{"id":"c1beb716-db78-42dd-9c44-f03ea457bb9f","arxiv_id":"2504.17845","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"In the Creutz ladder, eta-pairing states remain identifiable with logarithmic entanglement entropy even at band touching, but interband coupling breaks the exact tower structure and renormalizes the pairing energy.","lead":"This paper studies how the exact eta-pairing states of the Hubbard model are deformed when a flatband is coupled to other bands, using the Creutz ladder as a test bed. It finds the logarithmic entanglement entropy signature survives interband coupling, while energy spacings and doublon shapes acquire corrections that grow as the band gap shrinks.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 'modified η-pairing states' in Figs. 3–4 are never defined by a selection rule; without it, the log-EE and multi-η claims cannot be tied to η-pairing.","rationale":"I read the paper as attempting to show that η-pairing-like states survive in a flatband model with interband coupling, and that their entanglement entropy retains an approximately logarithmic form. The small-system exact-diagonalization data are real evidence, and the Creutz ladder is a sensible concrete model for this question. The most fragile link in the argument is the operational definition of the 'modified η-pairing state,' exactly as the reader identified. The Schrieffer-Wolff calculation provides an energy scale and suggests how the SGA is deformed, but it does not specify which eigenstates constitute the deformed η tower; the numerical plots therefore need an independent, reproducible state-selection rule. The reader's conditional verdict is appropriate: the paper should not be rejected outright, but the central numerical claims cannot be accepted unconditionally without stating and testing the selection criterion.","tokens_in":11544,"tokens_out":26549,"duration_ms":264125,"concrete_test":"For the L=12, t=1, U=-1 system used in Figs. 3 and 4, compute for n=1 and n=3 the normalized overlap O_n(ψ)=|⟨ψ|(η̄†)^n|0⟩| / ||(η̄†)^n|0⟩|| for every exact eigenstate |ψ⟩ in the N=2n sector, and check that the state actually plotted in Fig. 3(a) and Fig. 4(b) is the one with the largest O_n. Also verify that O_n→1 as t′→∞. If the plotted state is not the maximum-overlap state, or if O_n does not approach 1 in the large-gap limit, the identification of 'modified η-pairing states' is not established; if it is, the numerical claims become reproducible.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section IV states that the entanglement entropy of the 'modified η-pairing state' exhibits approximately logarithmic scaling, and Section V draws conclusions about doublon-doublon interactions from the same states, but the manuscript never specifies how these states are selected from the exact-diagonalization spectrum. This matters because for finite t′ the operator η̄† no longer generates exact eigenstates; the tower (η̄†)^n|vac⟩ is deformed and some operational criterion, for example maximum overlap with (η̄†)^n|vac⟩, is required to identify the deformed tower. The statement in Section III that the modified states converge to the exact η-pairing states as Δ→∞ is asserted but not demonstrated, and the Schrieffer-Wolff analysis alone cannot label eigenstates. Without this selection rule, the approximately logarithmic curves in Fig. 3(a) and the multi-η deviations in Fig. 4(b) could belong to generic low-entanglement eigenstates selected by an unspecified procedure, making the central claim unfalsifiable and unreproducible.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript studies the fate of η-pairing states in the Creutz ladder with an on-site Hubbard interaction, where the flat band touches a dispersive band. Using a Schrieffer–Wolff transformation, the authors derive an effective flatband Hamiltonian with a renormalized interaction U_eff = U − U²/(2Δ) and propose a modified spectrum generating algebra [H_eff, η̄†] = (U_eff/2 − 2μ)η̄†. Exact-diagonalization results for a single η pair show approximately logarithmic entanglement entropy that persists even at band touching, along with a spatially broadened doublon correlation function. For multiple η pairs, the paper reports repulsive doublon–doublon correlations and deviations from the exact η-pairing tower, while the entanglement entropy remains lower than and approximately logarithmic. The central claims are