{"id":"45dc94cf-b4c6-488c-b228-08937f9402f5","arxiv_id":"2607.05857","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"unknown","formal_verification":"none","parameter_count":7,"one_line_summary":"A Green-function framework derives Floquet polariton spectra in pumped quantum materials from their nonlinear optical susceptibilities, predicting flat bands, exceptional points, and parametric instability in graphene, hBN, and layered superconductors.","lead":"The paper develops a theoretical framework predicting how a strong laser pump reshapes polaritons—hybrid light-matter waves—into 'Floquet polaritons' in materials like graphene, boron nitride, and layered superconductors. A generalist might read it because it predicts experimentally accessible optical gain and instability effects in solid-state systems using only measurable nonlinear optical coefficients.","discovery_kind":"unclear","skeptic_critique":{"model":"glm-5.2","headline":"The uniform-pump assumption combined with the unresolved Fresnel reflection problem for gain media makes the Josephson plasmon reflectivity predictions (Fig. 5b) quantitatively unreliable, though the graphene and hBN cases are more robust.","rationale":"The reader's verdict of CONDITIONAL with HIGH confidence is appropriate. The uniform-pump assumption is indeed the most load-bearing simplification, and it is most consequential for the superconductor case where it compounds with the unresolved gain-medium Fresnel problem. The authors themselves flag both issues, which supports the CONDITIONAL rather than ACCEPT verdict. However, the framework's core construction (Eq. 9) is sound, the graphene and hBN examples are more robust (2D geometry avoids the Fresnel ambiguity, perturbative truncation is numerically verified), and the predictions are experimentally falsifiable. No adjustment to the verdict is needed. The paper's claim of being 'parameter-free' is slightly overstated — the hBN nonlinear coupling λ₁ comes from DFT (Ref. [122]) rather than direct experimental measurement, and the graphene χ^(3) formula (Eq. 11) assumes the kinetic regime (ω_p/ε_F ≈ 0.42 in Fig. 2, where interband corrections may be non-negligible) — but these are quantitative limitations rather than fundamental flaws in the framework.","tokens_in":39765,"tokens_out":4017,"duration_ms":248297,"concrete_test":"Solve the Josephson plasmon problem (Sec. V) with a finite pump penetration depth λ_p using the Fresnel-Floquet approach of Ref. [128], computing the probe reflectivity for a slab of thickness d >> λ_p, λ_probe. Compare the resulting reflectivity spectrum with Fig. 5(b). If the pump-induced peak positions shift by more than ~10% of ω_J or the peak lineshapes change qualitatively (e.g., from absorptive to dispersive), the uniform-pump predictions are not quantitatively reliable.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The reader correctly identifies the uniform-pump assumption as the weakest link, but the concern is sharper than stated. For the superconductor case (Sec. V), two problems compound: (1) the pump is assumed uniform with infinite penetration depth, and (2) the reflectivity in Fig. 5(b) is computed via the standard Fresnel formula R = |(1-√ε_eff)/(1+√ε_eff)|² for a semi-infinite medium. When χ_eff has negative spectral weight (which the paper shows occurs for the Floquet replica at -ω_J + 2ω_p), the medium becomes a gain medium, and the standard Fresnel formula is known to be ambiguous — one must specify the direction of energy flow and impose outgoing-wave boundary conditions consistently. The authors acknowledge this ('the reflection problem at the interface between vacuum and an infinitely deep gain medium described by Eq. (32) involves subtleties') but do not resolve it. This means the quantitative predictions for one of the three flagship examples are not yet on firm footing. For graphene and hBN, the concern is less severe: the 2D near-field reflection coefficient R_p = 1 - 1/ε_2D is well-defined even for gain media, and the perturbative truncation is checked against numerical Floquet diagonalization (Appendix C, cutoff N=5). The framework itself (Eq. 9) is a standard perturbative expansion of nonlinear response in a classical pump background and is internally consistent within its stated approximations.","agreement_with_reader":"agree"},"referee_report":{"model":"glm-5.2","summary":"This paper presents a theoretical framework for predicting Floquet polariton spectra in optically driven quantum materials. The central idea is to construct an effective linear susceptibility χ_eff from the material's nonlinear optical susceptibilities