{"id":"0dd62d18-0bc6-482b-a599-b3f4422ff068","arxiv_id":"2607.23985","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"At non-time-reversal-invariant momenta, a band's gradient is symmetry-forced to zero exactly when the little group's vector representation contains no trivial representation; this rule classifies all space groups in the single-group limit.","lead":"The paper gives a symmetry rule for when electron bands at special non-time-reversal-invariant momenta have zero slope, turning them into Van Hove singularities. It predicts the rule from the little group alone and maps which space groups enforce, allow, or forbid such critical points.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Degenerate-band criterion relies on a direction-dependent subband gradient; the per-subband zero-component count in Table II is not well-defined when Γvec intersects [Γ⊗Γ]_sym in multiple components.","rationale":"The nondegenerate-band criterion is solid: for 1D reps the Wigner-Eckart reduction to Γvec ⊃ Γ1 is rigorous, and the tight-binding calculations at W/K for 1D bands are consistent with that prediction. The reader's weakest_assumption correctly identifies the degenerate-band subband gradient as the least secure part of the central claim. The paper's Table II degenerate entries are plausible and likely correct when interpreted via the dimension of the invariant subspace of H^(1), but the text does not prove this; instead it relies on a 'per-subband gradient vector' that is direction-dependent in exactly the multi-component cases that dominate the degenerate rows. This is a genuine rigor gap, not a manufactured objection. It does not overturn the nondegenerate result, but it justifies the CONDITIONAL verdict: the degenerate half requires either a sharper definition of subband gradient or an explicit statement that Excluded/Param-dep classifications are determined by the intersection dimension of the symmetrized product with Γvec, not by counting zero components of a direction-dependent vector. The concrete test would settle whether any Table II entry changes under a correct formulation.","tokens_in":17563,"tokens_out":46940,"duration_ms":429850,"concrete_test":"For each degenerate irrep in Table II (D2d E, D3h E'/E'', D3 E, C3v E, Td T1/T2, T T), construct the V_i matrices from the little-group basis, form H^(1)(qhat) = Σ V_i qhat_i, and numerically sample a dense set of qhat directions. For each eigenstate ψ_α(qhat), compute the vector v_α(qhat) = (⟨ψ_α|V_x|ψ_α⟩, ⟨ψ_α|V_y|ψ_α⟩, ⟨ψ_α|V_z|ψ_α⟩). Check whether the number of identically zero components of v_α is constant across qhat; in particular, verify D2d E at W is well-defined while C3v/D3h/D3 E bands are not, and that every Excluded entry corresponds to an intersection dimension ≥2, independent of direction.","verdict_should_be":"UNCHANGED","load_bearing_attack":"For degenerate bands, Sec. II D constructs V_i and defines each subband's gradient vector by diagonalizing H^(1)(qhat). But when the symmetrized product [Γ⊗Γ]_sym has a multi-dimensional intersection with Γvec (e.g., E in C3v, D3h, D3, or T in Td, T), the eigenstates of H^(1)(qhat) rotate with qhat and no unique per-subband gradient vector ∇E_α exists. The 'zero components' count in the flowchart (Fig. 2) and Table II is then ill-defined. Section II D 3 concedes direction dependence but still uses the vector notion. This affects the rigor of the degenerate entries in Table II, especially the Excluded labels and the Param.-dep claim for D2d E at W. The final classifications may be recoverable by replacing subband counting with the dimension of the invariant subspace of H^(1) (the intersection dimension), but the paper does not supply that derivation, so the degenerate half of the central claim is not rigorously established as written.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a group-theoretic criterion for whether the linear term of a band dispersion vanishes at non-time-reversal-invariant momenta (non-TRIMs) in the single-group limit. For nondegenerate bands, it proves that ∇E is symmetry-forced to zero iff the little-group vector representation Γ_vec contains no trivial irrep Γ1; if Γ_vec contains Γ1 once, exactly one gradient component is allowed and the band is generically noncritical; if Γ_vec contains two or three Γ1 components, the band is placed outside the single-band VHS classification. For degenerate bands, the paper invokes Wigner–Eckart/Clebsch–Gordan analysis, constructs the matrices V_i, diagonalizes H^(1)(q̂) direction-by-direction, and classifies each subband by counting zero components of its gradient vector. The criterion is applied to space group 225 with analytic tight-binding phase diagrams and is then extended to all space groups containing non-TRIMs, with results collected in Table II.","tokens_in":17785,"tokens_out":13095,"duration_ms":141344,"significance":"The nondegenerate part of the criterion is an elegant and useful