that interband coupling deforms but does not destroy η-pairing signatures, and that the SW analysis quantitatively captures the leading corrections.","tokens_in":11744,"tokens_out":2219,"duration_ms":22643,"significance":"If substantiated, the paper would provide a concrete, parameter-free prediction for how band touching modifies η-pairing physics in a flatband system, connecting the exact η-pairing literature to realistic flatband models with band crossings. The analytical SW result is tested against exact diagonalization without parameter fitting, which is a strength. The work also speaks to the robustness of quantum many-body scars and off-diagonal long-range order in multiband systems. However, the lack of an operational definition of the 'modified η-pairing states' used in the entanglement and correlation calculations currently leaves the main numerical claims ambiguous; this issue must be resolved before the significance can be fully assessed.","major_comments":[{"comment":"The manuscript never specifies how the 'modified η-pairing states' are selected among the exact eigenstates. For finite t′ the operator η̄† does not generate exact eigenstates, so some criterion must be given—for example, maximal overlap with (η̄†)^n|vac⟩ or selection by energy spacing—to identify the deformed tower. Without such a rule, the approximately logarithmic curves in Fig. 3(a) and Fig. 4(b) cannot be tied to η-pairing deformation, and the multi-η conclusions in Section V are unfalsifiable. Please state the selection rule explicitly and, if possible, quantify the overlap of the selected states with the unperturbed η-pairing tower for several values of Δ.","section":"§IV and §V (Figs. 3 and 4)"},{"comment":"The second-order Schrieffer-Wolff correction in Eq. (20) is asserted with a single line and no derivation of the resulting energy shift ε = U/2 − U²/(4Δ) shown in Fig. 2. The projection onto the flatband subspace is nontrivial because the projected operators satisfy the nonlocal anticommutation relations in Eq. (10), and the reduction from the double sum over i,j,α,β to a local density-density term requires justification. Please provide the intermediate steps, including how the complementary-space operators ˜c are eliminated and how the SGA commutator in Eq. (21) follows from the full projected algebra.","section":"§III, Eq. (20) and Fig. 2"},{"comment":"The SW expansion is justified in the t ≪ t′ limit, but Fig. 3(a) claims logarithmic scaling 'even in the band-touching limit (t′ = 2)' with t = 1, where Δ is not large and the perturbative regime does not apply. The statement in Section III that the modified states converge to the exact η-pairing states as Δ → ∞ is asserted but not demonstrated. Please provide a quantitative convergence check (e.g., fidelity or energy spacing versus 1/Δ) and clarify which of the claims rely on the perturbative regime versus on the numerical selection rule.","section":"§III and §IV (band-touching claim)"}],"minor_comments":[{"comment":"There are numerous typos: 'moel' for model (p. 2), 'deontes' for denotes (p. 3), 'satify' for satisfy (p. 4), 'Schriffer-Wolff' in the Section II.C title, and the spelling of 'deontes' in the text around Eq. (14). Please proofread carefully.","section":"Throughout"},{"comment":"Several references are duplicated, including [18] and [19] (both Serbyn et al.), [21] and [22] (both Moudgalya et al.), and [23] (P. Sala et al., a third copy). Please consolidate the bibliography.","section":"References"},{"comment":"The hypergeometric distribution in Eq. (5) and the asymptotic formula in Eq. (6) are given without derivation or citation to the original source in [40–42]; a sentence explaining the quasiparticle picture would help readers.","section":"Eq. (5) and Eq. (6)"},{"comment":"The caption states 'for various values of t′ when t = 0' and then 'for t = 0.5' in the inset description. Please clarify the parameter values used for each panel and whether the analytical curve is the same in both.","section":"Fig. 2 caption"},{"comment":"The doublon creation operator D†_{i,j} sums over α,β,γ,δ without excluding the case where the two doublons share the same unit cell or the same spin; please specify whether i ≠ j and whether the sum runs over all spin combinations or only distinct ones.","section":"Eq. (24)"}],"recommendation":"major_revision","confidential_remarks":"The core idea is interesting and the SW prediction is a useful parameter-free result, but the missing selection rule for the modified η-pairing states is a substantive gap that compromises the main numerical claims. The paper is likely to be publishable after the authors clearly define how the states are chosen and provide the omitted SW derivation. I would not reject but would request a careful revision."