{"id":"0018074e-7948-490a-9cb3-1d70fa66efa6","arxiv_id":"1908.09320","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"Nonlocal response in graphene saturates the momentum of singular metasurface plasmons, and a local model with a conductivity offset reproduces the effect.","lead":"The authors model how nonlocal electron response changes terahertz plasmons in graphene metasurfaces with sharply suppressed conductivity. They find nonlocality prevents the extreme field compression that local models predict, and offer a simple local approximation with a fitted conductivity offset.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Eq. (4) assumes a spatially uniform nonlocal kernel; the singular modulation drives the local Fermi wavevector to zero, so g ≪ k_F(x) fails exactly at the singular point. The nonlocal spectra and β = 1.29 are thus conditional on an uncontrolled factorization.","rationale":"Good-faith reading: the paper is a physically motivated extension of singular metasurface theory to nonlocal graphene response, with a clear qualitative message and a practical local-analogue recipe. The local FEM benchmark supports the local-response implementation, and the nonlocal conductivity model itself is standard RPA in the relaxation-time approximation. The load-bearing step is the adiabatic factorization leading to Eq. (4), because all nonlocal spectra in Figs. 3 and 4 depend on it and there is no independent nonlocal benchmark. The factorization is not simply a benign use of L ≫ λ_F; since the modulation is doping, k_F(x) = k_F,0 ζ(x), the small parameter g/k_F becomes O(1) at the singular point. For the reported parameters, g/k_F,min ≈ 2 while the average value is 2 × 10^−3. Thus the validity condition stated by the authors is violated in exactly the region where the effect is claimed. The alternative would be a kernel whose characteristic momentum varies with position, which Eq. (4) does not capture. This does not invalidate the qualitative physics, since nonlocality opposing extreme field compression and Landau damping broadening modes are plausible, and the fitted β close to unity makes the local analogue appealing; but the quantitative spectra and β = 1.29 are conditional on a calculation that has not been demonstrated in the singular regime. Hence the reader's CONDITIONAL verdict stands.","tokens_in":9933,"tokens_out":12115,"duration_ms":139985,"concrete_test":"Recompute the Δ = 3 transmittance (Fig. 3c) with a Wigner-transformed nonlocal kernel that uses the local Fermi wavevector k_F(x) = k_F,0 ζ(x) in the RPA polarizability of Appendix A, instead of the single-kernel factorization behind Eq. (4). If the spectrum and the best-fit β change by more than the linewidth or by more than about 10%, the adiabatic factorization is not controlled in the singular limit and the quantitative claim is conditional.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim, that nonlocal response dominates and smears the singularity, rests on the mode-matching relation Eq. (4), which is obtained by factoring a scalar modulation ζ(x′) inside the nonlocal convolution in Eq. (2). The stated validity condition is g ≪ k_F and L ≫ λ_F. These conditions are not merely pushed to their edge near the singular point; they fail there. Because the modulation is realized by local doping, the local Fermi wavevector itself scales as k_F(x) = k_F,0 ζ(x). For the parameters of Fig. 3 (E_F = 0.4 eV, v_F = 9.5 × 10^5 m/s, L = 5 μm), k_F,0 ≈ 640 μm^−1 and g ≈ 1.26 μm^−1, so g/k_F,0 ≈ 2 × 10^−3; but at the singular point Δ = 3 gives ζ_min = 10^−3, hence k_F,min ≈ 0.64 μm^−1 and g/k_F,min ≈ 2. The condition g ≪ k_F is therefore violated by three orders of magnitude in the very region where the conductivity is suppressed and the field is most singular. Eq. (4) keeps a single nonlocal kernel σ(k + ng) evaluated at the average Fermi level and lets the doping modulation appear only as a scalar prefactor. A position-dependent doping should instead enter the nonlocal kernel's characteristic momentum, so the kernel itself, not just its amplitude, varies in space. No nonlocal benchmark protects against this; the FEM comparison is local only. The qualitative direction, that nonlocality opposes singular field concentration, is plausible and consistent with prior nonlocal plasmonics, but the specific spectra and the successful fit β ≈ 1.29 are not established unless the factorization is shown to be controlled in the singular limit.