combined with a classical pump field, then solve Maxwell's equations with this χ_eff to obtain the Floquet polariton dispersion. The framework is applied to three systems: (1) graphene plasmons driven via χ^(3) nonlinearity, exhibiting flat bands, exceptional points, and parametric instability; (2) hBN phonon polaritons driven via phononic nonlinearity, in both monolayer and thin-flake geometries; and (3) Josephson plasmons in layered superconductors driven by THz fields, producing new reflectivity peaks. The approach is grounded in a Keldysh path-integral formalism (Appendix A) whose classical saddle-point approximation yields the main-text equations. The graphene and hBN results are cross-checked against numerical Floquet diagonalization (Appendix C, cutoff N=5).","tokens_in":40046,"tokens_out":1700,"duration_ms":337964,"significance":"The paper addresses a timely problem — Floquet engineering of polaritons in quantum materials — and provides a practical, parameter-light framework that connects measurable nonlinear optical coefficients to Floquet polariton spectra. The construction of χ_eff from 1PI diagrams (Eq. 9) and its application to three distinct material platforms is a valuable contribution. Specific strengths include: (i) the parametric oscillator cross-check for graphene (Eq. 20) providing an independent derivation of the non-Hermitian band structure; (ii) the identification of negative spectral weight in Floquet replicas as a generic feature across all three platforms; (iii) the resummed effective susceptibility for hBN (Eq. 25) that goes beyond leading-order perturbation theory; and (iv) concrete experimental protocols with realistic parameters for near-field and far-field detection. The framework is largely non-circular: χ_eff is built from independently measurable nonlinear coefficients and an external pump field.","major_comments":[{"comment":"Sec. V, Eq. (32) and Fig. 5(b): The reflectivity R = |(1−√ε_eff)/(1+√ε_eff)|² is computed for a semi-infinite gain medium using the standard Fresnel formula. When χ_eff has negative spectral weight (which the paper shows occurs for the Floquet replica at −ω_J + 2ω_p), the standard Fresnel formula is known to be ambiguous for gain media — the sign of the imaginary part of √ε_eff must be chosen consistently with outgoing-wave boundary conditions and the direction of energy flow. The authors acknowledge this ('the reflection problem at the interface between vacuum and an infinitely deep gain medium described by Eq. (32) involves subtleties') but do not resolve it. This means the quantitative reflectivity predictions in Fig. 5(b) — one of the three flagship examples — are not on firm footing. The authors should either (a) restrict the Josephson plasmon predictions to the optical conductivity","section":null},{"comment":"Sec. V: The uniform-pump assumption (infinite penetration depth) is stated but its quantitative impact is not bounded. The authors note that the pump penetration depth is assumed much larger than that of the probe, but for layered superconductors the c-axis penetration depth at THz frequencies can be comparable to or smaller than the probe skin depth, and pump depletion/self-consistent penetration-depth effects may be significant. Since this assumption is load-bearing for the analytical results in Eq. (32), the authors should at minimum provide an estimate of the parameter regime (pump frequency, field strength, temperature) where the uniform-pump approximation is self-consistent, or cite the Fresnel-Floquet approach (Ref. 128) with a more specific statement of what corrections are expected.","section":null},{"comment":"Sec. III.B, Eq. (17): The parametric instability threshold is derived as ξ_c ≃ 0.52 for γ/ω_p = 0.05, corresponding to E_c ≈ 290 kV/cm. The paper states that 'in continuous-wave experiments, this unstable growth is expected to be cut off by nonlinear saturation that is beyond the present weak-probe treatment.' However, the weak-probe linearization itself may break down once the unstable mode grows to amplitudes comparable to the probe. The authors should clarify the timescale on which the linear-response treatment remains valid after the pump is turned on (i.e., the exponential growth time 1/Im[ω] versus the probe pulse duration), to establish that the predicted Im[R_p] features in Fig. 2(c) are observable before saturation.","section":null}],"minor_comments":[{"comment":"Eq. (14): The notation '(ω ± 2ω_p)' is used to denote two separate terms, but this convention is introduced only after the equation. A brief note before Eq. (14) would improve readability.","section":null},{"comment":"Fig. 