{"id":"c56a3bd5-d189-4549-b6a2-58504f92127e","arxiv_id":"2505.13086","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"Implementing matrix open boundary conditions in the CASCIE code removes coupler reflections and reveals that frequency-dependent reflections in a CERN-type accelerator structure are intrinsic to geometric inhomogeneity.","lead":"A modified code, CASCIE-M, adds open boundary conditions to an integral-equation solver for accelerator waveguide structures so waves can enter and exit without reflecting from couplers. The authors use it to show that reflections in a CERN-type tapered structure come from the geometry itself, and to test when a one-wave model is accurate.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Open-boundary validation rests entirely on a four-term truncation with no convergence test; the central reflection spectrum may include truncation artifacts, especially near cutoff.","rationale":"The reader's weakest assumption identifies exactly the load-bearing gap: the open boundary conditions are only validated by a single homogeneous-waveguide check, with no study of sensitivity to the four-term expansion truncation or to N_H. My review of the manuscript confirms this gap. The homogeneous test is good evidence of internal consistency, but it does not establish convergence near cutoff or for the strongly inhomogeneous CERN section, where evanescent fields are expected to be significant. Since the paper's main physical conclusion is drawn from the frequency dependence of R, a numerical artifact of the truncated boundary representation could masquerade as an inhomogeneity effect. The proposed test directly targets this by varying both the truncation order and the length of the uniform end sections; stability of R across those variations would resolve the concern, while instability would require the claim to be downgraded. This does not change the reader's CONDITIONAL verdict, because the same missing convergence evidence motivated that verdict.","tokens_in":7776,"tokens_out":3039,"duration_ms":35753,"concrete_test":"Recompute the benchmark cases with the number of expansion terms increased from 4 to 6, 8, and 10, and with N_H increased from 4 to 8 and 12, keeping all other settings fixed. Compare |R1| for the homogeneous waveguide and the full R(f) curves at the Figure 2/4 frequencies (11.994, 11.9464, 11.9168, 11.8984 GHz). If the homogeneous |R1| rises above about 1e-5, or if any point of the inhomogeneous R(f) shifts by more than about 1e-3, the open boundary conditions are not converged and the central interpretation is unsupported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is that CASCIE-M's reflected-wave spectrum is caused only by the waveguide inhomogeneity, not by reflections from the couplers. The implementation uses the open boundary conditions (Eqs. 10 and 11) with the longitudinal-field expansion truncated to four terms (footnote 3) and only N_H >= 4 uniform cells at each end. The only evidence that these conditions are reflectionless is a homogeneous-waveguide check reporting R1 < 1e-5 in the passband; no convergence study is reported in the truncation order or in N_H. This matters because R is extracted from the field in the first cell, so any error in the truncated evanescent fields near the boundary enters R directly. The risk is highest near the frequencies shown in Figure 2, where the incident mode is barely propagating or the right-hand section is cut off; there the assumption |lambda_+|<1 and |lambda_->|>1 becomes numerically marginal. Without evidence that the irregular R(f) shown in Figure 4 is stable as the truncation and the uniform end sections are increased, the interpretation that these reflections are due to inhomogeneity rather than to the open-boundary implementation is not fully established. The paper itself flags the four-term limitation in footnote 3 but does not test its convergence, and no independent solver comparison is provided.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports a modification of the CASCIE coupled-integral-equations code, called CASCIE-M, that replaces coupler-type boundary conditions with open (reflectionless) boundary conditions. The open conditions are constructed from eigenvector/Floquet solutions of semi-infinite uniform waveguides, as described in Eqs. (3)-(10). The code is tested on homogeneous disk-loaded waveguides, where the computed reflection coefficient is reported to be below about 1e-5 in the passband, and is then applied to a 26-cell CERN-type accelerating section with uniform cells added at each end. The main physical claim is that the frequency dependence of the reflection coefficient shown in Figure 4 lacks the regular maxima and minima associated with coupler resonances, indicating that the reflections are caused by the geometric inhomogeneity rather than by the couplers. The paper also uses CASCIE-M fields to assess the accuracy of a single-mode field representation based on the author's modified eigenfunctions.","tokens_in":8059,"tokens_out":4656,"duration_ms":47107,"significance":"If the open-boundary implementation is reliable, the paper offers a useful computational tool for studying finite structured waveguides without coupler-induced reflections, and it provides quantitative evidence relevant to the author's coupled-mode theory: the single-mode representation is reported to be accurate at several selected frequencies, while the