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper is worth a serious look, but not in its current form. What it does well: it shows that eta-pairing-like states with logarithmic entanglement entropy persist even when the flatband is not isolated, and it quantifies the deformation via a Schrieffer-Wolff correction to the spectrum generating algebra. The small-system exact diagonalization data in Figs. 2–4 look credible, and the agreement between the numerical symbols and the analytical energy shift ϵ = U/2 − U^2/(4Δ) in Fig. 2(a) is a good sign. The broadening of the doublon correlation function and the repulsive interaction between multiple eta pairs are also new and physically sensible.\n\nThe soft spots are real. The biggest one is the missing selection rule. Sections IV and V talk about \"the modified eta-pairing state\" without ever saying how it is picked from the spectrum. For finite t′ the tower (η̄†)^n|vac⟩ is deformed and is not exactly an eigenstate, so the curves in Fig. 3(a) and Fig. 4(b) could belong to any low-entanglement state. A maximum-overlap criterion with the η̄-tower would make the claim reproducible; as written, it is not. This is load-bearing, not cosmetic.\n\nSecond, the Schrieffer-Wolff step in Eq. (20) is asserted with a one-line jump to −U^2/(4Δ) Σ n̄ n̄, but the claimed SGA then uses U_eff = U − U^2/(2Δ), which is a factor of two off. Either the derivation or the SGA is wrong. The numerical agreement with the analytical shift suggests the SGA is correct, so Eq. (20) likely needs a factor of two. This needs to be fixed explicitly.\n\nThird, the statement in Section V that the multi-η states \"converge in the large band gap limit\" but \"do not coincide\" with the exact η-pairing states is confusing, since convergence usually implies coincidence. I think the authors mean the states approach but never exactly match for finite gap, but as written it is self-contradictory.\n\nThese are fixable. The paper's framing is honest, the physics is plausible, and the core observation—log EE surviving band touching—is worth publishing if the authors tighten the state identification and correct the SW factor. I would send it to peer review and ask the referee to demand both. For my own work, I would not cite it until those are sorted out.","headline":"Eta-pairing-like states likely survive band touching in a Creutz ladder, but the paper's central claim is under-specified because the 'modified eta-pairing state' is never defined by a selection rule, and the Schrieffer-Wolff derivation has a factor-of-two discrepancy that needs checking.","tokens_in":12266,"tokens_out":5091,"would_cite":false,"duration_ms":44490,"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":"Even when a flat band touches a dispersive band, a subset of eigenstates keeps near-logarithmic entanglement entropy and a deformed η-pairing tower, with energy spacing renormalized by virtual interband processes.","keywords":["eta-pairing","flat bands","entanglement entropy","Creutz ladder","many-body scars","spectrum generating algebra","Schrieffer-Wolff transformation","off-diagonal long-range order"],"falsifier":"Compute the overlap between each candidate eigenstate in the exact spectrum and the states $\\bar\\eta^{\\dagger n}|\\Omega\\rangle$ (or the SW-improved tower) at the parameters of Figs. 3 and 4. If the plotted logarithmic entanglement curves come from eigenstates with negligible overlap with the deformed η tower, the identification is wrong. A second check: measure the exponential decay length of $C_2(d)$ and compare it quantitatively with the SW prediction as a function of $\\Delta$ and $U$; a mismatch would indicate that the second-order algebra does not control the entanglement behavior.","tokens_in":11350,"feed_emoji":"⚛️","tokens_out":7369,"duration_ms":63473,"temperature":0.7,"pith_summary":"The paper asks whether η-pairing, the exact eigenstate tower of the Hubbard model that displays off-diagonal long-range order and logarithmic entanglement entropy, survives in flatband systems where the flat band