{"id":"ca609239-bdeb-4557-a780-00722bb8262b","arxiv_id":"2412.03139","paper_version":3,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"Biased bilayer and trilayer graphene are predicted to support exciton-derived surface polaritons with voltage-tunable confinement and a Lorentzian-based universal dispersion law.","lead":"A theory paper predicts that biased bilayer and trilayer graphene can host highly confined, electrically tunable infrared polaritons that come from interband excitons rather than free electrons. These graphene-exciton-polaritons could become a new, voltage-controlled platform for nanophotonics.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The load-bearing step is the ad hoc nonlocal extension of the Lorentzian exciton model: f_m(q), Γ_m(q), and the continuum are not computed, so the after-nonlocal confinement claims overreach.","rationale":"The reader identified the imported local conductivity model as the weakest assumption. I agree that the local model is load-bearing, but the sharper issue is the nonlocal step: even if Eq. 1 with parameters from refs [39,40] were exact at q=0, the paper uses that local model to predict modes at large q, then corrects them with a single quadratic energy shift. This correction is a standard effective-mass approximation, but the paper does not derive f_m(q), Γ_m(q), or the continuum contribution from a microscopic finite-q calculation. The stated validity condition ℏ²q²/(2m_X) ≪ E_m is not checked against the reported GEP solutions. The local-confinement numbers (up to 10³) are therefore illustrative, not quantitative, and the after-nonlocal numbers (still above graphene plasmons) inherit this uncertainty. The universal dispersion law Eq. 7 is only derived for a single-resonance, local Lorentzian; in the actual multi-resonance, continuum-bearing systems at finite q, it is an approximation, not a proven universal law. I gave credit where appropriate: the polariton dispersion derivation from Eq. 2 is standard, the TMM comparison for the local model is a useful internal consistency check, and the parameter-free part of the polariton argument is sound. The concern is not that the authors are wrong, but that the central quantitative claim rests on an unverified extrapolation that can be tested by a finite-q Bethe-Salpeter calculation. Since the reader already issued a conditional verdict, my stress test does not move the verdict; it sharpens the condition that should be imposed.","tokens_in":12166,"tokens_out":19512,"duration_ms":201117,"concrete_test":"Compute the finite-momentum exciton optical conductivity σ(q,ω) for BBLG and BRTG at V = 52/55 meV using the same Bethe-Salpeter/tight-binding approach as refs [39,40], for q ranging from 0 up to the local GEP wavevectors (~0.1–1 nm⁻¹). Extract E_m(q), f_m(q), Γ_m(q) and the continuum spectral weight, and insert these into the nonlocal dispersion relation (Eq. 10 with q-dependent parameters) or a TMM using σ(q,ω). If the recalculated confinement factors and inverse damping ratios in Figs. 2–3 change by less than ~20%, the central claim holds. If the modes shift in frequency, disappear, or confinement drops by more than a factor of ~2, the ad hoc ℏ²q²/(2m_X) model is insufficient and the quantitative GEP predictions need revision.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central quantitative claims—GEP confinement exceeding graphene plasmons and the claimed universal dispersion law—rest on a nonlocal continuation of a zero-momentum exciton conductivity. Eq. 8 is obtained by replacing E_m with E_m + ℏ²q²/(2m_X) in Eq. 3 while keeping f_m and Γ_m fixed. The paper states this replacement holds for ℏ²q²/(2m_X) ≪ E_m, but it never verifies that the solutions of Eq. 11 satisfy this condition, and the momenta corresponding to the reported confinement factors (10²–10³) place q in a regime where the exciton form factor and q-dependent oscillator strength matter. The 'before nonlocal' confinement of three orders of magnitude in Fig. 2 is explicitly outside the stated validity of the nonlocal correction and should not be used as evidence. Moreover, Eq. 1 contains only discrete Lorentzian resonances; the interband continuum is omitted. For mode energies above the band gap, the continuum can alter the sign and magnitude of Im σ, and therefore the existence of the GEP branch. Because the imported parameters come from zero-momentum absorption calculations, with partial same-group authorship, the finite-q extrapolation is the least secure step of the argument. If f_m(q) decays with q or the continuum contributes near the mode frequencies, the positive-Im σ window narrows and the after-nonlocal confinement figures in Fig. 3 are overestimates.