{"id":"f7f5fb3b-a6e8-4b2f-8fe8-475623cf882a","arxiv_id":"2505.19452","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A finite-element method uses nonsymmorphic space-group irreducible representations to impose symmetry-adapted boundary conditions, reducing photonic band-structure calculations to smaller sub-domains with matched results and lower cost.","lead":"This paper introduces a finite-element scheme that exploits nonsymmorphic symmetries such as glide reflections to shrink the simulated unit cell of photonic crystals. It reports matching band structures at lower computational cost across layer, plane and space group examples.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Exactness of the degenerate reduction is unproven: the off-diagonal O' blocks in Eq. (13) are called 'weak' rather than zero, and no eigenvalue-level comparison to standard FEM is given; if nonzero, NSA-FEM is approximate, not exact.","rationale":"The reader's conditional verdict is the right starting point. My stress-test agrees with the reader's weakest assumption but identifies a concrete mechanism by which it could fail: the text labels the off-diagonal blocks in Eq. (13) as 'weak coupling' instead of proving them zero. A nonzero coupling would make the claimed decoupling approximate, directly undermining the central claim of exact symmetry reduction. I considered whether this is merely a wording issue: the paper also writes Eq. (15) as the system to be solved, and if that full reduced system retains the coupling, the method could still be exact. The ambiguity is precisely why a numerical comparison of the truncated versus full reduced spectrum is the right test. The GitHub code is a concrete asset, and the three examples are encouraging, but visual agreement on selected bands does not probe missing or spurious eigenvalues, and the paper reports no error metric. Therefore the verdict should remain conditional pending this check; I do not see a basis for outright rejection, and acceptance would require the missing proof or the numerical test.","tokens_in":19647,"tokens_out":13360,"duration_ms":143071,"concrete_test":"Use the code and geometry of Model B at Lambda = (1/2, 1/4). Assemble the full FEM matrices A and B; form P from Eq. (14); compute the full reduced matrices A~ = P^T A P and B~ = P^T B P. Compare the spectrum of the full reduced eigenproblem with (i) the diagonal truncation in Eq. (13) and (ii) the standard FEM spectrum at the same k. Record the maximum relative frequency error and the number of missing or extra eigenvalues. If the truncated diagonal system reproduces standard FEM to machine precision, the exactness claim holds for this case; if not, the paper must either retain the off-diagonal blocks or explicitly downgrade the method to an approximation.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Central claim: exact decomposition into decoupled sub-tasks. For non-degenerate modes, Eq. (8) is split into A_{1,1} x_{1,1} = lambda B_{1,1} x_{1,1}; for degenerate Lambda modes, Eq. (12) is replaced by diagonal Eq. (13), with O' described as 'extremely sparse' and 'weak coupling terms among the diagonal block matrices.' Neither the paper nor the appendix proves that these off-diagonal blocks vanish after the congruence transform P^T A P (Eq. 15). If they are merely small, then the sub-tasks are not exactly decoupled and the eigenpairs of Eq. (13) are not guaranteed to coincide with the original FEM spectrum. The only validation reported is visual agreement of selected bands; no eigenvalue counts, frequency residuals, or tests at non-high-symmetry k are provided. Because the stated advantage is an exact symmetry reduction rather than a heuristic reduction of the domain, the unproven vanishing of the coupling blocks is the load-bearing issue.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes NSA-FEM, a symmetry-adapted finite element procedure for photonic band structures with nonsymmorphic space groups. It derives boundary conditions from the finite-dimensional irreducible representations of the little group at high-symmetry k, reduces the standard FEM eigenproblem Ax=λBx to subproblems on a minimal computational domain (MCD), and recovers full-cell fields via symmetry operations. The method is demonstrated on three structures: an AB-layer-stacked photonic crystal with rod group pmcm, a twisted quadrupole photonic crystal with a plane group, and a 3D metacrystal with space group P4/mbm. Comparisons to standard FEM are shown for band structures along high-symmetry paths, with DOF reductions by factors of 2–4 and reported speedups. The central claim is that the decomposition is exact rather than a heuristic domain reduction.","tokens_in":19864,"tokens_out":12110,"duration_ms":109885,"significance":"If exactness is established, this is a useful extension of symmetry-adapted FEM from symmorphic to nonsymmorphic groups. It offers a systematic route to exploit non-primitive translations and k-dependent irreps, provides symmetry labels for bands before solving, and enables parallel sub-task computations. The paper includes open-source MATLAB code on GitHub, which supports reproducibility, and benchmarks against standard FEM on three different structure types rather than a single toy model. However, the value of the contribution depends on proving that the reduced subproblems are exactly decoupled and spectrum-preserving; the manuscript currently leaves this at the level of assertion, so the significance is conditional.","major_comments":[{"comment":"The