negative-traveling component exhibits a complex spatial structure. The homogeneous-waveguide check is a good internal sanity test, and Figure 4 constitutes an explicit, falsifiable prediction that could be checked by an independent solver or by measurement. However, the numerical evidence currently lacks a convergence study and independent validation, so the significance of the central claim is conditional on the robustness of the open-boundary implementation.","major_comments":[{"comment":"The central validation of the open boundary conditions is a single homogeneous-waveguide test with the longitudinal-field expansion truncated to four terms and with N_H >= 4 uniform cells at each end, but no convergence study in the truncation order or in N_H is reported. Because the reflection vector R appears directly in Eq. (10) and is extracted from the field at the first cell, any error in the truncated evanescent fields near the boundary propagates directly into R. The claim that Figure 4's R(f) is caused by inhomogeneity rather than by the open-boundary implementation would be established only by showing that R(f) is stable as the number of expansion terms and the length of the uniform end sections are increased.","section":"Numerical Results (footnote 3 and Eq. (10))"},{"comment":"The homogeneous-waveguide test is reported only for passband frequencies, with R1 < 1e-5, while Figure 4 includes frequencies where the incident mode is barely propagating in the input section or evanescent in the output section, i.e., near the turning-point region of Figure 2. In such regimes the identification of Floquet coefficients satisfying |lambda_+| < 1 and |lambda_-| > 1 in Eq. (6) becomes numerically marginal, and the four-term truncation may be especially inaccurate. The manuscript does not report how the homogeneous-waveguide reflection behaves near cutoff or how the computed R(f) changes when those frequencies are approached, which is precisely the regime where the open-boundary conditions are most at risk.","section":"Numerical Results (Fig. 2 and Eq. (6))"},{"comment":"The central interpretation that the irregular frequency dependence of R in Figure 4 is 'associated with inhomogeneity' is not independently corroborated. No comparison is made with an independent solver, such as a finite-element model with absorbing ports or another mode-matching implementation, nor with measured data for the CERN section. The single-point agreement at the working frequency (|R1| approximately 0.083, compared with 0.079 for the coupler-terminated section) is suggestive but does not validate the full spectrum. An independent check is needed to rule out numerical artifacts of the truncated open-boundary implementation.","section":"Numerical Results (Fig. 4)"}],"minor_comments":[{"comment":"The title should read 'calculation of the characteristics' and the first section heading contains a typo: 'INRODUCTION' should be 'INTRODUCTION'.","section":"Title and Section 1 heading"},{"comment":"The matrix equation (1) is typeset ambiguously; the superscripts on T and Q are easily confused, and the definitions of T^{k}, T^{k+}, and T^{k-} should be stated explicitly in the text.","section":"Eq. (1)"},{"comment":"The actual numbers of expansion terms and the exact N_H values used for each reported calculation are not stated; only N_H >= 4 is mentioned. Please specify these parameters in the figure captions or in the text so that the results are reproducible.","section":"Numerical Results"},{"comment":"The caption of Figure 3 does not make clear whether the plotted quantities are the real part, imaginary part, magnitude, or phase of the complex longitudinal field; the labels 1-4 and 1a-4a should be defined in a legend.","section":"Figure 3 caption"},{"comment":"The notation E_z(0, z) is used without defining the first argument as the radial coordinate r = 0; this should be stated explicitly to avoid confusion.","section":"Eqs. (16)-(17)"}],"recommendation":"major_revision","confidential_remarks":"The paper relies heavily on the author's own prior works (refs. [7]-[10], [14]) for the modified eigenfunctions and the coupled-mode interpretation, and the numerical validation of those foundations is outside the present manuscript. This is not by itself a reason to reject, but it makes independent verification of the open-boundary implementation especially important. I would encourage the editor to request, as part of the revision, a convergence study and an independent solver comparison for the reflection spectrum."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague, quick read on arXiv:2505.13086.\n\nThe genuinely new thing here is the open-boundary-condition formulation inside the author's own CIEM code (CASCIE-M), plus a first look at reflection spectra of the CERN section without coupler contamination. That is a real extension: prior CASCIE versions had reflecting end boundaries, and the ability to study intrinsic inhomogeneity reflections is useful for accelerator structure designers. The homogeneous-waveguide test, with R1 ~ 1e-6, is a solid internal sanity check that the open boundaries are approximately reflectionless, and the paper is honest in footnote 3 that only four terms are retained.\n\nWhat the paper does well: it separates the single-mode expansion error into the full representation and the positive-travelling-only part, showing the latter is much larger. That is a useful caution for people building coupled-mode theories. The CERN structure example is practical.