{"id":"0a15066e-60b0-499a-83ef-3c569451b276","arxiv_id":"2507.10232","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"In a driven three-band PT-symmetric lattice, the Rabi frequency diverges near exceptional points, and total power oscillations can serve as an observable for self-orthogonality.","lead":"A driven three-band PT-symmetric lattice is shown to exhibit a Rabi frequency that diverges near exceptional points, with the slow oscillation of total power tracking that divergence. The paper proposes total power as a measurable signal for self-orthogonality in metamaterials.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The rotating-wave approximation used to derive Eq. (1) fails in the same ε→0 limit where Ω_R diverges, so the power-oscillation observable is not established in the divergent regime; the numerics only probe finite distances.","rationale":"The reader's weakest assumption identifies both the normalization independence and the RWA validity. I agree with the RWA part, but the normalization issue is less damaging: the L2 power formula (A16) is invariant under simultaneous rescaling of the right eigenvectors and the expansion coefficients, since the three terms each transform trivially, so the metric-tensor discussion is not needed for this particular observable. The load-bearing problem is that the derivation of Ω_R in Eq. (A13) and the power formula (A16) both assume the two-level truncation and the rotating-wave approximation. Near the EP, the counter-rotating amplitude diverges as ε^{-1}, so the approximation cannot be uniformly valid in the limit ε→0. The numerical agreement in Fig. 4(d) is real evidence but only at finite distance; the red crosses do not extend into the region where Ω_R becomes comparable to Ω_P. The central claim is therefore not disproven, but the divergent-regime prediction is an extrapolation from an approximation that fails there. A straightforward exact simulation at closer k would settle it, so the appropriate verdict remains conditional.","tokens_in":17564,"tokens_out":14238,"duration_ms":180277,"concrete_test":"Propagate the exact real-space Hamiltonian (2)/(B24), without any RWA, on a chain with N=240 (or with v/u tuned so that a momentum-grid point lies within δk<10^{-4} of k_EP) and extract the dominant power-oscillation frequency from the long-time evolution; compare with Eq. (1) at the same k. If the exact frequency deviates from Eq. (1) once Ω_R becomes comparable to Ω_P, or if the total power fails to follow the |α|^2 growth predicted by (A16), the proposed observable does not track the diverging Ω_R in the EP limit. An independent analytic check is to re-solve the two-level system (A9)-(A10) without dropping the counter-rotating term and compute the Floquet frequency as a function of ε.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central observable claim rests on the rotating-wave two-level reduction of Appendix A.1. Equations (A9)-(A10) are obtained from Eq. (A8) by discarding the counter-rotating term e^{-i(Δε+Ω_P)t}; the standard validity condition is that the off-diagonal matrix elements be small compared with Ω_P (or Δε). In the biorthogonal convention used in the paper, the target left eigenvector is scaled by ε^{-1}, so the relevant matrix element behaves as |<ψ_L^t|V|ψ_R^in>| ~ ε^{-1} while Ω_R ~ ε^{-1/2}. Hence the RWA condition fails for sufficiently small ε, exactly in the asymptotic regime where the divergence is predicted. The exact simulations in Fig. 4 are carried out at momenta where Ω_R is at most about 10 J_1(τ), still much smaller than Ω_P; the red crosses therefore test Eq. (1) only at finite distance and cannot separate the ε^{-1/2} divergence from a breakdown of the two-level truncation. Moreover, the RWA wavefunction (A14) carries a coefficient α ~ ε^{-1} multiplying a unit-norm target right eigenstate, so the L2 power in (A16) grows like ε^{-2}; once the neglected counter-rotating and extra-band terms are included, the stable PT-symmetric dynamics need not realize this growth. The claim that power oscillations provide a direct observable of self-orthogonality is thus unsupported precisely in the limit where Ω_R is supposed to diverge.