{"id":"26d69ee6-489c-4c03-bdae-6cd7fc099d10","arxiv_id":"2505.00188","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The paper derives exact thermodynamic relations for strongly coupled open systems under a zero-variance interaction condition, expressing free energy and internal energy through the interaction-energy distribution and a chi-squared divergence.","lead":"Open systems with strong coupling are hard to treat because the interaction energy matters. This paper shows that when the interaction does not fluctuate with the environment, thermodynamics can be expressed with the interaction-energy distribution instead of the difficult Hamiltonian of mean force.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Var_{E0}(e^{-βV})=0 forces V to be independent of the environment's fluctuating microstate, reducing the advertised strong-coupling regime to a slaved/mean-field interaction; the practical free-energy formula fails for generic strong coupling.","rationale":"The reader's weakest-assumption analysis identifies Var_{E0}(e^{-βV})=0 as the load-bearing condition; I agree and sharpen the consequence: zero variance implies the interaction is almost surely constant over the environment's fluctuating degrees of freedom, so the system-environment coupling is reduced to a deterministic system potential. This means the practical free-energy formulas in Eqs. (12)-(13) do not apply to generic strong coupling, only to the slaved/mean-field class where the HMF formalism is unnecessary. The universal-looking Eq. (31) is a definitional identity, not a nonequilibrium work theorem, so it does not rescue the strong-coupling claim. Independently, I noted a concrete algebraic error in the final relations: substituting Eq. (37) into Eq. (41) yields e^{-βμ}=⟨e^{-βW}⟩(1+χ^2)^{1/2}, whereas Eq. (42) states the exponent as -1/2; this should be corrected but is peripheral to the main scope concern. The paper's mathematical derivations are mostly correct within the stated condition, so the reader's CONDITIONAL verdict remains appropriate, pending an explicit statement of the slaved nature of the condition and a numerical demonstration of the regime where it applies.","tokens_in":10798,"tokens_out":23209,"duration_ms":225897,"concrete_test":"Evaluate the ratio R(xS)=⟨e^{-2βV}⟩_{E0}/⟨e^{-βV}⟩_{E0}^2 for a harmonic bilinear model with H_E=Σ ω_i x_i^2/2 and V=xS Σ c_i x_i, at finite β. The exact Gaussian calculation gives R(xS)=exp(β xS^2 Σ c_i^2/ω_i)>1 for any nonzero coupling and xS≠0. Then compare the two sides of Eq. (12) in this model: the left-hand side and right-hand side differ by the Pβ-weighted average ⟨R(xS)⟩_S, so Eq. (13) is violated. This single calculation settles that the central substitution holds only when the interaction is independent of the fluctuating environmental coordinates, i.e. when the coupling is decoupled or slaved.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Eq. (8), Var_{E0}(e^{-βV})=0, is the hinge of the practical claim. Since e^{-βV} is nonnegative, zero variance for fixed xS means e^{-βV(xS,xE)} is constant in xE almost surely on the support of PE0. Hence V(xS,xE)=v(xS) up to a null set: the interaction contains no fluctuating environmental degrees of freedom. The two 'strongly coupled' scenarios in the text are exactly this slaved/mean-field case, where H*_β=H_S+v and the system is effectively closed with a modified potential. In that regime Eq. (12) holds trivially and Eq. (13) is an ordinary FEP identity; the HMF machinery is not needed. For any interaction that genuinely depends on environmental microstates, e.g. V=xS Σ c_i x_i, Var_{E0}(e^{-βV})>0 and the substitution in Eq. (12) fails by the pointwise factor ⟨e^{-2βV}⟩_{E0}/⟨e^{-βV}⟩_{E0}^2 > 1. Eq. (31), by contrast, is an algebraic identity following from Pβ=e^{-βH*}/Z* and P_S0=e^{-βHS}/Z_S; it is valid for arbitrary coupling but is not a dynamical fluctuation theorem and does not by itself justify computing potentials from P(VSE). The paper's advertised ability to compute thermodynamic potentials from {Pβ(xS), P(VSE)} is therefore confined to a class where the environment is slaved, not generic strong coupling.