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper predicts a new class of surface polaritons, termed graphene-exciton-polaritons (GEPs), in biased bilayer and trilayer graphene (BBLG and BRTG). The starting point is an excitonic optical conductivity of the form sigma = 4i sigma0 sum_n f_n/(E - E_n + i Gamma_n/2), with parameters imported from prior calculations. Because Im sigma is positive near the main exciton resonance, the authors identify surface-polariton modes with hBN and derive an analytic dispersion (Eq. 2). Near the main resonance, a single-Lorentzian approximation yields approximate confinement, damping, and a square-root dispersion (Eqs. 5-7), which the paper presents as a universal law for 2D excitonic systems. A nonlocal correction, implemented as a momentum-dependent exciton energy, leads to Eq. 11 and to the claim that after nonlocal corrections the modes are less confined but still more confined than graphene plasmons, with propagation lengths accessible to cryo-SNOM.","tokens_in":12479,"tokens_out":7445,"duration_ms":69608,"significance":"If the prediction is correct, GEPs would provide a new electrically tunable platform for far-infrared polaritonic phenomena, and the single-resonance dispersion would be a useful organizing result for excitonic surface polaritons. The algebraic chain from Eq. 1 to Eqs. 4-7 is internally consistent, and the analytic dispersion matches the TMM simulations carried out with the same conductivity model; this agreement, however, validates the electrodynamics but not the underlying exciton model. The manuscript also gives explicit limitations (e.g., constant BRTG linewidth, local-model validity), which is commendable. The main open risk is that the quantitative confinement comparison with graphene plasmons rests on an unvalidated finite-momentum extension of the zero-momentum exciton conductivity and on the omission of the electron-hole continuum.","major_comments":[{"comment":"The nonlocal extension is introduced by replacing E_m with E_m + ℏ²q²/(2m_X), while f_m and Γ_m remain q-independent, and the text states that this holds for ℏ²q²/(2m_X) ≪ E_m. The manuscript never verifies that the solutions of Eq. (11) satisfy this condition, nor does it report q/q_c (q_c = sqrt(2m_X E_m)/ℏ) for the branches shown in Figs. 3(b)-(e). Since the confinement factors quoted there imply large q, the claim that GEPs remain more confined than graphene plasmons after nonlocal corrections is not yet established. Please report the validity parameter along each dispersion branch and, where it is not small, compute the momentum dependence of f_m and Γ_m or state the resulting uncertainty.","section":"§IV, Eq. (8)"},{"comment":"The conductivity model contains only discrete Lorentzian exciton resonances; the interband electron-hole continuum is not included. Positive Im σ, which is the condition for GEP existence, can be modified by continuum absorption when the mode energy is at or above the band gap. The paper does not state where the GEP branches sit relative to the gap or quantify the continuum contribution. Please add this check or provide a bound showing that continuum contributions are negligible in the frequency windows of Figs. 2 and 3.","section":"Eq. (1) and §III"},{"comment":"For BRTG, the linewidth is taken constant because phonon-exciton parameters are unavailable, as admitted in SI section C. This ad hoc assumption feeds directly into the inverse damping ratio and the 'moderate losses' claim for BRTG. Please include a sensitivity analysis over the plausible range of Γ_m(T), or otherwise show that the main conclusions are robust to this choice.","section":"§II / SI C"},{"comment":"Equation (7) is derived from the single-resonance Lorentzian approximation and is therefore a consequence of the assumed conductivity lineshape, not a relation established independently of the exciton model. The phrase 'universal dispersion law for all surface polaritons in 2D excitonic systems' overstates this status. I recommend either deriving the law from a more general microscopic starting point or explicitly defining the class of systems (Lorentzian response, negligible nonlocality, single dominant resonance) for which it holds.","section":"§III, Eq. (7)"}],"minor_comments":[{"comment":"There are several typographical errors, including 'hybridizatoin', 'denisty', 'affect', and 'Bern el'; a careful proofread is needed.","section":"Throughout"},{"comment":"The sentence defining the effective permittivity repeats εhBNxx twice and should read εhBNxx, εhBNzz, and εhBNef f.","section":"Eq. (2) text"},{"comment":"The