{"id":"d096007b-3b4e-46f0-89e9-d662c63720a3","arxiv_id":"2607.23260","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"A curl-based electric-field integral equation (Curl-EFIE) whose MoM discretization converges to the MFIE as the testing-disk radius shrinks, giving weakly singular kernels and stable conditioning.","lead":"This paper introduces a new way to solve electromagnetic scattering from metal objects by testing the electric field with small spinning disks, which recovers the magnetic-field equation in the limit of tiny disks. It offers simpler integrals and stable low-frequency behavior while avoiding some drawbacks of standard magnetic-field formulations.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The Curl-EFIE-to-MFIE limit is proven only for smooth fields; at sharp edges the Taylor expansion and H→0 limit fail, so the claimed edge accuracy rests on the empirical H≈h/100.","rationale":"The reader's weakest assumption identifies exactly the same load-bearing gap: the Taylor expansion and limit interchange require smoothness, which fails at sharp edges and corners. This is the right concern because the paper's novelty rests on two pillars: (i) a formal convergence to the well-conditioned MFIE, and (ii) improved accuracy on sharp-edged objects. Pillar (i) is established only for smooth manifolds; the derivation in Sec. II does not address edges. Pillar (ii) is demonstrated numerically but the theory does not explain why finite H works better than the H→0 limit—indeed, if H→0 converges to MFIE, the edge accuracy should approach the MFIE's known poor edge accuracy, contradicting the observed optimum at H≈h/100. The paper itself acknowledges the practical choice is 'bounded away from the vanishing limit,' but the abstract's broad convergence claim overstates the theoretical support. This justifies keeping the CONDITIONAL verdict: the central claim is plausible for smooth objects and the stability data support the MFIE-correspondence in that regime, but the sharp-edge superiority claim is not underpinned by the derivation. The proposed numerical sweep would directly test whether H-dependence is a tuning artifact or an intrinsic feature, settling whether the concern materially weakens the paper's main selling point.","tokens_in":16272,"tokens_out":4883,"duration_ms":50558,"concrete_test":"Re-run the 0.1λ square-pyramid experiment of Fig. 4 with the same mesh and quadrature, sweeping H = h, h/10, h/100, h/1000, h/10000, and compare the average RCS deviation against both the DEM-MFIE reference and the fine RWG-EFIE reference. If the RCS deviation does not monotonicly approach the DEM-MFIE value as H decreases, or if the minimum error occurs at H=h/100 while much smaller H gives notably worse accuracy, then the d→0 MFIE-convergence is not the operative mechanism for sharp-edge accuracy, confirming that finite-disk tuning drives the claimed edge performance.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central derivation in Sec. II (Eqs. 4–14) uses a multivariate Taylor expansion of the total electric field over the testing disk, then drops O(|ρ|^2) terms. This requires the field to be at least C^2 on the disk. For a point a distance ρ from a sharp edge or corner, the scattered field has a singular edge contribution scaling as ρ^{ν−1} (0<ν<1), so second derivatives are unbounded as ρ→0; the Taylor remainder is not uniformly small even as d→0 when ρ is comparable to d. The practical choice H≈h/100 places disk quadrature points at distances O(h) from edges while the disk radius is H, so H and the edge distance can be of the same order; the expansion then has no controlled accuracy. More importantly, the MFIE limit itself is known to be inaccurate at sharp edges—the factor 1/2 in the jump term (Eq. 19) fails at edges/corners. If the Curl-EFIE truly converges to a Galerkin MFIE as H→0, it would inherit that edge inaccuracy. The numerical results in Figs. 4–5 show the best accuracy at H=h/100, not at the smallest H, indicating the method's edge advantage is due to finite-H off-boundary testing, not the d→0 correspondence. Therefore the claimed MFIE-convergence and the resulting low-frequency/dense-grid stability are rigorously established only for smooth geometries; for the sharp-edged cases where the abstract claims superiority, the derivation does not apply and the empirical choice of H carries the result.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper introduces the Curl-EFIE, a new boundary integral equation for PEC electromagnetic scattering. The equation is obtained by enforcing a zero curl of the electric-field integral equation on the boundary via solenoidal, disk-shaped testing functions placed just inside the body. The authors claim that its MoM discretization converges to a Galerkin-discretized MFIE as the testing-disk radius d→0, which would combine the MFIE's favorable low-frequency and dense-mesh conditioning with the EFIE's weakly singular kernels and absence of a Gram-matrix requirement. Numerical experiments on sharp-edged objects, non-matching meshes, low frequencies, and dense meshes are presented, along with a CFIE-like combined formulation. The core mathematical claim is that the Curl-EFIE impedance matrix converges to the Galerkin MFIE matrix in the