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies a driven three-band PT-symmetric lattice model with a flat band and exceptional points (EPs), and claims that the Rabi frequency associated with resonant pumping diverges as the system approaches an EP due to the self-orthogonality of the coalescing eigenstates. It further proposes that the total power of the system, which oscillates in PT-symmetric systems, exhibits slow oscillations at this Rabi frequency and therefore provides a direct experimental observable for self-orthogonality. The manuscript provides an analytic derivation of the Rabi frequency from biorthogonal perturbation theory, a Floquet stability analysis of the driven model, and numerical simulations of a finite chain with periodic boundary conditions that are compared with the analytic formula.","tokens_in":17936,"tokens_out":6849,"duration_ms":76797,"significance":"If the central claim were fully established, the paper would offer an experimentally accessible signature of self-orthogonality at exceptional points in a tractable lattice model, which would be relevant for photonic and topolectrical metamaterial platforms. The work has clear strengths: the model is concrete, the algebraic derivations in Appendices A and B are internally consistent, the RWA-based formula in Eq. (1) is stated explicitly, and the numerical curves in Figs. 3 and 4 match the derived expressions at finite distances from the EP. The stability analysis via Floquet quasi-energies is a useful practical addition. However, the observable claim is weakened by unresolved normalization and rotating-wave issues, and the numerics do not probe the divergent regime; these concerns are load-bearing for the paper's main message rather than cosmetic.","major_comments":[{"comment":"The paper asserts in Appendix A.1 that physical observables must be computed with the metric tensor G defined in Eq. (A1) and that expectation values take the form of Eq. (A2) to be independent of the biorthogonal normalization. However, the power formula in Eq. (A16) is the plain L2 norm |ψ_k(t)|^2, not ⟨ψ|G|ψ⟩. Since the wavefunction in Eq. (A14) contains α ~ ε^{-1} while the right target state has unit norm, this L2 power grows like ε^{-2} as the EP is approached. The claimed normalization independence is therefore not actually used in the derivation of the proposed observable, and the power signal may be an artifact of the convention in which only left eigenstates are rescaled by ε^{-1}. This undermines the central claim that power oscillations provide a direct, convention-independent observable of self-orthogonality.","section":"Appendix A.1-A.2, Eqs. (A1)-(A2) and (A16)"},{"comment":"The rotating-wave approximation is applied by discarding the term e^{-i(Δε+Ω_P)t} in Eq. (A8), which is valid only when the off-diagonal matrix elements are small compared with Δε (or Ω_P). In the biorthonormal convention adopted in the paper, the matrix element ⟨ψ_L^t|V|ψ_R^in⟩ scales as ε^{-1} while the resulting Ω_R scales as ε^{-1/2}. Consequently, for sufficiently small ε the neglected counter-rotating term is no longer perturbative and the two-level truncation breaks down exactly in the regime where the divergence is predicted. The derivation of Eq. (1) and of the power formula in Eq. (5) therefore does not establish the observable divergence without an additional argument or an exact treatment that goes beyond the RWA.","section":"Appendix A, Eqs. (A8)-(A13)"},{"comment":"The red crosses in Fig. 4(d), obtained from the power-oscillation spectra of a finite N=60 chain, sample only momenta where Ω_R is at most about 10 J1(τ). With τ=10^{-3}, J1(τ)≈5×10^{-4} in units of the hopping, so the numerically accessed Ω_R remains much smaller than the pumping frequency Ω_P. These data lie in the RWA-valid regime and are consistent with the theoretical curve, but they do not probe the ε→0 limit; no error bars are shown, and no scaling analysis or extrapolation toward k_EP is presented. Thus the numerical evidence does not support the claimed divergence of the power-oscillation frequency at the exceptional point.","section":"Section IV, Fig. 4(d)"}],"minor_comments":[{"comment":"The Bessel-function arguments J0(τ/2) and J1(τ/2) appear in the Bloch Hamiltonian, but the relation of τ to the driving amplitude and the origin of the factor 1/2 are only explained later in Appendix B, Eq. (B25); a sentence in the main text defining τ would improve readability.","section":"Section III, Eq. (2)"},{"comment":"The schematic rows use the notation ⟨ψ_L^2|ψ_R^2⟩ = ε and ⟨ψ_L^2|ψ_R^2⟩ = 0 without clearly indicating which eigenstates are being plotted; adding explicit labels and consistent subscripts for the initial, target, and left states would make the figure less ambiguous.","section":"Figure 1"},{"comment":"The expression 'γϵ0 = 1√ϵ0 sqrt(...)' appears to contain a typographical ambiguity; it should read 1/√ϵ0, and the same notation should be used consistently in Eq. (B9).","section":"Appendix B, Eq. (B5)"},{"comment":"The notation |ψ(t)|^2 is used both for the L2 norm of the state and for the expectation value of a projector, which can confuse readers; using the norm symbol ∥|ψ(t)⟩∥^2 for the total power would be clearer.","section":"Section IV, Eq. (5)"}],"recommendation":"major_revision","confidential_remarks":"The paper is within the scope of the journal and the model is interesting, but the referee is concerned that the diverging Ω_R may be a consequence of the chosen left-vector normalization and that the proposed power observable is not computed with the metric tensor that the authors themselves argue is required. The authors should either prove that the L2 power is normalization-invariant in this setting, or compute the physically observable power from the metric-tensor prescription of Eq. (A2), and they should add an exact non-RWA Floquet analysis or an asymptotic scaling study near the EP to support the central claim."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear [Colleague],\n\nThe paper is worth a look if you work on driven PT-symmetric lattices. What's new: a three-band model with stable Floquet dynamics, a clear derivation that the biorthogonal Rabi frequency diverges as ε^{-1/2} near the EP, and a proposal to read that frequency off total-power oscillations. The numerics in Fig. 4(d) are clean and do show power oscillations tracking the computed Ω_R at finite distances from the EP. The stability analysis in App. C is careful, and the model is concrete enough to build on.\n\nThe soft spot is exactly where the stress-test lands, and it is load-bearing. The RWA used to derive Eq. (1) and Eq. (5) requires the off-diagonal coupling to be small compared to Ω_P. In the biorthogonal convention chosen, the target left eigenvector is scaled by ε^{-1}, so the matrix element ⟨ψ_L^t|V|ψ_R^in⟩ scales as ε^{-1}, while Ω_R only grows as ε^{-1/2}. The RWA condition therefore fails in the same limit where the divergence is predicted. The red crosses in Fig. 4 are all at distances where Ω_R is at most about ten times J_1(τ), still far below Ω_P, so the numerics do not test the divergent regime. The RWA wavefunction also carries α ~ ε^{-1}, making the L2 power in Eq. (A16) grow like ε^{-2}; that is an artifact of the truncation, not a property of the stable PT dynamics. The authors mention the metric tensor but never use it to define the power, and they don't discuss the RWA cutoff.\n\nThis is not a careless paper. The algebra in Apps. A and B is internally consistent, the finite-distance match is a legitimate result, and the robustness-to-detuning discussion in App. E is a nice touch. The problem is that the headline claim—a direct observable for self-orthogonality—is not supported exactly where it matters. The divergence is asymptotic, and the observable is defined in a regime where the model's own approximation breaks down.\n\nWho is it for: experimentalists and theorists working on EP physics in photonic or circuit lattices. They'll get a concrete model and a possible observable, but they should treat the divergence as a prediction to be tested at finite distance first. It deserves a serious referee: the core idea is interesting enough to engage with, but the authors need to either justify the RWA in the near-EP region, or soften the claim, or provide non-perturbative evidence that the power frequency still tracks Ω_R there.\n\nBest,\n[Your name]","headline":"A clean three-band PT lattice with a nice finite-distance match, but the RWA breakdown in the divergent regime undermines the power-oscillation claim at the EP.","tokens_in":18419,"tokens_out":7610,"would_cite":true,"duration_ms":83668,"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":"In a driven $\\mathcal{PT}$-symmetric three-band lattice, the Rabi frequency of resonant pumping diverges as the target band approaches an exceptional point, and this divergence shows up in the slow oscillations of the total power.","keywords":["exceptional points","parity-time symmetry","non-Hermitian systems","Rabi oscillations","self-orthogonality","driven lattice models","Floquet stability","total power oscillations"],"falsifier":"The claim would be falsified if, in a finite chain with $N=60$ unit cells and periodic boundary conditions at $\\tau = 10^{-3}$, the slow-oscillation frequency of the