{"id":"cadbe992-4ba3-41d5-94ad-38b1b2c7eeba","arxiv_id":"2501.19396","paper_version":3,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"low","formal_verification":"none","parameter_count":3,"one_line_summary":"The Gibbs state of the mean-field Bose gas at critical temperatures is proven to be, in trace norm, a coherent-state weighted average of Bogoliubov states governed by a one-mode Phi^4 condensate distribution.","lead":"For a gas of weakly interacting bosons at temperatures near the Bose-Einstein condensation point, this paper proves that the full quantum equilibrium state equals, to high precision, an average of coherent condensate states combined with Bogoliubov quasiparticle states. It is the first rigorous justification of Lee and Yang's 1958 positive-temperature refinement of Bogoliubov theory, and it yields concrete formulas for condensate fluctuations and particle correlations.","discovery_kind":"first_principles","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem statements omit the hypothesis ∠\\hat v(0)>0; for v≡0, Theorem 6's Gaussian claim is false and Theorem 5/Corollary 2.7 are singular, so the formal scope needs the nonzero-interaction assumption.","rationale":"The reader's weakest assumption is close: the proof's lower bounds discard positive terms proportional to ∠\\hat v(p), so ∠\\hat v≥0 is structurally essential. The sharper issue in the formal statements is that v=0 (equivalently ∠\\hat v(0)=0) is not excluded. The ideal gas is admitted by the hypotheses of Theorem 1 and is even the benchmark of Section 1.2. Under the condensed-phase condition N0≥N^{2/3}, Theorem 1 part (a) would define Γ via a Φ^4 distribution with vanishing quartic coefficient. If μ_BEC=μ0, this Γ is actually the ideal gas state (the P-representation of the zero-mode thermal state is an exponential mixture of coherent states), so Theorem 1 itself is plausibly true for v=0. However, the Gaussian-fluctuation statements that follow are not. Theorem 6's g(x) has variance N/(β∠\\hat v(0)); for v=0 this is infinite, and the true density of |z|^2 is exponential, not Gaussian. Similarly, Theorem 5(a) and Corollary 2.7 contain 1/∠\\hat v(0) and ln(∠\\hat v(0)) and are not meaningful at v=0. Since the theorems are stated for all v satisfying Theorem 1's hypotheses, this is a real overbreadth. The fix is explicit and local: add ∠\\hat v(0)>0 to the hypotheses of the results that use the Φ^4 condensate variance, or at least to Theorem 1 and then propagate. The nonzero-interaction proofs otherwise appear internally consistent; I spot-checked the Stahl-based correlation inequality, the free-energy matching, and the log-removal bootstrap in Remark 9.1, and found no circularity. Therefore the verdict should remain acceptance, but conditional on the hypothesis amendment.","tokens_in":81484,"tokens_out":22070,"duration_ms":212342,"concrete_test":"Take v≡0 and βN^{2/3}→κ>κ_c. For the ideal gas, compute ζ_G(z)=Tr[|z⟩⟨z|G_id]=(πN0)^{-1}exp(-|z|^2/N0) with N0∼N. The distribution of X=|z|^2 is exponential with mean N0, so its variance is N0^2∼N^2, whereas Theorem 6's g(x) would require variance N/(β∠\\hat v(0)), which diverges as ∠\\hat v(0)→0. Hence Theorem 6 as stated fails for v=0. The same example makes Theorem 5(a)'s first line and Corollary 2.7 singular. If the authors add ∠\\hat v(0)>0 (or ‘v not identically zero’) to the hypotheses of Theorems 1 and 5–7, the concern is resolved.