2(c): The color scale is stated to be logarithmic, which makes the plasmons 'appear broader than what they actually are.' Consider also providing a linear-scale inset or panel for direct comparison with experimental data.","section":null},{"comment":"Sec. IV.A, Eq. (24): The definition of χ^(2) in terms of the bare phonon propagator χ(ω) is stated, but the distinction between χ as used in Eq. (22) (the bare phonon susceptibility) and χ as used in Eq. (4) (the full linear susceptibility including polaritonic effects) could be clearer. A brief sentence distinguishing these two uses of χ would help.","section":null},{"comment":"Appendix C.1, Eq. (C7): The definition ε^F_2D = −M_eff/(ω(ω+iγ)) is introduced without explicit derivation of the prefactor. A one-line derivation connecting M_eff to the dielectric function would be helpful.","section":null},{"comment":"Fig. 5(a): The pump field values E_p = 10 kV/cm and 15 kV/cm are used, but the dimensionless pump strength φ_p = 2edA_p/(ℏc) corresponding to these values is not stated. Providing φ_p would help readers assess the perturbative regime.","section":null},{"comment":"Sec. IV.B: The out-of-plane dielectric ε_z = 3.0 is stated as 'a good approximation to hBN in this frequency range' without a reference. A citation or brief justification would strengthen this claim.","section":null},{"comment":"References: Several 2026 references (e.g., Refs. 50, 56, 80, 81, 86, 105, 132) appear to be preprints or very recent publications. The authors should verify that final published versions are cited where available.","section":null},{"comment":"Sec. II.B, Eq. (9): The '···' representing higher-order nested diagrams is mentioned but the truncation criterion is not quantified. A brief statement of the small parameter controlling the expansion (e.g., ξ² for graphene, E²_p λ²_1 for hBN) would help readers assess convergence.","section":null}],"recommendation":"major_revision","confidential_remarks":"The stress-test concern about the Fresnel formula for gain media in the Josephson plasmon case (Sec. V) is well-founded and is the primary reason for the major_revision recommendation. The graphene and hBN cases are more robust because the 2D near-field reflection coefficient R_p = 1 − 1/ε_2D is well-defined even for gain media, and the perturbative truncation is checked against numerical Floquet diagonalization. The Josephson plasmon case would benefit from either restricting to conductivity predictions or implementing the Fresnel-Floquet approach of Ref. 128. The paper is a solid contribution once this issue is addressed."},"author_rebuttal":{"model":"glm-5.2","summary":"We thank the referee for a careful and constructive report. The three major comments all identify legitimate limitations of our presentation. We address each below and describe revisions we will make.","responses":[{"response":"The referee is correct that the standard Fresnel formula for a semi-infinite gain medium is not on firm footing in the frequency ranges where ε_eff has negative imaginary part (i.e., where the Floquet replica at −ω_J + 2ω_p gives gain). We agree that the quantitative reflectivity values in Fig. 5(b) are not reliable in those specific frequency windows. We will make the following revisions: (1) We will explicitly state that the optical conductivity predictions in Fig. 5(a) are unaffected by this issue, since σ_eff is computed directly from ε_eff without invoking a reflection boundary condition. (2) We will add a shaded region or explicit annotation in Fig. 5(b) marking the frequency ranges where ε_eff corresponds to a gain medium, noting that the reflectivity there requires the Fresnel-Floquet approach (Ref. 128) for a quantitative treatment. (3) We will note that for the reflectivity peaks associated with the normal (non-gain) Floquet modes — specifically the fundamental mode and the replica at +ω_J + 2ω_p — the standard Fresnel formula remains valid, and those predictions are robust. (4) We will add a sentence clarifying that R > 1 in the gain windows is not a quantitative prediction of our current treatment. We note that a full resolution via the Fresnel-Floquet formalism, which self-consistently treats the spatial profile of both pump and probe fields, is beyond the scope of this paper but is a natural follow-up.","revision_made":"partial","referee_comment":"Sec. V, Eq. (32) and Fig. 5(b): The reflectivity R = |(1−√ε_eff)/(1+√ε_eff)|² for a semi-infinite gain medium is ambiguous when χ_eff has negative spectral weight. The sign of Im[√ε_eff] must be chosen consistently with outgoing-wave conditions. The authors acknowledge but do not resolve this."