{"id":"4b819b12-8b64-496b-828f-f23e4b7fce55","arxiv_id":"1908.02956","paper_version":2,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Local sheet resistance in epitaxial graphene on 6H-SiC varies by up to 270% at low temperature and correlates with the SiC substrate termination and the graphene-substrate distance.","lead":"Using scanning tunneling potentiometry, this paper maps the local electrical resistance of graphene grown on a silicon carbide substrate and finds that it varies by up to 270% from terrace to terrace at low temperature. The variation tracks the stacking of the SiC substrate and the distance between the graphene sheet and the substrate, which matters for understanding and engineering epitaxial graphene devices.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 270% sheet-resistance spread depends on unmeasured COMSOL current-density inputs (step resistivities 6/12/18 Ωμm), so errors in these inputs propagate into every reported ρ_sheet.","rationale":"The reader's weakest_assumption correctly identifies the most load-bearing concern: the quantitative extraction of local sheet resistance depends on a modeled, not measured, local current density, and the model inputs (step resistivities) are taken from prior work without in-paper validation. This is indeed the linchpin of the central claim, because if j_local is wrong, every reported ρ_sheet value and the headline 270% spread are wrong. The paper's own statement that local current density variations are indistinguishable from sheet-resistance variations underscores the severity: the only evidence for the required homogeneity is the COMSOL model. The concern is not merely a disagreement with consensus; it is a question of whether the argument's key premise is supported. The paper does provide some independent checks—such as the consistency of step height deviations measured by STM and AFM, and the qualitative difference in potential gradients across steps—but these do not quantify the uncertainty in j_local. A sensitivity analysis of the step-resistivity inputs is a straightforward and decisive test. The reader's CONDITIONAL verdict is therefore appropriate: the paper should be accepted only if such an analysis is provided and shows the results are insensitive to these inputs. Our stress-test does not change that verdict, hence UNCHANGED. The secondary concern about circular sorting in Supplementary Fig. 5 is real but less fundamental; it affects the distance-correlation interpretation, not the primary quantitative extraction. Agreement with the reader is 'agree' because we identify the same weakest assumption.","tokens_in":17539,"tokens_out":4615,"duration_ms":48473,"concrete_test":"Re-run the COMSOL simulation with step-resistivity inputs varied over a plausible range—e.g., (single, double, triple) = (0,0,0), (3,6,9), (6,12,18), (12,24,36) Ωμm—and recompute j_local and the resulting terrace sheet resistances at 300 K and 8 K (using temperature-scaled macroscopic resistances). If the extracted ρ_low/ρ_high ordering or the 270% spread changes by more than the reported terrace-to-terrace variation (±20 Ω), the quantitative claim is not robust to these inputs. Reporting the resulting spread as a function of assumed step resistivities would settle whether the conclusion stands.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central quantitative claim is that local sheet resistance varies by up to 270% at 8 K and separates into two distinct values ρ_high and ρ_low, governed by SiC termination and graphene-substrate distance. The STP measurement yields only the local potential gradient dV_STP/dx; converting this to ρ_sheet requires dividing by the local current density j_local. The paper acknowledges that spatial variations in j are experimentally inaccessible and indistinguishable from variations in ρ_sheet. To resolve this, the authors use a COMSOL finite-element model that takes as inputs the macroscopic ohmic resistance, sample geometry, and step resistivities of 6 Ωμm, 12 Ωμm, and 18 Ωμm for single, double, and triple substrate steps. These step resistivities are not measured in this work or derived in the text; they are taken from prior STP studies, several from the same group. If the true step resistivities differ—for instance, because PASG graphene has different step coupling than the conventional epitaxial graphene used to calibrate those values—then the modeled j_local is wrong. The paper claims j_local varies by at most 7% in selected regions, but this estimate is itself a model output, not a measurement. Because every reported sheet resistance is proportional to 1/j_local, any systematic error in j_local propagates directly into ρ_low and ρ_high and into the 270% low-temperature spread. The problem is especially acute at 8 K: the COMSOL model is described for T = 300 K, and no temperature-dependent step resistivities are provided, yet the headline 270% claim is made at 8 K. Without a sensitivity analysis of these inputs, the quantitative central claim is not yet independently supported.