{"id":"44bae81a-4151-436d-b021-8fc97fa7ce59","arxiv_id":"2411.15679","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"Lithium-doped NdNiO3 thin films switch from a lossy metal to a transparent insulator with a band gap of about 3 to 4 eV, enabling electrically tunable optics.","lead":"Researchers measured how lithium doping changes the optical properties of thin films of the quantum material NdNiO3. The doped films become transparent insulators with a large electronic band gap, suggesting new electrically tunable optical devices.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The reported Li-doped NdNiO3 optical constants rest on a sharp two-layer approximation to a measured Li gradient, leaving the quantitative n, kappa, and 3–4 eV gap not uniquely determined.","rationale":"The paper's central claim is that variable-angle spectroscopic ellipsometry, analyzed with a two-layer model, gives accurate complex refractive indices for Li-doped NdNiO3 and a 3–4 eV band gap. The qualitative picture is plausible and independently supported by prior SmNiO3 experiments and by theory, and the authors made a reasonable effort to validate their structural and doping characterization with XRD, RBS, and ToF-SIMS. The residual risk is quantitative and concerns the inversion model: the Li concentration profile is graded, but the model is a sharp bilayer, and the intermediate-doping regime is known from the cited literature to have a much smaller gap than the fully doped state. Since the fitted top-layer thickness and optical constants are correlated, the sharp-interface assumption can bias the extracted kappa and the Tauc gap. The consistency between the Nb:STO and LAO samples is reassuring but does not rule out a systematic bias because both samples are analyzed with the same two-layer approximation. This is exactly the weakest assumption identified by the reader, and a graded-model refit is a concrete, feasible check that would settle whether the reported numbers are robust. The verdict therefore remains CONDITIONAL/UNCHANGED: the paper is not fatally flawed, but the central numbers should be treated as provisional pending a graded-profile analysis.","tokens_in":13663,"tokens_out":6730,"duration_ms":67814,"concrete_test":"Calibrate the ToF-SIMS depth axis by measuring the SIMS crater depth with profilometry, then convert the Li and Sr count profiles into a physical Li concentration profile. Re-fit the published Ψ and Δ spectra (Fig. 2c for Nb:STO and Fig. 3e for LAO) with a graded multilayer: either a B-spline n(z) profile constrained by the calibrated Li profile or a Bruggeman effective-medium mixture of undoped and fully doped NdNiO3 whose filling fraction follows the SIMS profile. Compare the extracted top-layer n, kappa, and Tauc gap against the two-layer result. If kappa in the 0.4–2.5 µm range changes by more than about 0.05, or the Tauc gap by more than about 0.5 eV, the reported constants depend on the abrupt-interface assumption.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The load-bearing step is the conversion of ellipsometric data into optical constants of the Li-doped phase using a two-layer model (top fully doped insulator / bottom undoped metal). ToF-SIMS shows an exponentially decaying Li concentration, not a step profile, and the SIMS data are counts versus sputter time, so they do not provide an independent depth calibration for the fitted 20.7-nm top layer on Nb:STO. Reference 23 indicates that intermediate H-doping levels in SmNiO3 have gaps of only about 1.4 eV, so insufficiently doped material is not optically equivalent to undoped metallic NdNiO3. Because the top-layer thickness and refractive index are fitted simultaneously, the sharp-interface model can partially absorb the missing intermediate-doping region into the fitted n and kappa. Agreement between Nb:STO and LAO samples is helpful, but both fits use the same two-layer approximation, so it does not isolate this systematic error. No uncertainty or uniqueness analysis is reported.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports synthesis of crystalline NdNiO3 films, reversible Li-ion intercalation from a liquid electrolyte, and variable-angle spectroscopic ellipsometry (VASE) characterization of both undoped and Li-doped regions over 0.3-2.5 um wavelength. Because ToF-SIMS shows a Li concentration gradient, the authors use a two-layer model (fully doped top layer / undoped or lightly doped bottom layer) to extract the complex refractive indices of the two phases. They report that the undoped film is metallic, while the Li-doped insulator has n near 2, kappa below 0.1 in the visible and below 0.02 in the near-infrared, and Tauc analysis yields an optical gap of roughly 3-4 eV. The same qualitative behavior is reproduced on two substrates (Nb:STO and LAO).","tokens_in":13896,"tokens_out":6147,"duration_ms":59057,"significance":"The qualitative finding is credible and potentially valuable: electron doping of NdNiO3 produces a transparent insulating state, and the reported index contrast between the metallic and insulating phases is large. The paper is transparent about the doping gradient and provides detailed oscillator parameters in Table S1, which