{"id":"3f40f554-74c4-440b-b713-20121b85e6a0","arxiv_id":"2505.01220","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Mirror-symmetric inverse-designed disorder places multiple reflection zeros at real frequencies, yielding experimentally validated broadband quasi-perfect transmission of predefined wavefronts in a multichannel waveguide.","lead":"This paper shows that imposing mirror symmetry in disordered waveguides makes it much easier to design structures that transmit waves almost perfectly over broad frequency ranges, confirmed by microwave experiments. The approach could lead to compact passive filters and high-transmission structures that work across wide bandwidths without active tuning.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The equivalence in Eq. (4) requires t1 invertible; the paper never checks whether optimized structures keep the half-system transmission matrix non-singular across the target band, leaving the symmetry-reduction claim conditional.","rationale":"The strongest claim of the paper is not merely that certain devices work, but that left-right mirror symmetry reduces the optimization of a reflection zero to controlling Im(r'_1), and that this reduction is what makes the inverse design tractable. Equation (4) is the mathematical pivot of that claim. The derivation of Eq. (4) via Eq. (2) requires t1 invertible, as the SM acknowledges. The paper does not state whether this condition is satisfied for the optimized structures, and the optimizer is not constrained to avoid near-singular t1. This is more central than the CDA calibration concern: the CDA and its frequency dependence are validated by the COMSOL comparison in Fig. 1(b,c) and by the microwave experiments, whereas the t1 condition has no empirical check and directly affects the validity of the symmetry argument. It is not a fatal flaw; for a generic random medium t1 is invertible, and the demonstrated devices may well satisfy the condition. But because the central theoretical claim is phrased as a clean equivalence, a missing verification of the invertibility condition is the weakest link in the argument. The proposed test is cheap: it only reuses the CDA code already developed. If the test passes, the theoretical reduction is confirmed for the reported designs and the paper can be accepted with reporting improvements (error bars, baselines); if it fails, the paper must qualify the scope of Eq. (4), and the conditional verdict is justified. I therefore mark the reader's conditional verdict as unchanged.","tokens_in":14725,"tokens_out":16821,"duration_ms":171596,"concrete_test":"For each optimized configuration reported in Figs. 1(d), 2(a,c-f), 3(a) and 4(b,c), and in particular for the 600 MHz broadband plateau structure, compute the half-system transmission matrix t1(ν) from the CDA at a dense set of frequencies across the design band (e.g., every 10 MHz from 6.5 to 7.5 GHz). Evaluate the smallest singular value σ_min(t1) and the condition number κ=σ_max/σ_min. If σ_min stays above a threshold such as 0.05·σ_max over the entire band, the factorization in Eq. (2) is valid and the symmetry-based reduction in Eq. (4) is a faithful description; if σ_min approaches zero at any frequency, the design operates near a closed channel of the half-system and the paper should report and discuss this limitation, since Eq. (4) is then not sufficient for a reflection zero.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central result of the paper is the reduction of the reflection-zero condition to det(Im(r'_1(ν0)))=0, Eq. (4). This reduction is obtained through the factorization in Eq. (2), which is derived in the Supplemental Material under the explicit assumption that the half-system transmission matrix t1 is invertible. The SM itself states that when t1 is singular, the factorization cannot be written and a closed channel with perfect reflection exists. In that regime, the full system necessarily has at least one perfectly reflected channel, so it cannot possess a reflection zero; Eq. (4) may then be satisfied even though no physical transmission state exists. The paper never checks whether the optimized configurations (Figs. 2-4) maintain a non-singular t1 over the targeted frequency bands, nor whether the optimization trajectories avoid near-singular regions. If any optimized device has a frequency in the design band where det(t1) is zero or very small, the claimed simplification of the search space by symmetry is not operative at that frequency, and the success of the optimization would have to be attributed to the full transmission matrix rather than to the reduction validated by Eq. (4). This does not invalidate the experimental demonstrations, but it leaves a gap between the stated theoretical guarantee and the numerical experiments designed to illustrate it.