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the optical response of a graphene metasurface whose local conductivity is periodically suppressed, forming what the authors call a singular conductivity grating. It extends an earlier local-response treatment of such gratings by incorporating the nonlocal (momentum-dependent) conductivity of graphene. The central claims are: (i) in the singular limit the nonlocal response becomes the dominant effect, opposing the merging of surface-plasmon modes and effectively smearing the singularity; (ii) a simple local-analogue model, obtained by adding a constant conductivity offset with a phenomenological factor β ≈ 1.29, reproduces the fully nonlocal transmission spectrum. The nonlocal calculation is performed with a mode-matching method based on Eq. (4), which factorizes the spatially varying doping profile outside a nonlocal conductivity kernel, under an adiabatic approximation stated to require g ≪ k_F. The local limit of the method is benchmarked against COMSOL.","tokens_in":10324,"tokens_out":3468,"duration_ms":36594,"significance":"If the claims hold, the paper would establish a practically relevant platform for probing graphene's nonlocal response with far-field THz spectroscopy, and the proposed local-analogue model would be a useful tool for simulating complex graphene-based structures without a full nonlocal solver. The paper makes productive use of a standard nonlocal conductivity model (Appendix A, based on the Mermin/RPA polarizability), and the local-limit FEM benchmark is a genuine strength. However, the central nonlocal spectra rest on an uncontrolled adiabatic factorization near the singular point, and the key validation of the local analogue model is a fit rather than a prediction. These issues currently leave the quantitative predictions—the shape of the nonlocal spectra and the value β ≈ 1.29—unestablished.","major_comments":[{"comment":"The adiabatic factorization that produces Eq. (4) is stated to be accurate as long as g ≪ k_F. This condition fails exactly in the singular region that the paper studies. For the parameters of Fig. 3 (E_F = 0.4 eV, v_F = 9.5 × 10^5 m/s, L = 5 μm), the average Fermi wavevector is k_F,0 ≈ 640 μm^-1 and g = 2π/L ≈ 1.26 μm^-1, so g/k_F,0 ≈ 2 × 10^-3. However, the local doping profile is ζ(x) = 1 + ζ_1 cos(gx), so the local Fermi wavevector scales as k_F(x) = k_F,0 ζ(x). At the singular point for Δ = 3, ζ_min = 10^-3, giving k_F,min ≈ 0.64 μm^-1 and g/k_F,min ≈ 2. The stated validity condition is therefore violated by three orders of magnitude in the very region where the conductivity is suppressed and the field is expected to be most singular. Since Eq. (4) evaluates the nonlocal kernel at the average Fermi level and lets the modulation appear only as a scalar prefactor, the nonlocal spectra in Fig. 3(c) and the claimed smearing of the singularity are not established as consequences of the model.","section":"§2, Eq. (4)"},{"comment":"The local-analogue model is validated against the nonlocal spectrum by choosing β ≃ 1.29, as stated in the text: 'by choosing β≃1.29 this simple model is able to reproduce the entire transmission spectrum with remarkable accuracy, hereby validating the physical assumptions behind our local analogue model'. This is a fitting exercise, not an independent validation. The parameter β is free, and the agreement is therefore a measure of flexibility rather than physical content. To validate the model, the authors would need to determine β from a separate observable or regime (e.g., a different modulation strength, a different frequency range, or reflection data) and then show that the same value predicts the transmission spectrum without further adjustment.","section":"§3, Fig. 5 and Eq. (6)"},{"comment":"The nonlocal mode-matching method is benchmarked against FEM only in the local-response limit; no comparison with a nonlocal numerical solution is provided. The local benchmark validates the Bloch-mode implementation for a local conductivity, but it cannot validate the nonlocal factorization in Eq. (4), which is the only nonlocal ingredient in the calculation. Given the failure of the stated validity condition at the singular point (see first major comment), the authors should provide a direct numerical solution of the nonlocal integral equation (Eq. 2) for the grating, or an independent nonlocal solver (e.g., a full RPA-based electromagnetic simulation) for at least one of the singular cases, to establish that the predicted spectra are not an artifact of the factorization.","section":"§2–§3, method validation"}],"minor_comments":[{"comment":"The definition of Δσ(ω) is notationally confusing: the right-hand side contains 'i Im[σ_s − (1−ζ_1)σ_D(ω)]' multiplied by a complex factor, but the text refers to a 'positive surface conductivity offset' and to 'smearing of the imaginary part of the surface conductivity'. Please clarify whether the offset is purely imaginary, purely real, or complex, and show explicitly how the loss tangent is preserved by the second factor.","section":"§3, Eq. (6)"},{"comment":"The sentence 'The quasi-static dispersion relation of graphene plasmons is reads [20]:' contains a typo ('is reads' should be 'reads').","section":"§3, text before Eq. (5)"},{"comment":"The caption states that 'modes above each band gap become extremely broad, due to their stronger radiative coupling', but the plotted quantity is described only as the logarithm of the absolute value of the reflection coefficient. Please specify the axes and color scale more precisely, and indicate whether the broadening is attributed to radiative coupling or to material loss.","section":"Fig. 4 caption"},{"comment":"The abstract and introduction refer to the 'full nonlocal optical response' of graphene, but the calculation