notation 'n = m' for the dominant resonance is not defined in the main text; please clarify the Rydberg-series indexing.","section":"§III, Fig. 2"},{"comment":"The captions and text quote confinement factors in orders of magnitude, but no mode wavevectors or excitation energies are listed; reporting representative q and E values would improve reproducibility and make the validity check in the major comments possible.","section":"Figs. 2 and 3"},{"comment":"The BBLG exciton mass is given analytically, while BRTG is said to be computed numerically; a plot of m_X(V) for both systems, with a table of the used band parameters, would help readers reproduce the nonlocal results.","section":"Eq. (9)"}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a transparent model calculation that applies the standard 2D surface-polariton formalism to the excitonic conductivity of biased bilayer and trilayer graphene, and predicts large-confinement, electrically tunable 'graphene-exciton-polaritons'. The core physics is plausible, the algebra from Eq. 1 to the dispersion relations is clean, and the transfer-matrix comparison is reassuring. But the paper oversells the 'universal dispersion law' (Eq. 7), and the after-nonlocal confinement numbers rest on an unverified finite-q extrapolation.\n\nThe genuinely new content is the specific application: positive Im sigma in BBLG and BRTG from interband excitons, and the resulting surface modes that can be tuned with bias. That prediction is not in the cited literature, and the authors work out dispersion, confinement, and loss in one place with a useful TMM cross-check. The paper is honest about some limitations—the constant BRTG linewidth is flagged explicitly.\n\nThe soft spots are real but not fatal. Eq. 7 is a rearrangement of the standard surface-polariton dispersion for a single-resonance Lorentzian conductivity; presenting it as a new universal law overstates the case. More importantly, the nonlocal correction in Eq. 8 keeps f_m and Gamma_m fixed while shifting the exciton energy. The stated validity condition, hbar^2 q^2/(2m_X) << E_m, is never checked against the solutions of Eq. 11. At the q values corresponding to the reported two-to-three order-of-magnitude confinement, that inequality looks borderline. If f_m(q) decays with q or the continuum contributes near the mode frequencies, the after-nonlocal confinement figures in Fig. 3 are overestimates. Relatedly, the interband continuum is omitted from Eq. 1, which matters once the mode energy approaches the gap.\n\nNone of this sinks the qualitative message. As a theoretical proposal, it deserves peer review and likely publication after revision, with a toned-down universality claim and a quantitative check on where the nonlocal treatment is valid. The paper is useful for anyone working on infrared nanophotonics in 2D materials: it identifies a concrete platform and gives numbers to aim for in cryo-SNOM experiments. I would send it to referees and specifically ask them to test the nonlocal-validity requirement.","headline":"Solid model calculation, but the 'universal law' is a standard dispersion in disguise and the after-nonlocal confinement numbers rest on an unverified q-extension.","tokens_in":13044,"tokens_out":5801,"would_cite":true,"duration_ms":52996,"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":"This paper predicts that biased bilayer and trilayer graphene support graphene-exciton-polaritons—electrically tunable far-infrared surface polaritons whose confinement can exceed that of graphene plasmons.","keywords":["graphene-exciton-polaritons","biased bilayer graphene","rhombohedral trilayer graphene","interband excitons","surface polaritons","nonlocal corrections","far-infrared optics","electrically tunable polaritons"],"falsifier":"A cryo-SNOM measurement on hBN-encapsulated biased bilayer graphene at T = 4 K and a bias near 52 meV should show interference fringes whose spacing follows $q_{\\mathrm{GEP}}=2i\\varepsilon_0\\varepsilon_{\\mathrm{hBN}}^{\\mathrm{eff}}\\omega/\\sigma$; if no such propagating modes appear, or if the imaginary part of the extracted optical conductivity is not positive at the main exciton resonance, the central claim is wrong.","tokens_in":11984,"feed_emoji":"🔬","tokens_out":10887,"duration_ms":91021,"temperature":0.7,"pith_summary":"This paper predicts a new class of surface polaritons—hybrid light–matter waves that travel along an interface—supported by biased bilayer and trilayer graphene, and names them graphene-exciton-polaritons (GEPs). Unlike graphene plasmons, which come from intraband motion of free carriers, these modes