\n\nSoft spots. The stress-test concern is valid and it lands: the four-term truncation is never tested. The only evidence that the open boundary is clean is the homogeneous case, which is an internal consistency check, not a convergence study. Near cutoff, Figure 2 shows frequencies where the incident mode is barely propagating in the first homogeneous section and possibly cut off in the last; there the Floquet exponent assumptions get numerically marginal, and R is extracted from the field in the first cell, so truncation error bleeds directly into R. Without a sweep in the number of expansion terms and in N_H, the claim that the Figure 4 spectrum is inhomogeneity rather than boundary artifact is not fully established. Also, the abstract promises a comparison with the coupler case, but Figure 4 shows only the new spectrum; the actual difference from CASCIE-with-couplers is asserted, not displayed. The equations are badly typeset in the arXiv version, so verifying the algebra takes real effort. No code or data is shipped, so reproducibility rests on description only.\n\nBut I want to be fair: the circularity burden is low. The homogeneous test is exactly the right kind of check, and the single-mode comparison uses an independently computed reference field. The self-citations are to the author's own prior work, but that is natural for an incremental extension.\n\nConclusion: worth a serious referee, not a desk reject. The referee should demand a convergence study in truncation order and N_H, and a side-by-side with the old coupler-loaded boundary conditions. If those come back stable, this is a useful tool paper for the accelerator waveguide subfield.\n\nRecommendation: send to peer review; conditional on the missing convergence evidence.","headline":"Useful incremental extension of the author's CIEM code, with a plausible central claim that is under-supported by a missing convergence study.","tokens_in":8510,"tokens_out":2296,"would_cite":false,"duration_ms":23372,"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":"Coupled integral equations with matrix open boundary conditions can model a finite structured waveguide without reflections from its couplers, and tests on an accelerator-style chain attribute the remaining reflection to geometric…","keywords":["structured waveguides","coupled integral equations","open boundary conditions","reflection coefficient","disk-loaded waveguide","coupled mode theory","Floquet coefficients","accelerator structure"],"falsifier":"Rerun the same inhomogeneous chain with more than four expansion terms and with longer uniform end sections, especially at frequencies where only one end waveguide propagates the incident mode; if the reflection coefficient changes noticeably, the open boundary conditions are not converged.","tokens_in":7589,"feed_emoji":"📡","tokens_out":7916,"duration_ms":75100,"temperature":0.7,"pith_summary":"This paper reports a modification of the coupled integral equations method for structured waveguides, implemented in a code called CASCIE-M, that replaces closed waveguide boundary conditions with matrix open boundary conditions. The aim is to let an electromagnetic wave enter and leave a finite chain of cells, such as a disk-loaded accelerator section, without the extra reflections produced by input and output couplers. Testing on a prototype accelerator section shows that the open-boundary code yields a reflection coefficient whose frequency dependence has no regular alternation of minima and maxima, indicating that the remaining reflections come from the geometric inhomogeneity of the cells rather than from coupler resonances. The method also provides a way to assess the accuracy of the field expansion on which coupled mode theory for structured waveguides can be built.","feed_headline":"Open-boundary code strips coupler reflections from waveguide models","feed_subtitle":"Integral-equation method isolates reflections caused by a cell chain's own geometry.","key_machinery":"The central object is a set of matrix open boundary conditions appended to the coupled-cell matrix equations. Each end section is treated as a uniform, semi-infinite waveguide, and its eigenvectors and Floquet coefficients determine which modes are incoming, reflected, or transmitted. The resulting equations form a closed linear system that couples the reflection vector $R$, the transmission vector $V$, and the cell field coefficients $Q^{(k)}$, with the only exciting term being the amplitude of the incident fundamental mode. This machinery also supports a numerical check of coupled-mode truncation: the field coefficients from the open-boundary solution are expanded in modified eigenvector functions, giving relative errors for single-mode and positive-mode representations.","core_discovery":"The central claim is that matrix open boundary conditions can be formulated directly in the coupled-matrix description of a structured waveguide, and that when implemented they make a finite inhomogeneous structure behave as if it were embedded in two semi-infinite uniform waveguides carrying only the incident and transmitted modes. The boundary conditions are expressed through eigenvectors and Floquet coefficients of the transfer matrices of the uniform end waveguides, producing a closed linear system for the cell field coefficients together with reflection and transmission vectors. For a homogeneous