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This manuscript addresses the calculation of thermodynamic potentials for open systems strongly coupled to an environment. Starting from the Hamiltonian of mean force (HMF), the authors derive Eq. (7), express the condition under which the system-environment double average can be replaced by the coupled average, and identify it as Var_E0(e^{-βVSE})=0 in Eq. (8). Under this condition, they obtain Eq. (13) for the free energy and an expression for the internal energy, both in terms of Pβ(xS) and the distribution of VSE. They then derive Eq. (31), an equality relating e^{-βΔF_S} to ⟨e^{-βΔH}⟩_S and the chi-squared divergence, and combine it with the Jarzynski equality and Gaussian assumptions to obtain Eqs. (37)-(43). The central advertised results are the universal equality and the possibility of measuring thermodynamic potentials without the HMF.","tokens_in":11151,"tokens_out":13734,"duration_ms":124074,"significance":"If correct, the proposed shift from {Pβ(xS), H*_β} to {Pβ(xS), P(VSE)} would be practically valuable, since P(VSE) is more accessible than the HMF. The manuscript has genuine strengths: the derivation of Eq. (7) from Eq. (6) is careful, the consistency check in Eq. (11) is explicit, and the algebraic steps leading to Eq. (31) are correct. However, the zero-variance condition in Eq. (8) confines the practical formulas to a slaved/mean-field coupling regime rather than generic strong coupling, Eq. (31) is an algebraic identity with no explicit dependence on the interaction, and the final Gaussian relations in Eqs. (42)-(43) contain a sign error. These issues substantially reduce the reach of the paper as written, although the underlying algebra is mostly sound.","major_comments":[{"comment":"The variance condition in Eq. (8) is much more restrictive than the text suggests and does not describe generic strong coupling. Since e^{-βV}>0, the condition Var_E0(e^{-βV})=0 for a fixed system coordinate xS means that e^{-βV(xS,xE)} is constant in xE on the support of P_E0, so V(xS,xE)=v(xS) almost surely. Thus the interaction carries no fluctuating environmental degrees of freedom: the environment acts only as a fixed external field or is slaved to the system. The two 'strongly coupled' scenarios discussed in the text are precisely this slaved/mean-field situation, in which H*_β=H_S+v and the system is effectively closed with a modified potential. For any interaction that genuinely depends on environmental microstates, such as V=xS Σ c_i x_i, the variance is positive and the substitution ⟨⟨e^{-βV}⟩_{E0}⟩_S = ⟨e^{-βV}⟩_{S+E} fails; Eq. (12), Eq. (13), and the P(VSE)-based formulas are therefore not valid for generic strong coupling. Additionally, Eq. (8) must be read pointwise in xS; as written, Var_E0(e^{-βVSE}) is a function of xS, not a single global number.","section":"§5, Eq. (31)"},{"comment":"Eq. (31) is an algebraic identity rather than a new physical equality with independent content. Substituting e^{-βΔHβ(x)} = e^{-βΔF_S} Pβ(x)/P_S0(x) into the average over Pβ reproduces Eq. (31) by construction; the same is true of Eq. (33). No system-environment coupling, interaction energy, or environment distribution enters these relations. They hold for any two normalized distributions Pβ and P_S0 over xS with ΔH defined through their ratio. Consequently, the claim that Eq. (31) is 'analogous in importance to the Jarzynski equality' is overstated, and the equality does not by itself justify computing free energies from P(VSE). The physically nontrivial content resides in Eqs. (12)-(13), which require the restrictive condition Eq. (8).","section":"§6, Eqs. (42)-(43)"},{"comment":"There is a sign error in Eqs. (42) and (43). Let C=1+χ²(Pβ‖P_S0). From Eq. (37), exp(β²σ²/2)=C^{1/2}. Under the stated Gaussian assumption, ⟨e^{-βV}⟩_{S+E}=e^{-βμ} C^{1/2}. Inserting this into Eq. (41) gives e^{-βμ} = ⟨e^{-βW}⟩ C^{1/2}, not C^{-1/2} as printed in Eq. (42). Combining with Eq. (39) then yields e^{-βμ0} = ⟨e^{-βW}⟩ C^{-1/2}, not C^{1/2} as printed in Eq. (43). The two displayed formulas are mutually inconsistent and should be corrected, or the Gaussian/cumulant convention needs to be stated.","section":"§6, Eqs. (42)-(43)"}],"minor_comments":[{"comment":"Equation (19) writes P_S0(xS) exp(βH_S(xS)) in the denominator, but this quantity equals 1/Z_S and is independent of xS; writing it as an xS-dependent expression obscures the algebra and should be replaced by the constant.","section":"§2, Eq. (8)"},{"comment":"The