{"id":"61b37832-f369-4d83-b776-c19e13138b65","arxiv_id":"2411.12283","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"An isospectral reduction of the SSH3 lattice to a two-site energy-dependent model yields a winding-number bulk-edge correspondence, verified acoustically.","lead":"This paper maps a three-site SSH3 lattice onto a simpler two-site model with the same spectrum, and shows that the reduced model's winding number counts the topological edge states of the original chain. The approach is demonstrated in an acoustic experiment, where edge-state pairs are observed at two predicted frequencies.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The bulk-edge correspondence is inferred from semi-infinite solutions of ρ=0 without solving the two-boundary finite-chain problem, so the winding-number count could be altered by finite-size boundary conditions.","rationale":"The paper's isospectral reduction is algebraically transparent, the energy-dependent coupling ρ is a neat construction, and the experimental data agree with the predicted edge-state profiles. The main gap is that the bulk-edge correspondence is not proven: the derivation from ρ=0 gives evanescent bulk solutions, while the finite chain is treated by assuming independent terminations. This is a genuine soft spot because the central claim is a counting statement about the finite chain. However, the finite-chain spectra in Fig. 2(a) and the experiments for the deep-blue phase provide nontrivial support, so the correct status is conditional: the correspondence is plausible and partially verified but lacks a rigorous two-boundary derivation. The concrete test proposed (solving the finite-chain quantization condition and comparing with ν) would settle whether the missing derivation is a real flaw or merely a presentation gap. I therefore agree with the reader's weakest assumption and recommend keeping the CONDITIONAL verdict.","tokens_in":8809,"tokens_out":18888,"duration_ms":212265,"concrete_test":"Solve the finite-chain boundary value problem of the isospectrally reduced model: impose pC_0=0 and pA_{N+1}=0 on the general solution of the recurrence derived from H(ε)φ=ε^2φ, and derive the quantization condition for edge states as a function of N. For the three parameter points of Figs. 2(c)-(e), count the solutions with ζ>0 and energies in the bandgaps for N=31 and N=100, and compare with the total winding number ν. If the count matches ν in all phases, the finite-boundary assumption is benign; if not, the claimed bulk-edge correspondence needs revision.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that the winding number of ρ counts edge-state pairs (ν=0,2,4) rests on the step following Eq. (4), where edge states are inferred from complex wavevector solutions of ρ=0 with the evanescence condition ζ>0. This is a semi-infinite-chain argument: the ansatz (5), e.g. ϕ_l ∝ [1;0] e^{-(l-1)ζ}, satisfies the left boundary identically but does not satisfy the right boundary pA_{N+1}=0 for an N-cell chain. A finite-chain eigenstate must be a superposition of the two evanescent solutions so that both terminations are matched, which generally shifts the energies away from ±u, ±w and can in principle remove a solution. The paper states that 'there always exist a pair of eigenstates localized at the chain terminations' and that the total winding number counts them, but it does not solve this two-boundary problem or prove that the finite-chain spectrum has exactly ν states in the gaps. The numerical spectra in Fig. 2(a) demonstrate the count for a one-parameter path and for N=31, and the experiments confirm two edge states in the deep-blue phase, but the general bulk-edge correspondence is asserted rather than derived.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes an isospectral reduction (IR) of the three-site SSH3 lattice into an energy-dependent two-site model, whose off-diagonal coupling is ρ = uw + ε v e^{-iq}. The authors argue that the winding number of ρ on the three energy bands counts the number of topological edge-state pairs of the original chain: ν = 0, 2, or 4 for the trivial, one-pair, and two-pair phases. The transition criterion is |uw/(vε)| < 1, evaluated at the expected edge-state energies ±u and ±w. The claims are supported by numerical spectra for a 31-cell chain, by an acoustic implementation that measures band structure and edge-state profiles, and by a disorder-robustness calculation. The central theoretical step is the derivation of edge states from complex wavevector solutions of ρ = 0 with evanescence condition ζ > 0, which is used to infer the bulk-edge correspondence.","tokens_in":9030,"tokens_out":7398,"duration_ms":84929,"significance":"If the claimed correspondence is established rigorously, the paper provides a practical and intuitive invariant for a lattice that lacks the conventional symmetries usually invoked for bulk-edge correspondence, and it would extend isospectral reduction methods to topological wave systems. The exact Schur-complement reduction in Eq. (3) is clean and parameter-free, the acoustic