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents an inverse-design framework for achieving broadband quasi-perfect transmission in left-right mirror-symmetric disordered waveguides. The central theoretical result is that, for a system composed of two mirror-symmetric halves, a real-frequency reflection zero of the full system is equivalent to the condition det(Im(r'_1(ν0)))=0 on the half-system reflection matrix, reducing the number of real parameters that must be controlled. The authors implement this using a coupled-dipole-approximation model in a multichannel microwave waveguide, carry out gradient-based optimization of cylinder positions, and validate the results experimentally. They demonstrate a reflectionless exceptional point, single-mode bandpass filters for each of the four propagating modes, an ultra-broadband quasi-perfect transmission plateau over 600 MHz, and broadband transmission enhancement around a central barrier. Numerical CDA results are benchmarked against COMSOL and compared with microwave measurements.","tokens_in":14984,"tokens_out":7300,"duration_ms":74640,"significance":"The paper's strength is the combination of a clean symmetry reduction with direct experimental validation in a multimode waveguide. If the factorization leading to Eq. (4) holds, the reduction of the optimization space by a factor of two is a useful design principle for complex scattering media. The CDA model is carefully benchmarked (unitarity error below 10^-9) and the experimental agreement for the optimized devices is convincing. The demonstrations of an RL-EP, mode-selective filters, and a 600 MHz plateau go beyond prior single-frequency designs. The main caveats are the reliance on the invertibility of the half-system transmission matrix t1, which is not verified for the optimized structures, and the single-frequency calibration of the dipole polarizabilities used for broadband predictions. Both issues are addressable in revision.","major_comments":[{"comment":"The derivation of the central condition det(Im(r'_1(ν0)))=0 requires t1 to be non-singular; the SM explicitly states that when t1 is singular the factorization cannot be written and a closed channel with perfect reflection exists. The manuscript never checks whether the optimized structures in Figs. 2-4 keep det(t1) nonzero (or t1 well-conditioned) over the targeted frequency bands, nor whether the optimization trajectories avoid near-singular regions. At any frequency where t1 is singular, the full system necessarily has a perfectly reflected channel, so Eq. (4) may be satisfied without a physical reflection zero; the symmetry reduction would then not be operative. Please add a numerical assessment (e.g., the smallest singular value or condition number of t1 across the band for each optimized configuration) and discuss the implications for the theoretical claim.","section":"Theory, Eq. (2)-(4) and SM Section I"},{"comment":"The coupled dipole approximation uses bare polarizabilities α_Al = -6i and α_Teflon = 0.048i calibrated against COMSOL at a single frequency (7 GHz) and then treats them as frequency-independent throughout the 600 MHz target band. The SM states that \"the variations of α_n over the bandwidth considered in simulations are small\" but provides no quantitative data. Since the broadband quasi-perfect transmission plateau and the barrier-enhanced transmission are the central numerical results, please report the frequency dependence of the calibrated polarizabilities (or a sensitivity analysis) showing that the optimized spectra are stable under the expected drift. Without this, the reader cannot exclude the possibility that the broadband design exploits an artifact of the constant-polarizability model; the experimental agreement mitigates this concern but does not remove the need for quantitative support.","section":"SM Section II / Optimization procedure"}],"minor_comments":[{"comment":"The statement that \"real random matrices also exhibit reduced level repulsion between singular values\" cites distributions for general real and complex Gaussian matrices, whereas the object Im(r'_1) is real symmetric; the cited references are for non-symmetric random matrices. Please clarify that this is an analogy or supply the corresponding result for real symmetric matrices.","section":"Theory, paragraph after Eq. (4)"},{"comment":"The text says \"the smallest reflection eigenvalue 1−τ1(ν) shown in a dB scale reveals that the flat shape is due to the presence of three reflection zeroes,\" but the figure caption and legend do not explicitly identify which curve corresponds to 1−τ1. Please