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The proof is built on the mean-field interaction with v≥0 and ∠\\hat v≥0. Section 1.2 states ‘with v≠0’, but Theorem 1's formal hypotheses only require v to be nonnegative and even with ∠\\hat v≥0, which admits v≡0 (equivalently ∠\\hat v(0)=0). This matters: Theorem 6 asserts ζ_G(z) ≈ g(|z|^2) with g a Gaussian density in x=|z|^2 of mean N0 and variance N/(β∠\\hat v(0)); for v=0 this variance is undefined and the assertion is false. For the ideal gas at κ>κ_c, N0∼N, ζ_G(z)=(πN0)^{-1}e^{-|z|^2/N0}, so the density of x=|z|^2 is exponential, not Gaussian. Theorem 5(a)'s first line and Corollary 2.7 also contain 1/∠\\hat v(0) or ln(∠\\hat v(0)) terms and are singular for v=0. The intended restriction v≠0 is mentioned in Section 1.2 but never enters the formal theorem statements. Since several main results claim validity under ‘assumptions of Theorem 1’, the statements are overbroad; the proofs for ∠\\hat v(0)>0 do not establish the v=0 cases, and Theorem 6 is false there. This is a statement-scope gap, not a flaw in the nonzero-interaction argument.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the homogeneous mean-field Bose gas on the three-dimensional torus with interaction potential v/N at temperatures comparable to the critical temperature, i.e., βN^{2/3} → κ ∈ (0,∞). Its main result (Theorem 1) is a trace norm approximation of the grand canonical Gibbs state G_{β,N} by an explicit reference state Γ_{β,N}: in the condensed regime (N0(β,N) ≥ N^{2/3}) it is a convex combination of coherent states tensored with temperature-dependent Bogoliubov Gibbs states, weighted by a one-mode Φ^4 condensate distribution g_BEC; in the non-condensed regime it is the ideal gas Gibbs state. From this approximation the paper derives pointwise and trace-norm bounds for the one-particle density matrix, formulas for the two-particle density matrix, limiting distributions for the condensate particle number (Gaussian, interpolating, exponential, and geometric regimes), and an asymptotic expansion of the free energy. The proofs develop two new abstract correlation inequalities, one based on an infinite-dimensional version of Stahl's theorem, which are used to control second- and higher-order correlations in the true Gibbs state rather than only in trial states.","tokens_in":81633,"tokens_out":10566,"duration_ms":109804,"significance":"If the results hold as stated, this is a major contribution to the rigorous theory of Bose gases: it provides the first justification of Lee and Yang's positive-temperature extension of Bogoliubov theory in the mean-field regime, including a non-quasi-free reference state with condensate fluctuations described by a one-mode Φ^4 theory. The explicit formulas for the condensate number variance (of order N^{5/3}), the crossover of the limiting distribution at N0 ∼ N^{5/6}, and the free energy expansion are new and physically informative. The two abstract correlation inequalities are of independent interest and have already been used in follow-up work; the proof of the second-order inequality via Stahl's theorem is elegant. The manuscript is very detailed and essentially self-contained, with numerous technical lemmas relegated to appendices; the overall strategy is coherent and the estimates are stated with explicit rates. The only serious defect I found is a gap between the formal hypotheses of the main theorems and the cases actually covered by the proofs, which is fixable and does not affect the correctness of the nonzero-interaction results.","major_comments":[{"comment":"The formal hypotheses of Theorem 1 admit v ≡ 0 (v nonnegative, even, with nonnegative Fourier coefficients and finite (1+|p|)-weighted ℓ^1 norm), but the interacting-case results are false or singular for v = 0. The Gaussian density g in Theorem 6, Eq. (2.12), contains 1/β v̂(0); for the ideal gas at κ > κ_c the distribution ζ_G(z) = (πN0)^{-1} exp(−|z|^2/N0) makes x = |z|^2 exponential with variance N0^2, not the asserted Gaussian with variance N/(βv̂(0)). Similarly, the first line of Theorem 5(a) and Corollary 2.7, Eq. (2.31), contain 1/v̂(0) or ln(v̂(0)) and are singular for v = 0. The intended restriction v ≠ 0 appears informally in Section 1.2 but does not enter the theorem statements. Since the proofs for the interacting case rely on v̂(0) > 0 in several places (e.g., the Gaussian approximation of g_BEC in Section 9.4.1 and Lemma A.1), the