{"id":"81a97481-e7b7-4175-b954-fab0349de3fe","arxiv_id":"1908.03806","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Nonlocal phonon response, modeled by dispersive longitudinal optical phonons, qualitatively changes the optical spectra of nanoscale polar dielectrics and explains measured superlattice reflectance where local theory fails.","lead":"This paper develops a nonlocal theory of light-matter interaction in polar crystals, showing that standard local dielectric models fail for nanoscale phonon polaritons. It predicts new resonances in silicon carbide nanospheres and explains previously unexplained infrared spectra of atomic-scale nitride superlattices.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Acknowledged 795 cm⁻¹ mismatch in Fig. 5b undercuts the 'quantitatively correct down to few atomic lattices' claim; with β_AlN_L fit to the same data, the experimental validation remains conditional.","rationale":"The reader's weakest assumption flagged the quadratic approximation and ABC choice, and noted the 795 cm⁻¹ dip. My concern converges on the quadratic approximation's failure because it is not speculative: the authors state it directly, and it sits exactly on the quantitative claim they make. I considered the fitted-β circularity as an alternative, but the more fundamental point is that even with β fitted, the model fails in the narrowest structure; this is visible inside the paper. If the nonlocal model were mechanically correct, one would expect the extra fitting freedom to absorb or at least not worsen the dip, yet the model's own diagnostic (large-wavevector LO modes) shows the single-β quadratic approximation breaks down. The ABC choice is a real open question, but Appendix B argues Fuchs-Kliewer and Pekar-Ridley are practically indistinguishable; without numerical counterevidence this is less load-bearing. The nonlocal sphere predictions are parameter-free in the sense of using literature β values and therefore provide independent support for the qualitative phenomenon, which is why the paper should not be rejected; it should be accepted for the nonlocal phenomenology while the experimental verification claim is downgraded to conditional pending a dispersion-aware fit.","tokens_in":16027,"tokens_out":7235,"duration_ms":81943,"concrete_test":"Compute the reflectance of heterostructure B with the quadratic term β_L²k² in Eq. (2) replaced by the actual AlN LO phonon dispersion (e.g., DFT dispersion from Bungaro et al. [58] or a piecewise-constant β_L(k) as the authors suggest), refitting only δ and γ. If the 795 cm⁻¹ dip is then reproduced while retaining the 835/865 cm⁻¹ features and the ENZ redshift, the discrepancy is solely the quadratic approximation and the nonlocal framework can be preserved with a qualified dispersion model; if the dip persists or the other features shift, the fitted β_AlN_L was masking model error and the 'quantitative correctness' claim must be retracted or substantially qualified.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim—'first experimental verification that our theory provides quantitatively correct results for systems down to few atomic lattices in size'—is weakened by the authors' own Sec. IV.B admission: the quadratic dispersion approximation 'fails' to reproduce the dip at 795 cm⁻¹ in Fig. 5b because the 1.2/1.4 nm layers of heterostructure B probe wavevectors where constant β_L underestimates the LO frequency. Since heterostructure B is exactly the 'few atomic lattices' system used to make the quantitative claim, the model is demonstrably not quantitatively correct across the measured spectrum. This is compounded by Appendix F: β_AlN_L is one of three parameters fitted to the same reflectance data (local theory uses two), and no uncertainty or sensitivity analysis is given. Agreement at 835/865 cm⁻¹ and the redshifted ENZ mode is therefore partly a consequence of parameter adjustment; it does not independently confirm that a single quadratic β_L captures the nonlocal response at these wavevectors. The load-bearing condition is that Eqs. (2)-(3) with constant β_L describe the relevant LO branch; the paper itself shows that condition fails in the narrower structure.