central claim of exact decoupling is not established. In Eq. (13), the off-diagonal blocks O' are described as 'extremely sparse' and as representing 'weak coupling terms among the diagonal block matrices,' not as zero. No proof is given that these blocks vanish after the congruence transformation in Eq. (15), or that the spectrum of the reduced problem coincides with that of the original generalized eigenproblem in Eq. (7). If O' is nonzero, the sub-task eigenpairs are only an approximation, contradicting the abstract's claim of exact decomposition. Please provide a rigorous block-diagonalization argument (for example, equivariance of A and B under the symmetry representation and use of Schur's lemma), or, if the blocks are indeed nonzero, reframe the method as approximate and quantify the introduced error.","section":"II.D, Eq. (13)"},{"comment":"The sufficiency of the proposed MCD boundary conditions for the two-dimensional representation at Λ is not demonstrated. Equations (6a)-(6b) impose two scalar relations between pairs of boundary degrees of freedom, but the full 2×2 representation matrix [[0,1],[-1,0]] in Table I couples the two degenerate components in a way that is not shown to be fully captured by these two relations. The paper should prove that the constrained reduced problem has the same eigenvalues with the same multiplicities as the original full-cell problem, including at the corners and intersections of the MCD boundary, rather than assuming this.","section":"II.C, Eqs. (5)-(6)"},{"comment":"The numerical validation is visual rather than quantitative. In all three examples, agreement with standard FEM is claimed by overlaying band curves, but no eigenvalue residuals, eigenvalue counts, or tests of the degenerate sub-task at and around the star-k points are reported. Visual overlap of a few bands cannot certify that the reduced spectrum contains all eigenvalues with correct multiplicities, especially for the degenerate case where exactness is in question. Please add quantitative comparisons (for example, maximum relative frequency error, eigenvalue counts per sub-task) and a mesh-convergence test.","section":"III.A-C"}],"minor_comments":[{"comment":"The text identifies the second example as plane group P4g (number 12), but the caption of Table III labels it P4bm; these are different group names and the inconsistency should be corrected.","section":"Table III"},{"comment":"The caption of Fig. 2 labels subfigures (d) and (e), while the text refers to panels (a) and (b); the panel references and caption should be aligned.","section":"Fig. 2"},{"comment":"Equation (9) is not well defined as written: the right-hand side multiplies a representation matrix, a symmetry operator, and a finite-element vector; the action of the space-group operator on discrete DOFs should be defined explicitly.","section":"II.D, Eq. (9)"},{"comment":"The timing comparisons do not report the number of eigenvalues requested, the eigensolver settings, or whether the sub-tasks were run serially; these details are needed to interpret the reported speedups.","section":"III.A-C"},{"comment":"The abstract contains the sentence fragment 'our method fully accounting for non-primitive translations and nonstructural symmetries,' and Section II.A contains 'we can modal only half of the structure'; both should be rewritten.","section":"Abstract and II.A"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is a credible extension of the authors' earlier symmorphic symmetry-adapted FEM work, and the availability of open-source code is a plus. The decisive issue is the unproven exactness of the degenerate reduction in Eqs. (12)-(15); a proof or a quantitative estimate of the coupling terms, together with eigenvalue-level validation, is essential before publication. This is fixable within the manuscript's scope, so I do not recommend rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The genuinely new thing here is the systematic use of space-group irreps, including time-reversal-induced degeneracies, to derive MCD boundary conditions for glide-symmetric photonic structures. The authors extend their earlier symmorphic work to nonsymmorphic groups and handle the extra complication that at some k-points the glide operation requires a 2D representation. The three examples cover a rod group, a plane group, and a space group, with domain reductions up to 4x and speedups up to 4.8x. The code is promised open-source, and the comparison against standard FEM is consistent. Credit where due: this is a real step forward for computational practice.\n\nThe soft spot is in Section II.D. The authors claim the original problem is 'completely decoupled' into sub-tasks, but the off-diagonal blocks O' in Eqs. (12)-(13) are only described as 'weak coupling' and 'extremely sparse'—not zero. If those blocks do not vanish exactly, the reduced eigenproblem is not exactly equivalent to the original, and the claimed exact symmetry reduction becomes an approximation. The paper does not prove vanishing, and the validation consists of plotting NSA-FEM points on top of standard-FEM lines. No residual norms, no eigenvalue counts, no tests away from high-symmetry lines. The boundary conditions in Eqs. (5)-(6) are plausible from the character tables, but the sufficiency of those constraints for reproducing the full spectrum is asserted rather than shown.