{"id":"4b6888a1-9890-4811-8299-c3c755617486","arxiv_id":"2608.10580","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"high","formal_verification":"none","parameter_count":3,"one_line_summary":"Intercalating bilayer indium between graphene and SiC enhances dielectric screening to a claimed effective substrate constant around 600, an order of magnitude above prior intercalants, though the exact value is undermined by internal inconsistencies.","lead":"By sliding a two-layer film of indium between graphene and its silicon carbide support, researchers boosted the material's dielectric screening far beyond earlier attempts. The study uses photoemission to read out graphene's plasmaron signature and shows the second indium layer, a nearly free-electron metal, does the heavy lifting.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The reported αG=0.0089 is inconsistent with the stated calibration δE=αG^p: using the measured δE=0.0557 gives αG≈0.0045, and the 2ML point lies far outside the fitted range, so εs=622±49 is not supported.","rationale":"The reader's weakest_assumption pinpoints the same load-bearing issue: the αG extraction is internally inconsistent and rests on an unvalidated extrapolation. I verified the arithmetic explicitly: with p=0.534, δE=0.0557 yields αG≈0.0045, not 0.0089, and the reverse calculation gives δE≈0.080. This is a concrete, checkable discrepancy that directly undermines the paper's quantitative central claim. The qualitative conclusion—that a second In layer enhances screening—is supported by the direct ARPES comparison (δE drops from 0.17 to 0.056) and by the DFT layer-resolved mechanism, so I do not recommend rejecting the paper. However, the quantitative values of αG and εs cannot be trusted as reported; they need correction and a properly propagated uncertainty. The reader's verdict of CONDITIONAL is appropriate: the paper should be accepted only after the calibration analysis is corrected and the extrapolation is either justified or the claims are softened. I see no additional concern that would move the verdict further, and I do not find the qualitative physics internally inconsistent. Hence the verdict remains unchanged.","tokens_in":14534,"tokens_out":4207,"duration_ms":38974,"concrete_test":"Digitize the calibration data in Fig. 4b of Ref. [34] and refit with a two-parameter power law δE = A·αG^p, including parameter covariances. Then re-extract αG from the experimental δE=0.0557±0.0022 and propagate the full fit uncertainty. If the best-fit A is significantly different from 1, the published αG is simply wrong and εs must be recomputed. To test the extrapolation itself, compute G0W0-RPA δE for additional low-αG values (e.g., 0.01, 0.02, 0.03) using the same method as Ref. [34]; if the computed points deviate from the power-law extrapolation, the quoted εs is unreliable regardless of the prefactor.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The quantitative headline—αG=0.0089±0.0007 and εs=622±49—depends entirely on the empirical calibration δE(αG)=αG^p with p=0.534±0.017 (Fig. 4b). This stated functional form does not reproduce the reported values. Inserting the measured δE=0.0557±0.0022 gives αG=(0.0557)^(1/0.534)≈0.0045, a factor of two below the quoted 0.0089. Conversely, αG=0.0089 with the same formula predicts δE≈0.080, not the measured 0.0557. To match both numbers one needs an unstated prefactor A≈0.69 in δE=A·αG^p; the paper does not report or justify such a prefactor. Independently of this arithmetic slip, the 2ML In point (αG≈0.009) is roughly a factor of 5–10 below the lowest calibration point of Ref. [34] (αG≈0.05), so the extraction is a long extrapolation of a fitted power law whose functional form the authors themselves state is not known a priori. The quoted uncertainty on αG and εs reflects only the experimental δE error and ignores the calibration-form and parameter covariance. Because the abstract and conclusions present εs=622±49 as the central quantitative demonstration of 'unusual strong screening,' this inconsistency is load-bearing: if the stated calibration is used correctly, the quantitative claim changes substantially, and even a correctly implemented extrapolation risks large unquantified systematic error. The qualitative trend—2ML In gives much smaller δE than 1ML In—is supported by the raw spectra, but the specific numbers are not.