label the curves clearly.","section":"Fig. 3(a)"},{"comment":"The terms \"rainbow effect\" and \"rainbow-trapping effect\" are used to describe a phenomenon where the field pattern is highly correlated across frequencies, which is the opposite of conventional rainbow trapping (frequency-selective spatial localization). Consider renaming to avoid confusion.","section":"Abstract and Fig. 3 caption"},{"comment":"The renormalization formula \\tilde{t}_{mn} = t_{mn}\\sqrt{T^0_n} uses the empty-waveguide transmission; please clarify whether the same normalization is applied to the numerical data used for comparison in Figs. 1-4 or only to the experimental data.","section":"Eq. (6)"}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nWhat's actually new here is the inverse-design loop: they use left-right mirror symmetry to place several reflection zeros at real frequencies with prescribed input and output wavefronts, and they validate it in microwaves. The theoretical seed—det(Im(r'_1))=0 for a symmetric system—comes from earlier work, but the systematic design of RL-EPs, bandpass filters, a 600 MHz transmission plateau, and barrier enhancement is a real step forward. The CDA model is calibrated against COMSOL and matches full-wave simulations; the measurements line up with all four optimized structures. That is solid evidence.\n\nThe soft spots are mostly in the framing. The title and abstract say 'perfect' where the body honestly says 'quasi-perfect'—losses keep T below unity. The 'rainbow effect' is a loose analogy: unlike rainbow-trapping absorbers, their field patterns stay correlated across frequency rather than being sequentially trapped. The abstract also claims the optimized disorder outperforms random symmetric disorder, but I could not find the quantitative baseline in the main text; that comparison needs to be shown, or the sentence cut.\n\nThe stress-test note about t1 invertibility is a real but minor gap. Eq. (4) comes from a factorization requiring t1 non-singular; if t1 is singular at some in-band frequency, the equivalence fails and a closed channel with perfect reflection exists. The authors never check det(t1) for the optimized configurations. However, the optimization uses the full CDA scattering matrix, not Eq. (4), so the designs do not actually rest on that reduction; the experimental agreement is the real support. The technical claim should still be qualified, or det(t1) checked over the band.\n\nFinally, the polarizabilities α_Al=-6i and α_teflon=0.048i are fitted at 7 GHz and assumed constant over 600 MHz; the SM says the variation is small but gives no bound, and there are no experimental error bars. These are reporting deficiencies, not fatal flaws.\n\nMy take: a serious referee should engage with this. It's a coherent and reproducible demonstration, likely publishable after minor revisions. I would cite it in work on symmetry-enhanced wave transport, and it reads well in a group meeting.\n\nRecommendation: send it to review.","headline":"A credible inverse-design demonstration that mirror symmetry helps place multiple reflection zeros; the only real gap is some overclaiming and an unchecked technical assumption.","tokens_in":15540,"tokens_out":3672,"would_cite":true,"duration_ms":37933,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper establishes that in left-right mirror-symmetric disordered media a reflection zero at a real frequency reduces to $\\det(\\operatorname{Im}(r'_1(\\nu_0)))=0$, so inverse design only needs to tune the imaginary part of the…","keywords":["inverse design","mirror symmetry","broadband perfect transmission","disordered media","reflectionless scattering modes","exceptional points","coupled dipole approximation","microwave waveguide"],"falsifier":"Measure the transmission $T_{11}(\\nu)$ of the fabricated optimized waveguide from 6.7 to 7.3 GHz: if the CDA-planned flat plateau is not reproduced, for example if transmission drops far below unity where the model predicts $T_{11}\\approx 1$, then the claim that the half-system imaginary-reflection-matrix optimization transfers to a real device fails.","tokens_in":14524,"feed_emoji":"📡","tokens_out":10608,"duration_ms":105483,"temperature":0.7,"pith_summary":"This paper tries to establish that mirror symmetry turns the hard problem of broadband perfect transmission through a disordered medium into a much simpler real-matrix one. The load-bearing identity is that, for a left-right symmetric system, a reflection zero at frequency $\\nu_0$ exists exactly when the real symmetric matrix $\\operatorname{Im}(r'_1(\\nu_0))$ has a zero singular value. Optimizers therefore tune only the imaginary parts of the half-system reflection matrix, which makes it feasible