touches a dispersive band. Using the Creutz ladder, it claims that a distinct subset of eigenstates retains η-like behavior even in the band-touching limit: the entanglement entropy stays approximately logarithmic and the doublon pair order remains confined, though the doublon wavefunction spreads with an exponentially decaying tail. A Schrieffer–Wolff transformation quantifies the deformation: the effective interaction becomes $U_{\\rm eff} = U - U^2/(2\\Delta)$, and the spectrum generating algebra is modified to $[H_{\\rm eff}, \\bar\\eta^\\dagger] = (U_{\\rm eff}/2 - 2\\mu)\\bar\\eta^\\dagger$. For multiple η pairs, broadened doublons repel, so the ideal equal-spacing tower is lost except at $t = 0$, yet the logarithmic entanglement scaling persists. If right, strict flatband isolation is not a prerequisite for robust pairing signatures, which is relevant to cold-atom and designer flatband experiments.","feed_headline":"Eta pairing survives band touching with log-law entanglement","feed_subtitle":"Creutz-ladder numerics find deformed eta towers with renormalized spacing and persistent logarithmic entropy.","key_machinery":"The central object is the compact localized state (CLS) of the Creutz ladder and its projected fermion operator $\\bar c_{i,\\alpha,\\sigma} = (c_{i,A,\\sigma} - c_{i,B,\\sigma})/2$, which localizes the flatband degrees of freedom to single rungs. Acting on this, the Schrieffer–Wolff transformation systematically removes the coupling between the flatband subspace and the dispersive band, producing an effective Hamiltonian whose second-order term renormalizes the Hubbard $U$ to $U_{\\rm eff}$. The modified spectrum generating algebra $[H_{\\rm eff}, \\bar\\eta^\\dagger] = (U_{\\rm eff}/2 - 2\\mu)\\bar\\eta^\\dagger$ then does the explanatory work: it predicts the (approximately) equally spaced energies, the logarithmic entanglement entropy of the single-pair states, and the interaction-induced deviations in the multi-pair case.","core_discovery":"On the paper's own terms, the discovery is the deformation of an exact symmetry into a useful approximate one. In the Creutz ladder, the flatband projection makes $\\bar\\eta^\\dagger = \\sum_i \\bar c^\\dagger_{i,\\uparrow}\\bar c^\\dagger_{i,\\downarrow}$ an exact generator only when the dispersive band is infinitely far away. For finite gap $\\Delta$, virtual tunneling adds a second-order term $-U^2/(4\\Delta)\\sum \\bar n_{i,\\downarrow}\\bar n_{i,\\uparrow}$, so the interaction strength is renormalized to $U_{\\rm eff} = U - U^2/(2\\Delta)$ and the commutation relation $[H_{\\rm eff}, \\bar\\eta^\\dagger] = (U_{\\rm eff}/2 - 2\\mu)\\bar\\eta^\\dagger$ holds approximately. The paper shows numerically that the resulting modified η states follow the same logarithmic entanglement entropy curve as the exact η states, with deviations that grow as interband coupling strengthens, and that the pair correlation $C_2(d)$ develops an exponential tail that shrinks as $t'$ grows. In the many-pair sector the story is different: the pairs repel via $C_4(d)$, the tower spacing becomes nonuniform, and the multi-pair states never converge to the exact ones at finite gap except for $t=0$, even though their entanglement entropy retains a log law.","pith_inferences":["The same SW construction should apply to other band-touching flatband lattices such as Lieb and Kagome; an immediate check is whether their modified η towers obey the same $U_{\\rm eff} = U - U^2/(2\\Delta)$ renormalization with the appropriate gap.","Because the paper does not specify how the 'modified η-pairing states' are selected from the exact spectrum, the safest read is that its quantitative claims apply to that particular unidentified subset; a companion study that defines and tests the overlap selection rule would place the results on firmer ground.","If the identified states are indeed the most η-like, they should also be scar-like, meaning their nonthermal signatures (low entropy, confined correlations) should survive time evolution after a quench; this is a testable prediction the paper leaves implicit."],"forward_implications":["Single η-pair modified states show logarithmic entanglement entropy even when the flat band touches the dispersive band ($t'=2$), so log-law EE is not by itself evidence of exact η-pairing symmetry.","For $|U|\\ll\\Delta$, the energy tower is approximately equally spaced, so spectroscopic measurements of the