{"id":"084ae5ba-b9a2-429a-92a9-f5c812dbe88d","arxiv_id":"2507.20679","paper_version":1,"verdict":"REJECT","confidence":"MODERATE","novelty_score":2.0,"correctness_risk":"high","formal_verification":"none","parameter_count":0,"one_line_summary":"The paper asserts that domain-of-derivative corrections restore nonzero Berry curvature and modify the adiabatic theorem for parameter-independent Hamiltonians, but the derivation is flawed.","lead":"This paper claims that Berry curvature can be nonzero even when the Hamiltonian has no explicit dependence on the parameter, if one accounts for boundary terms from operator domains. The derivations contain sign errors and rely on ill-defined integrals, so the central claim is not supported.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The central construction is not sound: the paper's key formula for the corrected off-diagonal matrix element is asserted without a valid derivation, and its own free-particle and gauge-invariance checks are internally inconsistent.","rationale":"I agree with the reader's REJECT verdict. The reader's weakest_assumption correctly identifies Eq. 12 and Eq. 16 as the load-bearing step. My stress-test confirms and sharpens the concern: the paper's own examples are internally inconsistent. The free-particle calculation uses divergent integrals; the gauge-transformation argument contains a sign inconsistency in Eqs. 27-28; and the final curvature formula is never shown to equal the Berry curvature. The paper's central claim is therefore unsupported. I do not find any independent evidence that would rescue the argument: no machine-checked proof, no parameter-free derivation, and no reproducible numerical check. The verdict should remain REJECT.","tokens_in":13165,"tokens_out":1731,"duration_ms":19857,"concrete_test":"Re-derive Eq. 16 from first principles under explicit domain assumptions: either (a) compute <m|nabla_R H|n> = epsilon_m <m|nabla_R|n> + <H m|nabla_R n> and check whether the result reproduces Eq. 16; if <H m|nabla_R n> = epsilon_m <m|nabla_R n> is forced, then Delta_m,n vanishes and the central formula is an identity. Or (b) regularize the free-particle calculation by replacing the plane waves with a Gaussian wavepacket or an L^3 box, and check whether Eq. 33 and Eq. 37 agree after the divergent pieces are regulated; if they agree only after discarding equal and opposite infinite terms, the computation is not a legitimate derivation. Or (c) independently recompute the gauge transformation of Eq. 12 as written and check whether the sign in Eq. 27-28 holds.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is that the domain correction Delta_m,n (Eq. 12) restores nonzero Berry curvature and modifies the adiabatic theorem even when the Hamiltonian has no explicit parameter dependence. The load-bearing step is Eq. 16, which asserts that <m|nabla_R|n> = (<m|nabla_R H|n> + Delta_m,n)/(epsilon_n - epsilon_m). This formula is never derived from a well-defined operator-domain prescription. In fact, Eq. 12 is the difference of two expressions that are equal whenever both are well defined under the stated Hermiticity assumptions, so it is identically zero in that regime; the paper does not prove it is nonzero in any concrete, well-defined Hilbert space. The free-particle check (Eqs. 35-37) evaluates integrals of the form int d^3r vec{r} e^{i(k-k')·r}, which are divergent or distributional; the cancellation of (k^2 - k'^2) against the divergent integral is not a legitimate finite operation. The gauge-transformation step (Eqs. 26-28) also contains a sign inconsistency: the same derivation applied verbatim to Eq. 12 yields <m|H nabla n> - epsilon_m <m|nabla n>, which does not equal the expression stated in Eq. 28 unless one assumes the very domain condition that makes Delta nonzero. Finally, the Bloch curvature formula (Eqs. 49-50) is never shown to equal the standard Berry curvature nabla x A; the paper asserts the restoration of nonzero curvature rather than demonstrating it. Because the paper's own consistency checks fail, the central claim is unsupported.