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports an ARPES and DFT study of graphene on SiC(0001) with intercalated bilayer indium. It finds that the first In layer buffers the substrate and the second In layer hosts nearly free-electron bands, and it uses the energy separation between the hole and plasmaron bands (δE = 0.0557 ± 0.0022) to extract an effective graphene coupling constant α_G = 0.0089 ± 0.0007 and substrate dielectric screening ε_s = 622 ± 49 for 2ML In, compared with α_G = 0.07 and ε_s = 56.6 for 1ML In. The authors conclude that 2ML In intercalation is a powerful route to engineer the dielectric environment of epitaxial graphene. The paper also reports a large Rashba splitting of the first In layer and a nearly free-electron band in the second layer.","tokens_in":14913,"tokens_out":6618,"duration_ms":55115,"significance":"The qualitative finding—that bilayer In screening is much stronger than monolayer In or other intercalants—is potentially valuable for the graphene-on-SiC platform, and the combination of ARPES, STEM, and layer-resolved DFT is appropriate. The use of an external G0W0-RPA calibration (Ref. [34]) is a reasonable strategy, and the raw δE comparison (0.17 vs 0.056) is a clear falsifiable statement. However, the central quantitative claim (α_G and ε_s) is not supported by the calibration as stated; the reported numbers are internally inconsistent with the fit formula, and the extrapolation is far outside the calibrated range. With the quantitative headline corrected or substantially qualified, the remaining qualitative and structural conclusions would still be of interest, but the paper in its current form overstates the precision of its main result.","major_comments":[{"comment":"The stated calibration does not reproduce the quoted α_G. The text gives δE(α_G) = α_G^p with p = 0.534 ± 0.017; inserting the measured δE = 0.0557 ± 0.0022 yields α_G = (0.0557)^(1/0.534) ≈ 0.0045, not 0.0089 ± 0.0007. Conversely, α_G = 0.0089 predicts δE ≈ 0.080, well outside the experimental value. The same discrepancy appears for the 1ML In entry in Table I: δE = 0.17 with α_G = 0.07 would require δE ≈ 0.24 under the stated power law. This inconsistency is load-bearing because the abstract and conclusions base the 'unusual strong screening' claim on ε_s = 622 ± 49. The authors need to supply the actual fit (including any prefactor), show the calibration points, and recompute all derived quantities.","section":"Fig. 4b and the 'Graphene effective coupling constant' section"},{"comment":"The 2ML In point sits approximately an order of magnitude below the lowest calibration point of Ref. [34] (α_G ≈ 0.05), and the text itself notes that the functional form of δE(α_G) is not known a priori. The extrapolation of a fitted power law over this range is therefore a major source of systematic uncertainty, and the quoted ±0.0007 reflects only the experimental δE uncertainty. The manuscript should either restrict itself to the qualitative claim (2ML In has smaller δE than 1ML In) or provide an uncertainty budget that includes the calibration-form and parameter-covariance contributions.","section":"Fig. 4b extrapolation"},{"comment":"The conversion ε = e^2/(4πϵ0 α_G ℏ v_F) and ε_s ≈ 2ε − 1 requires a value of the graphene Fermi velocity and a relation between the effective coupling constant and the substrate dielectric constant. Neither the v_F value nor its uncertainty is stated; the reported ε = 312 ± 25 is numerically sensitive to v_F, and the approximation ε_s ≈ 2ε − 1 from Ref. [34] may not hold for a metallic bilayer intercalant. Please state the parameters used and test the sensitivity of ε_s to reasonable variations in v_F and to the ε_s(ε) relation.","section":"Conversion from α_G to ε_s"}],"minor_comments":[{"comment":"The figure caption labels panels 'a' and 'c' only, but the text refers to 'Fig. 4b' for the calibration plot; please fix the panel labeling.","section":"Figure 4 caption"},{"comment":"The description of the MDC fitting (Regions I–III) lacks the momentum and energy ranges, the number of spectra, and the goodness-of-fit values, so a reader cannot reproduce the δE extraction from the text alone.","section":"MDC fitting description"},{"comment":"Table I lists both δE and δk in the caption, but δk is never defined or used in the text; please remove it or define it.","section":"Table I caption"},{"comment":"The Data Availability section says 'To be published, WueData (2026)' rather than providing a repository identifier; a working link or DOI should be given if the data are meant to be openly available.","section":"Data Availability"}],"recommendation":"major_revision","confidential_remarks":"The central quantitative claim is internally inconsistent: the stated calibration formula does not reproduce the reported α_G and ε_s. I recommend asking the authors to redo the calibration analysis or remove the quantitative headline. The qualitative results—the 1ML vs 2ML comparison, the layer-resolved DFT picture, and the Rashba splitting—are solid and likely publishable after revision. Please also ask for the underlying fit data and a full uncertainty budget."