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This manuscript reports scanning tunneling potentiometry (STP) measurements on polymer-assisted sublimation growth (PASG) epitaxial graphene on 6H-SiC(0001). The authors convert STP potential gradients into local sheet resistances using a local current density obtained from a COMSOL finite-element model. They report two distinct room-temperature sheet resistances (ρ_high = 535 Ω, ρ_low = 460 Ω) that correlate with the S3 and S2 substrate terminations, and a strongly increased scatter at low temperatures, with a maximum variation of about 270% between extreme terrace values at 8 K. Step-height deviations are interpreted as variations of the graphene–substrate distance and are correlated with the local sheet resistance. STS data show nearly identical Dirac-point energies on the two terminations, leading the authors to attribute the resistance difference to mobility variations, and specifically to substrate phonon scattering and scattering by disorder in the 6×6 quasi-corrugation.","tokens_in":17825,"tokens_out":14692,"duration_ms":145201,"significance":"If the central claims hold, this work provides a quantitative, local view of how the SiC substrate termination and the graphene–substrate distance modulate charge transport in epitaxial graphene, a question of long-standing interest in the field. The study draws on a large dataset (more than 80 terraces), combines STP with STS, AFM, and ARPES on the same samples, and is corroborated by macroscopic van der Pauw measurements. The identification of two distinct local sheet resistances tied to the SiC stacking is a concrete and falsifiable result that would be of value to the graphene and SiC communities. However, the quantitative claims, especially the 270% figure, depend on a modeled current density whose temperature dependence and sensitivity to the very resistance inhomogeneities under study are not established in the manuscript.","major_comments":[{"comment":"The local current density used to convert STP potential gradients into ρ_sheet is specified only for T = 300 K: \"It is j_local = (0.89 ± 0.01) A/m for an applied voltage ... of 1V at T = 300 K.\" No statement is made about how j_local was obtained at 77 K and 8 K, although the macroscopic resistance changes substantially with temperature (Supplementary Fig. 3a). If the same 300 K current density was used at low temperatures, the reported ρ_sheet values at 77 K and 8 K are systematically overestimated, and the phonon contribution estimated in the Discussion as \"the difference between the mean sheet resistance at 300K and the highest measured values at 8K on terraces S3\" is biased. The authors should state how j_local was determined at each temperature and, if necessary, rerun the COMSOL simulation with the appropriate temperature-dependent input resistances.","section":"Homogeneity of the current density; Temperature-dependence of ρ_sheet"},{"comment":"The claim that the local current density varies by at most 7% in the selected regions is based on a resistor-network/COMSOL model that assigns a single, uniform sheet resistance to the graphene terraces. This model does not include the very terrace-to-terrace resistance differences that are the paper's central finding (14% at 300 K, and up to 178% for adjacent terraces at low temperature). It is not demonstrated that the 7% homogeneity remains valid when ρ_high and ρ_low are assigned to the S3 and S2 terraces, respectively. Since every reported ρ_sheet is inversely proportional to the assumed j_local, please provide a sensitivity analysis using the measured terrace-resistance distribution in the COMSOL model and show how the extracted ρ_sheet values, and the 270% low-temperature variation, change.","section":"Homogeneity of the current density; Fig. 1b; Supplementary Fig. 1"},{"comment":"The support for the distance–resistance correlation is weakened by a circular step: in Supplementary Fig. 5 the data are sorted \"under the assumption that a larger distance to the substrate leads to a reduction of the resistance,\" and the resulting ordering is then used as evidence for the correlation. The manuscript should present the unsorted scatter of ρ_sheet versus measured step height for all terraces, separately for S2 and S3, and quantify the correlation (e.g., a Spearman or Pearson coefficient) without imposing the hypothesis. The single exception (the yellow data set) should be discussed in this context rather than being presented as the only mismatch after sorting.","section":"Supplementary Fig. 5; Discussion"},{"comment":"The step resistivities used in the COMSOL model (6 Ωμm, 12 Ωμm, 18 Ωμm for single, double, and triple steps) are inputs to the calculation, yet the manuscript does not cite the source of these values or establish that they are valid for the PASG graphene samples studied here; presumably they are taken from earlier STP studies on other epitaxial graphene samples. A sensitivity analysis (e.g., varying the step resistivities by a reasonable factor and reporting the resulting range in j_local in the selected regions) would show whether the extracted ρ_sheet values and the 270% spread are robust. Without this, the quantitative sheet resistances inherit an unquantified systematic uncertainty from these inputs.","section":"Methods: Finite element simulation with COMSOL"}],"minor_comments":[{"comment":"There are a few typographical issues: \"for a the given surface geometry\" should be \"for the given surface geometry,\" and the abstract's first sentence is missing a comma after \"material.