{"id":"890248de-62dd-47d8-8bfa-dfbae341e7ea","arxiv_id":"1908.04764","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"high","formal_verification":"none","parameter_count":0,"one_line_summary":"The paper gives a torus-based Sommerfeld integral formalism for discrete lattice Helmholtz problems, with recursive formulas for the Green's function, an algebraic solution for the half-line, and an elliptic-function representation for the right angle.","lead":"This paper develops a discrete analogue of the Sommerfeld integral to solve three lattice wave problems: the Green's function on a plane, diffraction by a Dirichlet half-line, and diffraction by a Dirichlet right angle. The solutions are expressed as integrals on a torus, with recursive, algebraic, and elliptic-function forms respectively.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Transformant (70) has a residue normalization error: as written, AΨ has residue −2πi instead of the required −(2πi)⁻¹ at the incident-wave pole, so (44)+(70) does not solve the right-angle problem without correction.","rationale":"The reader's verdict is CONDITIONAL and identifies the right-angle transformant as the weakest point. My stress-test agrees that the sufficiency of A(p) is not established, but I found a more specific and checkable defect: formula (70) does not meet the residue normalization required by the functional problem in §4.3. Since Res[AΨ] enters the incident-wave amplitudes directly, the claimed solution is wrong as written unless the prefactor is a typo. This is a concrete algebraic check, not merely a missing verification. At the same time, the rest of the paper—especially the Green's function analysis and the half-line solution—is internally coherent, and the half-line example shows the intended residue technology works when properly normalized. Therefore the appropriate outcome is still a conditional one: the right-angle claim can be accepted only after the normalization is fixed and the Dirichlet boundary conditions and radiation/uniqueness are checked. The reader's weakest assumption was broader; my concern is a sharper instance of it, hence 'partial' agreement. The verdict should remain CONDITIONAL, meaning the reader's original CONDITIONAL verdict is unchanged but now with a specific corrective task identified.","tokens_in":21785,"tokens_out":7614,"duration_ms":81919,"concrete_test":"Compute the Laurent expansion of A(t) in (70) at t=t0+ω2 using the principal part of E given in (68). Since dt=Ψ, evaluate Res[AΨ] and compare with −(2πi)⁻¹ demanded in §4.3; if it equals −2πi, the prefactor in (70) must be corrected to 1/(2πi). After correcting the prefactor, additionally evaluate (44) by residue expansions at the infinity points for the corner-adjacent nodes (1,0), (0,1), (2,0), (0,2) and verify the Dirichlet conditions (65), as was done for the half-line in (58).","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is that (44) with A from (70) solves the Dirichlet right-angle problem. For this, A must satisfy the functional problem in §4.3: meromorphic on H3, poles only at the four incident-wave images, and Res[AΨ] = ∓(2πi)⁻¹ there. The construction uses E(t,ω1,3ω2), whose pole at t0+ω2 has principal part 1/(t−t0−ω2) by (68). Since t(p)=∫Ψ, so dt=Ψ in the t-coordinate, formula (70) gives A(t) ≈ −2πi/(t−t0−ω2) near that pole, hence Res[AΨ] = −2πi, not −(2πi)⁻¹. The same factor appears at the other three poles. Consequently, the contour integral (44) produces incident-wave contributions with amplitude (2πi)² times the prescribed one, so the total field cannot equal the physical scattered field unless the prefactor 2πi in (70) is a typo for 1/(2πi). Even after correcting this normalization, the paper does not check that the resulting field satisfies (65) or the radiation condition: for the half-line the Dirichlet condition was verified in (58), but the analogous computation for the right angle is absent. The right-angle solution is therefore asserted, not demonstrated.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops Sommerfeld-type integral representations for three discrete Helmholtz problems on a square lattice: the Green's function for a point source, diffraction by a Dirichlet half-line, and diffraction by a Dirichlet right angle. The authors introduce the dispersion surface H, which is a torus, and rewrite single-integral Green's function representations as contour integrals of a meromorphic 1-form over contours on H. For the half-line problem, they construct a two-sheeted covering H2 and solve a functional problem for the Sommerfeld transformant in terms of algebraic functions, verifying the Dirichlet condition and comparing the result with the Wiener-Hopf solution. For the right-angle problem, they introduce a three-sheeted covering H3 and propose a transformant built from elliptic functions, asserting that the resulting Sommerfeld integral solves the problem.","tokens_in":22118,"tokens_out":12362,"duration_ms":123534,"significance":"If the claims are correct, the paper gives a unified analytic framework for discrete diffraction, with the half-line solution independently checked against the Wiener-Hopf method and the right-angle solution being a new result in terms of elliptic functions. The derivation of recursive relations for the Green's function from the torus representation is elegant and potentially useful computationally. The paper contains no fitted parameters; the transformants