\n\nThat said, this is not a fatal flaw for the method's practical value. Even if the decoupling is approximate, the technique still gives large domain reductions with good visual agreement, and it provides a useful band-labeling scheme. The fix is straightforward: the authors should either prove the off-diagonal blocks vanish (which may follow from symmetry if the DOF partition is chosen correctly), or add quantitative error metrics and be explicit about the approximation. A referee should request that.\n\nThis deserves peer review. The method is important enough for computational nanophotonics, and the issues are addressable in revision.","headline":"Useful extension of symmetry-adapted FEM to glide-symmetric photonic structures, but the claimed exact decoupling of degenerate sub-tasks rests on an unproven vanishing of coupling blocks.","tokens_in":20398,"tokens_out":2874,"would_cite":false,"duration_ms":28064,"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":"This paper claims that the glide symmetry of photonic crystals lets one split the finite-element band-structure problem into smaller, decoupled, exactly equivalent subproblems, cutting the computational domain by a factor of two or four…","keywords":["nonsymmorphic space group","glide symmetry","symmetry-adapted finite element method","photonic crystals","band structure","irreducible representations","time-reversal symmetry","computational domain reduction"],"falsifier":"Run standard full-cell FEM and NSA-FEM on the same glide-symmetric structure over a dense set of wave vectors along an entire Brillouin-zone edge (or near a predicted accidental degeneracy) and check that the union of subtask eigenvalues matches the full-cell spectrum with no missing or spurious eigenvalues; a single mismatch would show the character-table boundary conditions do not suffice.","tokens_in":19475,"feed_emoji":"📡","tokens_out":8431,"duration_ms":70509,"temperature":0.7,"pith_summary":"Nonsymmorphic space groups—photonic crystals whose symmetry includes glide reflections, a mirror followed by a half-lattice shift—are harder to exploit numerically than ordinary point-group symmetries because the phase of the glide depends on the wave vector. The paper's central claim is that this difficulty can be turned into a systematic reduction: using the irreducible representations and character tables of the nonsymmorphic group, the full-cell finite element eigenvalue problem can be split into decoupled subtasks on a minimal computational domain (half or quarter of the unit cell), each with boundary conditions derived from the symmetry operation. If correct, band-structure calculations for glide-symmetric photonic crystals become more efficient and each computed band comes pre-labelled with its symmetry type, which reveals crossings, hidden degeneracies, and nodal lines directly from the subtask structure.","feed_headline":"Glide symmetry cuts photonic band calculations to half-cells","feed_subtitle":"Irreducible-representation boundary conditions split the FEM problem into exact, parallel subtasks, cutting degrees of freedom and runtime.","key_machinery":"The load-bearing machinery is the character table of the nonsymmorphic group at each high-symmetry wave vector: irreducible representations of the little group at $k$, and full-group irreps induced from the star $k^*$ when time-reversal symmetry is needed. The Seitz-operator glide with its non-primitive translation determines, through those characters, the ratio between the field on one side of the minimal domain and the field on the other side (a scalar $\\pm e^{i\\pi a}$ for non-degenerate modes, a $2\\times 2$ matrix for degenerate pairs), encoded in a transformation matrix $P$ that projects the full-cell system matrix onto the reduced degrees of freedom (Eqs. 10–11 and 14–15).","core_discovery":"The paper introduces the nonsymmorphic symmetry-adapted finite element method (NSA-FEM) and demonstrates that for a photonic structure whose unit cell carries a glide operation $G_x = \\{m_{01}|1/2,0\\}$, the original matrix eigenproblem $Ax = \\lambda Bx$ on the full cell can be reduced to independent subsystems on a minimal computational domain. For non-degenerate modes at a generic wave vector $\\Delta$, the glide imposes a scalar character boundary condition $E_\\Delta(r_R) = \\pm e^{i\\pi a}E_\\Delta(r_L)$ (Eq. 5), which lets one solve only the left half and reconstruct the right half through the relation $x_{1,2} = D^k_j(\\{R|\\tau+R_n\\})\\{R|\\tau+R_n\\}^{-1} x_{1,1}$. For degenerate modes at $\\Lambda$, where the glide has a two-dimensional representation and time-reversal symmetry couples the arms of the star $k^*$, the two modes form an inseparable pair whose coupled boundary conditions (Eq. 6) and transformation matrix $P$ (Eq. 15) yield the degenerate pair as a two-component subsystem. The band structures and modal fields obtained this way are reported to agree with standard full-cell FEM for a layer-group (rod group $pmcm$), a plane group ($P4g$), and a space group ($P4/mbm$) photonic crystal, while reducing degrees of freedom by one-half to one-quarter and wall-clock time in the examples shown.","pith_inferences":["Editorial extension: Because the decomposition uses only group algebra and Bloch phases, the same boundary-constraint construction should carry over to other Hermitian eigenproblems on glide-symmetric lattices, including acoustics, elasticity, and Schrödinger-type equations.","Editorial extension: A sharper validation than visual band agreement would be to compare the union of subtask eigenvalues with the full-cell spectrum over a dense set of wave vectors