{"id":"8ab49edd-be74-4f49-9389-0a9acbe94acf","arxiv_id":"2603.07635","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"The derivative of entanglement entropy with respect to region size equals the thermal entropy density, and a generalized Maxwell relation connects it to charge density — tested nonperturbatively in the 3D O(4) model.","lead":"This paper argues that in any quantum field theory, changing the size of a large measured region by one step changes the entanglement entropy by exactly the ordinary thermal entropy density — even when particles are packed at finite chemical potential. It also finds this thermodynamic imprint in computer simulations of a 3D O(4) model, so entanglement data could someday be used to read off the equation of state.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Sign error in the generalized Maxwell relation: Eq. (12) has +β²∂β n, but the standard Maxwell relation from Eq. (1) gives −β²∂β n.","rationale":"The reader's weakest_assumption focused on Eq. (9) being imported from ref. [12], which is a missing-derivation gap but not a demonstrated error. The present stress-test identifies a concrete, demonstrable sign error in the paper's central generalization: the stated Maxwell relation is incorrect, leading to an incorrect sign in Eq. (12) and Eq. (18). This is more load-bearing than the Eq. (9) concern because it directly affects the paper's headline claim about a generalized Maxwell relation. The numerical test in Fig. 3 actually uses the correct sign, so the data support a corrected version of the relation rather than the one written in the paper. Because the central theoretical formula is wrong as stated, the preprint in its current form should be rejected, though a corrected version may be viable. I disagree with the reader's assessment that the main weak point is the imported Eq. (9); the sign error is a more serious and immediate issue.","tokens_in":10538,"tokens_out":19976,"duration_ms":161466,"concrete_test":"Re-derive Eq. (18) from Eq. (15) alone: ∂_μ s_r = −∂_μ[ω_L(T)−ω_L(T/r)]/(T−T/r) = +[n(T)−n(T/r)]/(T−T/r) = +Δ^r_T n, using ∂ω_L/∂μ = −n. This confirms that Eq. (18) should have +Δ^r_T n, not −Δ^r_T n, and that Eq. (12) should have −β²∂_β n. Check Fig. 3: if the plotted bands and points have the same sign as listed in Eq. (27), the data verify the corrected relation and therefore disprove Eq. (12) as stated.","verdict_should_be":"REJECT","load_bearing_attack":"The paper's central generalized Maxwell relation, Eq. (12), has the wrong sign. From Eq. (1), dω_L = −s dT − n dμ, the mixed partial of ω_L yields the standard Maxwell relation (∂s/∂μ)_T = +(∂n/∂T)_μ. The paper instead states (∂s/∂μ)_T = −(∂n/∂T)_μ, and consequently Eq. (12) is written as (1/V⊥)∂_μ∂_ℓ S_EE = β² ∂_β n. Since T = 1/β, ∂n/∂T = −β²∂_β n, so the correct relation is (1/V⊥)∂_μ∂_ℓ S_EE = −β²∂_β n. The same sign error propagates into Eq. (18): differentiating Eq. (15) with respect to μ gives +Δ^r_T n, not −Δ^r_T n. Ironically, the numerical implementation (27) uses the correct sign: −2N_t(n(2N_t)−n(N_t)) equals +Δ^2_T n. Thus the data in Fig. 3 actually support the corrected relation and contradict Eq. (12) as written. This internal inconsistency affects the headline claim about thermodynamic response relations and must be resolved.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper argues that for slab-shaped entangling regions in the large-region limit, the size derivative of entanglement entropy equals the thermal entropy density, and that this relation leads to thermodynamic response identities, including a generalized Maxwell relation coupling chemical potential and charge density. The argument proceeds through the replica construction and uses a previously proposed identity, Eq. (9), to relate size derivatives of the replicated partition function to free-energy differences at scaled temperatures. The authors test an r=2 Rényi version of the relation by lattice simulations of the three-dimensional O(4) model at finite chemical potential, using a dual worm algorithm and a boundary-deformation method. They report agreement up to ξ_max/ℓ ≈ 0.5–1 and conjecture that the relation is generic in continuum QFTs.","tokens_in":10899,"tokens_out":8402,"duration_ms":77751,"significance":"If the central relation is correct, it provides a novel, nonperturbative bridge between entanglement entropy and equilibrium thermodynamics: variations of the entangling region encode equation-of-state data, and the generalized Maxwell relation would be a new universal sum rule. The lattice calculation is technically substantial: it uses a sign-problem-free dual representation, a boundary-deformation algorithm, and an internal-consistency check, and it explicitly tests a finite-density setting where such relations are difficult to access. However, the central identity is imported