are fixed by prescribed residue conditions, and the half-line result is cross-validated by an independent method, which are strengths. However, the right-angle construction, which is the paper's main advertised novelty, is asserted rather than demonstrated, and the transformant in equation (70) has a residue-normalization error. These issues bear directly on the central claim of Section 4.","major_comments":[{"comment":"The prefactor in equation (70) gives the wrong residues for the form AΨ. Since t(p) is defined by t = ∫Ψ in equation (66), we have dt = Ψ in the t-coordinate, and the function E in equation (68) has principal part 1/(t − a) at its pole. Therefore, at the pole t = t0 + ω2, the function (70) has principal part −2πi/(t − t0 − ω2), so Res[AΨ] = −2πi. The functional problem in Section 4.3 requires Res[AΨ] = −(2πi)^{-1} at this pole, with analogous requirements at the other three poles. The same factor 2πi appears at all four poles, so the polar contributions in the integral (44) are amplified by (2πi)^2 relative to the prescribed incident waves. As written, (44) with (70) does not solve the right-angle problem. The prefactor 2πi in (70) should presumably be 1/(2πi), and the residue normalization must be verified after this correction.","section":"Section 4.4, Eq. (70)"},{"comment":"The claim that 'the integral (44) with (70) provides the solution for the right-angled wedge problem' is asserted, not demonstrated. For the half-line problem the authors explicitly verify the Dirichlet condition in equation (58) and compare with the Wiener-Hopf solution in Section 3.6. For the right-angle problem, no analogous verification of the boundary conditions (65) on the two rays is given, no check of the radiation condition is provided, and no uniqueness argument is made. The transformant (70) is constructed as an ansatz satisfying a set of residue conditions; satisfying those conditions is not shown to be sufficient for the integral to satisfy the Dirichlet conditions. This gap is load-bearing because the right-angle solution is the main new result of the paper.","section":"Section 4.4, final paragraph"},{"comment":"The correspondence between the four physical pole locations and the t-plane offsets (t0 + ω2, 2ω2 − t0, t0 + 5ω2/2, ω2/2 − t0) is stated without proof. The construction only works if, for all admissible φin and K, these four points are pairwise incongruent modulo the period lattice of (ω1, 3ω2) and produce exactly the four required simple poles in one fundamental parallelogram. If two of the offsets coincide modulo the lattice for some parameter values, the residue conditions in Section 4.3 cannot be satisfied simultaneously by (70). The authors should either prove this non-coincidence for the stated parameter range or make explicit the generic restriction under which (70) is valid.","section":"Section 4.4, pole placement before Eq. (70)"}],"minor_comments":[{"comment":"The phrase 'triple periodic' should be 'doubly periodic', since the function is periodic with periods ω1 and 3ω2.","section":"Section 4.4, text after Eq. (70)"},{"comment":"The word 'Diffracton' in the section heading is a typo; it should be 'Diffraction'.","section":"Section 4 title"},{"comment":"The title of reference [2] contains typos: 'Theoretical about the difraction of x-rays' should likely read 'The theory of diffraction of X-rays' or similar.","section":"Reference [2]"},{"comment":"The statement that 'A(x) should be an algebraic function that can be expressed explicitly and should contain only square and cubic radicals' is not proved in the paper, and the final representation (70) is left in terms of elliptic functions. This claim should either be demonstrated or softened to avoid presenting an unproved structural assertion.","section":"Section 4.4, paragraph on algebraic functions"}],"recommendation":"major_revision","confidential_remarks":"The right-angle problem is the paper's advertised novelty, and the current manuscript does not establish that the proposed elliptic transformant solves it: there is a residue-normalization error in (70), and the Dirichlet and radiation conditions are not verified. The half-line section, by contrast, is careful and cross-checked against the Wiener-Hopf solution. I would be willing to reconsider after a major revision in which the normalization is fixed, the boundary conditions and radiation behavior are checked (or the result is explicitly labelled as a conjecture), and the pole-placement issue is resolved. I did not see signs of circularity or fitted parameters."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper gives a Sommerfeld integral formalism for lattice Helmholtz problems. The genuinely new and solid parts are: (i) the interpretation of the Green's function as a period of an elliptic integral on the dispersion torus, with recursive relations that avoid numerical integration; (ii) an algebraic-function Sommerfeld transformant for the Dirichlet half-line, checked explicitly against the Wiener-Hopf solution. Those parts are worth reading, and the half-line result in particular is an advance over the earlier elliptic-integral representation.