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: the qualitative result is likely real, and the 1ML vs 2ML ARPES comparison is convincing. The headline numbers are not. The stated calibration function δE=α_G^p with p=0.534 does not reproduce α_G=0.0089 from δE=0.0557; plugging the numbers in gives α_G≈0.0045. Matching the paper's value requires an unstated prefactor around 0.69. That is load-bearing, because the whole quantitative screening claim hangs on this extraction.\n\nWhat is genuinely new: the bilayer-indium intercalation system, the layer-resolved mechanism where the first In layer buffers the SiC potential and the second becomes a nearly free-electron metal, and the experimental demonstration that the second layer changes the plasmaron splitting dramatically. The DFT band structure matches the ARPES for the In-derived states, and the Rashba splitting on layer 1 versus layer 2 is a nice direct visualization of the buffering effect. The δE drop from 0.17 to 0.056 is a solid qualitative statement that screening is much stronger with two layers.\n\nThe soft spots beyond the arithmetic: the 2ML point is roughly a factor of 5–10 below the lowest calibration point of Ref. [34], so even with a correct prefactor the extraction is a long extrapolation of an empirical power law whose form the authors admit is not known a priori. The uncertainty on α_G and ε_s only propagates the δE error and ignores the calibration-form covariance. The conversion from α_G to ε_s does not obviously close as stated; the reader gets something near 491 for ε_s, not 622, unless additional assumptions are hidden. Finally, the data availability statement says 'to be published' rather than providing an actual link; for a quantitative claim like this, that is not good enough.\n\nNone of this kills the central physical picture. The qualitative trend is not in serious doubt. But the abstract and conclusions present ε_s=622±49 as the demonstration of 'unusual strong screening,' and that specific number is not currently supported by the paper's own analysis.\n\nWho should read it: people working on graphene/SiC intercalation, dielectric environment engineering, and plasmaron spectroscopy. The calibration method discussion is also useful for the ARPES community.\n\nRecommendation: send to peer review. A good referee should require the authors to redo the extraction honestly, report the raw δE and the calibration range, and downgrade the quantitative claims to match what the data support. The qualitative finding is worth publishing after a major revision.","headline":"Qualitative screening enhancement from bilayer In intercalation looks real, but the headline εs=622±49 is not reproducible from the paper's own calibration formula.","tokens_in":15531,"tokens_out":2671,"would_cite":false,"duration_ms":23911,"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":"Intercalating a bilayer of indium under epitaxial graphene creates a strongly screening interface, with the substrate dielectric constant extracted as $\\epsilon_s = 622 \\pm 49$ from plasmaron band separations.","keywords":["graphene","dielectric screening","plasmaron","indium intercalation","angle-resolved photoemission spectroscopy","epitaxial graphene on SiC","nearly free-electron metal","many-body interactions"],"falsifier":"Calculate the graphene spectral function with G0W0-RPA at the reported effective coupling ($\\alpha_G \\approx 0.0089$, equivalently substrate screening $\\epsilon_s \\approx 622$) and at the experimental carrier density, and compare the predicted hole–plasmaron separation with the measured $\\delta E = 0.0557$; a prediction that does not match $\\delta E$ would show the calibration extrapolation, and hence the quoted $\\epsilon_s$, is unsupported.","tokens_in":14293,"feed_emoji":"🛡️","tokens_out":10644,"duration_ms":92271,"temperature":0.7,"pith_summary":"This paper argues that inserting two layers of indium between epitaxial graphene and its silicon-carbide substrate turns the interface into an unusually strong dielectric screen, addressing a well-known weakness of the graphene-on-SiC platform. Using the energy separation between graphene's hole band and its plasmaron satellite in angle-resolved photoemission as a proxy for electron–electron coupling, the paper extracts an effective coupling $\\alpha_G = 0.0089 \\pm 0.0007$ and a substrate dielectric constant $\\epsilon_s = 622 \\pm 49$, an order of magnitude larger than earlier intercalant systems. Layer-resolved density functional theory attributes the effect to a division of labor: the first indium layer buffers the substrate potential, while the second forms a nearly free-electron metal that screens the graphene above. Comparison