{"id":"5b8f3ce9-94b9-4a82-9133-9217286dc874","arxiv_id":"1908.00640","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"A nonlocal Drude model predicts thickness-dependent splitting of ENZ modes in ultrathin films, causing two-exponential decay enhancement of up to three orders of magnitude.","lead":"This paper predicts that ultra-thin metal films support two distinct epsilon-near-zero plasmon modes instead of one, with the splitting tuned by film thickness. The effect couples strongly to nearby quantum emitters, giving thickness-controlled spontaneous-emission rates up to 1000 times faster than free space.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Eq. (1) as printed is inconsistent with Eq. (10): the plasma-frequency renormalization must read ωp(k)=ωp^3D/√(1+(ε1+ε2)/(ε k d)), not the printed √(1+(ε1+ε2)/(ε k d)) factor.","rationale":"The reader correctly identifies Eq. (1) as the load-bearing input, but the more serious issue is not merely that it is self-cited; the equation as printed is internally inconsistent with Eq. (10) and with the paper's own physical description. The reciprocal square-root form explains the stated red shift with decreasing d, the √k 2D plasmon dispersion, the √d thickness scaling of Eq. (3), and the algebra leading to Eq. (10). This is almost certainly a typographical error rather than a conceptual flaw, but it is not cosmetic: the magnitude and even the sign of the thickness dependence of the ENZ modes and the decay enhancement follow from the reciprocal form. The paper's quantitative claims (up to ~10^3 enhancement) and the specific pole locations would need to be re-confirmed with the corrected Eq. (1). This reinforces the reader's CONDITIONAL verdict: the core physics is plausible and internally consistent once Eq. (1) is fixed, but the manuscript should be corrected and the results re-verified before the central claim is taken as established. No experimental comparison or independent derivation of the Keldysh-Rytova plasma-frequency renormalization is provided, so the quantitative predictions remain conditional on the model input and on the typo correction.","tokens_in":10745,"tokens_out":10061,"duration_ms":81385,"concrete_test":"Re-derive ε(k,ω) by substituting both candidate forms of Eq. (1) into Eq. (2) and compare algebraically with Eq. (10) under the rescaling (7) with ε1=ε2=1. Confirm that Eq. (10) follows only from the reciprocal form; then recompute the Re ε=0 contour in Fig. 1(a) and the ENZ pole positions from Eq. (9) using the reciprocal form and verify the figures are unchanged.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central claims all reduce to Eq. (1), but the printed equation is internally inconsistent with the rest of the derivation. Substituting ωp(k)=ωp^3D√(1+(ε1+ε2)/(ε k d)) into Eq. (2) and using the rescaling (7) yields ε/ε = 1 − (A u + 2)/(A u^2 (u+iδ)) with A=ε k_p^3D d√(1+x^2) and ε1=ε2=1, whereas Eq. (10) is ε/ε = 1 − A/[(2 + A u)(u+iδ)]. These are not equal. The printed form also contradicts the text: for fixed k it increases as d decreases (blue shift), while the text states a red shift and a √k 2D dispersion; in the 2D limit the printed form gives ωp(k) ∝ 1/√k, not √k. The reciprocal form ωp(k)=ωp^3D/√(1+(ε1+ε2)/(ε k d)) reproduces Eq. (10) exactly, gives the stated √k dispersion, and yields the √d thermal-average behavior in Eq. (3). Since all subsequent pole analysis, mode splitting (Eq. (9)), and the biexponential decay formula (Eq. (14)) use Eq. (10), the intended model is the reciprocal form, but the foundational equation is misprinted. A reader implementing Eq. (1) literally would obtain opposite thickness dependence and altered ENZ poles, so the published central claim rests on an unstated correction.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript studies epsilon-near-zero (ENZ) modes in ultrathin metallic films in the transdimensional regime, using macroscopic QED and a confinement-induced nonlocal Drude dielectric function based on the Keldysh-Rytova interaction potential. From the pole structure of the p-polarized reflection coefficient of a symmetric planar film, the authors derive two ENZ dispersion branches x±(d) whose degeneracy is lifted as the film thickness d decreases (Eqs. (9) and (11)). Coupling of a point dipole emitter to these split modes is shown to produce a biexponential distance dependence of the spontaneous decay rate (Eq. (14)), with enhancements of two to three orders of magnitude over free space (Fig. 3). The effects are controlled by the thickness-dependent