aids reproducibility. If the quantitative n, kappa, and band-gap values hold, this work demonstrates electrically controlled, non-volatile, low-loss refractive-index tuning in a perovskite nickelate. However, the central quantitative claims rest on the two-layer ellipsometric model and on a single-Lorentz-oscillator representation of the doped layer, so the uncertainty in those values needs to be addressed before the results can be taken at face value.","major_comments":[{"comment":"The two-layer model for the Nb:STO sample fixes the bottom layer as undoped NdNiO3 and fits the top layer as fully doped, while the ToF-SIMS profile in Fig. 2e shows an exponential Li decay. The paper itself acknowledges in the Fig. S1 note that intermediate doping states are not modeled. Because Ref. 23 reports a gap of only about 1.4 eV at half doping in H-doped SmNiO3, partially doped material is not optically equivalent to undoped metallic NdNiO3. Consequently the fitted top-layer n and kappa, and the Tauc gap derived from them, can partially absorb the missing intermediate-doping region. Please quantify this systematic error by fitting a graded-layer or multi-layer profile (e.g., 3-5 layers with intermediate oscillator sets) and reporting the resulting range of n, kappa, and gap values.","section":"Section 2 (Fig. 2e and Note S1)"},{"comment":"The doped layer is represented by a single Lorentz oscillator, and the Tauc plots in Fig. 4 are computed from the absorption coefficient of that same fitted oscillator. A Lorentzian model imposes a specific absorption tail, so the linear extrapolation in the (alpha h nu)^0.5 plot may be largely determined by the oscillator center energy and broadening rather than by an independently measured electronic band edge. Please demonstrate robustness of the 2.9-3.2 eV (indirect) and 3.7-4.3 eV (direct) estimates by refitting with at least one alternative dispersion model (e.g., a Tauc-Lorentz model or a numerical inversion with Kramers-Kronig consistency) and by showing how much the extracted intercept changes.","section":"Table S1 and Fig. 4"},{"comment":"No figure-of-merit or confidence intervals are reported for the ellipsometric fits. The top-layer thickness and its optical constants are fitted simultaneously for the Nb:STO sample, and the LAO analysis only varies the bottom-layer thickness from 1 to 5 nm. The consistency between Nb:STO and LAO is a useful internal check, but both samples are analyzed with the same two-layer approximation, so it does not isolate the systematic model error. Please provide typical ellipsometric MSE values, parameter correlation information, or an explicit covariance estimate for the reported n and kappa.","section":"Sections 2 and 3 (Note S1)"}],"minor_comments":[{"comment":"The panel label reads \"60 s doping,\" but the text and Fig. S1b describe samples doped for 120 and 180 s; please correct the label.","section":"Figure S1c"},{"comment":"The legend entries \"Ref. [23] [23][16]\" are difficult to parse; please separate the references and align them clearly with the plotted symbols.","section":"Figure 4c"},{"comment":"The LAO substrate oscillator parameters list Br in eV while also giving wavenumbers in parentheses; please state the conversion used and keep units consistent throughout the table.","section":"Note S1 and Table S1"},{"comment":"The text says the single-layer model shows good agreement with the data, but no quantitative fit-quality metric is given; please add the MSE or equivalent for the undoped fits as well.","section":"Page 5, undoped NdNiO3 fitting"},{"comment":"The argument that the material differs from Co3O4 and CoxFe3-xO4 because absorption is substantial only below 400 nm is not by itself a test of whether the relevant transition is from the valence band maximum to the conduction band minimum; please support this claim with a comparison to the calculated band structure or at least state it as a model assumption.","section":"Tauc plot discussion (Section 4)"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is likely publishable in this journal if the authors can resolve the model-dependence issues. The novelty claim (first optical constants of both undoped and Li-doped NdNiO3) is plausible, and the qualitative metal-to-insulator transition is well supported by the transparency change on LAO and by the ellipsometric contrast. My main concern is quantitative reliability: the reported n, kappa, and 3-4 eV band gap are all derived from a sharp two-layer model with a single Lorentz oscillator, and no uncertainty analysis is provided. The same two-layer approximation is used for both substrates, so the internal consistency does not remove the systematic bias. I would like to see a graded-profile analysis and a robustness check of the Tauc intercept before accepting the numbers as quantitative."