condition in Eq. (8) should be stated as Var_{E0}(e^{-βV(xS,·)})=0 for each xS, with a note on the support of P_E0, to avoid ambiguity.","section":"§3"},{"comment":"The notation Pβ(VSE) is used before the distribution of the interaction energy is defined; a definition of P(VSE) as the marginal of the coupled equilibrium distribution is needed.","section":"§2"},{"comment":"The text claims 'three scenarios' following Eq. (8), but only the decoupled case V=0 and two nontrivial scenarios are listed; renumber or correct this.","section":"§6"},{"comment":"The Gaussian assumption is applied separately to the coupled and uncoupled ensembles; clarify whether both are assumed and whether the equality σ²=σ0² is an independent assumption or a consequence of Eqs. (37)-(38). There is also a typographical error, 'we combine or result', in the paragraph preceding Eq. (40).","section":"§6"}],"recommendation":"major_revision","confidential_remarks":"The central issue is scope: the practical claim depends on a zero-variance condition that reduces the problem to a closed system, and Eq. (31) is an algebraic identity. The arithmetic error in Eqs. (42)-(43) can be fixed, but the framing of the contribution needs substantial revision before the manuscript can be considered for publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper has one genuinely useful observation and one serious overclaim. The observation is that if e^{-βV} is constant over environment fluctuations, the HMF machinery can be bypassed and free energy and internal energy depend only on Pβ(xS) and the distribution of V. The overclaim is calling this strong coupling.\n\nWhat the paper does well: the derivation of Eq. (7) is correct, the variance condition is stated clearly, and the algebra leading to Eq. (31) is fine. Eq. (31) is a clean identity relating the FEP average to the chi-squared divergence, though it is essentially a reformulation of the Gibbs–Bogoliubov–Feynman bound. The idea of expressing thermodynamic variables in terms of P(V_SE) is worth recording, and the partition-function consistency check is a nice touch.\n\nThe soft spots are substantial. The stress-test note is correct: zero variance of a positive random variable means it is constant almost surely, so V is independent of the environment's microstate for each system coordinate. Therefore the two \"strongly coupled\" scenarios are slaved or mean-field; the environment does not fluctuate in response to the system. In that regime, the system is effectively closed with a modified potential, and the HMF is trivial. The practical claim about computing from P(V_SE) does not extend to generic strong coupling, which is precisely where HMF is needed. The Gaussian assumption for V_SE appears without rigorous support and is used to derive Eqs. (37)–(39); that is a minor-to-moderate weakness.\n\nThe abstract's analogy to the Jarzynski equality is not apt: Eq. (31) is an exact ensemble identity, not a non-equilibrium fluctuation theorem. The phrasing \"analogous in importance\" will raise expectations the paper does not meet.\n\nThis paper deserves a serious referee. The core algebra is sound, and the identity may be useful in some contexts, but the authors need to reframe the claims significantly, either dropping \"strong coupling\" or proving the variance condition holds in a genuinely strong-coupling case. I would send it to review, expecting major revision.","headline":"A clean identity and a restrictive condition: the variance condition actually forces the interaction to be independent of the environment's microstate, so the strong-coupling framing is overstated, but the algebra is sound enough to merit review.","tokens_in":11667,"tokens_out":3139,"would_cite":false,"duration_ms":32335,"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":"Under a zero-variance condition on the coupling energy, this paper proves that free energy and internal energy of a strongly coupled open system follow from two probability distributions alone, bypassing the Hamiltonian of mean force.","keywords":["strongly coupled open systems","Hamiltonian of mean force","interaction energy distribution","free energy perturbation","Jarzynski equality","chi-squared divergence","thermodynamic potentials","system-environment coupling"],"falsifier":"Take a solvable strongly coupled