come from interband excitons: bound electron–hole pairs whose collective response makes the imaginary part of the optical conductivity positive. The paper shows that GEPs follow a universal dispersion law for surface polaritons in 2D excitonic systems, and that their frequency, confinement, and loss can be tuned electrically by the bias voltage across the graphene layers. If the prediction holds, these systems would provide electrically controllable far-infrared modes with confinement up to three orders of magnitude, and—after nonlocal corrections—still more confined than graphene plasmons, with losses moderate enough for cryo-SNOM detection.","feed_headline":"Biased graphene hosts electrically tunable exciton polaritons","feed_subtitle":"Interband excitons make biased graphene a tunable source of deeply subwavelength far-infrared light.","key_machinery":"The central object is the excitonic optical conductivity $\\sigma = 4i\\sigma_0 \\sum_n f_n/(E - E_n + i\\Gamma_n/2)$, a multi-Lorentzian model taken from density-matrix calculations of biased bilayer and trilayer graphene. The positive imaginary part of $\\sigma$ near the main resonance, where the contribution of other resonances is negligible, is what permits a surface polariton; inserting this conductivity into the standard 2D surface-polariton dispersion relation $q = 2i\\varepsilon_0\\varepsilon_{\\mathrm{hBN}}^{\\mathrm{eff}}\\omega/\\sigma$ gives the GEP dispersion. The nonlocal extension replaces the exciton energy $E_m$ with $E_m + \\hbar^2 q^2/(2m_X)$, with $m_X$ given analytically for BBLG and numerically for BRTG, and leads to the quadratic equation whose physical root is the nonlocal dispersion. This machinery converts known excitonic parameters into a concrete, electrically tunable polaritonic dispersion.","core_discovery":"The central claim is that biased bilayer graphene (BBLG) and biased rhombohedral trilayer graphene (BRTG), when encapsulated in hBN under a perpendicular electric field, support surface polaritons whose positive imaginary conductivity is provided by the interband excitonic response rather than by free-carrier intraband response. The local dispersion is $q_{\\mathrm{GEP}}=2i\\varepsilon_0\\varepsilon_{\\mathrm{hBN}}^{\\mathrm{eff}}\\omega/\\sigma$ with $\\sigma$ the multi-Lorentzian excitonic conductivity, and near the main resonance this yields confinement factors of two to three orders of magnitude. In the lossless single-resonance limit the dispersion reduces to $\\omega=\\omega_m/2+\\sqrt{(2\\sigma_0 f_m/\\varepsilon_0\\varepsilon_{\\mathrm{hBN}}^{\\mathrm{eff}}\\hbar)\\,q+(\\omega_m/2)^2}$, which the paper proposes as a universal law for surface polaritons in 2D excitonic systems. Including nonlocal corrections via the exciton kinetic term $\\hbar^2 q^2/(2m_X)$ lowers the confinement by up to an order of magnitude but leaves the GEPs more confined than graphene plasmons and improves the inverse damping ratio by a factor of four, making observation in cryo-SNOM experiments plausible.","pith_inferences":["A testable extension: applying the same positive-imaginary-conductivity criterion to other biased few-layer graphene stackings, such as ABC-stacked trilayer or twisted bilayer graphene, could reveal GEPs at different frequencies and with different electrical tunability.","The nonlocal crossover at which $\\hbar^2 q^2/(2m_X)\\sim E_m$ sets an upper bound on GEP confinement; measuring the dispersion at large momentum would provide a direct probe of the exciton effective mass $m_X$.","Since the BRTG linewidth is held constant for lack of phonon-exciton parameters, the temperature-dependent loss predictions for trilayer graphene are less certain than for bilayer; a full phonon-scattering calculation would sharpen them and might shift the optimal operating temperature.","Because the lossless GEP dispersion has a square-root form, coupling GEPs to a cavity photon could produce an anticrossing similar to exciton-polariton physics, even though the uncoupled mode is a surface polariton."],"forward_implications":["Biased bilayer and trilayer graphene become electrically tunable far-infrared polaritonic platforms: the bias voltage changes the exciton energy, oscillator strength, and decay rate, so the GEP frequency and confinement can be set in situ.","The universal dispersion law derived for GEPs should apply to any 2D excitonic system that supports surface polaritons, giving a common design rule beyond the specific graphene systems considered here.","After nonlocal corrections, GEPs remain more confined than graphene plasmons while their inverse damping ratio improves by a factor of four, so the