with single-layer indium, which gives $\\epsilon_s \\approx 57$, supports the claim that the second layer is essential.","feed_headline":"Bilayer indium boosts graphene-on-SiC screening to 622","feed_subtitle":"Plasmaron-band measurements show one indium layer buffers SiC while a second, nearly free-electron layer screens the graphene.","key_machinery":"The central object is the plasmaron signature in ARPES: a satellite band created when a photoexcited hole propagates together with a plasmon, split off from the main hole band by an energy gap that grows with electron–electron coupling. The paper measures the normalized energy separation $\\delta E = (E_2 - E_1)/E_1$ at the Dirac point and uses it as a proxy for the effective graphene coupling $\\alpha_G$. The conversion is carried by an empirical power-law fit $\\delta E = \\alpha_G^p$, $p = 0.534$, calibrated to G0W0-RPA spectral-function calculations from the literature, and then by $\\epsilon_s \\approx 2\\epsilon - 1$ to a substrate dielectric constant. A second piece of machinery is layer-resolved DFT, which separates the first indium layer (Rashba-split buffer) from the second (nearly free-electron screening layer) and motivates why the bilayer, not the monolayer, screens so strongly.","core_discovery":"The paper's central claim is that the dielectric environment of graphene on SiC can be engineered by intercalating a bilayer of indium, and that this bilayer provides substrate screening $\\epsilon_s = 622 \\pm 49$, far beyond what previous intercalants achieve. The quantitative evidence is the normalized energy separation $\\delta E = 0.0557 \\pm 0.0022$ between the extrapolated hole and plasmaron bands at the Dirac point, measured by angle-resolved photoemission. Calibrating $\\delta E$ against G0W0-RPA spectral-function calculations through an empirical power law $\\delta E = \\alpha_G^p$ with $p = 0.534$ yields $\\alpha_G = 0.0089 \\pm 0.0007$, and the conversion $\\epsilon = e^2/(4\\pi \\epsilon_0 \\alpha_G \\hbar v_F)$ with $\\epsilon_s \\approx 2\\epsilon - 1$ gives $\\epsilon = 312 \\pm 25$. Layer-resolved DFT and ARPES identify the first indium layer as a buffer (Rashba splitting of 161 meV) and the second as a nearly free-electron metallic layer; graphene on a single indium layer shows much weaker screening ($\\epsilon_s = 56.6 \\pm 8.2$), confirming the second layer's essential role.","pith_inferences":["A natural transport check follows: if $\\epsilon_s \\approx 622$ is physically real, charged-impurity scattering in this heterostructure should be strongly suppressed, so a four-probe mobility measurement on 2ML indium-intercalated graphene would directly test the ARPES-derived screening.","The reported $\\alpha_G$ lies far below the lowest calibration point of the power-law fit, so a dedicated calculation of the plasmaron separation at $\\alpha_G \\approx 0.009$ would settle whether the extrapolated screening value is reliable.","The 'buffer plus nearly free-electron metal' design rule suggests that other two-layer intercalants whose first layer passivates the substrate potential and whose second layer forms a nearly free-electron band should reproduce the effect; this could be screened computationally before growth.","The same plasmaron-based calibration could become a general metrology for dielectric environments, turning ARPES tables like the paper's $\\epsilon_s$ ranking into a design library for 2D material heterostructures."],"forward_implications":["Graphene on bilayer indium/SiC should see sharply reduced long-range Coulomb scattering from the substrate: with $\\epsilon_s \\approx 622$, the substrate screening is roughly an order of magnitude larger than in previously studied intercalant systems.","The second indium layer is the operative ingredient, not a small correction: replacing it with a single layer drops the extracted screening from about $622$ to about $57$.","Because the normalized plasmaron separation is doping-independent, the same ARPES analysis can rank the dielectric quality of other graphene-substrate systems without retuning the carrier density.","The mechanism points to specific new materials: the paper identifies bilayer and trilayer gallium as predicted hosts of similar nearly free-electron states, and therefore as candidates for comparably strong screening.","The intercalation approach is compatible with wafer-scale epitaxial graphene on SiC, offering a scalable route to tailoring many-body interactions in large-area electronic devices."],"supporting_citations":[{"why":"Supplies the G0W0-RPA calibration of plasmaron band separation $\\delta E$ versus effective coupling $\\alpha_G$ and the conversion to substrate screening $\\epsilon_s \\approx 2\\epsilon - 1$.","marker":"[34]"},{"why":"Establishes