mapping is clearly explained, and the experimental measurements of the predicted edge states are a definite strength. The numerical and experimental evidence for the specific deep-blue phase (SC > SA, SB) is convincing. However, the general bulk-edge correspondence is currently asserted rather than proven; the main evidence is a single parameter path in Fig. 2(a) and three illustrative winding plots in Figs. 2(c)-(e). A rigorous finite-chain argument or a systematic parameter scan is needed before the central claim can be accepted as a general result.","major_comments":[{"comment":"The central bulk-edge correspondence is inferred from semi-infinite solutions of ρ = 0 with ζ > 0, but the finite N-cell boundary value problem is not solved. The ansatz in Eq. (5) satisfies the left termination by construction, yet it does not satisfy the right boundary condition pA_{N+1} = 0 in the C-site equation at l = N; a finite-chain eigenstate must be a superposition of the two evanescent solutions so that both terminations are matched. Such a superposition generally shifts the edge-state energies away from ±u and ±w and could, in principle, remove a state. Because the claim that the winding number ν counts edge-state pairs (ν = 0, 2, 4) rests on this step, the authors should either solve the finite-chain problem (e.g., by a transfer-matrix method) or provide a systematic numerical demonstration over the parameter space, rather than only the N = 31 path in Fig. 2(a).","section":"Eqs. (4)-(5) and the paragraph following Eq. (4)"},{"comment":"The winding number ν is presented as the bulk invariant of the IRM, but ρ(q) is evaluated along the three energy bands ε_n(q) of the original SSH3 lattice, so ν is a functional of the SSH3 band structure rather than an invariant of a fixed, energy-independent Hamiltonian. The paper should state this explicitly and give the definition of ν as a sum over bands, including how the second band can wind twice in Fig. 2(e). As written, the text asserts the correspondence after illustrating three parameter points; it does not prove that ν equals the number of in-gap states for general (u, v, w).","section":"Fig. 2(c)-(e) and the paragraph on winding numbers"}],"minor_comments":[{"comment":"For ε^+_ϕ = +u and ε^+_ψ = +w, Eq. (4) gives κ^+ = (2n+1)π, so e^{-iq} = -e^{-ζ}; the wavefunctions should therefore contain a factor (-1)^{l-1}. As written, Eq. (5) and Eq. (6) satisfy ρ = 0 only with κ = 0, which corresponds to the negative-energy solutions. Please correct the sign or clarify the convention for q.","section":"Eqs. (5)-(6)"},{"comment":"The phrase \"there distinct topological phases\" should read \"three distinct topological phases.\"","section":"Abstract"},{"comment":"The caption \"mode intercity\" should be \"mode intensity,\" and the text \"a perforation (radium 4 mm)\" should be \"a perforation (radius 4 mm).\"","section":"Fig. 1(e) and experimental setup"},{"comment":"The disorder study perturbs channel widths while preserving the deep-blue phase condition (SC remains larger than SA and SB); a sentence clarifying that robustness across the phase boundary is not tested would be useful.","section":"Fig. 4(a)"},{"comment":"The experiments demonstrate the deep-blue phase with two edge-state pairs; the one-pair and trivial phases are supported only by the numerical spectrum in Fig. 2(a). A brief statement acknowledging this limitation would improve the accuracy of the claims.","section":"Experimental verification"}],"recommendation":"major_revision","confidential_remarks":"For the editor: The isospectral reduction itself is correct and the experiments are convincing, but the paper's central claim — a general bulk-edge correspondence counted by the winding number — currently rests on a semi-infinite-chain argument and three illustrative parameter points. This is fixable within the manuscript by adding a finite-chain proof or a systematic numerical scan; I therefore recommend major revision rather than rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nWhat you should know about this paper: the isospectral reduction of SSH3 to a two-site model is exact, transparent, and genuinely useful, and the acoustic experiment is well executed. The headline claim—that the winding number of ρ counts edge-state pairs, with ν = 0, 2, or 4—is not actually proven. It is inferred from semi-infinite solutions of ρ = 0 and then illustrated on three parameter sets. That is a real gap, not a nitpick.\n\nThe genuinely new thing is the reduction itself. The Schur complement step, integrating out site B to get a two-site Hamiltonian with energy-dependent coupling ρ = uw + εv e^{-iq}, is clean and correct. The observation that the winding of ρ can change without a bandgap closing is also interesting and separates this from the usual SSH picture. The acoustic mapping from waveguide widths to u, w, v is standard and well handled; the measured edge-state profiles and the localization experiment in the two-pair phase agree with theory and simulation.