model, for instance a harmonic oscillator bilinearly coupled to a harmonic bath, and compute $\\operatorname{Var}_{\\mathcal{E}_0}(e^{-\\beta V_{SE}})$ while testing both sides of Eq. (31) with exact or well-converged simulation. If the variance is not zero yet the equality still holds, the paper's condition is not necessary; if the variance is zero but the equality fails, the central claim is false.","tokens_in":10581,"feed_emoji":"🌡️","tokens_out":9996,"duration_ms":85224,"temperature":0.7,"pith_summary":"The paper targets strongly coupled open systems, where system-environment interaction energy is too large to ignore. It claims that when a single condition holds, namely that $e^{-\\beta V_{SE}}$ has zero variance over the unperturbed environment, all thermodynamic potentials can be built from two accessible objects: the system's probability density $P_\\beta(x_S)$ and the probability distribution $P(V_{SE})$ of the interaction energy. This removes the need to construct the Hamiltonian of mean force, which is hard to compute and has a non-uniqueness problem. The central exact result is an equality connecting the free-energy shift $\\Delta F_S$ to an average of $e^{-\\beta\\Delta H_\\beta}$ divided by $1+\\chi^2(P_\\beta\\|P_{S0})$, valid at arbitrary coupling strength. A sympathetic reader should care because the result turns a hard effective-Hamiltonian problem into a statistics problem about directly measurable interaction energies.","feed_headline":"A new equality sets free energy at any coupling strength","feed_subtitle":"When interaction-energy fluctuations vanish, thermodynamics follows from two probability distributions alone.","key_machinery":"The load-bearing object is the ratio of the ensemble average $\\langle e^{-\\beta\\Delta H_\\beta}\\rangle_S$ to $1+\\chi^2(P_\\beta\\|P_{S0})$, with $\\chi^2$ the chi-squared divergence measuring how far the strongly coupled system distribution sits from the uncoupled reference. The variance condition $\\operatorname{Var}_{\\mathcal{E}_0}(e^{-\\beta V_{SE}})=0$ is the gateway: it makes $e^{-\\beta V_{SE}}$ constant over unperturbed environment fluctuations, so conditional and unconditional environment averages coincide and the Hamiltonian of mean force can be traded for $P(V_{SE})$. Free-energy perturbation supplies the reference-state averages, and Jarzynski's equality supplies the non-equilibrium link.","core_discovery":"The central claim is the equality $e^{-\\beta\\Delta F_S} = \\langle e^{-\\beta\\Delta H_\\beta(x)}\\rangle_S / (1+\\chi^2(P_\\beta\\|P_{S0}))$, presented as valid for arbitrary coupling strength. From it the authors derive the inequality $\\langle e^{-\\beta\\Delta H_\\beta}\\rangle_S \\ge e^{-\\beta\\Delta F_S}$, the same bound Jensen's inequality gives when applied to the Gibbs-Bogoliubov-Feynman inequality. Under the condition $\\operatorname{Var}_{\\mathcal{E}_0}(e^{-\\beta V_{SE}})=0$, they show the free energy and internal energy no longer require the Hamiltonian of mean force: $\\exp(-\\beta F_S^*)=\\langle e^{-\\beta V_{SE}}\\rangle_{S+E}/\\langle P_\\beta(x_S)e^{\\beta H_S(x_S)}\\rangle_S$ and $U_S=\\langle H_S\\rangle_S+\\langle V_{SE}\\rangle$. Combining the equality with the Jarzynski equality yields expressions in which the mean interaction energy is fixed by the average non-equilibrium work and the chi-squared divergence between the coupled and uncoupled system distributions.","pith_inferences":["The variance condition may hold for a broader class of environments than the paper's two strong-coupling examples, since any environment whose response is deterministic at fixed system coordinates would satisfy it.","If the variance condition fails, Eq. (31) might still stand as a free-energy identity while the interaction-energy formulas (12)-(13) break; this split is testable in models with inhomogeneous coupling.","When $P(V_{SE})$ is non-Gaussian, the exponential relations linking variance to $\\chi^2$ should become inequalities, giving a measurable deviation from the central-limit picture."],"forward_implications":["Free-energy differences in strongly coupled systems become measurable from two histograms: the system's configurational density and the interaction-energy distribution.","The ambiguity in the Hamiltonian of mean force, where infinitely many pairs reproduce the same system