to place several reflection zeros with prescribed input and output wavefronts in one structure. The authors use this to design a reflectionless exceptional point, single-mode bandpass filters, and a 600 MHz quasi-perfect transmission plateau in a four-mode microwave waveguide, and to lift transmission through a centered barrier from 0.28 to 0.99. If the method holds, broadband perfect transmission becomes something a structure can be designed to do, not a wavefront that must be shaped in real time.","feed_headline":"Mirror symmetry turns perfect transmission into a real-matrix problem","feed_subtitle":"Only the imaginary part of the half-system reflection matrix has to be tuned to get a flat lossless band.","key_machinery":"The machinery is the factored reflection matrix and its mirror-symmetric specialization. For two half-systems, $r = t_1^{-1}(1 - r_2 r'_1)^{-1}(r_2 - r'^*_1)t^{*-1}_1$, and when $r'_1 = r_2$ the zero-reflection condition is $\\det(\\operatorname{Im}(r'_1(\\nu_0))) = 0$. This identity carries the argument: it converts the condition that one eigenvalue of the full reflection matrix be zero into a real symmetric matrix becoming singular, which is exactly the quantity an optimizer can most easily push toward zero. It also explains why random mirror-symmetric media already show reflectionless states: the smallest singular value of a real random matrix approaches zero much faster than that of a complex random matrix.","core_discovery":"The central discovery is a symmetry reduction of the zero-reflection condition. Combining two half-systems with scattering matrices $S_1$ and $S_2$, the composite reflection matrix factorizes as $r = t_1^{-1}(1 - r_2 r'_1)^{-1}(r_2 - r'^*_1)t^{*-1}_1$, so $\\det(r(\\nu_0))=0$ becomes $\\det(r_2(\\nu_0)-r'^*_1(\\nu_0))=0$. Under left-right mirror symmetry $r'_1=r_2$, and the condition collapses to $\\det(\\operatorname{Im}(r'_1(\\nu_0)))=0$: a reflection zero is a singular value of a real symmetric matrix. This is the reduced optimization problem the paper exploits: only the $N(N+1)/2$ real entries of $\\operatorname{Im}(r'_1)$ need to be controlled instead of all complex entries of a full reflection matrix. The authors then show that multiple zeros can be positioned at chosen frequencies, each with a user-defined input and output wavefront, producing flattened exceptional-point lineshapes, bandpass filters, a 600 MHz plateau, and enhanced transmission around an opaque barrier. Experimental transmission matrices measured in the microwave waveguide reproduce the numerical spectra.","pith_inferences":["The factor-of-two reduction is generic: for $N$ propagating channels the search space has $N(N+1)/2$ real parameters, so the approach may scale to waveguides with many more modes than the four used here.","Any scalar wave system with mirror symmetry and unitary scattering, such as acoustics, elastic plates, or quantum wires, should admit the same real-symmetric zero-reflection condition, making the procedure portable outside electromagnetics.","The flat phase-delay time and frequency-correlated field maps suggest that the optimized structure behaves like a particle-like ballistic channel inside a scattering medium; a direct test would be to check whether the transmitted wavefront stays nearly identical across the whole plateau, which would make the structures useful as mode-preserving multiplexers.","For fixed obstacles, the paper's route suggests a design recipe: surround the obstacle with a mirror-symmetric disorder and tune $\\operatorname{Im}(r'_1)$ to be singular at the operating frequencies, potentially with reconfigurable elements as a tunable anti-reflection coating."],"forward_implications":["A single inverse-design run can place several reflection zeros with preselected wavefronts, so broadband quasi-perfect transmission can be engineered rather than searched for.","The optimization cost drops by a factor of two in parameter count, and each scattering-matrix evaluation takes milliseconds in the coupled-dipole model, so multi-frequency constraints over hundreds of megahertz are solvable in minutes.","The four fabricated single-mode bandpass filters, the 600 MHz plateau, and the barrier-enhanced transmission spectra all reproduce the numerical predictions in microwave measurements, indicating the designs transfer to physical devices.","Placing symmetric disorder around an otherwise opaque barrier raises average transmission from about 0.28 to about 0.99 at the target frequency and creates flat bands for individual modes, so the method works even when the central obstacle cannot be moved."],"supporting_citations":[{"why":"It supplies the composite scattering-matrix construction and the anti-reflection inverse-design