from a previous paper whose authors include two of the current authors, and the claimed Maxwell relation is written with an internal sign inconsistency. The numerical data appear to support the corrected sign, so the core idea is plausible, but the manuscript as written does not yet establish the headline relation.","major_comments":[{"comment":"The generalized Maxwell relation has the wrong sign. From dω_L = -s dT - n dμ, the mixed-partial relation is (∂s/∂μ)_T = +(∂n/∂T)_μ. Since ∂n/∂T = -β² ∂β n, Eq. (12) should read (1/V⊥)∂²S_EE/∂μ∂ℓ = -β² ∂β n. Correspondingly, differentiating Eq. (15) with respect to μ gives +Δ_T^r n, not -Δ_T^r n as in Eq. (18). The numerical implementation in Eq. (27), -2N_t[n(2N_t)-n(N_t)] = +Δ_T^2 n, has the sign required by the corrected relation. Thus Fig. 3 tests the corrected relation and contradicts Eqs. (12) and (18) as written. This internal inconsistency must be fixed before the thermodynamic-response claim can be assessed.","section":"Eqs. (12), (18), (27)"},{"comment":"The load-bearing input, Eq. (9), is not derived here; the text says 'we now use the argument presented in [12]'. Since [12] shares two of the current authors, the derivation is not independently established in this Letter. Moreover, the commutation of r→1 with ∂_ℓ, assumed below Eq. (8), is nontrivial. If Eq. (9) or the commutation fails at finite μ, the central relations (10), (12), (15), and (18) collapse. A numerical test at r=2 and one lattice spacing cannot by itself control the r→1 continuum limit. The authors should either provide a self-contained derivation of Eq. (9) with explicit hypotheses or clearly label it as an assumption and discuss its validity.","section":"Eqs. (8)–(10)"},{"comment":"The nonperturbative evidence is more limited than the text suggests. The simulation tests the r=2 step-scaling relation (27) at a single entangling-region width ℓ=17.5 and at N_s=12, and agreement is shown only for ξ_max/ℓ ≲ 0.5–1, i.e., away from the ξ_max ≪ ℓ limit in which Eq. (9) is supposed to hold. No r→1 extrapolation is attempted, although the headline statement (10) is an r→1 (von Neumann) result. The statement 'strong nonperturbative evidence' should be softened, or systematic checks toward r→1 and larger ℓ should be provided.","section":"Eq. (27) and Fig. 3"}],"minor_comments":[{"comment":"The sentence saying the derivative 'can equivalently be understood as a derivative with respect to spatial size of A' is confusing, since ∂_ℓ is already the derivative with respect to the slab width. Clarify what distinction is intended.","section":"After Eq. (10)"},{"comment":"The lattice action includes parameters κ, λ, and j, but the simulation parameters state κ=1.2 and j_3=0.2 without giving λ. Please specify λ or state that the linearized model is used.","section":"Eq. (20)"},{"comment":"The caption refers to ϕ4, while the text and the main discussion refer to the ϕ0 Goldstone mode. Align the notation.","section":"Fig. 1"},{"comment":"The companion paper is listed as 'in progress'; for publication, please provide a stable arXiv reference or summary of the relevant derivations that are invoked from it.","section":"Ref. [10]"}],"recommendation":"major_revision","confidential_remarks":"The sign error in (12)/(18) is correctable and the numerical data actually support the corrected sign, so I would not reject on that ground alone. My main concern is the reliance on Eq. (9) from a previous paper by the same group without derivation; a referee may want to verify that derivation separately. After the sign inconsistency is resolved and the status of Eq. (9) is clarified, the paper could be a valuable contribution."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: the core identification—size derivative of EE → thermal entropy density, and the μ-derivative → charge response—is the right thing to test, and the O(4) lattice data are the first nonperturbative finite-density evidence I know of for it. But the written Maxwell relation has the wrong sign, and the numerical plot actually supports the corrected sign. That has to be fixed before the equations can be taken at face value.\n\nWhat's genuinely new: the step-scaling Rényi identity (Eq. 15 with Eq. 18) and its lattice verification using the boundary-deformation method in a finite-density O(4) model. The move from the known Calabrese–Cardy entropy/thermal-entropy scaling to a μ-derivative Maxwell-type relation is natural but was not previously demonstrated nonperturbatively at finite density. The consistency check in Fig. 2 (∂μ∂ℓH2 versus -2N_t V ∂ℓñ) is a good example of algorithmic cross-validation.