},{"response":"We agree that this assumption deserves quantitative justification. We will add an estimate to Sec. V. For the parameters used in Fig. 5 (ω_J = 0.5 THz, ε_∞ = 4.5), the Josephson penetration depth is λ_J = c/(ω_J √ε_∞) ≈ 28 μm. When the pump frequency ω_p is near or below ω_J, the pump field is evanescent within the superconductor with penetration depth ~λ_J, which is comparable to the probe skin depth — in this regime the uniform-pump approximation is not self-consistent. However, when ω_p > ω_J (as is the case for the results in Fig. 5, where ω_p = ω_J and the pump is at the plasma edge), the pump penetrates more deeply. More generally, for pump frequencies well above ω_J, the pump propagates into the bulk with a penetration depth much larger than λ_J, while the probe (at frequencies near ω_J) still has skin depth ~λ_J; in this regime the uniform-pump approximation is well justified. We will state this explicitly and note that the results in Fig. 5, where ω_p = ω_J, represent a borderline case where corrections from finite pump penetration depth are expected to be moderate but not negligible. We will also cite Ref. 128 (Fresnel-Floquet) and Ref. 127 more specifically, noting that a self-consistent treatment would modify the effective pump strength as a function of depth and could shift the Floquet replica positions slightly.","revision_made":"partial","referee_comment":"Sec. V: The uniform-pump assumption (infinite penetration depth) is stated but its quantitative impact is not bounded. For layered superconductors the c-axis penetration depth at THz frequencies can be comparable to or smaller than the probe skin depth."},{"response":"This is a valid concern. We will add a quantitative discussion of the validity timescale. For the parameters in Fig. 2 (ω_p = 30 THz, γ = 1.5 THz, E_p = 600 kV/cm, corresponding to ξ ≈ 1.04), the maximum parametric growth rate at zero detuning (Q = 0) is Im[ω] ≈ ω_p κ − γ/2, where κ = 3ξ²/32 ≈ 0.10. This gives Im[ω] ≈ 3.0 − 0.75 ≈ 2.3 THz, corresponding to an exponential growth time τ_growth ≈ 1/(2π × 2.3 THz) ≈ 70 fs. Typical near-field probe pulses in the mid-infrared have durations of 100–500 fs, so the unstable modes can grow by a factor of e^1 to e^7 during the probe pulse, meaning the weak-probe linearization may indeed break down for the unstable modes at this pump strength. However, we note the following: (1) The frequency-domain linear response function Im[R_p] in Fig. 2(c) is mathematically well-defined regardless of instability — it is the retarded Green function of the linearized system, and its poles in the upper half-plane signal the instability. (2) For pump fields closer to threshold (ξ → ξ_c), the growth rate vanishes and the observable window extends to arbitrarily long times. (3) For few-cycle probe pulses (~50 fs), the linear-response features remain observable even at E_p = 600 kV/cm before saturation sets in. (4) The non-gain features of Im[R_p] — including the negative spectral weight of the n=1 replica away from the crossing point — are not affected by the instability timescale, since they do not involve exponentially growing modes. We will add this discussion to Sec. III.B and note that experimental observation of the flat-band/exceptional-point features is most favorable with pump strengths near threshold and short probe pulses.","revision_made":"partial","referee_comment":"Sec. III.B, Eq. (17): The weak-probe linearization may break down once the unstable mode grows to amplitudes comparable to the probe. The authors should clarify the timescale on which the linear-response treatment remains valid."}],"tokens_in":39759,"tokens_out":4044,"duration_ms":196026,"standing_objections":[]},"desk_editor":{"model":"glm-5.2","letter":"The main thing to know: this paper gives a clean, usable framework for computing Floquet polariton spectra from experimentally measurable nonlinear susceptibilities (χ^(2), χ^(3)) and the pump field, without ad hoc driving Hamiltonians. That is a real contribution. The central construction (Eq. 9) is a perturbative expansion of the pump-dressed linear susceptibility, built from 1PI diagrams in the Keldysh formalism. It is internally consistent and the classical saddle-point version in the main text is the right level for the experimental regimes discussed. The three applications — graphene plasmons, hBN phonon polaritons, Josephson plasmons — are worked out with concrete parameters and produce falsifiable predictions (near-field reflection coefficients, THz reflectivity peaks). The graphene calculation is cross-checked against numerical Floquet diagonalization (Appendix C, cutoff N=5), and the parametric oscillator picture (Eq. 20) gives an independent analytical handle on the flat-band and exceptional-point physics. The hBN monolayer and flake calculations use first-principles nonlinear coupling parameters from Ref. 122. The non-perturbative Bessel-function result for the Josephson case (Eq. 32) is a nice bonus. The negative-spectral-weight argument for why Floquet replicas act as gain media is physically clear and applies uniformly across all three platforms. The stress-test concern about the Josephson plasmon reflectivity (Fig. 5b) is real and correctly identified. When χ_eff has negative spectral weight, the standard Fresnel formula R = |(1−√ε_eff)/(1+√ε_eff)|² for a semi-infinite medium is ambiguous — one needs outgoing-wave boundary conditions specified consistently. The authors acknowledge this but do not resolve it. Combined with the uniform-pump assumption (which they also flag as problematic for superconductors where penetration depths differ), the quantitative predictions for the Josephson case are not yet on firm footing. This is a load-bearing issue for one of three flagship examples, not for the framework as a whole. For graphene and hBN, the concern is much less severe: the 2D near-field reflection coefficient R_p = 1 − 1/ε_2D is well-defined even for gain media, and the perturbative truncation is checked against numerical Floquet diagonalization. The linearized treatment of parametric instability (no saturation regime) is a standard limitation — acceptable for a theory paper, worth noting but not a reason to reject. No code or data shipped, but the analytical formulas are reproducible from the paper. The reader's verdict of CONDITIONAL is slightly too cautious. The framework itself is sound and the graphene/hBN predictions are quantitative. The Josephson reflectivity issue should be fixed or the claim softened, but the paper's core contribution does not depend on that one figure being quantitatively exact. This paper is for theorists and experimentalists in ultrafast nanophotonics and polaritonics who want concrete predictions for pump-probe experiments. It deserves a serious referee who can check the diagrammatic bookkeeping and assess whether the Josephson reflectivity problem can be patched in revision.","headline":"A practical framework for predicting Floquet polariton spectra from measurable nonlinear optical coefficients, applied to three material platforms — mostly solid, with one genuinely unresolved problem in the superconductor case.","tokens_in":40597,"tokens_out":731,"would_cite":true,"duration_ms":142954,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"glm-5.2","headline":"Pump-driven polaritons predicted from nonlinear optics alone","keywords":[],"falsifier":"If an experiment measured the Floquet polariton spectrum under a strong pump and found spectral features that cannot be reproduced by the effective susceptibility constructed from known nonlinear optical coefficients and the pump field, the framework would fail. Most directly, if the pump-induced flat bands and exceptional points predicted near band crossings were absent, or if their locations in frequency-momentum space differed systematically from the parametric-instability threshold condition, the core mechanism would be falsified.","tokens_in":40007,"feed_emoji":"🔦","tokens_out":875,"duration_ms":78526,"temperature":0.7,"pith_summary":"This paper presents a method to predict how a strong pump laser reshapes polariton spectra in quantum materials, without introducing any fitting parameters. The core idea is that the pump field, combined with the material's known nonlinear optical susceptibilities, generates an effective linear susceptibility. Solving Maxwell's equations with this effective susceptibility yields the Floquet polariton bands. The authors apply this to three systems: graphene plasmons driven by infrared light, hexagonal boron nitride phonon polaritons driven by mid-infrared light, and Josephson plasmons in layered superconductors driven by THz fields. In graphene and hBN, the pump creates Floquet replica bands that cross the original polariton branches. At these crossings, the hybridization involves one mode with positive spectral weight and one with negative spectral weight, producing non-Hermitian coupling. This leads to flat bands bounded by exceptional points and modes with negative damping, signaling parametric instability where two pump photons convert into two polaritons. In layered superconductors, the pump produces new reflectivity peaks at frequencies shifted by twice the pump frequency from the Josephson