{"id":"6cc805e3-9c36-4dc7-ae2a-97232c8ab188","arxiv_id":"2607.06286","paper_version":1,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"unknown","formal_verification":"none","parameter_count":4,"one_line_summary":"The derivative of entanglement entropy with respect to subregion volume equals the thermal entropy density in the large-subregion limit, verified via lattice simulations of the finite-density O(4) model using dual worm algorithms.","lead":"This paper shows that the derivative of entanglement entropy with respect to subregion size equals the thermal entropy density for large subregions, and verifies this on the lattice for the O(4) model at finite density. A smart generalist might read it because it opens a new route to extracting thermodynamic equations of state from entanglement measurements, potentially applicable to theories where standard thermodynamic methods face sign problems.","discovery_kind":"unclear","skeptic_critique":{"model":"glm-5.2","headline":"The Maxwell relation test verifies only the μ-derivative of the central claim, not the claim itself; a μ-independent offset at finite ℓ would be invisible to this test.","rationale":"The derivation in Section 3 is clean and relies on standard thermodynamic reasoning: in the ξ ≪ ℓ ≪ L limit, moving a slice between regions A and B changes the free energy by the difference of bulk free energy densities (Eq. 18). The r → 1 limit then recovers the thermal entropy density (Eq. 20). This argument is parameter-free and involves no fitting. The lattice verification via the Maxwell relation (Eq. 87, Fig. 15) is a genuine non-trivial check: it compares two independently computed quantities (the mixed derivative from replicated simulations vs. the step scaling of charge density from unreplicated simulations) and shows quantitative agreement up to ξ_max/ℓ ≈ 0.5. The dual-variable worm algorithm with boundary deformation is a real technical contribution, and the consistency between the two boundary update algorithms (Fig. 21) adds confidence. However, the Maxwell relation test is strictly weaker than a direct test of Eq. (27): it verifies ∂μ of the relation, not the relation itself. A μ-independent, temperature-dependent offset — which could arise from residual interface effects at finite ℓ — would pass the Maxwell test while invalidating the quantitative extraction of s from entanglement data. The authors are transparent about this limitation and frame their entropy extraction (Figs. 16–17) as estimates rather than precision measurements. For a first study, this is appropriate. The concern does not undermine the theoretical derivation (which is correct in the strict limit) but does identify the gap between what is derived and what is numerically verified. The reader's identified concern (shrinking window near μc) is valid but less load-bearing: the authors correctly restrict to μ ≤ 0.4 where the window is open, and the limitation is honestly stated. My concern is about what is tested within that valid regime. Verdict: UNCHANGED. The ACCEPT rating is appropriate for a first study with honestly acknowledged limitations. The theoretical result is parameter-free and the Maxwell relation test provides strong (if indirect) evidence. A direct comparison would strengthen the claim from 'strongly suggested' to 'verified,' but the current evidence is sufficient for acceptance.","tokens_in":37909,"tokens_out":5041,"duration_ms":267438,"concrete_test":"Compute the free energy density ω(β,μ) independently in unreplicated simulations at Nt and 2Nt (e.g., via the integral method or by integrating ⟨∂S/∂λ⟩ over a coupling λ), form s₂ = ω(2β,μ) − 2ω(β,μ) from Eq. (28), and compare directly with (1/V⊥)∂ℓH₂ at ℓ = ℓ* = 17.5 for several (Nt, μ) pairs with μ ≤ 0.4. If they agree within statistical errors, the μ-independent offset C(T) is ruled out and the central claim is directly verified. If a nonzero offset is found, its ℓ-dependence should be checked to confirm it vanishes as ℓ → ∞.