uses an adiabatic factorization that assumes a separation of length scales and evaluates the nonlocal kernel at the average Fermi level. This wording overstates the generality of the treatment; a more cautious phrasing, such as 'a nonlocal response model based on the adiabatic approximation', would be more accurate.","section":"Abstract and §2"}],"recommendation":"major_revision","confidential_remarks":"The paper addresses an interesting and timely problem, and the qualitative direction of the results (nonlocality opposes singular field concentration) is plausible and consistent with the broader nonlocal-plasmonics literature. However, the central quantitative claims are currently supported only by a mode-matching relation whose stated validity condition fails at the singular point, and the proposed local analogue is validated with a fitted parameter. The revision should focus on establishing the nonlocal results through a controlled numerical solution or a more rigorous derivation, and on reframing the local-analogue agreement as a fit rather than a validation. I am sympathetic to the authors' approach and believe the issues are addressable within the manuscript's scope."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe headline on this one: the authors extend their singular graphene metasurface work into the nonlocal RPA regime and find that Landau damping puts a floor under the conductivity, smearing the singularity. The qualitative physics is plausible, and the local-analogue conductivity offset in Eq. (6) is a genuinely handy tool for people running local FEM simulations of complex graphene devices.\n\nWhat is new: the specific prediction that nonlocal intraband Landau damping saturates the singular response, so the perfect singularity is unattainable, plus the offset recipe that mimics this in a local solver. The nonlocal conductivity model in Appendix A is standard Mermin/Lindhard, and the local-response benchmark against COMSOL in Fig. 3 is a good sanity check. The paper is clearly written and the authors state their assumptions.\n\nThe soft spots are not cosmetic. The central mode-matching relation, Eq. (4), factors the doping profile ζ(x') out of the nonlocal kernel and keeps a single kernel σ(k+ng) evaluated at the average Fermi level. The stated validity condition is g ≪ k_F and L ≫ λ_F. For Δ=3, ζ_min = 10^-3, so the local Fermi wavevector at the singular point is k_F,min ≈ 0.64 μm^-1 while g ≈ 1.26 μm^-1, giving g/k_F,min ~ 2. The condition fails in exactly the region where the field is most singular and nonlocality matters most. The authors do not check this, and the FEM validation is local only, so it does not protect the nonlocal calculation. This is load-bearing: if the adiabatic factorization is uncontrolled at large momenta, the specific spectra and the fitted β ≈ 1.29 are not established. The qualitative direction—nonlocality opposes extreme field concentration—is consistent with the rest of nonlocal plasmonics, so I expect the physics to survive, but the quantitative claims need more support.\n\nAlso, β in Eq. (6) is fitted to the very nonlocal spectrum it is then used to validate. That is not circular in a damning sense—the model is a fitting exercise, not an independent prediction—but the paper calls it validation, which oversells it.\n\nBottom line: this paper deserves a serious referee, but the referee should push on the factorization. I would send it to review expecting major revision, asking for either a rigorous justification of the adiabatic approximation in the singular limit or a numerical nonlocal benchmark covering the relevant momentum range. I would not cite it in its current form.\n\nBest.","headline":"Plausible and practically useful extension of singular graphene metasurface theory into the nonlocal regime, but the central nonlocal spectra rest on an adiabatic factorization that breaks down exactly at the singularity.","tokens_in":10864,"tokens_out":2548,"would_cite":false,"duration_ms":24046,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper claims that a singular graphene metasurface—a conductivity grating whose local doping is almost completely suppressed—couples THz light to plasmon momenta near the Fermi wavevector, where graphene's nonlocal electron response…","keywords":["graphene plasmonics","nonlocal optical response","singular metasurfaces","terahertz light harvesting","Landau damping","conductivity grating","local analogue model"],"falsifier":"A far-field THz transmission measurement on a graphene grating with period $L = 5\\,\\mu\\mathrm{m}$, Fermi level $E_F = 0.4\\,\\mathrm{eV}$, and modulation strength $\\Delta = 3$ would settle the claim: if the spectrum matches the local-response prediction and the two plasmon resonances continue to merge as the grating deepens, nonlocality is not playing the dominant role claimed here; if the spectrum shows the predicted blueshift, extra broadening, and saturation of the resonance positions, the nonlocal mechanism is confirmed.","tokens_in":9779,"feed_emoji":"⚡","tokens_out":19015,"duration_ms":160838,"temperature":0.7,"pith_summary":"This paper argues