plasma edge, with a non-perturbative Bessel-function expression for the effective dielectric.","feed_headline":"Pump-driven polaritons predicted from nonlinear optics alone","feed_subtitle":"A parameter-free framework shows how laser-pumped materials gain flat bands, exceptional points, and parametric instability in their polarit","key_machinery":"The effective linear susceptibility chi_eff, constructed by combining the pump field with the material's nonlinear optical susceptibilities via a diagrammatic self-energy expansion. The dressed polariton propagator G(omega,k) is obtained by replacing the equilibrium susceptibility chi with chi_eff in the bare propagator G_0. The poles of G give the Floquet polariton branches.","core_discovery":"The paper's central technical result is that the pump-contributed self-energy for the polariton propagator can be built entirely from the material's nonlinear optical susceptibilities and the pump field, yielding an effective linear susceptibility whose poles give the Floquet polariton spectrum. A key physical consequence is that Floquet replica bands carry negative spectral weight, because they originate from shifting negative-frequency modes to positive frequencies. When such a replica crosses an original positive-weight branch, their opposite spectral signs produce attractive non-Hermitian hybridization, creating flat bands with exceptional points and parametrically unstable modes that do","pith_inferences":[],"forward_implications":["Pump-probe near-field experiments on graphene with mid-IR pump fields around 290 kV/cm should reveal flat plasmon bands with gain signatures in the near-field reflection coefficient, testable with current table-top lasers.","Far-field THz pump-probe spectroscopy on layered superconductors should show new reflectivity peaks at frequencies shifted by twice the pump frequency from the Josephson plasma edge, providing a direct fingerprint of Floquet Josephson plasmons.","The framework extends to any polaritonic system with known nonlinear optical coefficients, including exciton polaritons in semiconductor microcavities, where similar parametric instability and flat bands are expected.","Circularly polarized pump on graphene generates a chirality-dependent Hall optical response, suggesting topological Floquet plasmon physics at sample boundaries that could be probed near-field."],"fun_headline_variants":["Floquet polaritons derived purely from nonlinear optical response","Nonlinear optics alone predicts laser-driven Floquet polariton bands","Floquet polariton spectra built from nonlinear susceptibilities and pump field","Pump-driven polaritons inherit flat bands and exceptional points from nonlinear optics","Negative-weight Floquet replicas create unstable modes at band crossings"],"cache_read_input_tokens":0,"weakest_assumption_plain":"The pump field is treated as a fixed classical background that does not receive self-consistent feedback from the material response, which neglects pump depletion and spatial variation of the pump inside the sample. This is most questionable in superconductors where pump and probe penetration depths can differ significantly, and the authors flag this as needing further study.","fun_headline_variants_meta":{"raw":{"variants":["Floquet polaritons derived purely from nonlinear optical response","Nonlinear optics alone predicts laser-driven Floquet polariton bands","Floquet polariton spectra built from nonlinear susceptibilities and pump field","Pump-driven polaritons inherit flat bands and exceptional points from nonlinear optics","Negative-weight Floquet replicas create unstable modes at band crossings"]},"model":"glm-5.2","effort":"high","cost_usd":0.0,"raw_usage":{"total_tokens":713,"prompt_tokens":642,"completion_tokens":71,"prompt_tokens_details":null},"tokens_in":642,"tokens_out":71,"duration_ms":18063,"temperature":1.0,"reasoning_tokens":null,"cache_read_input_tokens":0,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-08T22:15:30.289199+00:00","model_set":{"reader":"glm-5.2"},"falsifier":"If an experiment measured the Floquet polariton spectrum under a strong pump and found spectral features that cannot be reproduced by the effective susceptibility constructed from known nonlinear optical coefficients and the pump field, the framework would fail. Most directly, if the pump-induced flat bands and exceptional points predicted near band crossings were absent, or if their locations in frequency-momentum space differed systematically from the parametric-instability threshold condition, the core mechanism would be falsified.","supporting_citations":[],"review_version":1}