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim (Eq. 27) is (1/V⊥)∂ℓH_r = s_r(T,μ) for ξ ≪ ℓ ≪ L. What is actually tested on the lattice (Eq. 87, Fig. 15) is the μ-derivative of this relation: ∂μ[(1/V⊥)∂ℓH₂] = ∂μ[s₂(T,μ)]. If the true relation were (1/V⊥)∂ℓH₂ = s₂(T,μ) + C(T) for some μ-independent offset C(T), the Maxwell relation test would still pass perfectly. Such an offset could arise from subleading boundary/interface contributions at finite ℓ that are O(ξ/ℓ) but not exactly zero — for instance, a temperature-dependent surface tension at the A–B interface that survives the ℓ-derivative at finite ℓ. The plateau observed in Fig. 11 suggests these corrections are small, but their absolute magnitude is never directly bounded against an independent measurement of s₂. The authors acknowledge this gap explicitly ('we tested the proposed framework here through the Maxwell relation rather than by directly comparing s with (1/V⊥)∂ℓSEE'), so this is not a hidden flaw but it is the weakest link between the theoretical derivation and the numerical evidence. The reader's concern about the shrinking ξ ≪ ℓ ≪ L window near μc is valid but secondary: even within the valid regime (μ ≤ 0.4), the direct relation is not independently verified.","agreement_with_reader":"partial"},"referee_report":{"model":"glm-5.2","summary":"This paper establishes a relation between the derivative of entanglement entropy (and Rényi entropies) with respect to the entangling region width and the thermal entropy density for slab-shaped subregions in general QFTs, in the limit where all linear sizes exceed the longest correlation length. The authors then test this relation nonperturbatively in the (2+1)-dimensional O(4) model at finite density using a dual-variable worm algorithm that avoids the sign problem. The lattice verification proceeds by checking a Maxwell relation (Eq. 87) rather than a direct comparison of the entropy density with the entanglement derivative. The theoretical derivation (Section 3) is clean and parameter-free; the lattice methodology is technically sound with two independent boundary-update algorithms showing consistent results.","tokens_in":38551,"tokens_out":1353,"duration_ms":338592,"significance":"The paper makes a worthwhile contribution on two fronts. First, the theoretical relation (Eq. 20/27) is derived from first principles (replica trick, thermodynamic identities, bulk free energy density assumption) with no fitted parameters, and applies to general QFTs. Second, the lattice implementation is a genuine technical achievement: the dual-variable worm algorithm with boundary deformation in the replicated geometry is non-trivial, the sign problem is circumvented, and the authors provide two independent update schemes (plaquette and boundary worm) with fully consistent results (Fig. 21, Table II). The observation that the phase transition at μc is visible in ∂ℓH₂ (Fig. 13) is a nice physical result. The framework offers a novel route to extracting thermodynamic equations of state from entanglement data, which is of broad interest.","major_comments":[{"comment":"§7, Eq. (87) and Fig. 15: The central theoretical claim is (1/V⊥)∂ℓH_r = s_r(T,μ) (Eq. 27), but the lattice test verifies only its μ-derivative: ∂μ[(1/V⊥)∂ℓH₂] = ∂μ[s₂(T,μ)] (Eq. 87). A μ-independent offset C(T) at finite ℓ — for instance from subleading boundary/interface contributions that are O(ξ/ℓ) but not exactly zero — would be invisible to this test. The authors acknowledge this gap explicitly in §8 ('we tested the proposed framework here through the Maxwell relation rather than by directly comparing s with (1/V⊥)∂ℓSEE'), so it is not a hidden flaw. However, the abstract and conclusion could be read as claiming direct verification. I recommend the authors strengthen the presentation by: (a) stating more prominently in the abstract that the Maxwell relation (not the direct equality) is what is verified, and (b) providing a quantitative bound on the plateau value of ∂ℓH₂/V⊥ against ","section":null},{"comment":"§3, Eq. (18): The derivation assumes that when a thin slice is moved from deep inside region B to deep inside region A, the interface between A and B can be neglected and the bulk free energy densities ω(β,μ) and ω(rβ,μ) apply independently. This requires ξ ≪ ℓ ≪ L simultaneously. Near the critical point μc ≈ 0.5, ξ_max diverges (Eq. 81), shrinking the validity window. The lattice data in Fig. 15 confirms agreement only up to ξ_max/ℓ ≈ 0.5, and the authors restrict entropy