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops a macroscopic continuum nonlocal dielectric theory for polar dielectrics, in which the LO and TO phonon dispersions are treated as quadratic in wavevector (Eqs. 2 and 3) and an additional boundary condition (Fuchs-Kliewer) closes the interface problem. The theory is applied to three systems: SiC nanospheres, where it predicts extra extinction peaks and size-dependent damping; thin AlN films, where it predicts a redshift of the ENZ mode with decreasing thickness; and AlN/GaN atomic-scale superlattices, where it aims to reproduce unexplained reflectance features. The central claim is that the local dielectric description fails at nanometric scales and that the nonlocal theory provides the first experimental verification of quantitatively correct results down to a few atomic lattices. The paper includes detailed appendices deriving the continuum model, the additional boundary conditions, the nonlocal Mie theory, and the scattering-matrix treatment of layered systems.","tokens_in":16237,"tokens_out":3954,"duration_ms":40824,"significance":"If the central claim were fully supported, the paper would establish that local effective-medium descriptions of phonon polaritons break down below roughly 10 nm and would provide a computationally light continuum tool for nanophotonic design. The SiC nanosphere predictions are the strongest part: the parameters β_SiC_L and β_SiC_T are taken from independent ab initio phonon dispersion calculations, so those predictions are not circular, and the predicted extra peaks and Fano-like features are concrete and falsifiable. The appendix derivations are careful and the numerical methods are clearly described. However, the superlattice validation, which is the basis of the 'first experimental verification' claim, is substantially weakened by the fact that β_AlN_L, γ_AlN, and a thickness shift are all fitted to the same reflectance data, and by the paper's own admission that the quadratic dispersion fails at the 795 cm⁻¹ dip in Fig. 5b.","major_comments":[{"comment":"The authors explicitly acknowledge that the quadratic dispersion approximation fails to reproduce the dip at 795 cm⁻¹ in Fig. 5b, because the 1.2/1.4 nm layers of heterostructure B probe wavevectors where the constant β_L underestimates the LO frequency. Since heterostructure B is precisely the 'few atomic lattices' system used to support the conclusion that the theory provides quantitatively correct results down to few atomic lattices in size, this admission undercuts that claim. The authors should either restrict the quantitative claim to the features that are actually reproduced, or improve the dispersion model (e.g., using piecewise constant velocities) and refit the data.","section":"Sec. IV.B and Fig. 5b"},{"comment":"The nonlocal superlattice fit uses three parameters (β_AlN_L, γ_AlN, and the thickness shift δ) fitted to the same experimental reflectance data, whereas the local theory uses two parameters, and no uncertainties, confidence intervals, or sensitivity analysis are reported. Consequently, the improved agreement at 835/865 cm⁻¹ in heterostructure A and the redshifted ENZ mode in heterostructure B is partly a result of parameter adjustment rather than an independent confirmation that a single constant β_AlN_L captures the nonlocal response at the relevant wavevectors. The authors should provide an independent determination of β_AlN_L from phonon dispersion data, or at least quantify how the fit quality varies with β_AlN_L and show that the qualitative conclusions are robust.","section":"Appendix F and Fig. 5"},{"comment":"The abstract and the concluding section state that the paper provides 'the first experimental verification that our theory provides quantitatively correct results for systems down to few atomic lattices in size.' Given the acknowledged 795 cm⁻¹ mismatch and the fitted nature of β_AlN_L, this wording overstates what the comparison demonstrates. The authors should soften the claim to something like 'first experimental evidence consistent with the nonlocal theory' and explicitly delineate which spectral features are and are not reproduced.","section":"Abstract and Sec. V"}],"minor_comments":[{"comment":"The paragraph beginning 'By modelling recently published experimental data...' is a sentence fragment without a main verb; it should be rewritten as a complete sentence.","section":"Sec. V"},{"comment":"The phrase 'through it's surface' should read 'through its surface'.","section":"Appendix B"},{"comment":"The name 'Debarnardi' should be 'Debernardi' to match the cited reference [55] (Debernardi et al., 1999).","section":"Appendix C"},{"comment":"The caption refers to 'The dots' while the main text describes the n = 1 longitudinal mode as green circles; please make the notation consistent.","section":"Fig. 4 caption and text"},{"comment":"The fitting procedure would benefit from a statement of the uncertainty on δ, γ_AlN, and β_AlN_L, and from a discussion of the correlation between δ and β_AlN_L, since both affect the effective confinement length.","section":"Appendix F"}],"recommendation":"major_revision","confidential_remarks":"The strongest contribution is the nonlocal Mie and layered-system framework with independently parameterized SiC predictions; the weakest point is the superlattice 'verification' claim, which is partly circular and contradicted by the acknowledged 795 cm⁻¹ mismatch. I believe the paper can be brought to publishable form by moderating the claims, adding uncertainty/sensitivity analysis, and possibly presenting an independent β_AlN_L estimate from phonon dispersion. The current version is not acceptable as is, but the issues are fixable within the manuscript's scope."