that a 'singular graphene metasurface'—a periodic grating in which the local doping, and hence the local conductivity, is suppressed to near zero—can couple ordinary THz light to surface plasmons carrying momenta large enough to enter graphene's nonlocal regime. In a purely local description, deepening the grating would make the plasmon resonances merge into flat, degenerate bands with extreme field concentration. The paper shows that nonlocality opposes that merging: at momenta approaching the Fermi wavevector, the conductivity is enhanced and Landau damping—loss from matching the plasmon phase velocity to the electron velocity—sets in, so the singular spectrum blueshifts, broadens, and saturates instead. It then proposes a local analogue model—an added conductivity offset with a single fitted parameter, $\\beta \\simeq 1.29$—that reproduces the full nonlocal transmission spectrum, so complex graphene platforms can be modelled with ordinary local-response calculations.","feed_headline":"Graphene's nonlocality blocks mode merging in singular metasurfaces","feed_subtitle":"Deep THz gratings make plasmons blueshift and broaden instead of fusing, exposing graphene's electron-scale response.","key_machinery":"The central object is the nonlocal surface conductivity folded into a plane-wave coupling equation. Under the adiabatic approximation, the current Fourier amplitude reads $J_{n,x} = \\sigma(k+ng)[E_{n,x} + \\frac{\\zeta_1}{2}(E_{n+1,x}+E_{n-1,x})]$, valid when the grating wavenumber $g$ is much smaller than the Fermi wavevector $k_F$; this equation determines which Fourier harmonics of the plasmon field enter the lossy, Landau-damped regions of graphene's phase space. The nonlocal $\\sigma(k,\\omega)$ comes from the two-dimensional density-density response function in the relaxation-time approximation, which encodes Pauli-blocked interband transitions at low momentum and intraband Landau damping at high momentum. The second piece of machinery is the local analogue offset of Eq. (6), a positive conductivity shift $\\Delta\\sigma(\\omega)$ that saturates the local conductivity at $\\sigma_s = 2i\\epsilon_0 v_F/\\beta$, so that an ordinary local simulation mimics the nonlocal saturation of the plasmon wavevector.","core_discovery":"On its own terms, the paper's central discovery is that the singular limit of a graphene conductivity grating is the natural far-field probe of graphene's nonlocal response. Using a Fourier-mode expansion with the full momentum-dependent conductivity $\\sigma(k,\\omega)$, the authors find that for weak gratings the local and nonlocal transmission spectra coincide, but as the modulation strength $\\Delta = -\\log_{10}(1-\\zeta_1)$ grows to 3, nonlocality becomes dominant: the Fourier components with large momenta see an enhanced conductivity, the resonances blueshift, and the merging of modes predicted by local-response theory is prevented. The mechanism is that the high-momentum components enter the region of phase space where Landau damping from intraband transitions is active, effectively saturating the plasmon wavevector near the electronic one. The paper further claims that a constant surface-conductivity offset, Eq. (6), with $\\beta \\simeq 1.29$, reproduces the nonlocal spectrum in a local calculation, providing a practical route to including nonlocal effects in more complex setups.","pith_inferences":["Beyond the paper, the same saturation mechanism should operate in any two-dimensional conductor whose doping can be periodically suppressed, so singular gratings could serve as a general spectroscopy of nonlocal response in 2D materials.","Beyond the paper, the fitted coefficient $\\beta \\simeq 1.29$ is an experimentally accessible parameter: measuring how it varies with Fermi level, grating period, and temperature would test the simple offset model and yield quantitative information about the momentum at which Landau damping saturates.","Beyond the paper, the smearing of the singularity implies a practical upper bound on the achievable field enhancement and absorption bandwidth of singular graphene metasurfaces, because the conductivity cannot be driven below the nonlocal saturation value; the paper does not spell out this bound."],"forward_implications":["Far-field THz transmission through a deep graphene conductivity grating becomes a direct probe of graphene's nonlocal conductivity at momenta up to the Fermi wavevector.","Nonlocality sets a floor on the effective conductivity at the singular point, so the predicted ultra-flat degenerate bands and the associated field enhancement are partially washed out by intraband Landau damping.","The single-offset local analogue with $\\beta \\simeq 1.29$ reproduces the nonlocal spectrum, allowing local-response simulations of more complex graphene metasurfaces without solving the full nonlocal problem.","The difference between local and nonlocal behaviour is negligible for weak gratings and becomes decisive only as the singular limit is approached, so nonlocal corrections are needed precisely in the regime designed for maximum absorption and field concentration."],"supporting_citations":[{"why":"Proposed