metric-tensor-weighted power from Eq. (A2) disagreed with the plain $|\\psi(t)|^2$ frequency from Eq. (A16) as $k$ approaches $k_{\\mathrm{EP}}^+$; that disagreement would mean the proposed observable depends on the normalization choice.","tokens_in":17407,"feed_emoji":"⚛️","tokens_out":8633,"duration_ms":90563,"temperature":0.7,"pith_summary":"The paper establishes that self-orthogonality at an exceptional point—the vanishing overlap between a right eigenstate and its left partner—can be read off from the dynamics of a driven parity-time-symmetric lattice. In a three-band model where two bands meet at an exceptional point and a third band serves as the pumping partner, the Rabi frequency of resonant population transfer diverges as momentum approaches the exceptional point. The paper further argues that the total power of the system, which is far easier to measure than the full state, oscillates with this same diverging Rabi frequency, making power oscillations a direct observable of self-orthogonality. If correct, this gives metamaterial platforms such as photonic lattices or topolectrical circuits a simple route to probing exceptional-point physics without resolving individual eigenstates.","feed_headline":"Rabi frequency diverges at exceptional points","feed_subtitle":"Measurable total-power oscillations in a driven PT-symmetric lattice carry the self-orthogonality signature.","key_machinery":"The central object is the biorthogonal Rabi frequency defined in Eq. (1), $$\\Omega_R(k) = \\sqrt{|\\langle \\psi^L_{j;k}|V(k)|\\psi^R_{i;k}\\rangle| \\, |\\langle \\psi^L_{i;k}|V(k)|\\psi^R_{j;k}\\rangle|}.$$ It carries the argument because near the exceptional point the target right eigenstate coalesces with its partner, so $\\langle \\psi^L | \\psi^R \\rangle = \\varepsilon \\to 0$; after rescaling only the left eigenstates by $\\varepsilon^{-1}$, each matrix element in the product scales as $\\varepsilon^{-1/2}$, giving $\\Omega_R \\propto \\varepsilon^{-1/2}$. The companion mechanism is the power formula of Eq. (5), in which the cross-term $\\sin(\\Omega_R t)\\,\\mathrm{Re}[\\alpha e^{i\\Omega_P t} \\langle \\psi^R_t | \\psi^R_{in}\\rangle]$ makes the slow envelope of the total power oscillate at exactly $\\Omega_R$.","core_discovery":"The paper claims that in a driven $\\mathcal{PT}$-symmetric three-band sawtooth lattice, resonant pumping between a gapped band and a band that forms an exceptional point with a flat band produces Rabi oscillations whose frequency $\\Omega_R$ diverges as $\\varepsilon^{-1/2}$, because the target eigenstate becomes self-orthogonal, meaning $\\langle \\psi^L | \\psi^R \\rangle = \\varepsilon \\to 0$. It also claims that the total power $|\\psi(t)|^2$ of the $\\mathcal{PT}$-symmetric system oscillates with the same slow Rabi frequency, so measuring power oscillations provides a direct experimental observable for self-orthogonality. For a finite chain with $N=60$ unit cells and periodic boundary conditions at driving parameter $\\tau = 10^{-3}$, the frequencies extracted from the power oscillations match the analytically expected $\\Omega_R(k)$ curve, supporting the claim.","pith_inferences":["A testable extension is to repeat the calculation with a metric-tensor-weighted power, as Appendix A1 sets up, and check whether the slow-oscillation frequency still matches $\\Omega_R(k)$; if not, the proposed observable is tied to the plain $\\mathrm{L}^2$ normalization rather than to normalization-independent physics.","If the mechanism is generic, any non-Hermitian platform with a two-level transition coupled to a coalescing eigenstate, not only the sawtooth lattice, should show an $\\varepsilon^{-1/2}$ divergence of the Rabi frequency, which could be tested in topolectrical circuits, microwave resonators, or photonic waveguide arrays.","The detuning robustness near the exceptional point implies a practical feature: population transfer becomes increasingly insensitive to frequency errors the closer the drive is to the exceptional point, a non-Hermitian effect without a Hermitian counterpart."],"forward_implications":["The slow envelope of the total power oscillates at $\\Omega_R$, so a time-resolved measurement of total power at fixed momentum directly maps out the divergence of $\\Omega_R$ near the exceptional point.","Because the system is translationally invariant, power oscillations within a single unit cell follow the same Rabi frequency as the total power, giving a local observable usable in finite experimental geometries.","Near