\n\nThe soft spot is the right-angle section. The authors state that (44) with the transformant (70) solves the right-angle problem, but they do not verify the Dirichlet boundary condition (65), the radiation condition, or uniqueness. The half-line case had an explicit boundary check (58); the right-angle case has none. Moreover, there appears to be a concrete normalization error in (70): since t(p) is defined by dt = Ψ, the principal part of A at t0+ω2 is -2πi/(t-t0-ω2), so the residue of AΨ is -2πi rather than the required -(2πi)^{-1}. A similar factor appears at the other three poles. Unless the prefactor is a typo for (2πi)^{-1}, the contour integral produces incident waves with the wrong amplitude. Even with that corrected, the solution is asserted rather than demonstrated.\n\nThis is not a takedown: the first two-thirds of the paper are careful, and the right-angle issue looks fixable. But the central claim for the right-angle problem, as written, does not hold.\n\nWho should read it: people working on lattice scattering, fracture mechanics with discrete models, or computational benchmarking of Green's functions. They will get value from the half-line and Green's function sections. The right-angle section should be treated as a conjecture with a likely typo.\n\nRecommendation: send to peer review, but with a request for the authors to fix the normalization and to actually check the boundary condition and radiation condition for the right-angle case. It deserves referee time, not a desk reject.","headline":"Solid half-line and Green's function results, but the right-angle solution is asserted with a residue normalization error that as written breaks the incident-wave amplitude.","tokens_in":22555,"tokens_out":6941,"would_cite":true,"duration_ms":63169,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["39A14","30E20","33E05"],"pacs":[],"model":"deepseek-v4-flash","headline":"A torus of plane waves carries explicit solutions for three discrete diffraction problems.","keywords":["discrete Helmholtz equation","Sommerfeld integral","diffraction by half-line","diffraction by right angle","dispersion surface","torus","elliptic functions","Green's function"],"falsifier":"Evaluate the integral (44) with the transformant (70) at the nodes $(m,0)$ for $m>0$ and $(0,n)$ for $n>0$, using the residue-at-infinity procedure applied to the half-line problem; if any of these values is nonzero, or if the homogeneous equation (11) fails at a node adjacent to the corner, the claimed right-angle solution is incorrect.","tokens_in":21553,"feed_emoji":"📐","tokens_out":8798,"duration_ms":76249,"temperature":0.7,"pith_summary":"The paper develops a discrete analogue of the Sommerfeld integral for three problems on the square lattice: the point-source Green's function, diffraction by a Dirichlet half-line, and diffraction by a Dirichlet right angle. It shows that in all three cases the total field is a contour integral of the algebraic form $x^m y^n dx/(x(y-y^{-1}))$ over the dispersion surface $H$, which is a torus. This unified viewpoint yields an explicit Green's function determined by two starting values and a linear recurrence, an algebraic-function representation for the half-line problem, and an elliptic-function representation for the right-angle problem. The right-angle statement is the paper's central new application: an analytic solution to a problem the authors found unexamined in the literature.","feed_headline":"Torus integrals solve three discrete diffraction problems","feed_subtitle":"Green's function, half-line, and right-angle fields become explicit algebraic or elliptic expressions.","key_machinery":"The central object is the dispersion torus $H$ together with the analytic one-form $\\Psi=dx/(x(y-y^{-1}))$; plane waves $x^m y^n$ restricted to $H$ form the integrand. The mechanism is contour deformation on $H$ and on its coverings $H_2$ and $H_3$, where the observation angle $\\varphi$ has period $4\\pi$ and $6\\pi$, respectively; this enforces the reflection principle that converts Dirichlet boundaries into branched sheets. The only unknown is the Sommerfeld transformant $A(p)$: for the half-line it is an algebraic function with prescribed residues at two poles, and for the right angle it is assembled from the elliptic function $E(t,\\omega_1,3\\omega_2)$ at four pole locations.","core_discovery":"For the discrete Helmholtz equation $u(m+1,n)+u(m-1,n)+u(m,n+1)+u(m,n-1)+(K^2-4)u(m,n)=0$, the paper identifies the set of all plane waves, the dispersion surface $H$ cut out by $x+x^{-1}+y+y^{-1}+K^2-4=0$, as a torus, and shows that the one-form $\\Psi=dx/(x(y-y^{-1}))$ is analytic on it. The Green's function is a period of an elliptic integral of $\\Psi$ over four homotopic contours on $H$, and the same contour-integral representation, after passage to the two- and three-sheet coverings $H_2$ and $H_3$, gives the half-line and right-angle diffraction fields. The half-line Sommerfeld transformant is found in closed algebraic form, while the right-angle transformant is built from the elliptic function (68), and the paper states that the integral (44) with (70) provides the solution of the right-angled wedge problem.","pith_inferences":["A direct check not performed in the paper is to substitute the elliptic transformant (70) into the integral (44) and compute residues at the infinity points on the half-lines $n=0, m>0$ and $m=0, n>0$; vanishing of those residues would confirm the Dirichlet boundary conditions (65).","The construction suggests a natural family: a Dirichlet wedge of aperture $\\pi/2$ corresponds to the three-sheeted covering $H_3$, so a wedge of aperture $p\\pi/2$ would plausibly correspond to a $p$-sheeted covering with a transformant built from elliptic