pair-binding energy should see the SW renormalized value $U_{\\rm eff}/2 - 2\\mu$ rather than the bare value.","The doublon correlation length grows as the gap decreases; measuring $C_2(d)$ in cold atoms would reveal the exponential tail and its recovery at large $t'$.","With two or more η pairs, doublons repel regardless of the sign of $U$, so interacting multi-pair states deviate from the free-boson picture; this limits the validity of simple η-pair condensation scenarios in realistic flatband systems."],"supporting_citations":[{"why":"defines the η-pairing state and off-diagonal long-range order in the Hubbard model, the starting point of the paper.","marker":"[26]"},{"why":"establishes emergent SU(2) symmetry in isolated flat bands, the ideal limit that the paper extends to coupled bands.","marker":"[27]"},{"why":"classifies band-touching singularities of flat bands, justifying why the Creutz-ladder band touching spoils the exact symmetry.","marker":"[30]"},{"why":"formulates η-pairing through a spectrum generating algebra and its connection to many-body scars.","marker":"[38]"},{"why":"provides the reduced density matrix representation used to derive the entanglement entropy of η-pairing states.","marker":"[40]"},{"why":"gives the analytic logarithmic entanglement entropy formula for exact η-pairing states that the numerics compare against.","marker":"[42]"},{"why":"supplies the Schrieffer-Wolff transformation framework used to derive the effective Hamiltonian and modified spectrum generating algebra.","marker":"[43]"}],"fun_headline_variants":["Eta pairing survives band touching, entropy stays log","Interband coupling deforms eta-pairing, entropy stays log","Creutz ladder: eta-pairing modified, entanglement still log","Approximate eta-pairing shows logarithmic entanglement","Eta-pairing deformation yields log-law entanglement"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument depends on being able to pick out the 'modified η-pairing states' from the exact eigenstate spectrum, but the paper never states the selection rule, such as maximum overlap with $\\bar\\eta^{\\dagger n}$ acting on a reference state, used to produce the entanglement and correlation data.","fun_headline_variants_meta":{"raw":{"variants":["Eta pairing survives band touching, entropy stays log","Interband coupling deforms eta-pairing, entropy stays log","Creutz ladder: eta-pairing modified, entanglement still log","Approximate eta-pairing shows logarithmic entanglement","Eta-pairing deformation yields log-law entanglement"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000647,"raw_usage":{"total_tokens":3011,"prompt_tokens":1024,"completion_tokens":1987,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":640,"completion_tokens_details":{"reasoning_tokens":1906}},"tokens_in":640,"tokens_out":1987,"duration_ms":15973,"temperature":1.0,"reasoning_tokens":1906,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T10:32:17.087237+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the overlap between each candidate eigenstate in the exact spectrum and the states $\\bar\\eta^{\\dagger n}|\\Omega\\rangle$ (or the SW-improved tower) at the parameters of Figs. 3 and 4. If the plotted logarithmic entanglement curves come from eigenstates with negligible overlap with the deformed η tower, the identification is wrong. A second check: measure the exponential decay length of $C_2(d)$ and compare it quantitatively with the SW prediction as a function of $\\Delta$ and $U$; a mismatch would indicate that the second-order algebra does not control the entanglement behavior.","supporting_citations":[{"cited_title":"Rhim and B.-J","cited_arxiv_id":null,"evidence_quote":"classifies band-touching singularities of flat bands, justifying why the Creutz-ladder band touching spoils the exact symmetry."},{"cited_title":"Fan and S","cited_arxiv_id":null,"evidence_quote":"provides the reduced density matrix representation used to derive the entanglement entropy of η-pairing states."},{"cited_title":"Vafek, N","cited_arxiv_id":null,"evidence_quote":"gives the analytic logarithmic entanglement entropy formula for exact η-pairing states that the numerics compare against."},{"cited_title":"Bravyi, D","cited_arxiv_id":null,"evidence_quote":"supplies the Schrieffer-Wolff transformation framework used to derive the effective Hamiltonian and modified spectrum generating algebra."}],"review_version":1}