{"id":"aa178694-41ca-4cb8-a388-48d160114630","arxiv_id":"2607.12710","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"Thermally populating non-Abelian singlet-sector states of local SU(2)-invariant chains yields distillable entanglement (1/2)log2 N + O(1) ebits, even from a T=0 product ground state.","lead":"This paper shows that heating a quantum chain inside a locked total-spin (singlet) sector can turn a completely unentangled ground state into one whose distillable entanglement grows without bound with system size. Because each half of the chain must carry the same spin representation, locally measuring that label converts thermal fluctuations into standard entangled pairs.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 1's universal 1/2 log2 N coefficient rests on the unproved uniform complex-twist cluster expansion (Lemma B.1); the exact dimer result is secure, but the general theorem remains conditional.","rationale":"The reader's weakest_assumption identifies Lemma B.1 as the load-bearing point, and I agree. I independently checked the exact dimer solution: the sector-resolved form (F.5), the equality E_D = E_F = Y_N via convexity, the two-center saddle, and the Bessel crossover are internally consistent; the finite-size recursion and asymptotic constants check out. I also examined the QBP stability transfer: the bandedness property (E.7) follows from the support of the truncated operator, Haar averaging preserves that support, and choosing R = q log N makes truncation errors n^{-4}; this part is plausible. The only place where the general claim can fail is the assertion that the non-Hermitian complex-twist partition function has a uniformly analytic logarithm with boundary derivatives bounded through third order. The manuscript's proof is a sketch: it cites cluster expansions for real interactions and asserts the extension to a complex source without fully deriving the uniform bounds. Everything downstream — the Gaussian saddle, n^{-3/2} normalization, cubic tail, and stability transfer — depends on that lemma. This does not undermine the exact dimer chain, which is parameter-free and exactly solvable; I find no internal inconsistency, circularity, or invented entity. Therefore the reader's CONDITIONAL verdict should stand unchanged, with the recommendation that either Lemma B.1 be fully proven or Theorem 1 be explicitly labeled a conjecture supported by outline-level arguments.","tokens_in":18794,"tokens_out":25139,"duration_ms":266429,"concrete_test":"Independently re-derive Lemma B.1 by writing out the full polymer expansion for the complex interaction K_{beta,z} in Eq. (B.9), proving absolute convergence on a z-polydisc independent of n, and using Cauchy estimates to establish the uniform third-derivative bound on b_{n,beta}(z) (Eq. B.6). If this derivation cannot be completed without additional assumptions (e.g., restricting to real twists or imposing an extra smallness condition on Im z), Theorem 1 must be downgraded to a conjecture. As a secondary numerical cross-check, compute f_{n,beta}(e^{iX·S}) for the J1-J2 chain at N=10,12,14,16 and beta J1=0.5, and test whether log f = n phi(X) + b_n(X) has bounded |partial^alpha b_n| and a strictly negative Hessian for phi.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Lemma B.1 is the origin of the Gaussian saddle, the n^{-3/2} normalization, the cubic low-spin tail, and hence the Y_N = 1/2 log2 N + O(1) scaling for every finite-range SU(2)-invariant chain. Its proof in Appendix B is only a sketch: the polymer expansion of Refs. [17,18] is invoked for a non-Hermitian one-site source i z·M_n, and the asserted uniform bounds on the boundary term b_{n,beta}(z) and its derivatives through third order (Eqs. B.5-B.6) are not actually derived. The non-Hermitian twist changes the interaction norm and the analyticity domain; if b_n acquires non-uniform derivatives or the logarithm develops a branch issue for complex z, the two-center saddle and the subsequent QBP transfer in Lemma E.1 have no rigorous foundation. This is not an internal inconsistency — the dimer result (Eqs. 7-12) is exact and independent — but it means Theorem 1 is an outline-level claim rather than a proved theorem.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a mechanism by which thermalization constrained to a non-Abelian strong-symmetry sector creates distillable entanglement that diverges with system size. It defines a representation-space distillation protocol that measures the half-chain SU(2) labels; for any state supported in the invariant subspace this yields E_D ≥ Y_N with Y_N = Σ_j p_j log2(2j+1). For an exactly solvable dimer chain with a product-singlet ground state, the authors prove E_D = E_F = Y_N exactly at every finite size and derive the fixed-temperature asymptotic E_D = 1/2 log2 N + O(1), together with an exact Bessel crossover F(x) in the double-scaling limit. They further state Theorem 1 asserting Y_N = 1/2 log2 N + O_β(1) and E_D ≥ Y_N for all