problem that this paper extends to multiple zeros and broad bands.","marker":"[16]"},{"why":"It defines reflectionless modes as eigenfunctions of a non-Hermitian spectral problem, the operator whose real zeros are the reflection zeros this paper places at chosen frequencies.","marker":"[23]"},{"why":"It provides the push-through identity and the theory of reflectionless scattering modes used to factor the composite reflection matrix into the form the symmetry condition acts on.","marker":"[24]"},{"why":"It establishes the reflectionless exceptional point and its flattened quartic lineshape, which the paper engineers on demand by optimizing the symmetric disorder.","marker":"[25]"},{"why":"It supplies the mirror-symmetry constraint $r'_1 = r_2$ that turns the generalized eigenvalue condition into $\\det(\\operatorname{Im}(r'_1(\\nu_0)))=0$.","marker":"[43]"},{"why":"It demonstrates broadband transmission enhancement through symmetric diffusive slabs, the phenomenon this paper's optimization makes controllable.","marker":"[19]"},{"why":"It supplies the coupled-dipole approximation used to compute scattering matrices quickly during the inverse-design optimization.","marker":"[47]"}],"fun_headline_variants":["Symmetry makes broadband perfect transmission a real-matrix problem","Broadband perfect transmission via mirror-symmetric real-matrix design","Real-matrix shortcut to broadband perfect transmission","Symmetry reduces perfect transmission to a real-matrix condition"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole procedure relies on the coupled-dipole model, with polarizabilities calibrated at one frequency, remaining accurate over the entire target band while the scatterers stay at least three radii apart and the half-system transmission matrix stays invertible.","fun_headline_variants_meta":{"raw":{"variants":["Symmetry makes broadband perfect transmission a real-matrix problem","Broadband perfect transmission via mirror-symmetric real-matrix design","Real-matrix shortcut to broadband perfect transmission","Symmetry reduces perfect transmission to a real-matrix condition"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001241,"raw_usage":{"total_tokens":5084,"prompt_tokens":926,"completion_tokens":4158,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":542,"completion_tokens_details":{"reasoning_tokens":4093}},"tokens_in":542,"tokens_out":4158,"duration_ms":29246,"temperature":1.0,"reasoning_tokens":4093,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T04:23:34.458167+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the transmission $T_{11}(\\nu)$ of the fabricated optimized waveguide from 6.7 to 7.3 GHz: if the CDA-planned flat plateau is not reproduced, for example if transmission drops far below unity where the model predicts $T_{11}\\approx 1$, then the claim that the half-system imaginary-reflection-matrix optimization transfers to a real device fails.","supporting_citations":[{"cited_title":"Rothammer, C","cited_arxiv_id":null,"evidence_quote":"It supplies the composite scattering-matrix construction and the anti-reflection inverse-design problem that this paper extends to multiple zeros and broad bands."},{"cited_title":"Bonnet-Ben Dhia, L","cited_arxiv_id":null,"evidence_quote":"It defines reflectionless modes as eigenfunctions of a non-Hermitian spectral problem, the operator whose real zeros are the reflection zeros this paper places at chosen frequencies."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"It provides the push-through identity and the theory of reflectionless scattering modes used to factor the composite reflection matrix into the form the symmetry condition acts on."},{"cited_title":"Borcea and J","cited_arxiv_id":null,"evidence_quote":"It establishes the reflectionless exceptional point and its flattened quartic lineshape, which the paper engineers on demand by optimizing the symmetric disorder."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"It supplies the mirror-symmetry constraint $r'_1 = r_2$ that turns the generalized eigenvalue condition into $\\det(\\operatorname{Im}(r'_1(\\nu_0)))=0$."},{"cited_title":"F´ elix, and V","cited_arxiv_id":null,"evidence_quote":"It demonstrates broadband transmission enhancement through symmetric diffusive slabs, the phenomenon this paper's optimization makes controllable."},{"cited_title":"We note that the bare polarizabilities for our lossless scatter- ers are imaginary numbers that are positive for dielectric scatterers and negative for metallic scatterers","cited_arxiv_id":null,"evidence_quote":"It supplies the coupled-dipole approximation used to compute scattering matrices quickly during the inverse-design optimization."}],"review_version":1}