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Bottom line: this is a solid, useful measurement paper that does what it claims—first complex refractive indices of undoped and Li-doped NdNiO3 across 0.3–2.5 µm—and the qualitative finding (metal becomes a transparent insulator with n~2 and κ<0.1 in the visible) is convincing. The two-layer ellipsometric model is a reasonable first-cut treatment of a real Li gradient, and the authors are unusually candid about its limits in the SI. Still, the quantitative n, κ, and the 3–4 eV Tauc gap ride on the sharp-interface assumption. ToF-SIMS shows an exponential Li decay, not a step, and the intermediate-doping region is known to have a much smaller gap (~1.4 eV), so the fitted top-layer constants can absorb some of that transition region. The single-Lorentz oscillator for the doped layer also puts a strong prior on the absorption edge, which makes the Tauc extrapolation less clean than it looks. On the plus side, the agreement between Nb:STO and LAO substrates, the visible transparency photo, and the consistency with prior SmNiO3 results all support the central claim. The paper would be stronger with a sensitivity analysis (e.g., a graded-layer model or multi-oscillator fits) and explicit uncertainty bars on the extracted optical constants. Those are addressable, not fatal. I'd send this to a serious referee—it's exactly the kind of measurement the tunable-photonics community needs—and ask for the extra analysis in revision. Worth citing for the dataset when it appears.","headline":"First measured optical constants of electron-doped NdNiO3, with a caveat on the two-layer model; worth a serious referee.","tokens_in":14450,"tokens_out":2824,"would_cite":true,"duration_ms":26148,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Electron doping switches NdNiO3 from a metal to a 3–4 eV band-gap insulator with low optical loss.","keywords":["NdNiO3 thin films","lithium intercalation","electron doping","spectroscopic ellipsometry","complex refractive index","metal-insulator transition","optical band gap","tunable photonics"],"falsifier":"Intercalate a film thin enough, roughly below 10 nanometers, that ToF-SIMS shows lithium throughout the full thickness, then measure its ellipsometric Psi and $\\Delta$ and fit n and kappa directly with a single layer; if those optical constants differ from the two-layer values reported here, the abrupt-interface assumption is the source of bias. Equivalently, fit the existing data with a graded-index multilayer constrained to the measured SIMS profile and check whether the fit improves and the top-layer constants shift.","tokens_in":13474,"feed_emoji":"🔬","tokens_out":8169,"duration_ms":72931,"temperature":0.7,"pith_summary":"This paper reports the first measurements, to the authors' knowledge, of the complex refractive indices of undoped and Li-doped NdNiO3 films across 0.3 to 2.5 micrometers. It claims that electron doping by field-driven lithium intercalation turns NdNiO3 from a metal with no optical gap into an insulator with a band gap of roughly 3 to 4 eV, a refractive index of about 2, and an extinction coefficient below 0.1 in the visible and below 0.02 in the near infrared. The importance is that this large optical change is controlled by an electric field rather than by heat, making NdNiO3 a candidate for tunable, low-loss optical components. The paper also demonstrates an analysis method for inhomogeneously doped films, using a two-layer model justified by depth-resolved ion measurements.","feed_headline":"Lithium doping turns NdNiO3 metal into a low-loss insulator","feed_subtitle":"Doped films show refractive index ~2 and near-zero absorption from visible to infrared.","key_machinery":"The central mechanism is a two-layer ellipsometric model: a top fully Li-doped insulating layer with an opened band gap sits over a bottom undoped metallic layer, and the pair is fit to variable-angle spectroscopic ellipsometry data using Drude and Lorentz or Gaussian oscillators. Depth-resolved ToF-SIMS measurements showing an exponentially decaying lithium concentration justify the model, since prior first-principles work indicates the gap opens sharply only near full electron doping. The model is what extracts the refractive index and extinction coefficient of each phase together with the layer thicknesses, and Tauc plots of the extracted absorption are then used to estimate the band gap.","core_discovery":"The paper's central claim is that field-driven lithium intercalation switches nanoscale NdNiO3 from a metallic state with no optical band gap into an insulating state with a gap of roughly 3 to 4 eV, and that this switch produces a large, low-loss change in the complex refractive index across 0.3 to 2.5 micrometers. From ellipsometric fits, the undoped phase is metallic, with the extinction coefficient rising with wavelength, while the Li-doped phase has $n \\approx 2$, $\\kappa < 0.1$ from 0.4 to 0.7 micrometers, and $\\kappa < 0.02$ from 0.7 to 2.5 micrometers. Because time-of-flight secondary-ion mass spectrometry shows that the lithium profile is a gradient, the authors extract these values with a two-layer model rather than assuming uniform doping, and they verify the doped film's transparency visually on transparent LaAlO3 substrates. Tauc analysis places the gap at 2.9 to 3.2 eV assuming an indirect transition or 3.7 to 4.3 eV assuming a direct transition, which the paper summarizes as roughly 3 to 4 eV.","pith_inferences":["Editorial inference: if the central claim holds, a partially doped film is effectively a graded-index layer, so