extraction to μ ≤ 0.4. This is handled reasonably, but the manuscript would benefit from a brief quantitative discussion of the expected scaling of finite-ℓ corrections — e.g., whether the residual at ξ/ℓ ≈ 0.5 is consistent with O(ξ/ℓ) or O((ξ/ℓ)²) — to justify the choice of cutoff.","section":null}],"minor_comments":[{"comment":"§3, Eq. (18): The notation lim_{ℓ,L→∞, ℓ≪L} is slightly ambiguous about the order of limits. Clarifying whether the limit is ℓ→∞ at fixed L→∞ with ℓ/L→0, or a double limit with the constraint maintained, would improve precision.","section":null},{"comment":"§6: The simulation parameters (κ=1.2, j₃=0.2, Ns=12, Nx=36) are stated, but no continuum extrapolation is attempted. The authors note this is intentional. A brief remark on the expected size of lattice artifacts at these parameters would be welcome.","section":null},{"comment":"Fig. 15: The horizontal axis is ξ_max(μ)/ℓ, but the different temperature curves correspond to different μ ranges. It would help the reader if the μ values corresponding to the plotted ξ_max/ℓ range were indicated, perhaps as a secondary axis.","section":null},{"comment":"§7, Figs. 16–17: The entropy density extracted from ∂ℓH₂ is labeled s₂ (the r=2 step-scaling approximation), but the figure captions and axis labels sometimes refer to 'thermal entropy density s' without the subscript. Consistency in notation would avoid confusion about whether s or s₂ is being shown.","section":null},{"comment":"§4.2, Eq. (65): The partition function in dual variables is lengthy. A compact summary table of the dual variables and their physical meanings (which is currently spread across the text) would improve readability.","section":null},{"comment":"Reference [57] is cited as a companion letter. If it contains overlapping results, a brief statement of what is new in this paper versus [57] would help assess novelty.","section":null}],"recommendation":"minor_revision","confidential_remarks":"The stress-test concern about the Maxwell relation only testing the μ-derivative is valid and is the main substantive issue. However, the authors are transparent about this limitation, the theoretical derivation is sound, and the lattice methodology is solid. The concern is addressable through presentation changes (clarifying what is and isn't verified) rather than new computations. I do not think this rises to major revision. The paper is a good fit for the journal given its combination of theoretical derivation and nonperturbative lattice verification in a sign-problem-free formulation."},"author_rebuttal":{"model":"glm-5.2","summary":"We thank the referee for a careful and constructive report. Both major comments are well-taken and will be addressed in a revised manuscript. The first concerns the distinction between direct verification of Eq. (27) and verification of its μ-derivative (the Maxwell relation, Eq. (87)); we agree the abstract and conclusion should state this more prominently and will add a quantitative bound on the plateau value. The second concerns finite-ℓ corrections near the critical point and the expected scaling; we will add a quantitative discussion of the residual at ξ/ℓ ≈ 0.5.","responses":[{"response":"We fully agree with this assessment. The referee correctly identifies that our lattice test verifies the Maxwell relation (Eq. 87), i.e., the μ-derivative of Eq. (27), rather than the direct equality (1/V⊥)∂ℓH₂ = s₂(T,μ) itself. A μ-independent offset C(T) from subleading boundary/interface contributions would indeed be invisible to our test. We acknowledge this explicitly in §8 but agree that the abstract and conclusion could be misread as claiming direct verification. We will implement both recommendations: (a) We will revise the abstract to state more prominently that the Maxwell relation—not the direct equality—is what is verified on the lattice. The current abstract already mentions the Maxwell relation in its final sentence, but we will make the distinction sharper and more upfront. (b) We will provide a quantitative bound on the plateau value of ∂ℓH₂/V⊥ against s₂(T,μ). Concretely, from the data in Fig. 11, the plateau in ∂ℓH₂ is reached for ℓ ≥ 5 at the parameter values shown. At ℓ = ℓ* = 17.5, the residual ℓ-dependence is consistent