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The genuine new thing here is applying a hydrodynamic nonlocal dielectric model to polar dielectrics, using the negative LO phonon dispersion to predict propagative longitudinal modes inside the Reststrahlen region. That produces observable consequences: extra resonances in small SiC spheres, a redshift of ENZ modes in thin AlN films, and a Kreibig-like size-dependent damping. The sphere predictions use beta_SiC from independent ab initio phonon dispersion, so they are not circular. The appendices give a coherent derivation of the continuum model, boundary conditions, Mie theory, scattering matrix, and fitting procedure. This is a useful, reusable toolset for mid-infrared nanophotonics.\n\nThe soft spot is the experimental section. beta_AlN_L, gamma_AlN, and a thickness shift are all fitted to the same reflectance data, and the improvement over the local theory is not statistically quantified. The paper itself admits that the quadratic dispersion fails at the 795 cm-1 dip in Fig. 5b, which undercuts the abstract's claim of being 'quantitatively correct' for systems down to few atomic lattices, because heterostructure B is exactly that system. With three fitting parameters against two, the agreement at 835 and 865 cm-1 and the redshift in structure B is suggestive but not a clean confirmation. The abstract and conclusion state the quantitative claim without the caveat the authors themselves make in Sec. IV.B.\n\nThese are fixable. The authors could obtain beta_AlN_L independently from ab initio phonon dispersion, or show that the experimental features survive with beta fixed to a reasonable range, or at least rephrase the claim as 'consistent with nonlocal effects' rather than 'first experimental verification.' The ABC choice is left open, but the authors argue it does not affect these observables, which is plausible and not a fatal issue.\n\nThis paper is for the phonon-polariton and mid-infrared nanophotonics community. It deserves a serious referee. I would accept it with moderate revision, but the experimental validation language needs to be pulled back until the fitted parameters are better constrained.","headline":"Solid nonlocal theory for phonon polaritons with independent sphere predictions, but the 'first experimental verification' claim overreaches because the key nonlocal parameter is fit to the same data and the model misses a dip in the few-atomic-lattice sample.","tokens_in":16801,"tokens_out":1795,"would_cite":true,"duration_ms":20168,"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 demonstrates that the standard local dielectric approximation fails for phonon polaritons in nanometric polar dielectrics, and that a nonlocal continuum theory with quadratic phonon dispersion quantitatively reproduces…","keywords":["phonon polaritons","nonlocal optics","polar dielectrics","spatial dispersion","Reststrahlen band","epsilon-near-zero modes","AlN/GaN superlattices","Mie theory"],"falsifier":"A direct check would be to measure the infrared reflectance of an AlN/GaN superlattice with layer thicknesses around 1 nm and compare the position and shape of the dip near 795 cm$^{-1}$, where the paper itself states the quadratic approximation fails; a corrected nonlocal theory must reproduce that feature. Alternatively, measure the extinction spectrum of 5 nm radius 3C-SiC nanospheres and look for the discrete longitudinal peaks below the LO frequency that the nonlocal theory predicts: if no such peaks appear, the central mechanism is wrong.","tokens_in":1990,"feed_emoji":"🔬","tokens_out":4614,"duration_ms":110173,"temperature":0.7,"pith_summary":"The paper argues that the usual local dielectric model, which assumes the material response at a point depends only on the field at that point, breaks down for phonon polaritons in nanometric polar crystals. It develops a macroscopic nonlocal theory in which optical phonons have spatial dispersion, and shows that this changes the predicted optical response of nanospheres, thin films, and superlattices. The theory explains previously unexplained infrared reflectance features in AlN/GaN atomic-scale superlattices that local theory cannot capture. If correct, local models are inadequate for phonon-polariton structures below roughly ten nanometers, and a lightweight continuum theory can serve as a design tool for mid-infrared nanophotonics.","feed_headline":"Local optics fails in nanoscale polar crystals","feed_subtitle":"A nonlocal continuum model with dispersive phonons explains infrared spectra of atomic-scale AlN/GaN superlattices.","key_machinery":"The carrying object is a continuum