across the whole Brillouin zone; any missing or extra eigenvalue would pinpoint exactly where the character-table reduction breaks down.","Editorial extension: The symmetry-resolved subtask spectra could be used as a projection basis for topological invariants, automatically assigning parity or glide eigenvalues to each band during large-scale computations.","Editorial extension: The reported speedups are sequential; a direct benchmark of parallel subtask execution on a memory-bound problem would quantify the practical scaling advantage beyond the factor-of-domain reduction."],"forward_implications":["Every band of a glide-symmetric photonic crystal can be computed on a half- or quarter-cell domain with full fields reconstructed by the glide relation, giving exact band-structure agreement at the wave vectors tested while using a fraction of the degrees of freedom.","Bands are classified by irreducible representation from the outset, so band sticking, crossings, and degenerate points can be identified from the subtask decomposition rather than by post-processing the full spectrum.","Time-reversal-induced degeneracies are treated as coupled two-mode subtasks built from full-group irreps of the star $k^*$, keeping phenomena such as hidden-symmetry nodal points within the reduced computation.","The reduction procedure is systematic across group dimension: the same recipe is demonstrated for a layer group, a plane group, and a space group, and applies whenever a glide or screw operation is present.","The decoupled subtasks can be solved independently and in parallel, so the reported sequential runtime reductions do not yet include the additional wall-clock gain from parallel execution."],"supporting_citations":[{"why":"Provides the symmetry and degeneracy analysis of glide-symmetric periodic structures that motivates the irreducible-representation treatment.","marker":"[37]"},{"why":"Supplies the AB-layer-stacked photonic crystal with hidden symmetry used as the first validation model.","marker":"[38]"},{"why":"Gives the point-group symmetry-adapted FEM framework that this paper extends from symmorphic to nonsymmorphic groups.","marker":"[10]"},{"why":"Supplies the normal-subgroup induction method for space-group irreducible representations on which the boundary condition derivation rests.","marker":"[42]"},{"why":"Provides the tabulated irreducible representations of the plane and space groups used in the three examples.","marker":"[43]"},{"why":"Supplies the full-group irreducible representations induced from the star of k, needed to construct time-reversal-invariant degenerate subtasks.","marker":"[45]"}],"fun_headline_variants":["Glide symmetry-adapted FEM halves photonic band effort","Half-cell FEM with glide symmetry cuts runtime","Nonsymmorphic symmetry speeds photonic band calculations","Symmetry-adapted FE method shrinks photonic problem"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the symmetry-derived boundary conditions on the half- or quarter-cell domain yield exactly the same eigenvalues and eigenmodes as the full-cell problem, for both non-degenerate and degenerate modes; the paper shows visual agreement at selected wave vectors rather than a proof that the reduced spectrum coincides with the full spectrum.","fun_headline_variants_meta":{"raw":{"variants":["Glide symmetry-adapted FEM halves photonic band effort","Half-cell FEM with glide symmetry cuts runtime","Nonsymmorphic symmetry speeds photonic band calculations","Symmetry-adapted FE method shrinks photonic problem"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000706,"raw_usage":{"total_tokens":3256,"prompt_tokens":1094,"completion_tokens":2162,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":710,"completion_tokens_details":{"reasoning_tokens":2098}},"tokens_in":710,"tokens_out":2162,"duration_ms":14397,"temperature":1.0,"reasoning_tokens":2098,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T14:13:54.594827+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run standard full-cell FEM and NSA-FEM on the same glide-symmetric structure over a dense set of wave vectors along an entire Brillouin-zone edge (or near a predicted accidental degeneracy) and check that the union of subtask eigenvalues matches the full-cell spectrum with no missing or spurious eigenvalues; a single mismatch would show the character-table boundary conditions do not suffice.","supporting_citations":[{"cited_title":"Ikeda, M","cited_arxiv_id":null,"evidence_quote":"Provides the symmetry and degeneracy analysis of glide-symmetric periodic structures that motivates the irreducible-representation treatment."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the AB-layer-stacked photonic crystal with hidden symmetry used as the first validation model."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the point-group symmetry-adapted FEM framework that this paper extends from symmorphic to nonsymmorphic groups."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the normal-subgroup induction method for space-group irreducible representations on which the boundary condition derivation rests."},{"cited_title":"Zhang and C.-X","cited_arxiv_id":null,"evidence_quote":"Provides the tabulated irreducible representations of the plane and space groups used in the three examples."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the full-group irreducible representations induced from the star of k, needed to construct time-reversal-invariant degenerate subtasks."}],"review_version":1}