finite-range, uniformly bounded, locally SU(2)-invariant chains with reflection symmetry at sufficiently high temperature, and present finite-size exact-diagonalization data for a J1–J2 chain.","tokens_in":18900,"tokens_out":24398,"duration_ms":261589,"significance":"If the general theorem holds, this is a notable conceptual result: it shows that heating can convert O(√N) thermal spin fluctuations into an unbounded operational resource starting from an unentangled ground state, and it identifies the non-Abelian dimension as the source of the logarithmic coefficient. The dimer result is exact, derived without fitted parameters, and passes the β=0 check against the known maximally-mixed-sector result. The Bessel crossover is a concrete, falsifiable prediction, and numerical/ancillary code is provided. However, the universal statement for generic chains is currently conditional on Lemma B.1; the manuscript's own proof of that lemma is only an outline, so the theorem is not yet fully established.","major_comments":[{"comment":"Theorem 1's universal 1/2 log2 N coefficient is derived from Lemma B.1, but the proof of the lemma is not complete as written. The polymer expansion of Refs. [17,18] is invoked for the non-Hermitian one-site source K_{β,z} in (B.9), and the passage from convergence of the expansion to the decomposition (B.5) with the uniform derivative bounds (B.6) is asserted. In particular a nontrivial analyticity/uniformity argument is needed to show that the boundary term b_{n,β}(z) and its derivatives through third order are bounded independently of n; these bounds are later used in Lemma C.1 (C.2) and in the saddle normalization. Without a rigorous derivation, the n^{-3/2} Gaussian saddle and hence Theorem 1 are not proved for generic local chains. The exact dimer result (Sec. 2.2) is independent and is not affected. Please give a complete proof or a precise theorem with hypotheses covering this co","section":"Appendix B, Lemma B.1; Methods 4.4"},{"comment":"The QBP stability transfer is a second load-bearing step for Theorem 1. The lemma states exponential decay ∥η−η_R∥ ≤ Ce^{−μR}, the support/bandwidth property, and the moment bound (E.4), but it does not identify the QBP locality theorem being used or verify its hypotheses (finite range, high temperature, uniform constants independent of N) for the β interval of the theorem. Since the proof chooses R = q log N and needs n²e^{−μR} to be negligible, uniformity of μ matters. This is standard technology, but it should be made explicit before Theorem 1 can be accepted.","section":"Appendix E, Lemma E.1"}],"minor_comments":[{"comment":"The stated Hamiltonian H_dim = (Δ/2)∑P_r^{(1)} gives a triplet gap Δ/2, so the Boltzmann weight should be e^{−βΔ/2} unless Δ is defined as twice the physical gap. Please reconcile with y = e^{−βΔ} and with the crossover scale in Eq. (21).","section":"Sec. 2.2, Eq. (8) and Eq. (12)"},{"comment":"N = 4L in the dimer model, so replacing log2 L by log2 N shifts the additive constant by 1. The O(1) claim in the theorem is unaffected, but the displayed constant C(T) should be stated in one variable.","section":"Eq. (12) and Abstract"},{"comment":"The phrase \"universal Bessel-function crossover\" refers to the exact dimer model, not to all chains in Theorem 1. Please make this explicit to avoid overgeneralizing.","section":"Abstract and Sec. 2.3"},{"comment":"The identity ∑_j j(j+1)|a_j|² = ∫ ||∇f||² is used without derivation; a one-line integration-by-parts/Parseval statement would help.","section":"Sec. 4.5, Eq. (41)"}],"recommendation":"major_revision","confidential_remarks":"To the editor: The dimer-chain results are solid and the general mechanism is compelling. My recommendation of major_revision is driven by the sketch-level proof of Lemma B.1, which is the sole support for Theorem 1's universality. If the author supplies a full derivation (or states the lemma as an assumption and proves the theorem conditional on it), and fixes the small notational inconsistencies, I would be willing to accept. I do not see a fundamental error; the issue is completeness."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's my take. The dimer chain is the real content: exact E_D = E_F = Y_N at every finite size, zero at T=0, then 1/2 log_2 N + C(T) at any fixed T>0, with a closed Bessel crossover T_*(N)=Delta/(ln N + O(1)). I checked the beta=0 limit and the saddle calculation; they're consistent, and the recursion in Eq. (46) is checkable. That is a genuine counterexample to 'heating kills entanglement' inside a fixed symmetry sector, and the protocol is simple: measure the half-chain spin label, discard multiplicities, keep the singlet vector. The paper correctly credits [14,15] for the representation-space distillation bound; the new part is the finite-temperature activation and the exact model.