the singular graphene metasurface platform that this work extends from the local to the nonlocal regime.","marker":"[7]"},{"why":"Developed the singular graphene metasurface description in the local limit, providing the baseline spectra that the nonlocal calculation is compared against.","marker":"[13]"},{"why":"Supplies the nonlocal conductivity model for graphene, the mode-matching technique, and the quasi-static plasmon dispersion relation used throughout.","marker":"[20]"},{"why":"Reports experimental tuning of quantum nonlocal effects in graphene plasmons, establishing that nonlocal response matters at large plasmon momenta.","marker":"[15]"},{"why":"Shows graphene plasmons probing nonlocal response in metals, supporting the claim that singular gratings reach the nonlocal regime.","marker":"[16]"},{"why":"Provides the relaxation-time approximation for the density-density response function on which the nonlocal conductivity model is built.","marker":"[43]"},{"why":"Introduced the local-analogue approach for nonlocal response in metallic plasmonic nanostructures, which the paper adapts into a conductivity offset for graphene metasurfaces.","marker":"[32]"},{"why":"Extended nonlocal analysis to singular plasmonic metasurfaces, forming the conceptual bridge between hydrodynamic nonlocal models and the graphene case.","marker":"[31]"}],"fun_headline_variants":["Nonlocality halts plasmon merging in singular graphene metasurfaces","Singular gratings probe graphene's electron nonlocality","Local offset mimics nonlocal graphene response in metasurfaces","Blueshift and broadening expose graphene's nonlocality","Singular metasurface plasmons reveal electron-scale response"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the periodic doping modulation can be factored outside graphene's nonlocal response kernel, an adiabatic approximation whose validity requires the grating wavenumber $g$ to be far below the Fermi wavevector $k_F$, even though the singular regime drives the dominant plasmon momenta up to $k_F$, where that separation of scales is strained.","fun_headline_variants_meta":{"raw":{"variants":["Nonlocality halts plasmon merging in singular graphene metasurfaces","Singular gratings probe graphene's electron nonlocality","Local offset mimics nonlocal graphene response in metasurfaces","Blueshift and broadening expose graphene's nonlocality","Singular metasurface plasmons reveal electron-scale response"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000455,"raw_usage":{"total_tokens":2260,"prompt_tokens":894,"completion_tokens":1366,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":510,"completion_tokens_details":{"reasoning_tokens":1284}},"tokens_in":510,"tokens_out":1366,"duration_ms":10349,"temperature":1.0,"reasoning_tokens":1284,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T11:14:47.046389+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A far-field THz transmission measurement on a graphene grating with period $L = 5\\,\\mu\\mathrm{m}$, Fermi level $E_F = 0.4\\,\\mathrm{eV}$, and modulation strength $\\Delta = 3$ would settle the claim: if the spectrum matches the local-response prediction and the two plasmon resonances continue to merge as the grating deepens, nonlocality is not playing the dominant role claimed here; if the spectrum shows the predicted blueshift, extra broadening, and saturation of the resonance positions, the nonlocal mechanism is confirmed.","supporting_citations":[{"cited_title":"Galiffi, J","cited_arxiv_id":null,"evidence_quote":"Proposed the singular graphene metasurface platform that this work extends from the local to the nonlocal regime."},{"cited_title":"Galiffi, J","cited_arxiv_id":null,"evidence_quote":"Developed the singular graphene metasurface description in the local limit, providing the baseline spectra that the nonlocal calculation is compared against."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the nonlocal conductivity model for graphene, the mode-matching technique, and the quasi-static plasmon dispersion relation used throughout."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Reports experimental tuning of quantum nonlocal effects in graphene plasmons, establishing that nonlocal response matters at large plasmon momenta."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Shows graphene plasmons probing nonlocal response in metals, supporting the claim that singular gratings reach the nonlocal regime."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the relaxation-time approximation for the density-density response function on which the nonlocal conductivity model is built."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Introduced the local-analogue approach for nonlocal response in metallic plasmonic nanostructures, which the paper adapts into a conductivity offset for graphene metasurfaces."},{"cited_title":"Yang, Y.-T","cited_arxiv_id":null,"evidence_quote":"Extended nonlocal analysis to singular plasmonic metasurfaces, forming the conceptual bridge between hydrodynamic nonlocal models and the graphene case."}],"review_version":1}