statements are overbroad as written. The fix is to add the hypothesis v̂(0) > 0 (equivalently v not identically zero) to Theorem 1 and, consequently, to Theorems 4–8 and Corollary 2.7, or to state and prove separate v = 0 results.","section":"Theorem 1 (Section 1.4), Theorem 6 (Section 2.2), Theorem 5(a) (Section 2.1), Corollary 2.7 (Section 2.3)"}],"minor_comments":[{"comment":"The phrase 'with v, 0' should read 'with v ≠ 0'; this is exactly the hypothesis that is missing from the formal statements of the theorems.","section":"Section 1.2"},{"comment":"The text refers to 'Proposition 5.2' when bounding f^{BEC}(N0, N0^G); the only Proposition in Section 5 is Proposition 5.3, so this citation should be corrected (or the correct lemma, likely Lemma A.4, should be cited).","section":"Section 8.3"},{"comment":"The condition 'N0(β,N) = tN^{5/6} with some fixed t ∈ R' should be t > 0 (since N0 is nonnegative); the same correction applies to the 'some fixed t ∈ R' in the definition of σ.","section":"Theorem 7(b)"},{"comment":"The random variable N0 in (2.14) shares its symbol with the scalar N0(β,N) used throughout the paper; even though eN0 is later introduced, the notation is confusing and should be changed, e.g., to N_0^{ran} or another symbol.","section":"Section 2.2, Eq. (2.14)"},{"comment":"There is a misplaced dz inside the large parentheses after the third term; the measure dz should appear outside the bracket, as is clear from the surrounding terms.","section":"Eq. (1.23)"},{"comment":"The manuscript contains numerous typos and spacing errors, e.g., 'vaccuum', 'Gri ffi th', 'di fficulties', 'eN0' for 'eN0', and several missing spaces before parentheses. A careful proofreading pass is needed.","section":"General"}],"recommendation":"major_revision","confidential_remarks":"This is a strong and important paper. The main issue is a formal statement-scope gap: the hypothesis v̂(0) > 0 is missing from Theorem 1 and the theorems that depend on it, and several main results are false or singular for v = 0. This is easily fixed by adding the hypothesis, so I recommend major revision rather than rejection. The technical core appears sound, and the new correlation inequalities are valuable. The paper is long and technical, but the organization is clear. I would not insist on shortening it; however, the citation error to Proposition 5.2 and the notational overloading of N0 should be corrected. The paper fits well in a leading mathematical physics journal."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Bottom line: this is the real thing. It proves trace-norm approximation of the grand canonical Gibbs state for the homogeneous mean-field Bose gas on the critical temperature scale, with a non-quasi-free reference state built from a one-mode Phi^4 condensate distribution and Bogoliubov thermal states. That closes the main gap between the low-temperature quasi-free results and the critical-point classical field theory. The two abstract correlation inequalities, especially the one from Stahl's theorem, are clean and independently useful; the higher-order version is what makes the 2-pdm and variance statements possible. The free-energy upper and lower bounds match at order N^{5/8}, and the route from trace norm plus moment bounds to 1-pdm, 2-pdm, condensate distributions, and the second transition at N0 ~ N^{5/6} is coherent. I checked the structure and spot-checked several key inequalities; I did not verify every appendix estimate.\n\nMain soft spot: the formal theorem statements. The hypotheses of Theorem 1 admit v=0 (all Fourier coefficients zero). But Theorem 6's Gaussian density in |z|^2 has variance N/(\\beta \\hat v(0)) and is false for v=0: the ideal gas gives an exponential distribution in |z|^2. Theorem 5(a)'s first line and Corollary 2.7 similarly contain 1/\\hat v(0) or ln(\\hat v(0)) and are singular. The text says \"with v \\neq 0\" in Section 1.2, but that restriction never enters the theorem statements. This is a statement-scope gap, not a flaw in the nonzero-interaction proof. A clean revision would add \\hat v(0)>0 to every result that needs it and note that Theorem 1's v=0 case reduces to the ideal gas.