plasma frequency of the film, a hallmark of the nonlocal KR model, and the paper claims this provides a new thickness-based control knob for light-matter interactions.","tokens_in":11091,"tokens_out":8598,"duration_ms":77520,"significance":"If the predictions are correct, the paper would establish a useful generalization of Drexhage-type decay engineering and of local-Drude treatments of ultrathin films, with potential applications in tunable nanophotonics. The manuscript's explicit closed-form expressions, the clear physical interpretation of the two split modes, and the falsifiable predictions (decay-rate distance and thickness dependence) are strengths. However, the entire derivation reduces to the cited nonlocal Drude model, and the manuscript contains an internal inconsistency in the definition of the thickness-dependent plasma frequency that must be resolved before the predictions can be accepted. The central result is conditional on the correctness and applicability of the Keldysh-Rytova-based dielectric response at few-nanometer thicknesses, a condition the paper does not independently establish.","major_comments":[{"comment":"The printed Eq. (1) defines ωp(k) = ωp^3D √(1 + (ε1+ε2)/(ε k d)), but Eq. (10) is not derivable from this form. With the printed definition, substitution into Eq. (2) and the rescaling (7) yields ε/ε = 1 − (A u + 2)/(A u^2 (u+iδ)) with A = ε kp^3D d √(1+x^2), which differs from the printed Eq. (10). The reciprocal form ωp(k) = ωp^3D / √(1 + (ε1+ε2)/(ε k d)) reproduces Eq. (10) exactly. The printed form also contradicts the text: for fixed k it blue-shifts as d decreases, while the text states a red shift, and in the 2D limit it gives ωp(k) ∝ 1/√k, not the stated √k dispersion. Equation (1) is additionally inconsistent with the √d thickness behavior of Eq. (3). Since Eqs. (9), (11), (13), (14) and all figures explicitly use Eq. (10), a reader implementing the printed Eq. (1) would obtain opposite thickness trends and altered ENZ poles. This is a load-bearing error: the manuscript must be corrected to the reciprocal form (or an equivalent), and the authors should confirm that all subsequent results use that form.","section":"§II, Eq. (1); §III, Eq. (10)"},{"comment":"The biexponential formula (14) follows from the Lorentzian approximation (13), the second-order Maclaurin expansion used for Eq. (9), the b^2 >> c limit in Eq. (11), the condition Im x±/Re x± ≪ 1, and a first-order Taylor expansion of arctan. The manuscript does not provide a sensitivity analysis or error estimates for these steps. In particular, for the representative parameters used in Fig. 1 (ωp = 2.79 eV, u = 0.65, d = 10 nm), the x+ branch satisfies x+ ≈ 2.8, which is outside the domain x < 1 where the expansion is stated to be valid and where the poles contribute significantly in Eq. (8). The authors should demonstrate that the predictions are robust beyond that domain, or explicitly delimit the range of d, frequency, and dipole orientation for which Eq. (14) is quantitatively accurate. Without such a check, the two-to-three-orders-of-magnitude enhancement claim rests on an uncontrolled approximation.","section":"§III, §IV, Eqs. (9), (13), (14)"},{"comment":"The entire derivation is built on the confinement-induced nonlocal plasma frequency of Eq. (1)/(10), which is cited from prior work ([30]–[32]) without derivation. Given that the paper's new claims reduce to this expression, and given the inconsistency described above between the printed Eq. (1) and Eq. (10), the authors should provide a concise derivation of the reciprocal form from the Keldysh-Rytova potential, or at least state explicitly the approximations under which it holds at few-nanometer thicknesses. This is needed for readers to judge whether the model applies in the regime where the predicted splitting and decay enhancement are largest.","section":"§II, Eq. (1); §III, Eq. (10)"}],"minor_comments":[{"comment":"In the Concluding Remarks, 'transdimentional' should be 'transdimensional'.","section":"§V"},{"comment":"The black dotted lines are described as 'the real parts of the approximate modes of Eq. (9)', but the text also refers to Eq. (11) in connection with the ultra-thin limit; the authors should specify which approximate expression is used in the figure.","section":"Fig. 1(b)"},{"comment":"The parameters used for the figures (ωp = 2.79 eV, ε = 7.8, u = 0.65, δ = 0.01) are said to be typical of NV centers near a TiN surface, but the text does not explain how u and δ are obtained from the cited experiments; a brief