vanishing-disk limit.","tokens_in":16728,"tokens_out":11286,"duration_ms":98169,"significance":"If the convergence claim is rigorously established, the Curl-EFIE would be a practically valuable formulation: it promises MFIE-like conditioning and stability without strongly singular kernels or explicit Gram-matrix assembly, and it offers a natural extension to non-matching triangulations. The paper contains a clever first-moment testing construction, an internally consistent algebraic derivation for smooth fields, and extensive numerical evidence, including condition-number tables and RCS comparisons. The numerical results are reproducible in structure, and the paper is transparent about the practical choice of testing-disk diameter H≈h/100. However, the central mathematical claim is currently supported only by a formal asymptotic argument, and the edge-singularity issue directly affects the paper's headline advantage for sharp-edged geometries.","major_comments":[{"comment":"The derivation of the MFIE correspondence rests on the multivariate Taylor expansion in Eq. (4), which requires the electric field to be at least C^2 over the testing disk. For a PEC with sharp edges or corners, the scattered field behaves as O(ρ^{ν−1}) (0<ν<1) near the edge, so the remainder in Eq. (4) is not uniformly small as d→0. Equation (14) and the limit leading to the MFIE in Sec. III are therefore not justified for the nonsmooth geometries on which the paper claims superior accuracy. The numerical results in Figs. 4–5 show the best accuracy at H=h/100, not at the smallest H, which indicates that the sharp-edge benefit is tied to finite-H off-boundary testing rather than to the d→0 correspondence. The paper itself later states (Sec. V-C) that in practice H is \"bounded away from the vanishing limit,\" which is in tension with the abstract's convergence claim. The authors should eit","section":"Sec. II, Eqs. (4)-(14); Sec. III, Eqs. (15)-(19)"},{"comment":"The claim that the Curl-EFIE impedance elements in (28) converge to the Galerkin MFIE entries in (31) is only demonstrated for the local \"residue\" term Z_δ, which reproduces the Gram-matrix contribution. The convergence of the remaining, non-local CPV part of the MFIE matrix is not shown. The derivation of Eq. (33) is also not self-contained; it quotes a residue formula from [7] without deriving it in the context of the shrinking disk. A rigorous proof would need to show that the full double/triple integral in (28) converges to the CPV integral plus the jump term in (31) for all test/source pairs, not just for the singular coincident interaction. As it stands, the central theorem of the paper is a formal correspondence rather than a proven statement.","section":"Sec. IV, Eqs. (33)-(35)"},{"comment":"The Curl-CFIE is claimed to be interior-resonance-free because the Curl-EFIE corresponds to the MFIE in the d→0 limit. However, for any finite H—including the recommended H=h/100—the Curl-EFIE is not exactly the MFIE, so the resonance-free property does not follow automatically. The numerical evidence in Fig. 9 covers a limited range of electrical sizes (up to 0.9λ) and does not constitute a proof. The authors should either prove that the combination in (36) is resonance-free for finite H (or for H→0 with explicit error bounds) or temper the claim and provide a more extensive numerical study, particularly at frequencies where standard EFIE is known to have spurious interior resonances.","section":"Sec. IV, Eq. (36); Sec. V-D"}],"minor_comments":[{"comment":"The symbol d_∩ is used for the testing-disk radius, but the paper also uses d for the general disk radius and h for the mesh size. The notation is confusing and should be made consistent, especially in the limit statements.","section":"Eq. (33)"},{"comment":"The claim that the disk integral in (25) vanishes is stated as \"it is satisfied [19]\". This is an important identity; it should be proven inline using the divergence theorem and the fact that ψ_m is tangential to ∂D_m, rather than deferred to a conference paper.","section":"Between Eqs. (25) and (26)"},{"comment":"The phrase \"As demonstrated in section III\" appears before the residue derivation. The actual demonstration of convergence is only given in Sec. IV for the Gram term, so the cross-reference is misleading. Please rephrase to reflect what is actually proven.","section":"Sec. IV, after Eq. (35)"},{"comment":"The paper calls Curl-EFIE a first-kind equation while also emphasizing its MFIE-like conditioning. This is not contradictory, but the distinction should be clarified: the finite-H discretization is first-kind in form, while its spectrum and conditioning mimic the second-kind MFIE only as H→0. A sentence explaining this would prevent reader confusion.","section":"Sec. I and Sec. IV"},{"comment":"There are typographical/formatting issues in the equations (e.g., inconsistent bold/hat for unit vectors, missing spaces, and the use of both 𝑢 and 𝒖̂ in Eq. (34)). A careful editorial pass is needed, but these do not affect the technical content.