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper claims that when the parameter derivative of an eigenstate, |∇_R n⟩, lies outside the domain of a Hermitian Hamiltonian H, the conventional relation between off-diagonal matrix elements and Hamiltonian derivatives acquires a boundary correction Δ_{m,n} (Eq. 12). This correction is inserted into the adiabatic transition amplitude in Eq. (16), leading to a modified Berry curvature (Eq. 17) and a modified adiabatic condition (Eq. 19). The authors apply the construction to a free particle and to Bloch solids, asserting that it restores nonzero Berry curvature even when the Hamiltonian has no explicit parameter dependence, and they discuss gauge invariance in Section 3.","tokens_in":13502,"tokens_out":8654,"duration_ms":100273,"significance":"If the central construction were valid, the paper would address a real subtlety in the operator-domain treatment of Berry phases and adiabatic transport in periodic solids. The authors correctly identify that in Bloch theory the crystal momentum enters through boundary conditions and operator domains, not only through an explicit k-dependent Hamiltonian. However, the manuscript as it stands does not establish its central claim: the defining quantity Δ is either undefined or identically zero under the paper's own assumptions, the gauge-invariance check contains a sign error, the free-particle consistency check uses invalid distributional manipulations, and the final Bloch curvature formula is asserted rather than derived from the standard ∇×A definition. No concrete model with a well-defined Hilbert-space domain is provided, and the paper relies on the authors' prior work for a key vanishing result. The significance of the potential insight is therefore not matched by the rigor of the derivation.","major_comments":[{"comment":"The central correction Δ_{m,n} is not well defined under the paper's own hypotheses. Eq. (12) requires applying H to |∇_R n⟩, which the authors explicitly state may lie outside the domain of H; in that case the first term ⟨m|H|∇_R n⟩ is not defined. If |∇_R n⟩ does lie in the domain, Hermiticity gives ⟨m|H|∇_R n⟩ = ⟨Hm|∇_R n⟩, so Δ_{m,n}=0 by Eq. (12). The manuscript never constructs a concrete Hilbert space and operator domain in which Δ is both finite and nonzero. Since Eq. (16) is the load-bearing formula from which the modified Berry curvature and the revised adiabatic theorem follow, the central derivation is unsupported.","section":"Section 1, Eqs. (12) and (16)"},{"comment":"The gauge-transformation check contains a sign error. Under |n'⟩ = e^{iβ_n}|n⟩ and |m'⟩ = e^{iβ_m}|m⟩, with H|m⟩ = ε_m|m⟩, the bracket in Eq. (27) reduces to ⟨m|H∇n⟩ + ε_m⟨m|∇n⟩, which equals ⟨m|H∇n⟩ + ⟨Hm|∇n⟩. But the original correction in Eq. (12) is ⟨m|H∇n⟩ − ⟨Hm|∇n⟩. Equation (28) therefore identifies Δ with an expression of the opposite sign in the second term, so the claimed gauge covariance of Δ and the resulting gauge invariance of Ω_n in Eq. (30) are not established.","section":"Section 3, Eqs. (26)--(28)"},{"comment":"The free-particle check is not a legitimate calculation. The integral ∫ d³r r e^{i(k−k')·r} is a distribution (the gradient of a delta function), not an ordinary function. Equation (37) cancels the factor (k²−k'²) against this distribution, but (k²−k'²) vanishes on a codimension-one set, so the cancellation is not a well-defined finite operation. The apparent agreement between Eqs. (33) and (37) therefore does not validate Eq. (16); it instead illustrates that the formula is being manipulated outside its domain of validity.","section":"Section 4, Eqs. (33)--(37)"},{"comment":"The final Bloch-curvature formula is not derived. Even accepting Eq. (48), Eq. (49) simply writes Ω_n as a sum over products of Δ terms and never proves that this quantity equals the standard Berry curvature ∇×A. Moreover, the essential input that Δ^{H(k)}_{n',n}=0 in three dimensions is asserted in Appendix C by extrapolating a one-dimensional boundary-periodicity argument; the periodic cell functions u_{n,k} are not periodic in k, and no three-dimensional proof is given. The paper also defers this vanishing to Refs. [4,5] rather than demonstrating it here. Consequently, the central claim that the domain correction reproduces the correct Berry curvature in Bloch solids rests on an unproved assumption and an unshown equivalence.","section":"Section 5 and Appendix