justification would improve reproducibility.","section":"§IV"},{"comment":"The text states that only the p-evanescent coefficient rp_2− can have poles on the real axis; the reasoning for excluding s-polarized poles and propagating-wave poles is compressed. A sentence making the argument explicit would aid the reader.","section":"§III, Eq. (5)-(6)"}],"recommendation":"major_revision","confidential_remarks":"The key issue is the inconsistency between Eq. (1) and Eq. (10). It appears to be a typographical error in the direction of the plasma-frequency renormalization, but its correction is essential because Eq. (10) and all subsequent formulas use the reciprocal form. I recommend asking the authors to confirm the intended form, correct Eq. (1) accordingly, and recheck the downstream derivation and figures. The approximation analysis for Eq. (14) should also be strengthened, as the x+ branch can leave the formally valid domain x<1 for the parameters shown. The model itself is self-cited from the authors' prior work; independent validation of Eq. (1) at the relevant thicknesses would increase confidence."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this paper has a real new result and a load-bearing typo. The intended physics is recoverable, but the printed Eq. (1) contradicts the equation the rest of the derivation actually uses. Once that is fixed, the central argument holds up; until then, the paper is not reproducible as written.\n\nWhat is new: the application of the Keldysh-Rytova nonlocal Drude response to ENZ modes in ultrathin films, predicting degeneracy lifting of the two ENZ modes and a biexponential distance dependence of spontaneous decay with two-to-three orders of magnitude enhancement. Equations (9), (11), and (14) are new relative to the local-Drude ENZ papers [12,13], and the derivation through the QED Green tensor and pole analysis is clearly presented. Figure 1 shows the approximate modes track the full reflection coefficient well. That is honest, solid theoretical work.\n\nWhere it gets soft. First, the stress-test note is correct. Eq. (1) as printed gives a blue shift with decreasing d and does not produce Eq. (10); the reciprocal form reproduces Eq. (10) and the stated sqrt(k) 2D dispersion. Since Eq. (1) is the microscopic input from which everything else follows, this is not a cosmetic typo. Second, that input is cited to the authors' previous work [30] and not re-derived here. Self-citation is not automatically a flaw, and Eq. (3) has some empirical anchor in TiN measurements, but the new quantitative claim of up to 1000x enhancement rests on one parameter set with no sensitivity analysis. Third, the Lorentzian approximation assumes Im x+/-, / Re x+/- << 1, but no error estimate is given.\n\nBottom line: the reader's conditional verdict is fair, and I would make the condition more concrete. This deserves peer review, but the reviewers should require a corrected Eq. (1), a sensitivity analysis, and ideally a comparison to existing ENZ emission measurements.","headline":"A plausible and useful extension of ENZ thin-film physics, but Eq. (1) as printed is wrong and makes the paper unreproducible until corrected.","tokens_in":11678,"tokens_out":6102,"would_cite":false,"duration_ms":56074,"reading_group":"maybe","serious_thinker":"no","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"In transdimensional metal films, thickness alone splits the epsilon-near-zero plasmon mode and raises emitter decay rates up to 1000-fold.","keywords":["transdimensional plasmonics","epsilon-near-zero modes","nonlocal dielectric response","Keldysh-Rytova potential","spontaneous decay enhancement","ultrathin metallic films","confinement-induced nonlocality","surface plasmon polaritons"],"falsifier":"Measure the spontaneous-emission rate of an ensemble of identical emitters, such as nitrogen-vacancy centers, as a function of distance from an atomically smooth silver or titanium-nitride film whose thickness is varied from about 5 nm to 100 nm. If the decay curve is a single exponential rather than a sum of two exponentials, or if the enhancement does not rise steeply as the film thins, then Eq. (1) cannot be the correct low-frequency response.","tokens_in":1626,"feed_emoji":"⚛️","tokens_out":2598,"duration_ms":77169,"temperature":0.7,"pith_summary":"This paper argues that in ultrathin, few-monolayer metallic films the plasma frequency is not a fixed material constant but depends on film thickness and in-plane momentum, because vertical confinement changes the way electrons interact. Using