\n\nWhere it's soft: First and most important, the sign. From Eq. (1), the Maxwell relation is (∂s/∂μ)_T = +(∂n/∂T)_μ, which translates to 1/V⊥ ∂²S_EE/∂μ∂ℓ = −β² ∂β n. The paper writes +β² ∂β n in Eq. (12). The same sign error propagates into Eq. (18); the correct step scaling is +Δ^r_T n, not −Δ^r_T n. The irony is that Eq. (27) uses the correct sign for the lattice observable (−2N_t(n(2N_t)−n(N_t)) = +Δ^2_T n), so the agreement in Fig. 3 actually supports the corrected relation and contradicts Eq. (12) as written. This is a load-bearing typo in the formal section, but it's fixable.\n\nSecond, Eq. (9) is imported from [12] without derivation, and [12] has two of the same authors. The identity is plausible and consistent with earlier scaling arguments, but the formal thrust of the paper depends on it. The numerical test at one r=2 and a single ℓ=17.5 is not a full verification of the r→1, continuum limit. The assumption that the r→1 limit commutes with ∂ℓ, stated below Eq. (8), is also just asserted. The companion paper [10] will hopefully carry the derivations and algorithmic details.\n\nNone of these are fatal. The sign error is embarrassing but self-correcting in the numerics. The rest are addressable gaps. The paper is worth a serious referee, and the authors should fix the sign and either derive or justify Eq. (9) more carefully. I'd cite the lattice result once the equations are corrected.","headline":"The finite-density lattice demonstration is a genuine step, but the central Maxwell relation (Eq. 12) has a sign error that the Fig. 3 data actually contradict—fixable, but it has to be resolved before the formal claims as written can stand.","tokens_in":11377,"tokens_out":6788,"would_cite":true,"duration_ms":54286,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper argues that in the large-region limit the size-derivative of entanglement entropy approaches the thermal entropy density, so entanglement variations can be used to read off thermodynamic response and equation-of-state data.","keywords":["entanglement entropy","thermal entropy","Rényi entropy","finite chemical potential","Maxwell relation","O(4) model","lattice field theory","equation of state"],"falsifier":"These relations can be tested in solvable models by computing, for a free scalar or free fermion at finite temperature in 2+1 dimensions, the direct entanglement-spectrum derivative ∂_ℓ S_EE for slabs with ℓ≫ξ and checking whether (∂_ℓ S_EE)/V⊥ approaches s(T); if the ratio deviates, the identity (10) fails. Alternatively, compute the Rényi step-scaling relation (18) at r=3 and r=4 on the same O(4) lattices and check whether s_r converges to s as r→1, or verify (12) at r→1 using exact diagonalization or tensor-network methods for a finite-density lattice system.","tokens_in":1554,"feed_emoji":"🌡️","tokens_out":2586,"duration_ms":61689,"temperature":0.7,"pith_summary":"The paper aims to establish a precise, nonperturbative link between entanglement entropy and bulk thermodynamics: for a slab-shaped entangling region much wider than any correlation length, the derivative of entanglement entropy with respect to the slab width approaches the thermal entropy density, independent of microscopic details. The same statement holds for Rényi entropies of any integer order, with the thermal entropy replaced by a discrete step-scaling approximation in temperature. At finite chemical potential, these relations imply a generalized Maxwell relation linking mixed derivatives of entanglement entropy to the temperature derivative of the charge density. The paper demonstrates these relations nonperturbatively in the three-dimensional O(4) model at finite density, using a dual-variable worm algorithm to compute Rényi entropies and charge densities, and conjectures that they are generic features of continuum quantum field theories. If the conjecture is correct, entanglement measurements provide a route to directly extracting equation-of-state information.","feed_headline":"Entanglement entropy growth equals thermal entropy density","feed_subtitle":"If right, entanglement data directly read off the equation of state, even at finite chemical potential.","key_machinery":"The carrying mechanism is the replica method for entanglement entropy, expressed as a limit of replicated partition functions. The pivotal input is an identity (Eq. 9), imported from earlier work, stating that for ξ≪ℓ≪L the ℓ-derivative of the replicated free energy satisfies (1/V⊥) ∂_ℓ log Z̃(ℓ,r) → −[ω(rβ, μ) − r ω(β, μ)], where ω is the dimensionless grand-canonical free-energy density. Combining this with the definition of S_EE as the r→1 derivative of log tr ρ^r_A yields the thermal entropy result. For numerical access, the paper uses the boundary-deformation method to compute the ℓ-derivative of the second Rényi entropy H_2 as a log-ratio of replicated partition functions, circumventin","core_discovery":"The core claim is the identity (Eq. 10): in the limit where the slab width ℓ and the total spatial extent L both go to infinity with ℓ≪L, the derivative of entanglement entropy with respect