{"id":"32feb051-c757-4c85-9255-2854c2d928aa","arxiv_id":"2412.11353","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"In a 3D woodpile photonic crystal, joining shifted unit cells creates interface and hinge states, and a Wilson loop based partial Chern number calculation gives a sign-flip rule for hinge state emergence.","lead":"This paper numerically finds interface and hinge light states in a three-dimensional woodpile photonic crystal built from two shifted versions of the same unit cell. It also proposes a Wilson loop based method for computing topological invariants on arbitrary surfaces, and offers a sign rule for where hinge states appear. The work aims toward topological waveguides in three-dimensional photonic crystals.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The quadrant Wilson-loop windings are equated to partial Chern numbers through an invalid step in Appendix A, and the per-band assignment for the two connected bands is unjustified; the hinge-state selection rule inherits this unsupported identification.","rationale":"The reader identified the same load-bearing weakness: the partial Chern numbers and the per-band separation are assumed rather than derived. My stress-test sharpens this into a concrete mathematical defect: Eq. (12) in Appendix A contains an unjustified equality that is not merely missing detail but appears wrong as written, and the diagonalization of the 2×2 Wilson loop for two connected bands does not guarantee gauge-invariant individual windings. This matters because every downstream conclusion—the number of interface states, the existence of isolated hinge states in the gap, and the selection rule of Fig. 5—rests on those quadrant winding numbers being valid partial Chern numbers. The numerical observation of localized hinge states in Fig. 4 is plausible and counts as some evidence, but it does not by itself establish topological protection; the claimed mechanism is precisely the partial-Chern selection rule. The correct response is to keep the CONDITIONAL verdict: the authors should supply a corrected derivation and a direct computation of the partial Chern numbers. If those checks fail, the topological interpretation would need to be abandoned or substantially weakened. I therefore set verdict_should_be to UNCHANGED relative to the reader's verdict.","tokens_in":13626,"tokens_out":8000,"duration_ms":80375,"concrete_test":"Compute, for the primitive unit cell, the flux of the two-band subspace Berry curvature through each of the quadrants of the surface BZ using a gauge-invariant lattice discretization (e.g., Fukui–Hatsugai–Suzuki), and compare each quadrant flux with the winding numbers extracted from Eq. (7)/Fig. 2. If they disagree, the partial-Chern identification fails. In addition, extract the Wilson-loop eigenvalues after applying a randomly chosen smooth gauge; if the band-1 and band-2 windings change while the product/total winding stays fixed, the per-band assignment used for the selection rule is an artifact.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central claim—that 1D hinge states are selected by sign flips of partial Chern numbers—requires the quadrant winding numbers in Fig. 2 to be genuine partial Chern numbers. That link is not established. In Appendix A, W(kz) is defined as a Wilson phase around a loop L at fixed kz, and Eq. (12) asserts (1/2π)∫ dkz ∂kz W(kz) = (1/2π)∫_L ∂_L W(L)dL. The left side is a kz boundary term (W at the upper and lower kz), while the right side is a loop integral of the phase in the surface BZ; the equality does not follow and is dimensionally inconsistent. Thus the derivation does not connect the quadrant winding of W to an integral of Berry curvature over that quadrant. Moreover, the two bands below the gap are connected by degeneracy points (Sec. II). The procedure in Eqs. (4)–(6) diagonalizes the 2×2 Berry connection and assigns the resulting eigenvalues separately to 'band 1' and 'band 2'; for touching bands such individual eigenvalues are not smooth or gauge invariant, and the statement in Sec. IV that the bands are 'orthogonal polarization' is asserted without a proof of separability. The observed red hinge states in Fig. 4 may be real, but without a valid partial-Chern invariant the selection rule in Fig. 5 and the topological protection claim are unsupported.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper numerically studies a three-dimensional woodpile photonic crystal with before- and after-shifting unit cells, proposes a Wilson-loop method for arbitrary surfaces using the primitive unit cell, and uses the resulting quadrant winding numbers to define partial Chern numbers. It then attributes the observed topological interface states and second-order hinge states to differences in these partial Chern numbers, and proposes a sign-flip selection rule for the hinge states. The numerical observations include supercell band structures and field distributions for interface and hinge states.","tokens_in":13894,"tokens_out":5517,"duration_ms":49684,"significance":"If the topological-invariant identification were rigorous, the paper would offer a practical Wilson-loop procedure for arbitrary surfaces in 3D photonic crystals and a concrete selection rule for hinge states, which would be valuable for waveguide applications. The supercell simulations are explicit, the Wilson-loop computation and the supercell eigenmode searches are independent (so there is no circular fitting), and the predicted hinge-state locations are falsifiable. However, the central