a nonlocal Drude dielectric response built on that thickness-dependent plasma frequency, the authors show that the epsilon-near-zero surface-plasmon mode of a thin film splits into two distinct branches as the film becomes transdimensional. They then show that a quantum emitter placed near the film couples to both split branches, producing a spontaneous-decay rate whose distance dependence is a sum of two exponentials and whose magnitude is two to three orders of magnitude above the free-space rate. The consequence is that film thickness alone becomes a control knob for light-matter interaction in plasmonic nanostructures.","feed_headline":"Ultrathin metal films split plasmons and boost emitter decay 1,000x","feed_subtitle":"Two split epsilon-near-zero modes give a biexponential distance law with rates 100–1000 times free space.","key_machinery":"The load-bearing object is the confinement-induced nonlocal Drude dielectric response, Eq. (2), with the momentum- and thickness-dependent plasma frequency of Eq. (1). The physical origin is the Keldysh-Rytova electron interaction potential: in a film only a few monolayers thick, the field outside the film makes electron-electron repulsion stronger than in bulk, shifting the plasma frequency to the red and giving it $\\sqrt{k}$ spatial dispersion. This single substitution changes the pole structure of the Fresnel reflection coefficient: a zero of the dielectric function combined with finite thickness creates two ENZ modes, Eq. (9), whose analytic ultrathin forms are Eq. (11). A Lorentzian approximation to the spectral function then converts the Green tensor into the biexponential expression of Eq. (14), which directly produces the predicted decay enhancement.","core_discovery":"The paper's central claim is that confinement-induced nonlocality, not just geometry, governs the optical response of ultrathin metal films. In the model, vertical confinement of electrons turns the ordinary Coulomb potential into the Keldysh-Rytova electron interaction potential, making the in-plane plasma frequency $\\omega_p(k)=\\omega_p^{3D}\\sqrt{1+(\\varepsilon_1+\\varepsilon_2)/\\varepsilon k d}$ depend on thickness $d$ and momentum $k$. Inserting this plasma frequency into the Drude dielectric function and locating the poles of the p-polarized reflection coefficient yields two epsilon-near-zero (ENZ) branches, $x_+(d)$ and $x_-(d)$, whose degeneracy is lifted as $d$ shrinks. The upper branch reproduces the long-range plasmon of earlier local-Drude analyses, while the lower branch is the new transdimensional feature. A dipole emitter above the film couples to both branches, so the spontaneous-decay rate contains two exponentials with opposite thickness dependences; for a 10 nm film the enhancement reaches over three orders of magnitude. The authors present this as a generalization of the classic Drexhage emitter-near-interface problem to the transdimensional regime.","pith_inferences":["If Eq. (1) survives scrutiny at few-nanometer thickness, the same nonlocal dielectric response should also modify other near-field observables, such as resonant energy transfer between emitters or forces between two ultrathin films; the paper does not compute these.","One testable extension is a film patterned with a thickness gradient: the two split branches would shift along the film, potentially acting as a nanoscale frequency-selective coupler or a way to separate emitters by transition frequency in space.","The biexponential decay law itself could serve as a metrology tool: fitting the two decay lengths at fixed thickness would extract the effective plasma frequency and the dielectric environment, quantities that are otherwise hard to measure directly."],"forward_implications":["Film thickness becomes a practical tuning parameter for the epsilon-near-zero frequency and for spontaneous emission rates, supplementing or replacing material choice.","Ultrathin films support two coexisting surface-plasmon branches below the plasma frequency, and the lower branch keeps the enhanced decay over emitter-film distances much larger than the short-range branch.","Perpendicular (z-oriented) dipoles decay faster than parallel dipoles, consistent with the mirror-charge picture of an emitter near a conducting plane.","Above the plasma frequency, the usual half-wavelength film modes disappear in ultrathin films and are replaced by a sharp, nearly flat long-range mode.","As the film thickens toward the local-Drude limit, the mode splitting and the biexponential signature vanish, recovering