\n\nThe general Theorem 1 is where I get cautious. The proof rests on Lemma B.1, the uniform complex-twist cluster expansion. That lemma does all the work: it gives the Gaussian saddle around the two central elements, the n^{-3/2} normalization, the cubic low-spin tail, and hence the 1/2 log N coefficient. But the appendix only sketches it, citing [17,18] and asserting the uniform derivative bounds in (B.5)-(B.6). The non-Hermitian twist i z·M_n changes the analyticity domain and interaction norm; those bounds are exactly what needs proving. Lemma E.1 (QBP stability transfer) is similarly presented at outline level, citing [19,20]. I am not saying the theorem is false — the mechanism is physically sensible and the dimer result secures the phenomenon — but as written Theorem 1 is an outline-level claim, not a fully proved theorem. The reader's 'conditional' verdict is about right.\n\nNothing wrong with the numerics apart from being limited to N<=16; the ED code being 'available upon request' rather than archived is a small but fixable issue. The citation pattern is honest.\n\nWho is this for? Anyone working on symmetry-resolved entanglement or thermal LOCC distillation. The exact dimer result deserves a serious referee; the general theorem needs the missing lemmas filled in before I'd call it proven. I'd send it to peer review and require either full derivations of Lemmas B.1 and E.1 or an explicit 'conjecture' label for Theorem 1.","headline":"Exact dimer-chain activation is solid and elegant; the universal theorem is a well-argued claim resting on a sketched technical lemma that needs either a full proof or a conjecture label.","tokens_in":19593,"tokens_out":3893,"would_cite":true,"duration_ms":37021,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["03.67.Mn","05.30.-d"],"model":"deepseek-v4-flash","headline":"This paper shows that heating, when confined to a non-Abelian strong-symmetry sector, can produce distillable entanglement that diverges with system size, starting from a zero-temperature state with no entanglement across the cut.","keywords":["non-Abelian strong symmetry","thermal entanglement","distillable entanglement","SU(2) singlet sector","representation-space distillation","quantum belief propagation","spin chain","Bessel crossover"],"falsifier":"In the dimer chain, start from the factorized singlet-dimer ground state, heat within the total-spin-zero sector to a fixed T > 0, and measure the half-chain spin labels via a Schur transform: the yield should grow as 1/2 log2 N + C(T). A plateau or saturation in E_D as N increases would refute the activation claim. For the generic theorem, exact diagonalization of a small SU(2)-invariant chain can test whether the cumulative distribution of j/sqrt(N) approaches an order-one limiting shape and whether Y_N - 1/2 log2 N stays bounded; a spin distribution width that collapses to O(1) or a residua","tokens_in":18508,"feed_emoji":"🔥","tokens_out":7060,"duration_ms":68803,"temperature":0.7,"pith_summary":"The paper claims that thermalization inside a fixed non-Abelian charge sector can create distillable entanglement rather than destroy it. In an exactly solvable dimer chain whose ground state has zero entanglement across the central bipartition, every fixed positive temperature gives E_D = 1/2 log2 N + C(T) + o(1), so the resource grows without bound as the chain grows. The mechanism is that thermal fluctuations give each half-chain a total spin of order sqrt(N), the global singlet constraint forces both halves to share the same spin-j representation, and the representation dimension 2j+1 converts directly into ebits when the halves measure their spin labels. The paper also proves a general theorem: for finite-range, uniformly bounded, local SU(2)-invariant chains with reflection symmetry, the global-singlet thermal state supports a representation-distillation yield of 1/2 log2 N + O(1) throughout a nonzero high-temperature interval. If correct, this overturns the usual intuition that heating can only degrade entanglement, and it gives an operational LOCC protocol for extracting unbounded entanglement from a thermal equilibrium state.","feed_headline":"Heating an unentangled spin chain yields boundless ebits","feed_subtitle":"Sector-constrained heat turns thermal noise into a distillable resource that grows with system size.","key_machinery":"The carrying object is the representation-space distillation identity: for any invariant-subspace state, E_D >= sum_lambda p_lambda log2(d_lambda), where p_lambda is the probability that both halves carry representation lambda and d_lambda = dim V_lambda. For SU(2), d_j = 2j+1, so the yield is Y_N = sum_j p_j log2(2j+1). The thermodynamic input is a uniform complex-twist cluster expansion (Lemma B.1): log Tr exp(-beta H_A + i X . M_n)/Z_A = n phi_beta(X) + b_n,beta(X) with derivatives through third order bounded uniformly in n. This produces Gaussian saddles around the two central elements of SU(2), giving a half-chain spin distribution of width sqrt(N), a cubic low-spin tail, and a linear C","core_discovery":"The central discovery is that sector-constrained thermal fluctuations