the plasmaron satellite bands in doped graphene as the experimental observable being analyzed.","marker":"[47]"},{"why":"Provides the theoretical G0W0-RPA spectral function of graphene from which the $\\delta E$–$\\alpha_G$ calibration is computed.","marker":"[48]"},{"why":"Supplies the one-monolayer indium intercalation procedure, the buffer-layer behavior of the first In layer, and the comparison system used to isolate the second layer's role.","marker":"[32]"},{"why":"Shows stability and confinement of ultrathin In and Ga layers at the graphene/SiC interface and predicts nearly free-electron states in bilayer Ga, the proposed analogue.","marker":"[37]"},{"why":"Documents indium films on Si(111) as a nearly free-electron two-dimensional metal, the physical precedent for the second In layer's screening behavior.","marker":"[58]"},{"why":"Provides the first-principles description of indium on SiC (indenene) that underlies the layer-resolved DFT interpretation of the buffer layer.","marker":"[43]"}],"fun_headline_variants":["Bilayer indium intercalation yields graphene screening of 622","Two indium layers provide graphene 622-fold dielectric screening","Nearly free-electron indium layer gives graphene screening at 622","Indium bilayer produces 622-fold screening in graphene","Buffer layer plus metallic indium screens graphene to 622"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The reported $\\alpha_G = 0.0089$ and $\\epsilon_s = 622$ hang entirely on the empirical calibration curve between plasmaron band separation and graphene coupling, which is fitted at much stronger coupling and then extrapolated far outside its fitted range; the paper's written curve $\\delta E = \\alpha_G^p$ with $p = 0.534$ does not by itself reproduce the reported $\\alpha_G$ from the measured $\\delta E = 0.0557$, so the calibration is the load-bearing and least-supported step.","fun_headline_variants_meta":{"raw":{"variants":["Bilayer indium intercalation yields graphene screening of 622","Two indium layers provide graphene 622-fold dielectric screening","Nearly free-electron indium layer gives graphene screening at 622","Indium bilayer produces 622-fold screening in graphene","Buffer layer plus metallic indium screens graphene to 622"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001131,"raw_usage":{"total_tokens":4730,"prompt_tokens":1002,"completion_tokens":3728,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":618,"completion_tokens_details":{"reasoning_tokens":3644}},"tokens_in":618,"tokens_out":3728,"duration_ms":20607,"temperature":1.0,"reasoning_tokens":3644,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T14:13:06.142634+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Calculate the graphene spectral function with G0W0-RPA at the reported effective coupling ($\\alpha_G \\approx 0.0089$, equivalently substrate screening $\\epsilon_s \\approx 622$) and at the experimental carrier density, and compare the predicted hole–plasmaron separation with the measured $\\delta E = 0.0557$; a prediction that does not match $\\delta E$ would show the calibration extrapolation, and hence the quoted $\\epsilon_s$, is unsupported.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the G0W0-RPA calibration of plasmaron band separation $\\delta E$ versus effective coupling $\\alpha_G$ and the conversion to substrate screening $\\epsilon_s \\approx 2\\epsilon - 1$."},{"cited_title":"Bostwick, F","cited_arxiv_id":null,"evidence_quote":"Establishes the plasmaron satellite bands in doped graphene as the experimental observable being analyzed."},{"cited_title":"Polini, R","cited_arxiv_id":null,"evidence_quote":"Provides the theoretical G0W0-RPA spectral function of graphene from which the $\\delta E$–$\\alpha_G$ calibration is computed."},{"cited_title":"Schmitt, J","cited_arxiv_id":null,"evidence_quote":"Supplies the one-monolayer indium intercalation procedure, the buffer-layer behavior of the first In layer, and the comparison system used to isolate the second layer's role."},{"cited_title":"Briggs, B","cited_arxiv_id":null,"evidence_quote":"Shows stability and confinement of ultrathin In and Ga layers at the graphene/SiC interface and predicts nearly free-electron states in bilayer Ga, the proposed analogue."},{"cited_title":"Yoshizawa, H","cited_arxiv_id":null,"evidence_quote":"Documents indium films on Si(111) as a nearly free-electron two-dimensional metal, the physical precedent for the second In layer's screening behavior."},{"cited_title":"Bauernfeind, J","cited_arxiv_id":null,"evidence_quote":"Provides the first-principles description of indium on SiC (indenene) that underlies the layer-resolved DFT interpretation of the buffer layer."}],"review_version":2}