conventional thin-film optics."],"supporting_citations":[{"why":"supplies Eq. (1), the confinement-induced nonlocal plasma frequency on which the entire pole analysis and decay calculation rest.","marker":"[30]"},{"why":"introduces the Keldysh-Rytova electron interaction potential that gives the enhanced in-plane Coulomb interaction in confined films.","marker":"[33]"},{"why":"previous application of the Keldysh-Rytova dielectric response to transdimensional plasmonic films, establishing the model used here.","marker":"[31]"},{"why":"earlier local-Drude analysis of ENZ modes in thin films that this work extends and generalizes.","marker":"[12]"},{"why":"provides the medium-assisted QED formula for the spontaneous decay rate used in Eq. (4).","marker":"[38]"},{"why":"provides the diagonalized scattering Green tensor and reflection-coefficient decomposition used in Eq. (5).","marker":"[39]"},{"why":"the classic emitter-near-interface experiment that the paper claims to generalize to transdimensional films.","marker":"[34]"},{"why":"experimental data on thickness- and temperature-dependent plasma frequency of TiN films used to support the thermal average of Eq. (1).","marker":"[37]"}],"fun_headline_variants":["Nonlocal optics split ENZ modes, amplify decay 1000x","Ultrathin films split plasmons, boost emitter decay 1000x","Transdimensional ENZ modes lift degeneracy, boost decay","Confined electrons create split ENZ modes, speed up decay"],"cache_read_input_tokens":13696,"weakest_assumption_plain":"The entire chain, from mode splitting to the two-to-three-order decay enhancement, rests on Eq. (1), the confinement-induced plasma frequency $\\omega_p(k)=\\omega_p^{3D}\\sqrt{1+(\\varepsilon_1+\\varepsilon_2)/\\varepsilon k d}$, which is taken from the authors' earlier work and not re-derived here; if that formula is wrong or inapplicable for few-nanometer films, the predicted effects disappear.","fun_headline_variants_meta":{"raw":{"variants":["Nonlocal optics split ENZ modes, amplify decay 1000x","Ultrathin films split plasmons, boost emitter decay 1000x","Transdimensional ENZ modes lift degeneracy, boost decay","Confined electrons create split ENZ modes, speed up decay"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00022,"raw_usage":{"total_tokens":1404,"prompt_tokens":863,"completion_tokens":541,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":479,"completion_tokens_details":{"reasoning_tokens":465}},"tokens_in":479,"tokens_out":541,"duration_ms":5325,"temperature":1.0,"reasoning_tokens":465,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T15:41:05.606705+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the spontaneous-emission rate of an ensemble of identical emitters, such as nitrogen-vacancy centers, as a function of distance from an atomically smooth silver or titanium-nitride film whose thickness is varied from about 5 nm to 100 nm. If the decay curve is a single exponential rather than a sum of two exponentials, or if the enhancement does not rise steeply as the film thins, then Eq. (1) cannot be the correct low-frequency response.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"supplies Eq. (1), the confinement-induced nonlocal plasma frequency on which the entire pole analysis and decay calculation rest."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"introduces the Keldysh-Rytova electron interaction potential that gives the enhanced in-plane Coulomb interaction in confined films."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"previous application of the Keldysh-Rytova dielectric response to transdimensional plasmonic films, establishing the model used here."},{"cited_title":"Campione, I","cited_arxiv_id":null,"evidence_quote":"earlier local-Drude analysis of ENZ modes in thin films that this work extends and generalizes."},{"cited_title":"Vogel and D.-G","cited_arxiv_id":null,"evidence_quote":"provides the medium-assisted QED formula for the spontaneous decay rate used in Eq. (4)."},{"cited_title":"Sambale, D.-G","cited_arxiv_id":null,"evidence_quote":"provides the diagonalized scattering Green tensor and reflection-coefficient decomposition used in Eq. (5)."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"the classic emitter-near-interface experiment that the paper claims to generalize to transdimensional films."},{"cited_title":"Vertchenko, L","cited_arxiv_id":null,"evidence_quote":"experimental data on thickness- and temperature-dependent plasma frequency of TiN films used to support the thermal average of Eq. (1)."}],"review_version":1}