\n\nSecond, the proof really relies on \\hat v(p) \\ge 0 and weighted \\ell^1 summability; positive Fourier modes are load-bearing in the lower bound and a priori estimates. That is a genuine modeling restriction, worth advertising more prominently. The paper is also explicitly mean-field and fixed-volume, so it does not resolve the thermodynamic-limit transition; the authors do not overclaim this.\n\nMinor issues: the trace-norm rate N^{-1/48} is slow, and there are a few deferred computations plus a typo in Theorem 7's hypothesis referring to Theorem 8. None of these affect the main conclusions.\n\nAudience: mathematical physicists working on BEC and Bogoliubov theory. This deserves a serious referee. I would send it out and expect a revision that repairs the v=0 edge case; the nonzero-interaction results are strong enough that this should not be a desk reject.","headline":"A benchmark proof of Lee-Yang-Bogoliubov theory for the mean-field Bose gas at critical temperature; the main theorems need a small but real repair excluding v≡0 from statements where 1/\\hat v(0) appears.","tokens_in":82364,"tokens_out":3471,"would_cite":true,"duration_ms":41630,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["82B10","82B26","81V70"],"pacs":["05.30.-d","03.75.Hh","05.70.Fh"],"model":"deepseek-v4-flash","headline":"For the homogeneous mean-field Bose gas at BEC critical temperatures, the interacting grand canonical Gibbs state is approximated in trace norm to N^{-1/48} by an explicit Φ^4-weighted Bogoliubov state.","keywords":["mean-field Bose gas","Bose–Einstein condensation","Bogoliubov theory","Gibbs state","trace norm approximation","one-mode Φ^4 theory","correlation inequalities","critical temperature"],"falsifier":"A numerical computation of the variance of $a_0^*a_0$ in the condensed phase for a finite-$N$ mean-field gas with nonnegative Fourier coefficients would settle the claim: the paper predicts $\\mathrm{Var} = N/(\\beta \\hat v(0)) + O(N^{5/3-1/6})$, and a result with a different $N$-scaling of this variance, or with the Gaussian condensate distribution of Theorem 6 failing, would refute the trace norm approximation.","tokens_in":81119,"feed_emoji":"❄️","tokens_out":8645,"duration_ms":87101,"temperature":0.7,"pith_summary":"The paper proves that for a homogeneous mean-field Bose gas on the unit torus at temperatures of order $N^{{-2/3}}$, the scale of the Bose–Einstein condensation critical temperature, the grand canonical Gibbs state is approximated in trace norm by an explicit reference state. The reference state is a convex combination, weighted by a one-mode $Φ^{4}$ Gibbs distribution, of coherent states for the condensate tensored with Bogoliubov thermal states for the excited modes. This gives a rigorous justification of Lee and Yang's positive-temperature analogue of Bogoliubov theory, and it yields the limiting condensate number distributions, the one- and two-particle density matrices, and a free energy expansion. A reader should care because the result identifies the actual mechanism of the interacting BEC phase transition: the Gibbs state is not quasi-free, and the condensate's particle-number fluctuations are governed by a one-mode $Φ^{4}$ measure.","feed_headline":"Bose gas at BEC criticality pinned to N^{-1/48}","feed_subtitle":"One formula gives the interacting Gibbs state, the condensate counting statistics, and the free energy across the BEC transition.","key_machinery":"The load-bearing object is the reference state Γ_{β,N} = ∫ |z⟩⟨z| ⊗ G_Bog(z) g_BEC(z) dz, a non-quasi-free convex combination. The coherent state |z⟩ implements the c-number substitution for the condensate; G_Bog(z) is the Gibbs state of