functions of period $p\\omega_2$.","The linear recurrence (77) for the Green's function has coefficients depending on $K$, so its numerical stability as $|m|+|n|$ grows is not automatic; a stability analysis would determine the practical range of the recursion."],"forward_implications":["The lattice Green's function can be computed from $u(0,0)$ and $u(2,0)$ alone, with all other values generated by the recurrence (77), so repeated numerical integration is unnecessary.","For the half-line problem, explicit formulae (59) and (60) express every node value as an algebraic function of $K$ and the incidence data, and the Dirichlet condition on the half-line is verified by residue evaluation.","The same contour formalism applies to the right-angle problem, where the field integral can be evaluated by residues at the twelve infinity points of $H_3$ once the elliptic transformant is computed numerically.","The continuous Sommerfeld construction is thereby transferred to discrete lattices, with the dispersion torus replacing the tube of propagation angles and the coverings $H_2$, $H_3$ replacing multi-sheeted angle strips."],"supporting_citations":[{"why":"Supplies the original Sommerfeld half-plane integral that the discrete construction parallels.","marker":"[1]"},{"why":"One of the Wiener-Hopf solutions of the lattice half-line problem that the new Sommerfeld solution must reproduce.","marker":"[17]"},{"why":"Earlier lattice crack/half-line solution establishing the known result for the half-line diffraction problem.","marker":"[15]"},{"why":"Provides the Wiener-Hopf factorization and the comparison solution (95) used to verify the half-line Sommerfeld result.","marker":"[18]"},{"why":"Underlies the reduction of the Green's function integral to a combination of basic elliptic integrals, yielding the recursive relations.","marker":"[21]"},{"why":"Supplies the saddle-point method that identifies the real-waves line and the far-field asymptotics of the Sommerfeld integral.","marker":"[22]"},{"why":"Provides the theory of the elliptic function E used to build the right-angle transformant (70).","marker":"[25]"},{"why":"Supplies the theory of analytic 1-forms on tori that justifies contour deformation on H and its coverings.","marker":"[19]"}],"fun_headline_variants":["Torus integrals crack three discrete diffraction problems","Discrete Helmholtz solved via Sommerfeld torus integrals","Explicit solutions for discrete diffraction on torus contours","Torus method makes discrete diffraction fields explicit"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"For the right-angle problem, the load-bearing premise is that the elliptic-function transformant (70) enforces the Dirichlet boundary conditions (65) and the radiation condition; the paper asserts the solution property without checking these conditions, so if the poles and periods of $A$ fail to make the integral vanish on the two half-lines, the right-angle result would not stand.","fun_headline_variants_meta":{"raw":{"variants":["Torus integrals crack three discrete diffraction problems","Discrete Helmholtz solved via Sommerfeld torus integrals","Explicit solutions for discrete diffraction on torus contours","Torus method makes discrete diffraction fields explicit"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000838,"raw_usage":{"total_tokens":3616,"prompt_tokens":871,"completion_tokens":2745,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":487,"completion_tokens_details":{"reasoning_tokens":2694}},"tokens_in":487,"tokens_out":2745,"duration_ms":19467,"temperature":1.0,"reasoning_tokens":2694,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T15:07:04.874895+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Evaluate the integral (44) with the transformant (70) at the nodes $(m,0)$ for $m>0$ and $(0,n)$ for $n>0$, using the residue-at-infinity procedure applied to the half-line problem; if any of these values is nonzero, or if the homogeneous equation (11) fails at a node adjacent to the corner, the claimed right-angle solution is incorrect.","supporting_citations":[{"cited_title":"Sommerfeld","cited_arxiv_id":null,"evidence_quote":"Supplies the original Sommerfeld half-plane integral that the discrete construction parallels."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"One of the Wiener-Hopf solutions of the lattice half-line problem that the new Sommerfeld solution must reproduce."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Earlier lattice crack/half-line solution establishing the known result for the half-line diffraction problem."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the Wiener-Hopf factorization and the comparison solution (95) used to verify the half-line Sommerfeld result."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Underlies the reduction of the Green's function integral to a combination of basic elliptic integrals, yielding the recursive relations."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the saddle-point method that identifies the real-waves line and the far-field asymptotics of the Sommerfeld integral."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the theory of the elliptic function E used to build the right-angle transformant (70)."},{"cited_title":"Hurvitz and R","cited_arxiv_id":null,"evidence_quote":"Supplies the theory of analytic 1-forms on tori that justifies contour deformation on H and its coverings."}],"review_version":1}