test waveguide, the computed reflection coefficient in the passband is below about $10^{-5}$, which the paper takes as validation of the open boundary procedure. For an inhomogeneous accelerator-type chain, the reflection coefficient versus frequency no longer shows the regular minima and maxima characteristic of coupler resonances; the paper interprets the residual reflection as intrinsic to the inhomogeneity. The same tool is then used to test single-mode expansions of the field in terms of modified eigenvector functions, finding the single-mode representation accurate while the positive-traveling-wave part alone is not.","pith_inferences":["Because the open-boundary code removes coupler resonances, the remaining reflection spectrum should track the intrinsic band structure of the cell chain; a direct test would be to overplot the reflection curve on the dispersion curves of the end and central uniform waveguides and check that reflection peaks align with passband edges or turning points.","The paper's four-term truncation of the field expansion is untested by convergence studies, so the reported single-mode errors should be considered preliminary; re-running the same cases with more terms would either confirm or shift the two-orders-of-magnitude gap between single-mode and positive-mode representations.","The open-boundary reflection and transmission coefficients could serve as reference data for calibrating reduced coupled-mode differential equations, letting the single-mode approximation be validated cell by cell without designing physical couplers.","The ambiguous 'left-traveling' component seen in inhomogeneous waveguides could be reinterpreted through local Bloch-mode interference; computing its phase derivative locally at the tested frequencies would show whether it really behaves as a backward wave or as a modulated forward wave."],"forward_implications":["For homogeneous disk-loaded waveguides, the open-boundary code produces a passband reflection coefficient below about $10^{-5}$, meaning artificial reflections introduced by the boundaries are negligible.","For the tested inhomogeneous accelerator-type chain, the reflection coefficient versus frequency has no regular minima and maxima, so the residual reflections are attributed to the structure's geometric inhomogeneity rather than to coupler resonances.","At the working frequency, the reflection coefficient obtained with open boundaries is practically the same as that of the section with couplers, showing that coupler reflections do not dominate the working-frequency match.","The single-mode representation of the field is accurate for the tested geometry, while the positive-traveling-wave part alone is roughly two orders of magnitude less accurate, showing that the second component is essential.","Open boundary conditions allow frequency and size scans of a finite structured waveguide without retuning couplers at every point.",""],"supporting_citations":[{"why":"Supplies the coupled-cell matrix equations and the eigenvector/Floquet-coefficient solution used to derive the open boundary conditions.","marker":"[4]"},{"why":"Provides the original CASCIE code whose boundary conditions are replaced in CASCIE-M.","marker":"[6]"},{"why":"Introduces modified eigenvector functions used for the field expansion in inhomogeneous waveguides.","marker":"[7]"},{"why":"Defines the single-mode representation and relative-error formulas that the paper tests against CASCIE-M.","marker":"[8]"},{"why":"Introduces the phase method and prior analysis of the left-traveling component in inhomogeneous accelerating sections.","marker":"[9]"},{"why":"Supplies the geometric and RF parameters of the prototype accelerator section used in all numerical tests.","marker":"[15]"}],"fun_headline_variants":["Open boundaries end coupler echoes in waveguide modeling","Integral equations strip coupler reflections from waveguide sims","New boundary method isolates waveguide's own reflections","Coupled-integral code removes coupler spurious reflections"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The treatment assumes each end is a uniform semi-infinite waveguide in which only the incident mode propagates, and the implemented version keeps only four field-expansion terms at the ends, so accuracy near cutoff or for strongly evanescent end modes is unverified.","fun_headline_variants_meta":{"raw":{"variants":["Open boundaries end coupler echoes in waveguide modeling","Integral equations strip coupler reflections from waveguide sims","New boundary method isolates waveguide's own reflections","Coupled-integral code removes coupler spurious reflections"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000439,"raw_usage":{"total_tokens":2203,"prompt_tokens":896,"completion_tokens":1307,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":512,"completion_tokens_details":{"reasoning_tokens":1245}},"tokens_in":512,"tokens_out":1307,"duration_ms":10568,"temperature":1.0,"reasoning_tokens":1245,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T20:19:55.078695+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Rerun the same inhomogeneous chain with more than four expansion terms and with longer uniform end sections, especially at frequencies where only one end waveguide propagates the incident mode; if the reflection coefficient changes noticeably, the open boundary conditions are not converged.","supporting_citations":[],"review_version":1}