with zero within errors for μ ≤ 0.4 (where ξ_max/ℓ ≲ 0.5). We will extract a numerical bound on |∂ℓH₂/V⊥ − s₂| at fixed T and μ from the plateau region and include it in the revised §7, along with a discussion of the expected O(ξ/ℓ) scaling of any residual offset. We note that a fully direct comparison of ∂ℓH₂/V⊥ with an independently computed s₂ requires a separate lattice determination of the thermal entropy density in the O(4) model, which to our knowledge does not yet exist in the literature. We state this explicitly in §8 and identify it as an important next step.","revision_made":"yes","referee_comment":"§7, Eq. (87) and Fig. 15: The central theoretical claim is (1/V⊥)∂ℓH_r = s_r(T,μ) (Eq. 27), but the lattice test verifies only its μ-derivative. A μ-independent offset C(T) at finite ℓ would be invisible to this test. The authors acknowledge this gap in §8, but the abstract and conclusion could be read as claiming direct verification. Recommendations: (a) state more prominently in the abstract that the Maxwell relation (not the direct equality) is what is verified, and (b) provide a quantitative bound on the plateau value of ∂ℓH₂/V⊥ against s₂."},{"response":"This is a fair point. The derivation of Eq. (18) requires ξ ≪ ℓ ≪ L, and near μc the divergence of ξ_max (Eq. 81) shrinks the validity window. We will add a quantitative discussion of the expected finite-ℓ corrections. The interface contributions that are neglected in going from Eq. (17) to Eq. (18) arise from the boundary between regions A and B. For a slab geometry, these are surface terms scaling as the area of the entangling surface, i.e., O(ξ^d−2) relative to the bulk O(ℓ · ξ^d−2) contribution, giving corrections of order O(ξ/ℓ). The data in Fig. 15 shows agreement up to ξ_max/ℓ ≈ 0.5 at the lowest temperature and up to ξ_max/ℓ ≈ 1.0 at the highest temperature (where thermal effects shorten the effective correlation length below ξ_max). We will extract the residual deviation as a function of ξ_max/ℓ from the data underlying Fig. 15 and discuss whether it is consistent with linear O(ξ/ℓ) scaling or whether the data favor a faster, e.g., O((ξ/ℓ)²), falloff. This analysis will be included in the revised §7 to justify the choice of μ ≤ 0.4 as the cutoff for entropy extraction. We note that the current data, while clearly showing the onset of deviation as ξ_max/ℓ increases, may not have sufficient resolution to cleanly distinguish O(ξ/ℓ) from O((ξ/ℓ)²) across the full range; we will be transparent about this limitation.","revision_made":"yes","referee_comment":"§3, Eq. (18): The derivation assumes ξ ≪ ℓ ≪ L simultaneously. Near μc ≈ 0.5, ξ_max diverges, shrinking the validity window. The lattice data in Fig. 15 confirms agreement only up to ξ_max/ℓ ≈ 0.5. The manuscript would benefit from a brief quantitative discussion of the expected scaling of finite-ℓ corrections — e.g., whether the residual at ξ/ℓ ≈ 0.5 is consistent with O(ξ/ℓ) or O((ξ/ℓ)²) — to justify the choice of cutoff."}],"tokens_in":37777,"tokens_out":1202,"duration_ms":172528,"standing_objections":[]},"desk_editor":{"model":"glm-5.2","letter":"Two things to know about this paper. First, the theoretical relation (1/V⊥)∂S_EE/∂ℓ = s for slab entangling regions in the large-ℓ limit is derived cleanly from standard ingredients — the replica trick, thermodynamic identities, and the assumption that bulk free energy densities apply when ξ ≪ ℓ ≪ L. No fitted parameters, no circularity. The algebra from Eq. (16) through Eq. (20) is correct. Second, the lattice work — dual-variable worm algorithm for the O(4) model at finite density with boundary deformation for Rényi entropy — is a genuine technical contribution. The sign-problem-free dual formulation and the two boundary update schemes (plaquette and worm) are well-engineered, and the cross-check in Fig. 14 confirming consistency of the two evaluation methods for the mixed derivative is convincing. The authors are also honest about limitations: only r=2 Rényi is tested, no continuum extrapolation, critical region excluded. The stress-test concern about the Maxwell relation only testing the μ-derivative is valid but I'd put it in perspective. The authors explicitly acknowledge this gap — they say they tested via the