become a distillable resource. For a state supported in the invariant subspace of a compact group, the yield Y_N = sum_j p_j log2(2j+1) is always a lower bound on distillable entanglement, because the unique invariant vector in each matched representation pair is maximally entangled with Schmidt rank 2j+1. In the dimer chain this bound is tight at every finite size: E_D = E_F = Y_N exactly, with E_D = 0 at T=0 and E_D = 1/2 log2 N + 1/2 log2[2 kappa(beta)] + (1 - gamma/2)/ln 2 + o(1) at every fixed T>0. The general theorem extends the leading coefficient to a broad class of local SU(2)-invariant chains at sufficiently high","pith_inferences":["The paper does not claim this, but the same sector-constrained mechanism plausibly extends to higher-rank compact groups, where the Weyl dimension formula would replace 2j+1 and could change the universal prefactor from 1/2 to a rank-dependent constant.","A testable experimental probe in a simulator with SU(2) symmetry would be to prepare a global singlet, couple it weakly to a symmetry-preserving heat bath, and Schur-sample both halves; the predicted rescaled distribution of j/sqrt(N) should converge to h(u) = 4 u^2 e^{-u^2}/sqrt(pi) in the dimer case.","If the protocol is repeated on many copies, the distilled ebits could serve as a private-key resource, linking this thermal-entanglement effect to quantum cryptographic tasks; the paper does not discuss that connection.","The exact dimer equality E_D = E_F = Y_N may be special to the decoupled structure; in generic chains the lower bound E_D >= Y_N is likely the robust statement, and the entanglement of formation could exceed the representation yield."],"forward_implications":["In the dimer chain the limits do not commute: fixing any T > 0 and increasing N gives unbounded distillable entanglement, while taking T -> 0 first leaves E_D = 0.","For a broad class of local SU(2)-invariant chains, global-singlet thermal states contain at least ~1/2 log2 N distillable ebits across an equal bipartition throughout a nonzero high-temperature window, independent of microscopic coupling details.","The resource is operationally extractable: local measurement of the half-chain spin label followed by discarding the multiplicity spaces is an LOCC distillation protocol requiring no knowledge of the Hamiltonian.","The finite-size onset in the dimer chain is controlled by the activity x = L e^{-Delta/T}, so an order-one resource appears already at T_*(N) ~ Delta / ln N, i.e., at temperatures exponentially small relative to the gap.","Abelian strong symmetries cannot produce this effect, because their irreducible representations have dimension one; the growing Schmidt rank comes specifically from the non-Abelian dimension 2j+1."],"fun_headline_variants":["Heat creates divergent distillable entanglement in a spin chain","Thermalizing an unentangled chain yields ebits that grow with size","Sector-constrained heat turns thermal noise into unbounded ebits","Heating a symmetry sector generates a divergent entanglement resource","Unentangled at zero temperature, heated chain yields unbounded ebits"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The broad theorem rests on a high-temperature cluster expansion that stays uniformly convergent when the half-chain is twisted by a complex SU(2) rotation; if that uniformity fails, the universal 1/2 log2 N coefficient for generic local chains is unproven (the exactly solved dimer chain does not depend on this assumption).","fun_headline_variants_meta":{"raw":{"variants":["Heat creates divergent distillable entanglement in a spin chain","Thermalizing an unentangled chain yields ebits that grow with size","Sector-constrained heat turns thermal noise into unbounded ebits","Heating a symmetry sector generates a divergent entanglement resource","Unentangled at zero temperature, heated chain yields unbounded ebits"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000245,"raw_usage":{"total_tokens":1402,"prompt_tokens":801,"completion_tokens":601,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":545,"completion_tokens_details":{"reasoning_tokens":513}},"tokens_in":545,"tokens_out":601,"duration_ms":6483,"temperature":1.0,"reasoning_tokens":513,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-02T06:24:21.404750+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"In the dimer chain, start from the factorized singlet-dimer ground state, heat within the total-spin-zero sector to a fixed T > 0, and measure the half-chain spin labels via a Schur transform: the yield should grow as 1/2 log2 N + C(T). A plateau or saturation in E_D as N increases would refute the activation claim. For the generic theorem, exact diagonalization of a small SU(2)-invariant chain can test whether the cumulative distribution of j/sqrt(N) approaches an order-one limiting shape and whether Y_N - 1/2 log2 N stays bounded; a spin distribution width that collapses to O(1) or a residua","supporting_citations":[],"review_version":2}