result: it reduces a material-specific question to a single, parameter-free property of the little group, and the SG225 phase diagrams illustrate the predicted critical/noncritical dichotomy. The paper is also honest about its single-group, paramagnetic scope. However, the degenerate-band branch, which is an essential half of the central claim and of Table II, is not rigorously defined as written. The per-subband gradient vector does not generally exist for multi-component linear k·p Hamiltonians, so the degenerate classifications are not well-founded. With a rigorous reformulation of the degenerate case, the paper would be a solid contribution; in its present form it needs substantial revision.","major_comments":[{"comment":"The degenerate-band procedure is not well-defined. When H^(1)(q)=Σ V_i q_i has more than one nonzero matrix V_i, the branches of the dispersion are generally not differentiable at q=0. For example, H=v_x q_x σ_x + v_y q_y σ_y has eigenvalues ±√(v_x² q_x²+v_y² q_y²), so no per-subband gradient vector ∇E_α exists; the eigenstates of H^(1)(q̂) rotate with direction, making the zero-component count in Fig. 2 basis- and direction-dependent. This is not merely cosmetic: for the T representation of T_d, H=v q·J gives one branch with identically zero energy and zero gradient according to Eq. (14), yet Table II labels T_1,T_2 as Excluded. The degenerate rows of Tables I and II therefore need a rigorous definition of subband criticality, e.g., via the invariant subspaces or DOS behavior of the k·p Hamiltonian, rather than a per-subband gradient vector. Section II D 3 concedes direction dependence","section":"Sec. II D 3, Eq. (10), Fig. 2, Table II"},{"comment":"The multiplicity-2/3 branch for nondegenerate bands is presented as an unconditional 'Non-VHS (Excluded)'. The derivation only shows that multiple gradient components are symmetry-allowed; it does not show they are nonzero. The Wigner–Eckart matrix elements are parameter-dependent functions, so on codimension-one surfaces one allowed component can vanish accidentally — exactly as the paper itself finds for the multiplicity-one case at K, where v_z=0 along the dashed lines. Consequently, rows with Γ_vec containing two or three Γ1 components should be labeled 'generically Excluded, parameter-dependent' rather than flatly Excluded. As written, the claimed completeness of the trichotomy in Fig. 2 and Table II is an overstatement.","section":"Sec. II B, Fig. 2, Table II (C_s, C_1 rows)"}],"minor_comments":[{"comment":"Scanning (r_1, s_1) in [-1,1] with t_1=-1 does not exhaust the full (r_1/|t_1|, s_1/|t_1|) plane; values with |r_1/t_1|>1 or |s_1/t_1|>1 are not sampled. Please state the exact parameter domain that was scanned, or justify that all qualitative phase boundaries lie inside this square.","section":"Appendix A and Fig. 3"},{"comment":"The phase diagram at W is computed for the A_1 representation only. Since the criterion for nondegenerate bands is representation-independent, showing a second irrep (e.g., B_2 or A_2) would make the numerical verification more directly representative of the claim.","section":"Fig. 3"},{"comment":"The sentence about 'Γ⊗Γ≠Γ1' is confusing because the relevant object for the Hermitian first-order Hamiltonian is the symmetric square [Γ⊗Γ]_sym, not the full product. Please rephrase to avoid implying that the ordinary product is the selection-rule object.","section":"Sec. II D, after Eq. (5)"},{"comment":"There is a typo at the start of Sec. IV: ' .We have scanned' should be 'We have scanned.' It would also help to define 'Param.-dep.' explicitly in the Table II caption and to state that degenerate 'Excluded' entries are generic classifications, given the issue raised in the major comments.","section":"Sec. IV"}],"recommendation":"major_revision","confidential_remarks":"The nondegenerate result is likely correct, cleanly derived, and useful. The main risk is the degenerate-band half, which is not rigorously formulated and even appears internally inconsistent for the T_d T representation. Because this flaw is localized and fixable within the manuscript's scope, I recommend major revision rather than rejection. The authors should also be asked to clarify the exact parameter domain of the phase diagrams and to make Table II's 'Excluded' labels parameter-aware."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe core nondegenerate-band result is exactly what it looks like: Neumann's principle applied to the band gradient, and for that part the paper is clean and correct. What is genuinely new is the systematic extension to all non-TRIM high-symmetry points and the all-230-space-group table (Table II), plus a two-tier framework that separates symmetry-enforced vanishing of ∇E from parameter-dependent higher-order character. The tight-binding verification for SG 225 is honest and analytic, and the phase diagrams do confirm the dichotomy for 1D bands. Self-citation to [11] just supplies the N/S/T/M taxonomy, not the condition; no circularity.