paper's predicted modes should be observable with cryo-SNOM.","The analytic dispersion relations, both local and nonlocal, match full transfer-matrix simulations, so the closed-form expressions can be used directly to design GEP-based devices."],"supporting_citations":[{"why":"Supplies the multi-Lorentzian excitonic conductivity model and bias-dependent parameters for biased bilayer graphene.","marker":"[39]"},{"why":"Supplies the excitonic conductivity parameters for rhombohedral trilayer graphene, including the constant linewidth assumption used for BRTG.","marker":"[40]"},{"why":"Provides experimental evidence of tunable excitons in bilayer graphene, grounding the existence of the interband excitons.","marker":"[41]"},{"why":"Establishes biased bilayer graphene as a semiconductor with an electric-field-tunable gap, the prerequisite for interband excitons.","marker":"[45]"},{"why":"Supplies the anisotropic hBN effective permittivity used in the GEP dispersion relation.","marker":"[58]"},{"why":"Shows that positive imaginary conductivity from excitons produces in-plane propagating exciton-polaritons in 2D semiconductors, the mechanism GEPs extend.","marker":"[54]"},{"why":"Provides the nonlocal-correction comparison between exciton and plasmon polaritons used to interpret the reduced confinement and improved damping.","marker":"[66]"}],"fun_headline_variants":["Electrically tunable exciton polaritons in biased graphene","Biased graphene's interband excitons enable tunable polaritons","Graphene exciton polaritons get a voltage knob","Voltage-tunable light confinement via graphene excitons"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The prediction stands on the assumption that the optical conductivity of biased bilayer and trilayer graphene is accurately described by the multi-Lorentzian excitonic model with the parameters imported from the cited calculations, and that no other contribution—free-carrier Drude response, phonons, disorder, or the electron-hole continuum—changes the sign or size of the imaginary conductivity in the frequency windows used.","fun_headline_variants_meta":{"raw":{"variants":["Electrically tunable exciton polaritons in biased graphene","Biased graphene's interband excitons enable tunable polaritons","Graphene exciton polaritons get a voltage knob","Voltage-tunable light confinement via graphene excitons"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000548,"raw_usage":{"total_tokens":2681,"prompt_tokens":1069,"completion_tokens":1612,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":685,"completion_tokens_details":{"reasoning_tokens":1540}},"tokens_in":685,"tokens_out":1612,"duration_ms":12298,"temperature":1.0,"reasoning_tokens":1540,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T22:43:37.804747+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A cryo-SNOM measurement on hBN-encapsulated biased bilayer graphene at T = 4 K and a bias near 52 meV should show interference fringes whose spacing follows $q_{\\mathrm{GEP}}=2i\\varepsilon_0\\varepsilon_{\\mathrm{hBN}}^{\\mathrm{eff}}\\omega/\\sigma$; if no such propagating modes appear, or if the imaginary part of the extracted optical conductivity is not positive at the main exciton resonance, the central claim is wrong.","supporting_citations":[{"cited_title":"Horng, Y.-H","cited_arxiv_id":null,"evidence_quote":"Supplies the multi-Lorentzian excitonic conductivity model and bias-dependent parameters for biased bilayer graphene."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the excitonic conductivity parameters for rhombohedral trilayer graphene, including the constant linewidth assumption used for BRTG."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides experimental evidence of tunable excitons in bilayer graphene, grounding the existence of the interband excitons."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Establishes biased bilayer graphene as a semiconductor with an electric-field-tunable gap, the prerequisite for interband excitons."},{"cited_title":"2D Semiconductors Superlattices as Hyperbolic Materials","cited_arxiv_id":"2411.14785","evidence_quote":"Supplies the anisotropic hBN effective permittivity used in the GEP dispersion relation."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Shows that positive imaginary conductivity from excitons produces in-plane propagating exciton-polaritons in 2D semiconductors, the mechanism GEPs extend."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the nonlocal-correction comparison between exciton and plasmon polaritons used to interpret the reduced confinement and improved damping."}],"review_version":1}