mathematical link between the computed quadrant windings and partial Chern numbers is not established, and the per-band decomposition for the two connected bands is asserted rather than proved. These are load-bearing gaps: the hinge selection rule inherits them. The paper is therefore promising but not yet convincing as a topological explanation.","major_comments":[{"comment":"The derivation of Eq. (12) does not establish the claimed equality between the kz-integral of ∂kz W(kz) and the winding number of the Wilson loop around the boundary of a slice. The left-hand side, (1/2π)∫ dkz ∂kz W(kz), is a boundary term W(kz,upper)−W(kz,lower), whereas the right-hand side is a loop integral of ∂L W(L) in the surface plane; these are different objects and are not related by Eqs. (10) and (11). Thus the identification of the quadrant winding numbers in Fig. 2 with partial Chern numbers is unsupported. Please provide a correct derivation, or compute the partial Chern numbers directly from the Berry curvature on the quadrant patches and verify that they match the winding numbers.","section":"Appendix A, Eq. (12)"},{"comment":"The two bands below the gap are connected by degeneracy points (Section II), yet Eqs. (4)-(6) diagonalize the 2×2 overlap matrix and assign the resulting eigenvalues separately to 'band 1' and 'band 2'. For touching bands, individual Wilson-loop eigenvalues are not guaranteed to be smooth or gauge-invariant, and the statement in Section IV that 'these two connected bands have orthogonal polarization' is asserted without proof. Please demonstrate the separability of the two bands in a consistent gauge, or compute the invariant for the two-band subspace using a non-Abelian Wilson loop.","section":"Section III, Eqs. (4)-(6), and Section IV"},{"comment":"The selection rule for hinge states is built entirely on the partial Chern number differences shown in Fig. 5(a), which in turn rely on the quadrant winding numbers. Since the link between those windings and genuine partial Chern numbers is not established (comment 1) and the per-band assignment is not justified (comment 2), the hinge-state selection rule lacks a rigorous topological foundation. The red hinge states in Fig. 4 may be real, but their topological protection is not demonstrated by the present invariant calculation. Please either repair the invariant computation or provide an independent robustness check (for example, perturb the structure or vary the termination position) and adjust the claims accordingly.","section":"Section V, Figs. 4 and 5"}],"minor_comments":[{"comment":"There are multiple typos and OCR-like artifacts, including 'primitve' (Section II), 'AfUFS' in Fig. 2, and 'half-integrer' (Section III); please proofread carefully.","section":"Throughout"},{"comment":"The labels 'Band 1 Band 2 Band 2Band 1' and the garbled text in the figure panels make it difficult to associate the winding numbers with the correct band and surface; please redraw the figure with clear per-panel labels.","section":"Fig. 2"},{"comment":"Equation (7) defines w_n via an integral of ∂L W_n over a closed loop L, but the notation ∂L W_n is not defined; state explicitly how the discrete Wilson-loop data are converted into a winding number and how the quadrant boundaries in Fig. 2 are chosen.","section":"Eq. (7)"},{"comment":"The statement 'the number of interface states equals to the partial Chern number difference [59]' cites the standard bulk-edge correspondence for Chern numbers; since partial Chern numbers on finite patches are not the same as global Chern numbers, please cite a source that establishes this relation for patch invariants, or derive it.","section":"Section IV"},{"comment":"The conclusion that the Wilson-loop method is 'first-time introduced in this paper' is too strong given Refs. [44,54]; please rephrase to describe it as an application or generalization of existing Wilson-loop approaches to the woodpile PhC with arbitrary surfaces.","section":"Section VI"},{"comment":"The statement 'we try to choose a gauge choice which avoid the Γ point' is vague; specify the gauge-fixing procedure and explain how it interacts with the connected bands and with the gauge invariance of the individual Wilson-loop eigenvalues.","section":"Section III"}],"recommendation":"major_revision","confidential_remarks":"This is a borderline major-revision case. The numerical results and the proposed selection rule are interesting and likely worth publishing if the invariant computation can be repaired or the claims appropriately weakened. The central flaw is in Appendix A and in the per-band separability argument; it is not a matter of taste but a load-bearing mathematical gap. I would not reject because the observed states and the selection rule may survive a corrected treatment."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nShort version: this paper reports plausible numerical evidence for 2D interface and 1D hinge states in a 3D woodpile photonic crystal, and it proposes a Wilson-loop recipe for arbitrary surfaces. But the central topological interpretation rests on an invalid derivation, so the \"selection rule\" and the protection claim are not supported as written.\n\nWhat's actually new: the volume-conserving choice of Γ3 for computing Wilson loops on arbitrary planes is a useful practical step, and the application to a realistic GaAs woodpile structure with an embedded boundary is concrete. The observation of two isolated hinge states in the complete gap between interface states is interesting, and the field profiles look believable. The numerical evidence of the states themselves is the paper's real asset.