\n\nThe soft spot is the bulk-edge correspondence. The paragraph after Eq. (4) says \"there always exist a pair of eigenstates localized at the chain terminations\" based on evanescent solutions of ρ = 0 in the infinite chain. But a finite chain of N cells has two boundaries, and a true eigenstate is generally a superposition of left- and right-decaying solutions satisfying both ends. The paper does not solve that two-boundary problem, nor does it prove that the finite-chain spectrum contains exactly ν edge states in the gaps. The numerical spectra for one parameter path (N = 31) and the experiment in one phase are suggestive, but they do not establish a general theorem. This is the reader's weakest-assumption point, and it lands on reading the text: the assertion is exactly where the proof is missing.\n\nTwo smaller issues. The IRM is nonlinear in ε, so the winding number of ρ is an energy-dependent object; it needs some argument for why it is a robust invariant of the original linear Hamiltonian under perturbations. And there are no error bars or raw data for the measured frequencies and amplitudes, which makes verification harder than it should be.\n\nThat said, the mathematics up to the IRM is exact, and the acoustic realization is careful. The paper deserves a serious referee. I would send it to review and ask for a proof or at least systematic numerical verification of the bulk-boundary count across multiple parameter values and chain lengths, plus a clear statement of the conditions under which the semi-infinite argument applies to a finite chain.\n\nFor a reading group: worth a look, mostly to discuss how one turns a suggestive winding-number picture into a real index theorem.","headline":"A clean isospectral reduction and a well-done acoustic experiment, but the central bulk-edge correspondence is asserted more than proven.","tokens_in":9544,"tokens_out":1889,"would_cite":true,"duration_ms":23888,"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":"An SSH3 trimer chain can be reduced to a two-site model whose winding number counts its topological edge-state pairs.","keywords":["isospectral reduction","SSH3 trimer lattice","bulk-edge correspondence","winding number","topological edge states","acoustic topological insulator","energy-dependent coupling"],"falsifier":"Measure the full spectrum of a finite N-cell SSH3 chain (or simulate it) with couplings in the predicted two-pair regime, e.g. $u=0.36$, $w=0.42$, $v=0.69$, and count midgap states: the claim fails if the number of pairs is not 2 or if their energies are not $\\pm u$ and $\\pm w$ within experimental resolution.","tokens_in":8622,"feed_emoji":"🔊","tokens_out":5255,"duration_ms":52086,"temperature":0.7,"pith_summary":"The paper proposes that every SSH3 trimer chain---three sites per unit cell with staggered couplings---can be mapped without changing its spectrum onto a two-site model by integrating out the middle site. In that reduced model all topological information is carried by a single energy-dependent coupling $\\rho = uw + \\varepsilon v e^{-iq}$. The paper argues that the winding number of $\\rho$ around the origin, summed over bands, equals twice the number of topologically protected edge-state pairs: $\\nu = 0,2,4$ for zero, one, and two pairs. A reader should care because the SSH3 chain lacks the conventional symmetries that usually make bulk-edge correspondence well-defined, and the paper supplies a concrete counting rule that is verified in an acoustic experiment.","feed_headline":"A two-site model counts SSH3 edge-state pairs: 0, 2, or 4","feed_subtitle":"An energy-dependent coupling in the reduced model yields the trimer chain's bulk-edge correspondence.","key_machinery":"The central object is the isospectrally reduced model (IRM), obtained by eliminating the B site from the $3\\times 3$ SSH3 Hamiltonian. The effective $2\\times 2$ operator acting on $(\\tilde{p}_A, \\tilde{p}_C)$ is $H = \\begin{pmatrix} u^2 & \\rho \\\\ \\rho^* & w^2 \\end{pmatrix}$ with $\\rho = uw + \\varepsilon v e^{-iq}$; it is nonlinear in $\\varepsilon$ because the eliminated site feeds back through its Green's function. Its role is to turn a three-band lattice problem into a two-site problem whose only nontrivial coupling $\\rho$ winds around the origin. The winding number of $\\rho$ supplies the topological invariant, and the condition $|uw/(v\\varepsilon)| < 1$ marks where $\\rho = 0$ solutions become evanescent, giving edge states.","core_discovery":"The central claim is that the SSH3 lattice admits an isospectral reduction to a two-site 'fringe' model with Hamiltonian $H = \\begin{pmatrix} u^2 & \\rho \\\\ \\rho^* & w^2 \\end{pmatrix}$, where $\\rho = uw + \\varepsilon v e^{-iq}$, and the original three-band spectrum is recovered exactly. When $\\rho = 0$ with complex wavevector $q = \\kappa - i\\zeta$ and evanescent condition $\\zeta > 0$, the reduced model has an on-site inversion symmetry that protects pairs of edge states at energies $\\varepsilon = \\pm u$ and $\\varepsilon = \\pm w$. The winding number of $\\rho$ as $q$ runs over the Brillouin zone counts these pairs: $\\nu = 2$ for one pair and $\\nu = 4$ for