\n\nThe soft spots are real but localized. First, the degenerate-band branch is not rigorous as written. As the stress-test note says, when the symmetrized product [Γ⊗Γ]_sym has a multi-dimensional intersection with Γvec, the eigenstates of H^(1)(qhat) rotate with direction, and there is no unique per-subband gradient vector to count zeros of. Section II D 3 concedes the direction dependence and then continues using the vector notion anyway. That makes the degenerate entries in Table II—especially the “Excluded” labels and the C3v/D3h E-band rows—not well-defined. This is a genuine gap, not a quibble; the nondegenerate part does not depend on it, so the central claim survives, but the degenerate half needs a sharper definition (e.g., the dimension of the invariant subspace of H^(1)) before it is rigorous.\n\nSecond, the phase diagrams in Fig. 3 are called “complete,” but the parameter scan is only the box |r1/t1| ≤ 1, |s1/t1| ≤ 1. There is no scaling argument showing that this box exhausts the possible ratios. That is an overclaim, though it is a minor one because the decisive features (measure-zero critical lines at K) are visible in the box.\n\nThird, Table II's all-space-group coverage should be checked against an automated irrep database (Bilbao, SpaceGroupIrep) before being used for screening. That is a verification task, not a flaw in the method.\n\nWho gets value: anyone doing VHS engineering in 3D and high-throughput screening. The nondegenerate criterion is a useful lookup table even before revision. I would send it to peer review; it deserves referee time, but I would ask the referees to pin down the degenerate-subband definition and the completeness claim.","headline":"The nondegenerate VHS criterion is correct and useful, but the degenerate-band classifications in Table II are not rigorously anchored as written; fix that and the phase-diagram sweep, and this is a solid screening tool.","tokens_in":18285,"tokens_out":2547,"would_cite":true,"duration_ms":27784,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["20C35","82D25"],"pacs":["71.20.-b","61.50.Ah"],"model":"deepseek-v4-flash","headline":"Group theory alone predicts which band points are Van Hove critical.","keywords":["Van Hove singularity","band gradient","non-time-reversal-invariant momentum","little group","vector representation","Wigner-Eckart theorem","space group 225","Fermi surface topology"],"falsifier":"Find a single-group non-TRIM point where the little-group vector representation lacks the trivial representation, yet a tight-binding or ab initio calculation shows a one-dimensional band with nonzero ∇E at that point; or, conversely, a point where Γ_vec contains Γ_1 exactly once yet a one-dimensional band has all three gradient components zero over a finite parameter region. In space group 225, scanning the full (r1, s1) plane at W and checking whether any one-dimensional band ever acquires a nonzero gradient, or at K checking whether the vz component can be tuned to zero only on measure-zero","tokens_in":17433,"feed_emoji":"⚛️","tokens_out":4441,"duration_ms":42654,"temperature":0.7,"pith_summary":"This paper establishes a symmetry criterion for whether the band gradient ∇E vanishes at high-symmetry momenta that are not time-reversal invariant (non-TRIMs). For any nondegenerate band, ∇E is forced to zero if and only if the little group's vector representation lacks the trivial representation; if the trivial representation appears once, exactly one gradient component is symmetry-allowed and the band is generically noncritical. The same logic, extended through Clebsch–Gordan coefficients, classifies degenerate bands at the subband level. The result is parameter-independent and turns a band-structure question into a one-line group-theory check, verified with complete phase diagrams for space group 225 and tabulated for all space groups with non-TRIMs.","feed_headline":"For non-TRIM points, one representation check fixes band criticality","feed_subtitle":"Classification covers every space group: absence of the trivial representation forces the band gradient to vanish.","key_machinery":"The central object is the vector representation Γ_vec of the little group at the non-TRIM, decomposed into irreducible representations; the controlling number is the multiplicity of the trivial representation Γ_1 within Γ_vec. For nondegenerate bands, the Wigner–Eckart theorem collapses to Γ_vec ⊃ Γ_1 because the band representation cancels, making the criterion independent of the band's own irreducible representation. For degenerate bands, the symmetrized product [Γ⊗Γ]_sym must intersect Γ_vec, and the Clebsch–Gordan coefficients give the first-order Hamiltonian matrices V_i whose directional eigenvalues define the subband gradients. The two-tier hierarchy—symmetry fixes the linear term, pa","core_discovery":"At non-time-reversal-invariant