C, Eqs. (47)--(50)"}],"minor_comments":[{"comment":"The phrase 'Berry curvature (as given by Eq. (19))' is incorrect: Eq. (19) is the adiabaticity condition, not a curvature formula. The intended reference is likely Eq. (17) or Eq. (49).","section":"Section 5, text before Eq. (49)"},{"comment":"The text refers to 'the earlier result obtained in the 1D case (Eq. (23))', but Eq. (23) is a pair of terms in the divergence proof; the 1D result is Eq. (21).","section":"Section 5, after Eq. (47)"},{"comment":"The vector potential A appears with couplings e/c in Eq. (13), but in the later Bloch calculations, Eqs. (39) and (46), this term is dropped without comment. The relation between the two conventions should be stated explicitly.","section":"Section 1, Eq. (13)"},{"comment":"The notation for Δ is inconsistent: it is a vector in Eqs. (13)--(17) and (26)--(29), but a scalar in Eqs. (18)--(21). The manuscript should clarify which quantity is meant in each occurrence, especially in Eq. (21).","section":"Throughout"},{"comment":"There are several typographical errors: 'Therefore. the non-Hermitian term' has a misplaced period, and 'restores the the meaning' contains a duplicated article. In addition, calling Δ 'non-Hermitian' is misleading in light of the paper's own claim that Δ does not indicate non-Hermiticity of H.","section":"Section 3, paragraph after Eq. (24)"}],"recommendation":"reject","confidential_remarks":"The manuscript addresses a real and subtle topic, but the central construction is not sound and the internal consistency checks fail. The paper also relies heavily on the authors' own earlier work for a key vanishing result and does not benchmark against a concrete model with a well-defined domain. In my view, a publishable version would need to reformulate the argument in a rigorous domain-theoretic framework, provide a concrete example where Δ is nonzero and well defined, and demonstrate that the resulting curvature coincides with the standard Berry curvature. Those are substantial changes beyond a routine revision."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nThe short version: this paper has a real kernel of an observation, but the derivation does not hold up. The claim that domain subtleties can restore nonzero Berry curvature for parameter-independent Hamiltonians is not established, and the paper's own checks – the free-particle calculation and the gauge-transformation step – contain invalid manipulations.\n\nWhat is genuinely new is the attempt to formalize the boundary term Δ_{m,n} = <m|H|∇_R n> - <Hm|∇_R n> as a correction to the off-diagonal matrix element, and to apply it to the Bloch problem. That is a legitimate point: parameter derivatives of eigenstates can leave the domain of H, and a surface term can in principle matter. The final curvature formula (50) is also a recognizable rearrangement of the standard off-diagonal momentum/position expressions for Bloch bands, so the authors are pointing to a real expression.\n\nBut the load-bearing step, Eq. (16), is asserted rather than derived, and Δ as defined in Eq. (12) is zero whenever both terms are well-defined under Hermiticity. The paper never exhibits a concrete Hilbert space where Δ is genuinely nonzero. The free-particle test does the opposite of what the authors claim: integrals like ∫ d³r r e^{i(k-k')·r} are distributional, and canceling (k² - k'²) against a divergent integral is not legitimate. The gauge-transformation calculation in Eqs. (26)-(28) has a sign error: working through the transformation gives 2ε_m <m|∇n> in place of Δ, so the claimed gauge invariance is not proven. Finally, the vanishing of Δ in the H(k) formulation is imported from the authors' prior work, and the higher-dimensional extension of the periodicity argument is handwaving.\n\nThe paper is not a good candidate for peer review in its current form. The central construction is unsupported, and the internal consistency checks fail. At most, the observation about domain corrections could become a footnote in a rigorous treatment of adiabatic theory, but that would require a concrete example and a valid derivation.\n\nI would not cite it, and I would not send it to a referee unless the authors first fix the sign error and provide a well-posed example. The topic is real, but this paper does not advance it.