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"The paper addresses a worthwhile problem and the numerical results are promising. The main issue is that the central mathematical claim—MoM convergence to Galerkin MFIE in the vanishing-disk limit—is not proven; the current derivation is formal and, more importantly, breaks down exactly in the sharp-edge cases that the paper highlights. The authors should be asked to either provide a rigorous asymptotic analysis (or a precise statement of the regularity assumptions) or substantially weaken the claims. The empirical H≈h/100 choice also needs a theoretical explanation or at least a clear statement that it is an empirical parameter. I would not reject the paper: the construction is novel and the numerical evidence is strong enough to warrant a major revision."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"You should know this paper is a real contribution: it introduces the Curl-EFIE, a first-kind integral equation whose MoM discretization avoids the MFIE's strongly singular kernels and Gram matrix while keeping its low-frequency and dense-mesh stability. The construction—testing the interior electric field with solenoidal disks and taking the d→0 limit to extract a curl component—is new, and the numerical evidence in the stability tables and non-matching mesh figures is convincing.\n\nThe paper does several things well. The algebra in Sec. II is consistent, the residue correspondence with the Galerkin MFIE Gram term in Eqs. (33)–(35) is plausible, and the expm1 fix for the excitation integral at low frequencies is a nice practical touch. The Curl-CFIE combination is a natural and useful extension.\n\nWhere it's soft: the d→0 convergence is shown at the level of matrix-element limits, not as an operator statement, and the Taylor expansion in Sec. II requires the field to be smooth over the disk. At sharp edges and corners—where the paper claims improved accuracy—that condition is not met. The stress-test note is right that the derivation doesn't justify the edge advantage; the paper itself shows best accuracy at H = h/100, not in the limit, so the edge improvement appears to come from off-boundary testing at finite H. That is a legitimate observation, but it means the accuracy claim for sharp edges rests on an empirical parameter choice. The paper is honest about this, but a rigorous justification or a parameter selection criterion would strengthen it. Also, no code or data is shipped, so independent replication requires reimplementation.\n\nOverall: this is a solid, useful paper for the computational electromagnetics community. The central idea is worth engaging with, and the numerical results support the main claims about stability and the Gram-matrix-free first-kind structure. The theoretical concerns do not undermine the contribution; they just delimit it. I would send this to peer review, and I'd recommend the authors add a discussion of the H-dependence and a convergence study for smooth geometries.","headline":"A genuinely new first-kind EFIE variant that numerically mimics MFIE stability with weakly singular kernels; the theory is formal and the sharp-edge advantage is empirical, but it deserves a serious referee.","tokens_in":17149,"tokens_out":4566,"would_cite":true,"duration_ms":42026,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["78A45","65R20"],"pacs":[],"model":"deepseek-v4-flash","headline":"A curl-based integral equation for scattering from metal objects turns a first-kind electric-field equation into a magnetic-field-style, stable one.","keywords":["electromagnetic scattering","boundary integral equations","method of moments","electric field integral equation","magnetic field integral equation","low-frequency stability","perfect conductor","weakly singular kernel"],"falsifier":"For a sharp-edged target, compute the Curl-EFIE impedance matrix and RCS as the disk radius goes to zero (H = h/10^3, h/10^4, ...) and compare against a high-precision Galerkin-MFIE reference. If the matrices fail to converge to the MFIE limit near the edges, or if the RCS error grows as H shrinks, the claimed equivalence is broken. A simpler check: measure the condition number for a mesh refined into a re-entrant corner; the paper predicts it should match the MFIE's ~3.5 value, so any significant departure would falsify the correspondence.","tokens_in":1316,"feed_emoji":"🌀","tokens_out":1641,"duration_ms":56150,"temperature":0.7,"pith_summary":"This paper presents a new integral equation, the Curl-EFIE, for computing electromagnetic scattering from perfectly conducting objects. Instead of forcing the electric field itself to vanish on the surface, it forces the field's rotation (curl) to vanish, evaluated on small swirling disks placed just inside the body. The central claim is that as the disks shrink, the discretized equation approaches the well-conditioned magnetic-field integral equation (MFIE), so it keeps MFIE's stability at low frequencies and on dense meshes. At the same time the equation relies on the same weakly singular kernels as the standard electric-field integral equation, which are far easier to evaluate, and it needs no Gram matrix, so it works with mismatched neighboring meshes. The paper shows numerical evidence of better accuracy than the MFIE on sharp-edged objects and a resonance-free combined version, Curl-CFIE.","feed_headline":"Disk-tested curl equation fixes low-frequency scattering","feed_subtitle":"New first-kind integral equation