\"","section":"Homogeneity of the current density; Abstract"},{"comment":"The 270% variation is reported as a range between extreme values. Please also provide a measure of the distribution width (e.g., standard deviation or interquartile range) for ρ_sheet at each temperature, since range statistics are sensitive to outliers and the headline claim would be more robust with a dispersion measure.","section":"Temperature-dependence of ρ_sheet"},{"comment":"The reference to Supplementary Fig. 9 in the sentence \"A systematic difference in ρ_high for S3 compared to S3* and in ρ_low for S2 compared to S2* was not observed\" appears to point to the height-calibration figure, which does not display such a comparison; please correct the cross-reference.","section":"Discussion"},{"comment":"In the estimation of the phonon contribution as the difference between the mean 300 K sheet resistance and the highest measured values at 8 K on terraces S3, please justify the use of the highest rather than the mean or median 8 K value, or present the estimate as an upper bound.","section":"Discussion"}],"recommendation":"major_revision","confidential_remarks":"The manuscript states that it is a preprint of an article published in Nature Communications (see Acknowledgements). The editor may wish to consider whether prior publication affects the novelty assessment for this venue. The technical concerns raised above are addressable, but the temperature dependence of the current-density model and the circularity in Supplementary Fig. 5 require substantive revision before the quantitative claims can be accepted as stated."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe short version: this is a careful STP study with a real finding—two distinct terrace sheet resistances in monolayer PASG graphene, tied to the 6H-SiC termination—and, at low temperature, a much larger terrace-to-terrace spread. It deserves peer review. The quantitative side has a soft spot: the absolute resistances and the headline 270% spread rely on a COMSOL current-density model whose step-resistance inputs are taken from the same group's earlier STP work, not measured here. That's addressable, not fatal.\n\nWhat's genuinely new: earlier STP work looked at step resistances and defects. This paper uses bilayer-free PASG, argues the current density is nearly uniform, and quantifies two terrace sheet resistances differing by 14% at 300 K, with a temperature-dependent spread that reaches ~270% at 8 K. The measurements span more than 80 terraces, with STM/STS/ARPES cross-checks. The correlation with SiC stacking (S2 vs S3) and with graphene-substrate distance is plausible and supported by several lines of evidence. The step-height deviations measured in CCT and AFM are a nice structural hook.\n\nThe main concern is the conversion from STP voltage gradient to rho_sheet. The paper states the local current density varies by at most 7% in selected regions, but that's a model output, not a measurement. The model uses step resistivities of 6, 12, and 18 Ohm-micrometer for single, double, and triple steps, from previous work that is partly self-cited. If those inputs are wrong, the absolute rho_sheet values shift, and if the error is spatially nonuniform, the terrace-to-terrace spread could be distorted. The paper doesn't give a sensitivity analysis, and the COMSOL model is described at 300 K while the 270% claim is at 8 K, where the step resistances could have different relative weights. That said, the relative difference between S2 and S3 at 300 K is directly visible in the raw potential slopes (Fig. 1d-f), so that part does not depend on the model. The low-T spread, being a ratio, would survive a uniform rescaling of j; the risk is mainly from a nonuniform j error, which the 7% estimate is meant to bound.\n\nA smaller weakness: the distance-resistance correlation is tested by ordering the data under the hypothesis (Supplementary Fig. 5). The approach would be stronger if the step heights were used to make a prediction before looking at resistances.\n\nWho's it for: researchers using STP or similar local probes on epitaxial graphene and anyone interested in substrate proximity effects. The paper should go to peer review; the COMSOL sensitivity question is exactly what a referee should push on.\n\nI'd accept it for review and ask for a sensitivity analysis on the step resistivities.","headline":"Solid experimental study of local resistance variation in PASG graphene; the main effect is real, but the absolute scale depends on an unvalidated current-density model—referee it.","tokens_in":18436,"tokens_out":3685,"would_cite":true,"duration_ms":36548,"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":"Epitaxial graphene's local sheet resistance varies by up to 270% at 8 K, following the SiC substrate's terrace stacking and the graphene–substrate distance.","keywords":["graphene","epitaxial graphene","scanning tunneling potentiometry","sheet resistance","6H-SiC","graphene-substrate interaction","nanoscale transport","polymer-assisted sublimation growth"],"falsifier":"Simultaneously image the local current density and the electrochemical potential on the same PASG terraces, using a probe with