\n\nBest,\n[Your name]","headline":"A well-intentioned but mathematically unsound attempt to reinterpret Berry curvature through operator-domain corrections; the central derivation collapses under its own consistency checks.","tokens_in":14003,"tokens_out":3894,"would_cite":false,"duration_ms":43452,"reading_group":"no","serious_thinker":"yes","would_accept_peer_review":false},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper argues that nonzero Berry curvature can arise even when the Hamiltonian has no explicit parameter dependence, because the curvature is carried by the eigenstates and appears through a boundary correction that standard…","keywords":["Berry curvature","adiabatic theorem","operator domains","Bloch solids","crystal momentum","surface terms","geometric phase"],"falsifier":"Compute the exact matrix element $\\langle m|\\nabla_k n\\rangle$ from numerically known eigenstates of a finite periodic potential and compare it with the right-hand side of the corrected formula using the surface integral for $\\Delta$; any disagreement would show that $\\Delta$ does not capture the domain mismatch. Alternatively, simulate adiabatic transport with a Hamiltonian that has no explicit parameter dependence and check whether the transition probabilities follow the corrected condition rather than the standard one.","tokens_in":12909,"feed_emoji":"🌀","tokens_out":5478,"duration_ms":60618,"temperature":0.7,"pith_summary":"This paper tries to establish that the standard derivation of Berry curvature silently assumes all parameter-differentiated eigenstates lie in the Hamiltonian's domain. When that fails, as it does for Bloch states in periodic solids, a boundary term must be added to the off-diagonal matrix elements, and this term restores nonzero Berry curvature even for a Hamiltonian with no explicit parameter dependence. The same correction changes the adiabatic theorem's no-transition condition. If this is right, Berry curvature is fundamentally a property of eigenstate geometry, not of the Hamiltonian's parameter dependence, and standard Bloch calculations that ignore domain subtleties are missing a contribution.","feed_headline":"Even parameter-free Hamiltonians can carry Berry curvature","feed_subtitle":"A boundary correction restores nonzero curvature and changes the adiabatic condition for Bloch electrons.","key_machinery":"The central object is the domain-correction term $\\Delta_{m,n} = \\langle m|H|\\nabla_R n\\rangle - \\langle Hm|\\nabla_R n\\rangle$, which measures the failure of the familiar Hermitian-transfer identity when $|\\nabla_R n\\rangle$ is not in the domain of $H$. In position representation it becomes the flux of a generalized current density through the boundary, a surface term that carries the entire argument. This term enters the off-diagonal matrix element, the Berry curvature, and the adiabatic condition, and its vanishing in the explicitly $k$-dependent formulation is what reconciles the two standard forms of the Bloch Hamiltonian.","core_discovery":"The central claim is that the identity used to derive the standard Berry curvature formula, $\\langle m|H|\\nabla_R n\\rangle = \\epsilon_m\\langle m|\\nabla_R n\\rangle$, is invalid whenever the parameter derivative of an eigenstate lies outside the domain of $H$. The paper defines the mismatch as $\\Delta_{m,n} = \\langle m|H|\\nabla_R n\\rangle - \\langle Hm|\\nabla_R n\\rangle$, shows it is a surface term of a generalized current density, and inserts it into the off-diagonal matrix element so that $\\langle m|\\nabla_R n\\rangle = (\\langle m|\\nabla_R H|n\\rangle + \\Delta_{m,n})/(\\epsilon_n - \\epsilon_m)$. The resulting Berry curvature is gauge invariant and satisfies $\\sum_n \\Omega_n = 0$, and it remains nonzero when $\\nabla_R H = 0$. Applied to Bloch solids, the correction reduces to a combination of off-diagonal momentum and position matrix elements, giving curvature even for $H = p^2/2m + V$, and the adiabatic condition becomes $|(\\langle m|\\dot H|n\\rangle + \\Delta_{m,n})/(\\epsilon_n - \\epsilon_m)| \\ll 1$.","pith_inferences":["This suggests that common tight-binding or $k\\cdot p$ schemes, which use an explicitly $k$-dependent $H(k)$, already incorporate the domain effect implicitly, while first-principles codes using the parameter-independent $H$ need an explicit boundary term to match.","Because $\\Delta$ is a surface term, finite-size or open-boundary systems may show enhanced geometric-phase corrections, making the effect observable in nanostructures or edge states.","The same correction logic should apply to other eigenstate-derivative quantities such as the quantum metric, the shift vector, or higher Chern numbers, where $\\nabla_R n$ appears.","One direct test: in a one-dimensional superlattice, slowly sweep $k(t)$ and measure interband transition rates; the corrected adiabatic condition predicts transitions even when the Hamiltonian is time-independent except through $k(t)$."],"forward_implications":["In Bloch solids, nonzero Berry curvature no longer requires explicit $k$-dependence in $H$; band-geometry calculations using only $\\nabla_k H$ are incomplete without the $\\Delta$ term.","Adiabatic transport can occur even when $\\langle m|\\dot H|n\\rangle = 0$, because $\\Delta$ supplies the off-diagonal coupling.","The two standard forms of the Bloch Hamiltonian, $H = p^2/2m + V$ and $H(k) = (p+\\hbar k)^2/2m + V$, are inequivalent at the level of operator domains, and Eq. (47) gives the exact difference.","The corrected curvature remains gauge invariant and satisfies the sum rule $\\sum_n \\Omega_n = 0$, so topological interpretations survive the correction.","For Bloch bands, curvature can be expressed solely through off-diagonal momentum and position matrix elements, which may simplify numerical evaluation."],"supporting_citations":[{"why":"Defines the geometric phase and the standard Berry connection and curvature that the paper extends.","marker":"[1]"},{"why":"Documents the controversy over sufficiency conditions for adiabaticity that motivates the revised adiabatic condition.","marker":"[3]"},{"why":"Earlier work by the authors introducing emergent non-Hermitian contributions, used here for the vanishing of the correction in the explicitly k-dependent formulation.","marker":"[4]"}],"fun_headline_variants":["Berry curvature from boundary terms, not parameters","Parameter-free Hamiltonians still bend Berry phase","Adiabatic theorem gets a boundary correction","No parameters? Berry curvature persists","Berry curvature fixed by domain boundaries"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument assumes that the parameter derivative of an eigenstate can fall outside the Hamiltonian's domain, and that the resulting mismatch is faithfully captured by the boundary term $\\Delta$; if $\\Delta$ is not well defined or does not represent the true action of $H$, the corrected curvature and adiabatic formulas do not follow.","fun_headline_variants_meta":{"raw":{"variants":["Berry curvature from boundary terms, not parameters","Parameter-free Hamiltonians still bend Berry phase","Adiabatic theorem gets a boundary correction","No parameters? Berry curvature persists","Berry curvature fixed by domain boundaries"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000146,"raw_usage":{"total_tokens":1151,"prompt_tokens":880,"completion_tokens":271,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":496,"completion_tokens_details":{"reasoning_tokens":209}},"tokens_in":496,"tokens_out":271,"duration_ms":3400,"temperature":1.0,"reasoning_tokens":209,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T13:22:40.106956+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the exact matrix element $\\langle m|\\nabla_k n\\rangle$ from numerically known eigenstates of a finite periodic potential and compare it with the right-hand side of the corrected formula using the surface integral for $\\Delta$; any disagreement would show that $\\Delta$ does not capture the domain mismatch. Alternatively, simulate adiabatic transport with a Hamiltonian that has no explicit parameter dependence and check whether the transition probabilities follow the corrected condition rather than the standard one.","supporting_citations":[{"cited_title":"Marzlin and B","cited_arxiv_id":null,"evidence_quote":"Documents the controversy over sufficiency conditions for adiabaticity that motivates the revised adiabatic condition."}],"review_version":1}