two, with $\\nu = 0$ for the trivial phase. This establishes a bulk-edge correspondence for the SSH3 chain without requiring band-gap closing, and measurements on a 31-cell acoustic chain confirm the predicted edge-state profiles and energies.","pith_inferences":["Editorial extension: the same integration-out procedure should apply to any chain with a site coupled only to its cell neighbors, so the IRM construction likely generalizes to tetramer or pentamer lattices whose reduced models will contain more energy-dependent couplings and possibly higher winding numbers.","Editorial extension: because the coupling $\\rho$ depends on $\\varepsilon$, a single physical chain could be switched between trivial and topological response by driving at different frequencies, offering frequency-selective edge-state routing.","Editorial extension: the evanescent-counting argument assumes the two terminations are independent; for short chains the finite-length boundary problem may mix the two ends and shift edge-state frequencies away from $\\pm u$ and $\\pm w$, which is a checkable finite-size prediction."],"forward_implications":["The SSH3 phase diagram is set by the inequalities $v > u$ and $v > w$, not by band-gap closures; topological transitions occur when $|uw/(v\\varepsilon)| = 1$.","Each topological phase carries a definite number of edge-state pairs at fixed energies $\\pm u$ and $\\pm w$, so the edge spectrum is predictable from bulk data.","The IRM's energy-dependent coupling means the effective hopping strength changes with eigenstate energy, so topological phase boundaries can be crossed by changing frequency in a single sample.","Measured acoustic spectra and amplitude profiles at 4.12 and 4.31 kHz match the reduced-model wavefunctions, and the states survive random disorder up to $\\delta = 1$ mm."],"supporting_citations":[{"why":"Defines isospectral reduction as a graph operation preserving eigenvalues and eigenvectors, the method the paper applies to the SSH3 Hamiltonian.","marker":"[35,36]"},{"why":"Established the existence of edge states in trimer lattices, the phenomenon the paper re-derives through the reduced model.","marker":"[28]"},{"why":"Earlier bulk-edge correspondence for the trimer SSH model based on sublattice Zak phase; the paper's IRM provides an alternative invariant.","marker":"[29]"},{"why":"Associates SSH3 edge states with a Chern number in a synthetic dimension, another benchmark for counting edge states.","marker":"[30]"},{"why":"Provide the acoustic waveguide mapping that lets the authors realize the SSH3 Hamiltonian with tube widths and measure its edge states.","marker":"[48,49]"}],"fun_headline_variants":["Isospectral reduction: SSH3 becomes a two-site model","Two-site reduction of SSH3 yields exact bulk-edge correspondence","Winding number counts SSH3 edge-state pairs: 0, 2, or 4","Energy-dependent couplings let a two-site model mimic SSH3 topology","Isospectral reduction maps trimer chain to a two-site model with windings"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The count of edge-state pairs rests on assuming that every complex-wavevector solution of $\\rho = 0$ with $\\zeta > 0$ in the infinite chain corresponds to one physical eigenstate localized at a termination, with the two ends independent, rather than on solving the finite-length chain's boundary value problem.","fun_headline_variants_meta":{"raw":{"variants":["Isospectral reduction: SSH3 becomes a two-site model","Two-site reduction of SSH3 yields exact bulk-edge correspondence","Winding number counts SSH3 edge-state pairs: 0, 2, or 4","Energy-dependent couplings let a two-site model mimic SSH3 topology","Isospectral reduction maps trimer chain to a two-site model with windings"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000447,"raw_usage":{"total_tokens":2241,"prompt_tokens":910,"completion_tokens":1331,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":526,"completion_tokens_details":{"reasoning_tokens":1235}},"tokens_in":526,"tokens_out":1331,"duration_ms":11043,"temperature":1.0,"reasoning_tokens":1235,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T17:43:17.416338+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the full spectrum of a finite N-cell SSH3 chain (or simulate it) with couplings in the predicted two-pair regime, e.g. $u=0.36$, $w=0.42$, $v=0.69$, and count midgap states: the claim fails if the number of pairs is not 2 or if their energies are not $\\pm u$ and $\\pm w$ within experimental resolution.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Established the existence of edge states in trimer lattices, the phenomenon the paper re-derives through the reduced model."},{"cited_title":"Anastasiadis, G","cited_arxiv_id":null,"evidence_quote":"Earlier bulk-edge correspondence for the trimer SSH model based on sublattice Zak phase; the paper's IRM provides an alternative invariant."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Associates SSH3 edge states with a Chern number in a synthetic dimension, another benchmark for counting edge states."}],"review_version":1}