\n\nThe soft spots are where the paper tries to explain those states. Appendix A equates (1/2π)∫dk_z ∂_z W(k_z) with (1/2π)∮dL ∂_L W(L). Those are not the same object: the left is a k_z boundary term, the right is a circulation in the surface BZ. The step is dimensionally inconsistent, and the conclusion that the quadrant winding of W equals the partial Chern number of that quadrant does not follow. The per-band decomposition is also unproven: the two bands are connected by degeneracies, and assigning individual Wilson-loop eigenvalues to \"band 1\" and \"band 2\" requires gauge invariance that touching bands don't have. The assertion of \"orthogonal polarization\" in Section IV is asserted, not derived. So the hinge-state selection rule built on those partial Chern numbers has no rigorous bulk-boundary correspondence. The states might be real—nothing in the numerics suggests they are artifacts—but the paper's claim that they are topologically protected by partial Chern numbers is not established.\n\nThe citation pattern is fine; prior Wilson-loop work is cited, and the group's own related papers are appropriately referenced. There's no code or data, and no convergence tests, which is a minor omission for a numerical methods paper.\n\nWho gets value from this: researchers actively working on 3D photonic topological states, especially woodpile structures. The numerical recipe for Wilson loops on oblique surfaces could be useful, and the observed hinge states are worth knowing about. But the topological explanation needs a serious rework before it can be trusted.\n\nFor peer review: I'd send it to a competent referee, but with the expectation of heavy revision. The flaws are identifiable and possibly fixable—the authors could either supply a real derivation of the partial-Chern identification or reframe the paper as a numerical observation of states without the protection claim. As it stands, the central argument doesn't hold.","headline":"Plausible numerical states in a realistic woodpile photonic crystal, but the Wilson-loop derivation equating quadrant windings to partial Chern numbers is invalid and the hinge-state selection rule needs a rigorous basis.","tokens_in":14450,"tokens_out":3444,"would_cite":false,"duration_ms":30445,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["42.70.Qs"],"model":"deepseek-v4-flash","headline":"This paper claims that the 3D woodpile photonic crystal supports topologically protected 1D hinge states for light inside the complete band gap, with a selection rule based on sign flips of partial Chern number differences.","keywords":["woodpile photonic crystal","Wilson loop","partial Chern number","topological hinge states","higher-order topology","photonic band gap","topological waveguide","3D photonic crystal"],"falsifier":"Compute the Berry curvature directly by quadrature over each quadrant of the surface Brillouin zone for the two bands below the gap and check that the integrated values equal the ±2 windings of the Wilson loops; any mismatch or gauge dependence would falsify the selection rule.","tokens_in":13371,"feed_emoji":"💡","tokens_out":5813,"duration_ms":48004,"temperature":0.7,"pith_summary":"This paper claims that a 3D woodpile photonic crystal, built from dielectric blocks in a diamond-like arrangement, hosts one-dimensional hinge states for light that sit inside the complete photonic band gap, well separated from bulk and surface modes. It further claims that these hinge states appear only when the partial Chern number differences of the two interfaces meeting at the hinge have opposite signs, a concrete selection rule. To get there, the authors introduce a numerical Wilson-loop procedure that assigns a winding number to each quadrant of a surface Brillouin zone and treats those windings as partial Chern numbers. If the claims hold, woodpile photonic crystals become a practical platform for topologically protected routing of light in three dimensions.","feed_headline":"3D woodpile crystal hosts topologically protected hinge light states","feed_subtitle":"Hinge light modes sit inside the complete band gap, isolated from bulk and surface waves.","key_machinery":"The central object is the Wilson loop on an arbitrary surface of the 3D crystal, generalized by replacing the primitive-cell Brillouin zone with a parallelepiped of equal volume built from the surface periodicity vectors. The winding number of this loop around each quadrant of the surface Brillouin zone is interpreted as a partial Chern number. This object carries the argument because the difference in partial Chern numbers between the before-shifting and after-shifting unit cells predicts interface states, and the sign flip of that difference between two intersecting surfaces predicts which hinges host hinge states.","core_discovery":"By shifting the origin of the woodpile unit cell by a0/4 in x, a0/4 in y, and a0/2 in z, one obtains two unit cells with identical band structures but opposite topological properties. The Wilson-loop spectra on the kykz and kxkz surfaces reveal that each of the two occupied bands has winding numbers of magnitude 2 on each quadrant, so the partial Chern number difference between the two unit cells is ±2. These differences produce eight interface states per surface band gap and two doubly degenerate hinge states in a supercell; the hinge states localize only at the upper-right and lower-left hinges, because only there the sign of the