controlling intercalation time or voltage could yield continuously tunable refractive index rather than only the two endpoint states.","Editorial inference: the combination of large refractive-index contrast and $\\kappa < 0.02$ in the near infrared suggests a testable extension to electrically switchable phase shifters or tunable resonators, for example by measuring the resonance shift of a doped NdNiO3 metasurface.","Editorial inference: the reproducible exponential lithium profile could itself be treated as a programmable gradient-index coating, with its depth shape set by doping time and voltage.","Editorial inference: the gap estimate depends on choosing an indirect or direct Tauc exponent, so a measurement on a uniformly doped film using angle-resolved photoemission or direct absorption would settle which assignment is correct."],"forward_implications":["If the extracted constants are correct, NdNiO3 becomes a visible and near-infrared optical phase-change material whose refractive index changes substantially on lithium intercalation while the doped phase stays low-loss.","The reported gap of roughly 3 to 4 eV is consistent with several first-principles predictions for hydrogen- or lithium-doped rare-earth nickelates, suggesting the electronic mechanism is not specific to one ionic species.","Because the doping is field-driven and persists at room temperature, NdNiO3-based optical devices could be tuned electrically without heating and without continuous power.","The two-layer analysis strategy should extend to other nanoscale materials with inhomogeneous ion or electron concentration profiles, where uniform-doping fits would fail.","The low extinction coefficient of the doped phase across the near infrared opens a route to electrically switchable optical components at telecommunication wavelengths."],"supporting_citations":[{"why":"Supplies the prior demonstration of colossal resistance switching and band-gap modulation in nickelates by electron doping, which this work extends to optical constants.","marker":"[14]"},{"why":"Provides first-principles predictions of lithium-induced metal-insulator transitions in rare-earth nickelates used to motivate the expected gap.","marker":"[15]"},{"why":"Provides first-principles band gaps for hydrogen-doped nickelates, one of the theoretical datasets compared with the measured 3 to 4 eV gap.","marker":"[16]"},{"why":"Establishes lithium-ion intercalation as an electrical doping route in strongly correlated perovskites.","marker":"[17]"},{"why":"Prior ellipsometry of doped SmNiO3 that this work complements by using a doping-profile-aware model for NdNiO3.","marker":"[18]"},{"why":"Supplies the spectroscopic ellipsometry formalism used to fit the measured Psi and Delta spectra.","marker":"[22]"},{"why":"Shows the band gap opens abruptly only near full electron doping, which is the justification for treating the doped film as two layers.","marker":"[23]"},{"why":"Provides the Tauc plot method and its known limitations, from which the roughly 3 to 4 eV gap is estimated.","marker":"[24]"}],"fun_headline_variants":["Lithium doping flips NdNiO3 from metal to a 3-4 eV bandgap insulator","Electron doping switches NdNiO3 to a transparent insulator","Lithium ion gating opens a 3-4 eV optical gap in NdNiO3","NdNiO3 films become transparent insulators via electron doping","Field-driven lithium turns NdNiO3 metal into a low-loss window"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing assumption is that the exponential lithium concentration gradient can be modeled as two abrupt layers, a fully doped insulator on top and an undoped metal below, with no partially doped region in between; if intermediate doping states matter, the extracted refractive indices, extinction coefficients, and 3 to 4 eV band gap could be biased.","fun_headline_variants_meta":{"raw":{"variants":["Lithium doping flips NdNiO3 from metal to a 3-4 eV bandgap insulator","Electron doping switches NdNiO3 to a transparent insulator","Lithium ion gating opens a 3-4 eV optical gap in NdNiO3","NdNiO3 films become transparent insulators via electron doping","Field-driven lithium turns NdNiO3 metal into a low-loss window"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000786,"raw_usage":{"total_tokens":3475,"prompt_tokens":961,"completion_tokens":2514,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":577,"completion_tokens_details":{"reasoning_tokens":2410}},"tokens_in":577,"tokens_out":2514,"duration_ms":16092,"temperature":1.0,"reasoning_tokens":2410,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T14:00:49.860194+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Intercalate a film thin enough, roughly below 10 nanometers, that ToF-SIMS shows lithium throughout the full thickness, then measure its ellipsometric Psi and $\\Delta$ and fit n and kappa directly with a single layer; if those optical constants differ from the two-layer values reported here, the abrupt-interface assumption is the source of bias. Equivalently, fit the existing data with a graded-index multilayer constrained to the measured SIMS profile and check whether the fit improves and the top-layer constants shift.","supporting_citations":[],"review_version":1}