to ℓ, divided by the transverse area V⊥, equals the thermal entropy density s(T, μ) at the system's temperature and chemical potential. For the Rényi entropy H_r of integer order r≥2, the same limit gives a discrete approximation s_r(T, μ) to the thermal entropy, constructed as a step-scaling finite-difference in temperature with scaling factor r, which reduces to ordinary entropy as r→1. At finite chemical potential, differentiating with respect to μ produces a generalized Maxwell relation: the mixed der","pith_inferences":["The derivation relies only on extensivity of the free energy and the slab geometry, so the relations likely extend to other entangling-region shapes and to interacting theories with gauge fields; testing spherical regions in conformal field theories would sharpen this claim.","The numerical evidence uses only r=2 and a single slab width; a direct test of the r→1 limit would require a different estimator, and free-field theories could provide an analytic check for all r, which would be a strong test of the conjectured genericity.","If the conjecture is correct, entanglement entropy measurements on quantum simulators could serve as a thermometer or densitometer for many-body systems without coupling to a heat bath, though extracting s from a single measurement requires controlling the region-size derivative.","The generalized Maxwell relation implies an integrability condition on entanglement data; checking these cross-relations in experiments or simulations could confirm the thermodynamic interpretation without ever computing entropies directly."],"forward_implications":["For slab-shaped regions much wider than any correlation length, the growth of entanglement entropy with region size is the thermal entropy density times the added volume, so the UV-divergent area term drops out of size derivatives.","At finite chemical potential, entanglement entropy satisfies thermodynamic response relations, including a generalized Maxwell relation linking mixed μ–ℓ derivatives to the temperature derivative of the charge density.","Rényi entropies of any integer order give the same physics, with the thermal entropy replaced by a discrete step-scaling approximation in temperature; as r→1 these converge to the exact relations.","In the 3D O(4) model, the predicted equality holds within errors for ξ_max/ℓ up to about 0.5–1 depending on temperature, and the entanglement-derived observable clearly resolves the finite-density phase transition.","The relation connects entanglement entropy to bulk thermodynamics nonperturbatively, opening a route to extract equation-of-state information from entanglement data."],"fun_headline_variants":["Entanglement entropy yields thermal entropy density","Derivative of EE equals thermal entropy in large-region limit","Entanglement entropy density matches thermal entropy at finite density","EE derivative approaches thermal entropy for large slabs","EE derivative gives entropy density at any chemical potential"],"cache_read_input_tokens":12672,"weakest_assumption_plain":"The argument rests on an imported identity (Eq. 9 from earlier work) stating that for large slabs the ℓ-derivative of the replicated free energy equals ω(rβ, μ) − rω(β, μ), together with the assumption that the r→1 limit commutes with the ℓ-derivative; the paper takes the identity as given, and the numerical test checks only the r=2 case at one slab width.","fun_headline_variants_meta":{"raw":{"variants":["Entanglement entropy yields thermal entropy density","Derivative of EE equals thermal entropy in large-region limit","Entanglement entropy density matches thermal entropy at finite density","EE derivative approaches thermal entropy for large slabs","EE derivative gives entropy density at any chemical potential"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000389,"raw_usage":{"total_tokens":1848,"prompt_tokens":664,"completion_tokens":1184,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":408,"completion_tokens_details":{"reasoning_tokens":1112}},"tokens_in":408,"tokens_out":1184,"duration_ms":7897,"temperature":1.0,"reasoning_tokens":1112,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-02T18:34:41.096576+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"These relations can be tested in solvable models by computing, for a free scalar or free fermion at finite temperature in 2+1 dimensions, the direct entanglement-spectrum derivative ∂_ℓ S_EE for slabs with ℓ≫ξ and checking whether (∂_ℓ S_EE)/V⊥ approaches s(T); if the ratio deviates, the identity (10) fails. Alternatively, compute the Rényi step-scaling relation (18) at r=3 and r=4 on the same O(4) lattices and check whether s_r converges to s as r→1, or verify (12) at r→1 using exact diagonalization or tensor-network methods for a finite-density lattice system.","supporting_citations":[],"review_version":1}