear MZI simulator probes thermometry's true limit","Dispersive thermometry: exact QFI kills the N-squared claim","Quantum simulator shows atom-number linearity in temperature sensing","Approximation exposed: thermometry precision is standard-limited"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000277,"raw_usage":{"total_tokens":1740,"prompt_tokens":1124,"completion_tokens":616,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":740,"completion_tokens_details":{"reasoning_tokens":534}},"tokens_in":740,"tokens_out":616,"duration_ms":6680,"temperature":1.0,"reasoning_tokens":534,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T19:53:55.941913+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A concrete decisive check is to compute the exact quantum Fisher information of the full joint state after the dispersive interaction — field modes of both interferometer ports together with the $M$-atom thermal register — without tracing anything out. If in any regime (for example, near the best working point $N_\\mathrm{eff} \\approx \\pi/2$ at low temperature) that joint QFI grows faster than linearly in $N$ or $M$, the paper's general claim that the joint evolution is standard-quantum-limited is false; if it remains linear, the single-port result is confirmed and extended. A simpler comparison experiment is to measure photon statistics at the dark port and correlations with the atomic register and check whether the total classical Fisher information exceeds the bright-port QFI of Eq. (20).","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the spectral-decomposition formula for the quantum Fisher information and the Cramér-Rao bound formalism used throughout the exact calculation."},{"cited_title":"Jahnke, S","cited_arxiv_id":null,"evidence_quote":"The quantum-circuit framework used to build and simulate the digital implementation of the nonlinear Mach-Zehnder thermometer."},{"cited_title":"O’Connor, S","cited_arxiv_id":null,"evidence_quote":"The quantum platform on which the experimental runs were executed, supplying the hardware data reported for the single-photon case."}],"review_version":1}