partial Chern number difference flips between the two surfaces. The paper states this as a selection rule: hinge states emerge at a hinge if the partial Chern number differences on the two interfaces forming the hinge have opposite signs.","pith_inferences":["If the partial Chern numbers are genuinely gauge-invariant, the same quadrant-winding calculation could be used to search for hinge states in other dielectric photonic crystals with broken inversion symmetry but no net Chern number.","The sign-flip selection rule for hinges suggests a natural extension to zero-dimensional corner states: at a corner where three interfaces meet, a corner state should appear when the three sign matrices have a consistent mismatch; that prediction is not tested in the paper.","A direct experimental test would fabricate the before-shifting/after-shifting woodpile junction and measure transmission along each hinge; the hinge pair with opposite-sign interfaces should transmit in the gap frequency window while the other pair should not."],"forward_implications":["A woodpile photonic crystal with a shifted unit-cell boundary will exhibit two-dimensional interface states inside the complete photonic band gap, eight per surface when the folded Brillouin zone is accounted for.","One-dimensional hinge states appear in the complete gap between interface states, so light can be guided along a hinge without leaking into bulk or surface modes.","Whether a given hinge hosts a state is decided by the signs of the partial Chern number differences on the two interfaces: opposite signs produce a hinge state and identical signs do not.","The Wilson-loop procedure proposed here can be applied to compute topological invariants on arbitrary crystal surfaces in other 3D photonic crystals, not only cubic ones."],"supporting_citations":[{"why":"Supplies the Zak-phase/Wilson-loop numerical scheme for 3D photonic crystals that this paper extends to arbitrary surfaces and to the woodpile geometry.","marker":"[50]"},{"why":"Establishes the valley-Chern-number sign-flip reasoning for corner states in 2D honeycomb photonic crystals that the paper adapts into its hinge selection rule.","marker":"[27]"},{"why":"Provides the bulk-edge correspondence that lets the paper count interface states from partial Chern number differences.","marker":"[59]"},{"why":"Gives the winding-number/Chern-number equivalence used to identify quadrant windings of the Wilson loop as partial Chern numbers.","marker":"[58]"},{"why":"Describes techniques for computing topological invariants in 3D photonic crystals, invoked here to handle gauge choices away from the Gamma point.","marker":"[54]"},{"why":"Supports the Appendix relation between Berry-curvature flux and Wilson-loop winding used to justify the partial Chern number.","marker":"[61]"}],"fun_headline_variants":["Wilson loop reveals 3D photonic hinge states","Topological hinge light states in 3D woodpile crystal","3D woodpile crystal: Wilson loop maps topological light","Hinge states in 3D photonic crystal from Wilson loop"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole selection rule rests on treating the winding of the Wilson loop around each quadrant of the surface Brillouin zone as a well-defined partial Chern number, and on assuming the two bands below the gap contribute independently; if that identification fails, the hinge rule is unsupported.","fun_headline_variants_meta":{"raw":{"variants":["Wilson loop reveals 3D photonic hinge states","Topological hinge light states in 3D woodpile crystal","3D woodpile crystal: Wilson loop maps topological light","Hinge states in 3D photonic crystal from Wilson loop"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000157,"raw_usage":{"total_tokens":1174,"prompt_tokens":848,"completion_tokens":326,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":464,"completion_tokens_details":{"reasoning_tokens":257}},"tokens_in":464,"tokens_out":326,"duration_ms":3277,"temperature":1.0,"reasoning_tokens":257,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T15:01:51.942684+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the Berry curvature directly by quadrature over each quadrant of the surface Brillouin zone for the two bands below the gap and check that the integrated values equal the ±2 windings of the Wilson loops; any mismatch or gauge dependence would falsify the selection rule.","supporting_citations":[{"cited_title":"Takahashi, Y","cited_arxiv_id":null,"evidence_quote":"Supplies the Zak-phase/Wilson-loop numerical scheme for 3D photonic crystals that this paper extends to arbitrary surfaces and to the woodpile geometry."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Establishes the valley-Chern-number sign-flip reasoning for corner states in 2D honeycomb photonic crystals that the paper adapts into its hinge selection rule."},{"cited_title":"Hatsugai, Chern number and edge states in the integer quan- tum hall effect, Phys","cited_arxiv_id":null,"evidence_quote":"Provides the bulk-edge correspondence that lets the paper count interface states from partial Chern number differences."},{"cited_title":"Chen, C.-H","cited_arxiv_id":null,"evidence_quote":"Gives the winding-number/Chern-number equivalence used to identify quadrant windings of the Wilson loop as partial Chern numbers."},{"cited_title":"Devescovi, A","cited_arxiv_id":null,"evidence_quote":"Describes techniques for computing topological invariants in 3D photonic crystals, invoked here to handle gauge choices away from the Gamma point."}],"review_version":1}