the quadratic Bogoliubov Hamiltonian for the excited modes; and g_BEC(z) is a Gibbs distribution of a one-mode $Φ^{4}$ theory describing condensate number fluctuations. The proof is carried by two new abstract correlation inequalities: a second-order bound whose proof uses an infinite-dimensional version of Stahl's theorem to obtain convexity of the Duhamel two-point function, and a higher-order bound for moments that uses a commuting operator X dominating B and [[B,A],B]. These inequalities turn first-order Griffith-type bounds on perturbed Gibbs states into the sharp second- and higher-moment estimates needed to control the terms neglected by Bogoliubov theory.","core_discovery":"In the limit N→∞ with β $N^{{2/3}}$→κ∈(0,∞), the paper establishes the trace norm bound ||G_{β,N} − Γ_{β,N}||_1 ≤ C $N^{{-1/48}}$. When the ideal-gas condensate number N0(β,N) satisfies N0 ≥ $N^{{2/3}}$, Γ_{β,N} = ∫_C |z⟩⟨z| ⊗ G_Bog(z) g_BEC(z) dz, where |z⟩ is a coherent state, G_Bog(z) is the Gibbs state of a temperature-dependent Bogoliubov Hamiltonian with dispersion ε(p) = $\\sqrt$($p^{2}$−µ0) $\\sqrt$($p^{2}$−µ0 + 2 vhat(0) N0/N), and g_BEC(z) is the Gibbs distribution of the one-mode $Φ^{4}$ theory exp(−β vhat(0)|z|^4/(2N) − β µ_BEC |z|^2). When N0 < $N^{{2/3}}$, the interacting state is instead within the same trace norm of the ideal gas Gibbs state. From this approximation the paper derives the 1- and 2-particle density matrices, the variance and full limiting distribution of the number of condensate particles, and the free energy expansion F = F_Bog + vhat(0)N/2 + F_BEC(N0) + O($N^{{5/8}}$).","pith_inferences":["Because the approximation is in trace norm, it automatically transfers to all bounded observables; one can read off higher correlation functions beyond the 2-pdm from the reference state, although the paper computes only the 1- and 2-pdm.","The abstract correlation inequalities are likely to be the transferable core: any model where a Griffiths-type first-order bound for perturbed Gibbs states is available should admit the same second- and higher-moment control, including lattice or local Φ^4 settings.","The N0 ∼ N^{5/6} crossover in the condensate number distribution points to a genuine second fluctuation phase transition in the one-mode Φ^4 measure itself, with the interaction vhat(0) becoming irrelevant below that scale.","A natural testable extension is the canonical ensemble, where the condensate number is fixed; the grand canonical formulas here predict the characteristic function of N0 up to order N^{-1/48}, which could be compared with numerics at moderate N."],"forward_implications":["In the condensed phase the interacting Gibbs state is determined to leading order by the Φ^4 condensate measure times Bogoliubov thermal states, and it is not quasi-free because the condensate number fluctuates on scale N^{5/3}.","The BEC critical temperature is given to leading order by the ideal-gas value, and the 1-pdm has pointwise formulas for each momentum mode with error N^{-1/96}.","The variance of the condensate particle number is N/(β vhat(0)) in the condensed phase, and the full condensate distribution crosses over from Gaussian to exponential to geometric as N0 shrinks across N^{5/6}.","The free energy has the expansion F_Bog + vhat(0)N/2 + F_BEC(N0) + O(N^{5/8}), including interaction-dependent N^{2/3} log N and N^{2/3} terms in the condensed phase.","In the non-condensed phase the interacting Gibbs state is stable close to the ideal gas Gibbs state despite the added interaction."],"supporting_citations":[{"why":"Supplies the positive-temperature extension of Bogoliubov theory whose predictions the paper justifies.","marker":"[71]"},{"why":"Establishes BEC and the Bogoliubov spectrum in the mean-field limit, providing the starting point the present proof extends to critical temperatures.","marker":"[105]"},{"why":"Gives the low-temperature trace norm approximation of Gibbs