model of the polar crystal as an ionic displacement field $\\mathbf{X}$ coupled to Maxwell's equations, with the equation of motion $\\left[\\omega_T^2 - \\omega(\\omega+i\\gamma) + \\beta_L^2\\nabla(\\nabla\\cdot) - \\beta_T^2\\nabla\\times\\nabla\\times\\right]\\mathbf{X} = (\\mu/\\rho)\\mathbf{E}$. Fourier transforming this equation yields the longitudinal and transverse nonlocal dielectric functions $\\varepsilon_L(\\omega,k)$ and $\\varepsilon_T(\\omega,k)$, whose zeros define dispersive LO phonon modes. The system is closed by an additional boundary condition at interfaces, taken here as the vanishing of the ionic displacement at a dielectric-vacuum boundary, equivalent to specular reflection. This machinery lets longitudinal modes be excited at boundaries and propagate inside the nanostructure, coupling to transverse photonic modes and producing the predicted spectral features.","core_discovery":"The central claim is that LO phonons in polar dielectrics acquire spatial dispersion that becomes optically relevant at the nanoscale, so the local dielectric function $\\varepsilon_{\\mathrm{LRA}}(\\omega)$ must be replaced by wavevector-dependent functions $\\varepsilon_L(\\omega,k)$ and $\\varepsilon_T(\\omega,k)$ of the form given in Eqs. (2) and (3), with quadratic terms $\\beta_L^2 k^2$ and $\\beta_T^2 k^2$. Because optical phonons have negative dispersion, propagative longitudinal phonon modes coexist with the negative-dielectric Reststrahlen region, unlike in metals where longitudinal plasma waves are evanescent. This leads to discrete longitudinal resonances that couple to the Fröhlich mode, Fano-like interference, a small redshift of the main resonance, and size-dependent damping. Applied to AlN/GaN superlattices, the nonlocal model reproduces reflectance peaks and redshifts that local theory misses, with fitted parameters $\\beta^{\\mathrm{AlN}}_L = 5.1\\times10^5\\ \\mathrm{cm\\,s^{-1}}$ and $\\gamma_{\\mathrm{AlN}} = 10.3\\ \\mathrm{cm^{-1}}$, giving a nonlocal skin depth near 1.5 nm and indicating validity down to a few atomic layers.","pith_inferences":["The same continuum nonlocal formalism could be applied to other polar dielectrics with negative LO dispersion, such as quartz or hexagonal boron nitride, predicting analogous discrete longitudinal resonances and nonlocal redshifts in their Reststrahlen bands.","The fitted value of $\\beta^{\\mathrm{AlN}}_L$ could be checked independently against measured AlN phonon dispersion; a significant discrepancy would indicate that the single-quadratic-term approximation absorbs other physics and may not extrapolate to untested geometries.","Replacing the constant $\\beta_L$ with a wavevector-dependent piecewise-constant velocity, which the paper mentions as possible, would extend the model across the full Brillouin zone and should also capture the 795 cm$^{-1}$ dip where the quadratic approximation currently fails.","Because the nonlocal effects rely on boundary-induced coupling between longitudinal and transverse modes, patterned or roughened surfaces could be used to tune the strength of these features without changing material composition."],"forward_implications":["Below about 10 nm, local-response simulations of phonon-polariton structures will systematically miss extra resonances and mispredict mode frequencies; nonlocal terms must be included.","The extra peaks in small SiC nanospheres and in AlN/GaN superlattices are discrete longitudinal optical phonon modes confined by the particle or film boundaries, hybridizing with the Fröhlich or epsilon-near-zero resonances.","Nonlocal damping gives a size-dependent broadening of the form $\\gamma_{\\mathrm{NL}} = \\gamma + A\\beta_L/R$ with $A \\approx 0.03$ for 3C-SiC, meaning smaller particles lose more energy to LO phonon emission even at radii where the local model appears sufficient.","In thin films, confined LO phonons shift the effective edge of the Reststrahlen band, so epsilon-near-zero resonances follow the quantized mode frequency $\\omega_n$ rather than the zone-centre LO frequency for films thinner than about 10 nm.","A continuum nonlocal theory can reproduce experimentally observed infrared spectra of atomic-scale superlattices with only one additional fitted parameter, the longitudinal phonon velocity $\\beta_L$."],"supporting_citations":[{"why":"Supplies the experimental reflectance data for AlN/GaN superlattices that the nonlocal model reproduces and local theory cannot.","marker":"[31]"},{"why":"Demonstrates computationally that short-period AlN/GaN superlattices require density-functional perturbation theory, establishing the failure of local response that the paper addresses.","marker":"[30]"},{"why":"Provides the dispersive dielectric-function form for longitudinal and transverse optic vibrations used in Eqs. (2) and (3).","marker":"[33]"},{"why":"Establishes the nonlocal