probability density, disappears under the variance condition because thermodynamics is fixed by $P_\\beta(x_S)$ and $P(V_{SE})$.","Internal energy can be computed as $\\langle H_S\\rangle_S+\\langle V_{SE}\\rangle$ instead of requiring the temperature derivative of the Hamiltonian of mean force.","Combining with the Jarzynski equality gives the mean interaction energies $\\mu$ and $\\mu_0$ in terms of the average non-equilibrium work and $\\chi^2(P_\\beta\\|P_{S0})$, making the mean interaction energy an observable in non-equilibrium experiments.","If $V_{SE}$ is Gaussian, the variance of the interaction energy is tied to the chi-squared divergence, so distribution width and free-energy shift determine each other."],"supporting_citations":[{"why":"Supplies the Hamiltonian-of-mean-force definitions and documents the computational difficulty at strong coupling.","marker":"[17]"},{"why":"Provides the HMF relations and the non-uniqueness argument for free energy from a given system PDF.","marker":"[11]"},{"why":"Supplies the free-energy perturbation framework used for the reference-state averages.","marker":"[36]"},{"why":"Gives the FEP formula for free-energy differences used in Eq. (18).","marker":"[35]"},{"why":"States the Jarzynski equality that the paper combines with its main result.","marker":"[52]"},{"why":"Extends fluctuation theorems to open systems and supports the strong-coupling reading of Jarzynski.","marker":"[23]"},{"why":"Defines the chi-squared divergence used in Eq. (30).","marker":"[46]"},{"why":"Supplies Jensen's inequality behind the derived inequality.","marker":"[38]"},{"why":"Supplies Hölder's inequality used to prove the chi-squared term is non-negative.","marker":"[45]"}],"fun_headline_variants":["Equality tames strong-coupling thermodynamics","Free energy at any coupling: a new equality","Strong coupling? New equality fixes free energy","When fluctuations vanish, thermodynamics simplifies","Beyond Jarzynski: equality for strong coupling"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"Everything rests on the premise that for every fixed system coordinate, $e^{-\\beta V_{SE}}$ has exactly zero variance over the unperturbed environment; if that variance is not zero, the substitution that removes the Hamiltonian of mean force does not go through.","fun_headline_variants_meta":{"raw":{"variants":["Equality tames strong-coupling thermodynamics","Free energy at any coupling: a new equality","Strong coupling? New equality fixes free energy","When fluctuations vanish, thermodynamics simplifies","Beyond Jarzynski: equality for strong coupling"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000642,"raw_usage":{"total_tokens":2987,"prompt_tokens":1009,"completion_tokens":1978,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":625,"completion_tokens_details":{"reasoning_tokens":1912}},"tokens_in":625,"tokens_out":1978,"duration_ms":13360,"temperature":1.0,"reasoning_tokens":1912,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T04:50:26.826267+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a solvable strongly coupled model, for instance a harmonic oscillator bilinearly coupled to a harmonic bath, and compute $\\operatorname{Var}_{\\mathcal{E}_0}(e^{-\\beta V_{SE}})$ while testing both sides of Eq. (31) with exact or well-converged simulation. If the variance is not zero yet the equality still holds, the paper's condition is not necessary; if the variance is zero but the equality fails, the central claim is false.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the Hamiltonian-of-mean-force definitions and documents the computational difficulty at strong coupling."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the HMF relations and the non-uniqueness argument for free energy from a given system PDF."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the FEP formula for free-energy differences used in Eq. (18)."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"States the Jarzynski equality that the paper combines with its main result."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Extends fluctuation theorems to open systems and supports the strong-coupling reading of Jarzynski."},{"cited_title":"Colla, F","cited_arxiv_id":null,"evidence_quote":"Defines the chi-squared divergence used in Eq. (30)."}],"review_version":1}