states by quasi-free states, the framework generalized here to the non-quasi-free critical regime.","marker":"[74]"},{"why":"Contributes the classical-field description of the condensate and the first-moment-to-second-moment correlation inequality that the paper sharpens.","marker":"[73]"},{"why":"Introduces the one-mode Φ^4 condensate Gibbs distribution used in the trial state and in the reference state.","marker":"[20]"},{"why":"Provides Stahl's theorem, whose infinite-dimensional version yields the Laplace representation behind convexity of the Duhamel two-point function.","marker":"[110]"},{"why":"Supplies the rigorous coherent-state c-number substitution used throughout the lower-bound argument.","marker":"[81]"},{"why":"Gives the original Bogoliubov theory that the paper extends to positive temperatures.","marker":"[21]"}],"fun_headline_variants":["Bose gas Gibbs state pinned to N^{-1/48} at BEC criticality","N^{-1/48} trace norm bound for BEC Gibbs states","Condensate counting from Φ^4 theory in Bose gas","Coherent states plus Bogoliubov give BEC Gibbs state","Lee-Yang extension proven for mean-field Bose gas"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof assumes the interaction is nonnegative in position space with nonnegative Fourier coefficients $\\hat v(p)\\ge 0$ and a finite weighted sum $\\sum_p (1+|p|)\\hat v(p)$; if any Fourier mode is negative, the free-energy lower bound and the correlation estimates that carry the approximation no longer follow.","fun_headline_variants_meta":{"raw":{"variants":["Bose gas Gibbs state pinned to N^{-1/48} at BEC criticality","N^{-1/48} trace norm bound for BEC Gibbs states","Condensate counting from Φ^4 theory in Bose gas","Coherent states plus Bogoliubov give BEC Gibbs state","Lee-Yang extension proven for mean-field Bose gas"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000909,"raw_usage":{"total_tokens":3940,"prompt_tokens":1010,"completion_tokens":2930,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":626,"completion_tokens_details":{"reasoning_tokens":2837}},"tokens_in":626,"tokens_out":2930,"duration_ms":23160,"temperature":1.0,"reasoning_tokens":2837,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-09T20:16:25.621721+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A numerical computation of the variance of $a_0^*a_0$ in the condensed phase for a finite-$N$ mean-field gas with nonnegative Fourier coefficients would settle the claim: the paper predicts $\\mathrm{Var} = N/(\\beta \\hat v(0)) + O(N^{5/3-1/6})$, and a result with a different $N$-scaling of this variance, or with the Gaussian condensate distribution of Theorem 6 failing, would refute the trace norm approximation.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the positive-temperature extension of Bogoliubov theory whose predictions the paper justifies."},{"cited_title":"Seiringer, The excitation spectrum for weakly interacting bosons, Commun","cited_arxiv_id":null,"evidence_quote":"Establishes BEC and the Bogoliubov spectrum in the mean-field limit, providing the starting point the present proof extends to critical temperatures."},{"cited_title":"Lewin, P","cited_arxiv_id":null,"evidence_quote":"Gives the low-temperature trace norm approximation of Gibbs states by quasi-free states, the framework generalized here to the non-quasi-free critical regime."},{"cited_title":"Lewin, P","cited_arxiv_id":null,"evidence_quote":"Contributes the classical-field description of the condensate and the first-moment-to-second-moment correlation inequality that the paper sharpens."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides Stahl's theorem, whose infinite-dimensional version yields the Laplace representation behind convexity of the Duhamel two-point function."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the rigorous coherent-state c-number substitution used throughout the lower-bound argument."}],"review_version":1}