Maxwell relation rather than by direct comparison. A μ-independent offset at finite ℓ is possible in principle, but the plateau behavior in Fig. 11 suggests such corrections are small. The real question for future work is whether a direct |∂ℓH₂/V⊥| vs. s₂ comparison can be done, and the authors say as much. The reader's concern about the shrinking ξ ≪ ℓ window near μc is real but secondary — the data at μ ≤ 0.4 where the window is valid shows clean agreement. This is a solid first study. The theoretical derivation is the main contribution; the lattice verification is supportive rather than definitive. The paper is for lattice field theorists and QFT researchers interested in entanglement-thermodynamics connections. It deserves a serious referee — the technical content is substantial and the central claim is clearly stated and partially verified. The main thing a referee should push on is whether the authors can strengthen the case beyond the Maxwell relation test, even with existing data, by bounding the absolute offset more tightly.","headline":"Paper derives a clean entanglement-entropy/thermal-entropy relation and tests it on the lattice via a Maxwell relation — the test is indirect but the derivation is sound","tokens_in":38752,"tokens_out":536,"would_cite":true,"duration_ms":120308,"reading_group":"no","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"glm-5.2","headline":"Entanglement entropy yields thermal entropy density in large slabs","keywords":[],"falsifier":"Compute (1/V⊥) ∂S_EE/∂ℓ and the independently measured thermal entropy density s in the same theory at the same parameters; if they disagree in the regime ξ ≪ ℓ ≪ L, the identity (Equation 20) is falsified. The lattice test via the Maxwell relation (Equation 87) is an indirect version of this check.","tokens_in":38096,"feed_emoji":"🔥","tokens_out":1126,"duration_ms":182543,"temperature":0.7,"pith_summary":"This paper proves that for slab-shaped entangling regions in general quantum field theories, the derivative of entanglement entropy with respect to the slab width, divided by the cross-sectional area, equals the thermal entropy density s, provided all linear sizes exceed the longest correlation length. The argument proceeds by decomposing the replicated free energy into bulk contributions from the entangling region and its complement, showing that the ℓ-derivative isolates the difference of free energy densities ω(rβ, μ) − rω(β, μ), whose r→1 limit is exactly s. The authors verify this relation on the lattice for the three-dimensional O(4) model at finite chemical potential, using a dual-variable worm algorithm that avoids the sign problem. They confirm that the mixed derivative of the second Rényi entropy with respect to slab width and chemical potential satisfies the same Maxwell relation as the thermal entropy density, up to the point where the correlation length grows comparable to the slab width near the finite-density phase transition.","feed_headline":"Entanglement entropy derivative equals thermal entropy density","feed_subtitle":"A general QFT identity lets lattice entanglement data extract bulk thermodynamics, verified in the O(4) model at finite density.","key_machinery":"The derivation hinges on the observation that when a thin slice is moved from deep inside region B to deep inside region A (both regions being much larger than the correlation length), the interface between A and B can be neglected and the bulk free energy densities ω(β, μ) and ω(rβ, μ) apply independently. This reduces the ℓ-derivative of the replicated free energy to a difference of free energy densities, which upon taking the r→1 limit yields the standard thermodynamic expression for entropy. The lattice verification uses a dual-variable representation of the O(N) model that eliminates the sign problem, combined with a boundary-deformation worm algorithm to compute Rényi entropyderivativ.","core_discovery":"The central result is Equation (20): in the regime where all correlation lengths are much smaller than the slab width and system size, the quantity (1/V⊥) ∂S_EE/∂ℓ is identically the thermal entropy density s. This is not an approximation or a scaling relation but an exact identity derived from the replica trick. The paper then