Mie theory for spheres with longitudinal modes that the paper adapts to dielectric nanospheres.","marker":"[21]"},{"why":"Supplies the generalized nonlocal response framework and size-dependent damping formalism for plasmonic nanostructures that motivates the analogous phononic treatment.","marker":"[24]"},{"why":"Provides the specular-reflection additional boundary condition used to close the nonlocal interface problem.","marker":"[35]"},{"why":"Gives ab initio phonon dispersion of SiC, from which the nonlocal velocities $\\beta^{\\mathrm{SiC}}_L$ and $\\beta^{\\mathrm{SiC}}_T$ are obtained.","marker":"[49]"},{"why":"Provides ab initio phonon dispersions of wurtzite AlN and GaN used to fix the GaN longitudinal velocity in the superlattice fits.","marker":"[58]"},{"why":"Introduces the size-dependent damping mechanism that the paper generalizes to phonon-polariton systems through its $\\gamma_{\\mathrm{NL}} = \\gamma + A\\beta_L/R$ fit.","marker":"[45]"}],"fun_headline_variants":["Phonon dispersion breaks local optics at the nanoscale","Negative phonon dispersion reveals hidden modes","Nonlocal phonon response reshapes infrared nanophotonics","Local optics fails, nonlocal phonon theory steps in"],"cache_read_input_tokens":18944,"weakest_assumption_plain":"The load-bearing premise is that the true optical-phonon dispersion in layers only one or two nanometers thick is well approximated by a single quadratic term with a constant velocity, and that the chosen interface boundary condition is the correct one; if the real dispersion bends away from that quadratic curve at large wavevectors, the predicted spectra will not generalize.","fun_headline_variants_meta":{"raw":{"variants":["Phonon dispersion breaks local optics at the nanoscale","Negative phonon dispersion reveals hidden modes","Nonlocal phonon response reshapes infrared nanophotonics","Local optics fails, nonlocal phonon theory steps in"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000945,"raw_usage":{"total_tokens":4042,"prompt_tokens":959,"completion_tokens":3083,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":575,"completion_tokens_details":{"reasoning_tokens":3019}},"tokens_in":575,"tokens_out":3083,"duration_ms":23471,"temperature":1.0,"reasoning_tokens":3019,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T14:01:18.010751+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A direct check would be to measure the infrared reflectance of an AlN/GaN superlattice with layer thicknesses around 1 nm and compare the position and shape of the dip near 795 cm$^{-1}$, where the paper itself states the quadratic approximation fails; a corrected nonlocal theory must reproduce that feature. Alternatively, measure the extinction spectrum of 5 nm radius 3C-SiC nanospheres and look for the discrete longitudinal peaks below the LO frequency that the nonlocal theory predicts: if no such peaks appear, the central mechanism is wrong.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the experimental reflectance data for AlN/GaN superlattices that the nonlocal model reproduces and local theory cannot."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Demonstrates computationally that short-period AlN/GaN superlattices require density-functional perturbation theory, establishing the failure of local response that the paper addresses."},{"cited_title":"& Arakawa, Y","cited_arxiv_id":null,"evidence_quote":"Provides the dispersive dielectric-function form for longitudinal and transverse optic vibrations used in Eqs. (2) and (3)."},{"cited_title":"& Mortensen, N","cited_arxiv_id":null,"evidence_quote":"Establishes the nonlocal Mie theory for spheres with longitudinal modes that the paper adapts to dielectric nanospheres."},{"cited_title":"A., Raza, S., Wubs, M., Sndergaard, T","cited_arxiv_id":null,"evidence_quote":"Supplies the generalized nonlocal response framework and size-dependent damping formalism for plasmonic nanostructures that motivates the analogous phononic treatment."},{"cited_title":"& Kliewer, K","cited_arxiv_id":null,"evidence_quote":"Provides the specular-reflection additional boundary condition used to close the nonlocal interface problem."},{"cited_title":"& Strauch, D","cited_arxiv_id":null,"evidence_quote":"Gives ab initio phonon dispersion of SiC, from which the nonlocal velocities $\\beta^{\\mathrm{SiC}}_L$ and $\\beta^{\\mathrm{SiC}}_T$ are obtained."},{"cited_title":"& Bernholc, J","cited_arxiv_id":null,"evidence_quote":"Provides ab initio phonon dispersions of wurtzite AlN and GaN used to fix the GaN longitudinal velocity in the superlattice fits."},{"cited_title":"& Fragstein, C","cited_arxiv_id":null,"evidence_quote":"Introduces the size-dependent damping mechanism that the paper generalizes to phonon-polariton systems through its $\\gamma_{\\mathrm{NL}} = \\gamma + A\\beta_L/R$ fit."}],"review_version":1}