momenta, time-reversal symmetry does not constrain the linear term of the dispersion, so the presence or absence of a nonzero band gradient is decided entirely by the little group. For nondegenerate bands, the intra-band Wigner-Eckart condition reduces to the purely geometric condition that the vector representation Γ_vec contain the trivial representation Γ_1: if Γ_vec does not contain Γ_1, ∇E must vanish and the band is symmetry-enforced critical; if Γ_vec contains Γ_1 exactly once, exactly one gradient component survives and the band is generically noncritical; if it contains two or three, the band is excluded from the single-band VHS classification. For deg","pith_inferences":["The same Γ_vec∋Γ_1 counting could serve as a high-throughput screening rule: any material with a non-TRIM whose little group lacks a trivial vector component is guaranteed at least one symmetry-forced critical band, without requiring first-principles calculations.","Extending the criterion to double groups or magnetic groups should follow the same multiplicity logic, but the relevant vector representation and trivial representation change; the paper leaves that extension open, and the degenerate-band subband definition needs additional care.","A testable refinement for degenerate bands would be to compute the full angular dependence of the first-order Hamiltonian's eigenvalues; the paper concedes that subband wavefunctions can be direction-dependent, so the per-subband zero-component count may need a stability check against the direction of approach."],"forward_implications":["In space group 225, the W point is symmetry-enforced critical for all nondegenerate bands regardless of hopping parameters, while K and U are generically noncritical with a single allowed gradient component.","The critical/noncritical dichotomy at any non-TRIM can be read off the little-group vector representation without any band-structure calculation; Table II provides this classification for every space group containing non-TRIMs in the single-group limit.","Complete two-dimensional phase diagrams at W and K verify the prediction: the W diagram contains only critical phases, while the K diagram is dominated by noncritical phases with criticality confined to measure-zero parameter lines.","Parameter tuning cannot turn a nondegenerate band at W into a noncritical one, nor can it remove the single allowed gradient component at K and U except by accidental zeroing on lower-dimensional boundaries.","The specific VHS subtype (ordinary versus higher-order) remains parameter-dependent; symmetry alone determines only whether the linear term vanishes."],"fun_headline_variants":["One group-theory check predicts Van Hove criticality at non-TRIMs","Symmetry alone decides if band gradient vanishes at non-TRIM points","Missing trivial rep forces band gradient to vanish at non-TRIMs","Non-TRIM criticality: a single representation check decides","Group-theory criterion explains Van Hove criticality at non-TRIMs"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"For degenerate bands, the paper assumes that diagonalizing the first-order Hamiltonian direction-by-direction yields a well-defined per-subband gradient vector whose zero components can be counted, even though the paper itself notes that subband wavefunctions and gradient components can depend on the direction of approach; the nondegenerate-band half of the criterion does not rely on this assumption.","fun_headline_variants_meta":{"raw":{"variants":["One group-theory check predicts Van Hove criticality at non-TRIMs","Symmetry alone decides if band gradient vanishes at non-TRIM points","Missing trivial rep forces band gradient to vanish at non-TRIMs","Non-TRIM criticality: a single representation check decides","Group-theory criterion explains Van Hove criticality at non-TRIMs"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001037,"raw_usage":{"total_tokens":4232,"prompt_tokens":809,"completion_tokens":3423,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":553,"completion_tokens_details":{"reasoning_tokens":3324}},"tokens_in":553,"tokens_out":3423,"duration_ms":20343,"temperature":1.0,"reasoning_tokens":3324,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-31T23:20:50.352189+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Find a single-group non-TRIM point where the little-group vector representation lacks the trivial representation, yet a tight-binding or ab initio calculation shows a one-dimensional band with nonzero ∇E at that point; or, conversely, a point where Γ_vec contains Γ_1 exactly once yet a one-dimensional band has all three gradient components zero over a finite parameter region. In space group 225, scanning the full (r1, s1) plane at W and checking whether any one-dimensional band ever acquires a nonzero gradient, or at K checking whether the vz component can be tuned to zero only on measure-zero","supporting_citations":[],"review_version":1}