demonstrates it nonperturbatively by showing that the r=2 Rényi analogue satisfies the corresponding Maxwell relation (Equation 87) on the lattice, with agreement persisting up to ξ_max/ℓ ≈ 0.5–1.0 depending on temperature.","pith_inferences":["If entanglement entropy derivatives encode the equation of state, then measuring S_EE at multiple temperatures and chemical potentials could in principle reconstruct the full thermodynamic phase diagram from purely information-theoretic observables, though the practicality depends on whether the correlation-length constraint can be satisfied across the full parameter space.","The breakdown near criticality suggests a natural complementarity: entanglement entropy is most useful as a thermodynamic probe away from phase transitions (where it gives bulk thermodynamics) and most useful as a diagnostic tool near phase transitions (where it detects critical behavior but not the equation of state).","The dual-variable worm algorithm and boundary-deformation technique developed here could be adapted to other theories admitting dual formulations, potentially including CP(N−1) models or abelian gauge-Higgs systems, extending the reach of entanglement-based thermodynamic extraction beyond O(N) models."],"forward_implications":["If the relation holds broadly, entanglement entropy measurements on the lattice could serve as an alternative route to extracting the full equation of state of interacting quantum field theories, including those where direct thermodynamic measurements are difficult.","The method is most cleanly applicable in confining or gapped phases where correlation lengths are finite; near critical points where correlation lengths diverge, the window of validity shrinks and the relation breaks down, as confirmed by the lattice data near μ_c ≈ 0.5.","The framework extends to any Rényi order r, where the derivative yields a step-scaling approximation s_r of the entropy density, converging to s as r→1; this means even r=2 Rényi data (which are far more accessible on the lattice than von Neumann entropy) carry genuine thermodynamic information.","The relation works in reverse: in theories where thermal entropy is easier to compute than entanglement entropy, one can use thermodynamic data to infer entanglement properties."],"fun_headline_variants":["Entanglement entropy derivative yields exact thermal entropy density","Exact link between entanglement derivative and thermal entropy","Thermal entropy density extracted from entanglement entropy derivative","Entanglement derivative equals thermal entropy density in large limit","Derivative of entanglement entropy recovers bulk thermodynamics"],"cache_read_input_tokens":0,"weakest_assumption_plain":"The derivation assumes that when a thin slice is transferred between the two regions, the interface between them can be neglected and the bulk free energy densities apply independently. This requires the correlation length to be simultaneously much smaller than the slab width and the system size, a condition that fails near the critical point where the correlation length diverges.","fun_headline_variants_meta":{"raw":{"variants":["Entanglement entropy derivative yields exact thermal entropy density","Exact link between entanglement derivative and thermal entropy","Thermal entropy density extracted from entanglement entropy derivative","Entanglement derivative equals thermal entropy density in large limit","Derivative of entanglement entropy recovers bulk thermodynamics"]},"model":"glm-5.2","effort":"high","cost_usd":0.0,"raw_usage":{"total_tokens":787,"prompt_tokens":443,"completion_tokens":344,"prompt_tokens_details":null},"tokens_in":443,"tokens_out":344,"duration_ms":15639,"temperature":1.0,"reasoning_tokens":322,"cache_read_input_tokens":0,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-08T10:50:14.804923+00:00","model_set":{"reader":"glm-5.2"},"falsifier":"Compute (1/V⊥) ∂S_EE/∂ℓ and the independently measured thermal entropy density s in the same theory at the same parameters; if they disagree in the regime ξ ≪ ℓ ≪ L, the identity (Equation 20) is falsified. The lattice test via the Maxwell relation (Equation 87) is an indirect version of this check.","supporting_citations":[],"review_version":1}