the exceptional point, the divergence of $\\Omega_R$ makes the resonant pumping increasingly robust against detuning of the pump frequency.","The exceptional points of the Floquet Hamiltonian occur at the same momenta as those of the static Hamiltonian, so the diverging Rabi frequency persists in the stably driven regime."],"supporting_citations":[{"why":"Supplies the biorthogonal formalism and the definition of self-orthogonality for non-Hermitian Hamiltonians used throughout.","marker":"[1]"},{"why":"The prior prediction that the Rabi frequency diverges near an exceptional point in optical subwavelength systems, which this paper transfers to PT-symmetric lattice models.","marker":"[46]"},{"why":"Provides the metric-tensor formalism for defining probabilities in non-Hermitian systems that Appendix A1 introduces.","marker":"[47]"},{"why":"The biorthogonal normalization and metric-tensor prescription used to define observables when only left eigenstates are rescaled by $\\varepsilon^{-1}$.","marker":"[48]"},{"why":"The appendix containing the detailed derivations of the Rabi frequency, the power oscillations, the driven Bloch Hamiltonian, and the stability analysis.","marker":"[49]"},{"why":"The Jacobi-Anger expansion technique used to derive the effective driven Bloch Hamiltonian with Bessel-function couplings.","marker":"[50]"},{"why":"The static three-band PT-symmetric model with a flat band and analytically located exceptional points to which the driven model reduces at $\\tau = 0$.","marker":"[51]"},{"why":"The result that total power in PT-symmetric systems oscillates around an average value, used as the basis for proposing power as the observable.","marker":"[52]"}],"fun_headline_variants":["Diverging Rabi frequency reveals exceptional-point self-orthogonality","Power oscillations directly probe self-orthogonality","Rabi divergence in driven PT lattice signals self-orthogonality","Measurable power oscillations expose exceptional-point signature","Exceptional point makes Rabi frequency diverge"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The result stands on the assumption that the measured power signal is independent of how the left eigenstates are normalized, and that the two-level rotating-wave description remains valid at the point where the Rabi frequency and oscillation amplitude diverge.","fun_headline_variants_meta":{"raw":{"variants":["Diverging Rabi frequency reveals exceptional-point self-orthogonality","Power oscillations directly probe self-orthogonality","Rabi divergence in driven PT lattice signals self-orthogonality","Measurable power oscillations expose exceptional-point signature","Exceptional point makes Rabi frequency diverge"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001055,"raw_usage":{"total_tokens":4362,"prompt_tokens":810,"completion_tokens":3552,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":426,"completion_tokens_details":{"reasoning_tokens":3473}},"tokens_in":426,"tokens_out":3552,"duration_ms":28934,"temperature":1.0,"reasoning_tokens":3473,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T17:36:51.451581+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"The claim would be falsified if, in a finite chain with $N=60$ unit cells and periodic boundary conditions at $\\tau = 10^{-3}$, the slow-oscillation frequency of the metric-tensor-weighted power from Eq. (A2) disagreed with the plain $|\\psi(t)|^2$ frequency from Eq. (A16) as $k$ approaches $k_{\\mathrm{EP}}^+$; that disagreement would mean the proposed observable depends on the normalization choice.","supporting_citations":[{"cited_title":"Alfassi, O","cited_arxiv_id":null,"evidence_quote":"The prior prediction that the Rabi frequency diverges near an exceptional point in optical subwavelength systems, which this paper transfers to PT-symmetric lattice models."},{"cited_title":"[53, 54]","cited_arxiv_id":null,"evidence_quote":"The appendix containing the detailed derivations of the Rabi frequency, the power oscillations, the driven Bloch Hamiltonian, and the stability analysis."},{"cited_title":"Ramezani, Phys","cited_arxiv_id":null,"evidence_quote":"The static three-band PT-symmetric model with a flat band and analytically located exceptional points to which the driven model reduces at $\\tau = 0$."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"The result that total power in PT-symmetric systems oscillates around an average value, used as the basis for proposing power as the observable."}],"review_version":1}