inherits MFIE conditioning while keeping EFIE's easy kernels.","key_machinery":"The central object is a tangent solenoidal testing disk, a small disk lying just inside the body at each boundary point whose testing function swirls around the disk center (psi = (1/N_D) c-hat x rho''). Testing the electric field with this pattern and keeping only the linear term of a Taylor expansion singles out the field's curl, D_E = c-hat dot (curl E); the identity connects the electric-field condition to the magnetic-field condition via Faraday's law. As the disk shrinks, all higher-order terms die out and the first-kind Curl-EFIE matrix approaches the second-kind Galerkin-MFIE matrix, which is the load-bearing mechanism of the paper.","core_discovery":"The central discovery is that a method-of-moments discretization of the Curl-EFIE converges, as the testing-disk radius tends to zero, to a Galerkin discretization of the MFIE. The mechanism is the curl extraction: a solenoidal testing pattern over the disk isolates the component of the electric field's curl tangential to the boundary, and by Faraday's law that component is proportional to the interior magnetic field. Consequently the null electric-field condition becomes the null magnetic-field condition, giving the Curl-EFIE the spectral behavior of a second-kind integral equation even though it is written as a first-kind one. The practical payoff is an impedance matrix that stays well-con","pith_inferences":["The disk radius H is effectively a free regularization parameter; the paper's theory fixes only the limit H to 0, so choosing a good H for arbitrary geometries, especially with corners, may require a priori information or adaptive strategies the paper does not provide.","The same curl-extraction idea might apply to other first-kind boundary integral equations (for example, dielectric or impedance surfaces), converting a first-kind operator into a second-kind-like one to improve conditioning.","The Taylor expansion of the Green's function around the disk center suggests that disk testing could be replaced by point evaluation plus local derivatives, potentially connecting to asymptotic or quadrature-based approaches in boundary element methods.","The method may be seen as a way to regularize the EFIE without changing the unknown current expansion, which could allow the use of arbitrary, non-divergence-conforming basis functions in a wider class of problems."],"forward_implications":["Engineers can build method-of-moments solvers that combine MFIE-level conditioning with EFIE-level kernel simplicity, lowering the barrier to accurate low-frequency and dense-mesh scattering analysis.","Non-matching and locally-refined meshes become straightforward for magnetic-field-quality equations, since no Gram matrix is needed across overlapping facets.","The Curl-CFIE offers a resonance-free combined equation that requires only electric-field integral operators, avoiding magnetic-field operators entirely.","A fixed four-point disk quadrature plus a Taylor expansion of the Green's function keeps the computational cost comparable to the MFIE while retaining simpler singularity treatment.","The method's stability is demonstrated numerically for very small spheres and sharp-edged objects, where standard EFIE fails."],"fun_headline_variants":["Curl-EFIE matches MFIE stability, uses weak kernels","Disk-shrink links Curl-EFIE to MFIE, fixes low-freq","Weak-singularity curl EFIE gets MFIE conditioning","Curl-EFIE: first-kind, low-frequency stable, no Gram matrix"],"cache_read_input_tokens":18432,"weakest_assumption_plain":"The proof that the disk testing extracts exactly the curl assumes the electric field is smooth enough across the disk for a first-order Taylor expansion; sharp edges and corners, where the paper claims the method works best, are precisely where the field is singular, so the limit argument doesn't strictly apply and the chosen disk size does the real work.","fun_headline_variants_meta":{"raw":{"variants":["Curl-EFIE matches MFIE stability, uses weak kernels","Disk-shrink links Curl-EFIE to MFIE, fixes low-freq","Weak-singularity curl EFIE gets MFIE conditioning","Curl-EFIE: first-kind, low-frequency stable, no Gram matrix"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000161,"raw_usage":{"total_tokens":1062,"prompt_tokens":726,"completion_tokens":336,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":470,"completion_tokens_details":{"reasoning_tokens":254}},"tokens_in":470,"tokens_out":336,"duration_ms":3874,"temperature":1.0,"reasoning_tokens":254,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-31T23:54:22.503130+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"For a sharp-edged target, compute the Curl-EFIE impedance matrix and RCS as the disk radius goes to zero (H = h/10^3, h/10^4, ...) and compare against a high-precision Galerkin-MFIE reference. If the matrices fail to converge to the MFIE limit near the edges, or if the RCS error grows as H shrinks, the claimed equivalence is broken. A simpler check: measure the condition number for a mesh refined into a re-entrant corner; the paper predicts it should match the MFIE's ~3.5 value, so any significant departure would falsify the correspondence.","supporting_citations":[],"review_version":1}