current-density sensitivity, and test whether the potential slope divided by the true local current density still shows a ~270% terrace-to-terrace variation at 8 K.","tokens_in":17318,"feed_emoji":"⚡","tokens_out":9382,"duration_ms":97111,"temperature":0.7,"pith_summary":"On graphene grown by polymer-assisted sublimation growth (PASG) on 6H-SiC(0001), the paper measures the local sheet resistance terrace by terrace using scanning tunneling potentiometry. Because this growth method produces bilayer-free, laterally homogeneous monolayers, the current density is almost uniform, so a measured potential gradient can be converted into a quantitative local sheet resistance. The paper finds that the sheet resistance comes in two distinct levels—about 460 Ω and 535 Ω at 300 K—that alternate with the S2 and S3 terminations of the SiC substrate, and that at 8 K the spread between lowest and highest terrace values reaches about 270%. Larger graphene–substrate distances correlate with lower resistance, while the Dirac point shifts by only about 5 meV between terrace types, pointing to mobility rather than doping as the carrier of the variation. A reader would care because it gives a concrete, quantitative picture of how a substrate controls transport in a two-dimensional material at the nanoscale.","feed_headline":"Graphene resistance swings 270 percent with the substrate below","feed_subtitle":"Nanoscale potential maps tie the resistance contrast to SiC terrace stacking and graphene–substrate distance.","key_machinery":"The central machinery is scanning tunneling potentiometry (STP), which records the local electrochemical potential $V_{\\mathrm{STP}}(x,y)$ while a bias is applied across the sample, combined with the extraction identity $\\rho_{\\mathrm{sheet}} = (dV_{\\mathrm{STP}}/dx)/j$ for each terrace. What makes the identity usable here is the claim that PASG graphene has a nearly uniform local current density $j_{\\mathrm{local}}(x,y)\\approx j$, verified by a finite-element resistor-network simulation of the measured surface geometry that puts residual current-density variations at or below 7% in the selected regions; this converts a potential gradient into a resistance without knowing the current-density distribution independently. The analysis also leans on the identification of terrace types S2/S3 in the 6H-SiC stacking sequence, on step-height deviations as a proxy for graphene–substrate distance, and on Fourier-filtered topographies of the $6\\times6$ quasi-corrugation, whose long-wavelength disorder is invoked as the source of low-temperature potential scattering.","core_discovery":"The paper's central claim is that the local sheet resistance of PASG epitaxial graphene is spatially inhomogeneous and is governed jointly by the termination of the 6H-SiC substrate and by the local distance between the graphene layer and the substrate. On more than 80 terraces, STP potential maps yield two reproducible resistance levels, $\\rho_{\\mathrm{high}}\\approx 535\\,\\Omega$ and $\\rho_{\\mathrm{low}}\\approx 460\\,\\Omega$ at 300 K (a $(14\\pm1)\\%$ contrast), which the authors assign to substrate terminations S3 and S2. At 8 K the two means drop to roughly 365 Ω and 250 Ω, but the terrace-to-terrace scatter grows to a maximum of about 270% between the lowest and highest values. Step-height deviations from SiC bilayer multiples are read as a varying graphene–substrate separation, and the trend is that larger separations give lower resistance; simultaneous tunneling spectroscopy shows nearly identical Dirac energies on S2 and S3, so the resistance contrast is assigned to mobility differences rather than doping differences. The paper concludes with a two-channel scattering model, $\\rho(T,d)=\\rho_{\\mathrm{el-phonon}}(T)+\\rho_{\\mathrm{el-defect}}(T,d)$, where the low-temperature spread comes from potential scattering by the imperfect $6\\times6$ interfacial corrugation and interface states.","pith_inferences":["If the distance–resistance correlation is causal, deliberately changing the graphene–substrate separation—by intercalation or by choosing different SiC polytypes—should reproduce the same monotonic decrease in resistance, which the same STP protocol can check.","Because the 270% spread appears only when phonon scattering is frozen out, the low-temperature resistance map is effectively an image of the interfacial potential landscape; combining it with Fourier-filtered topographies could quantify how particular modulation wavelengths scatter electrons.","The small Dirac-energy difference between S2 and S3 implies that any doping-based explanation must be minor; a local Hall or capacitance measurement on individual terraces would test whether the mobility contrast is as large as the resistance contrast implies.","A direct measurement of the step resistivities on the same sample would remove the main external input in the current-density model; if the true step resistivities differ from the values taken from earlier work, the absolute values of $\\rho_{\\mathrm{high}}$ and $\\rho_{\\mathrm{low}}$ will shift, though the S2/S3 contrast may survive."],"forward_implications":["Epitaxial graphene on SiC(0001) cannot be assigned a single local sheet resistance; transport measurements that average over terraces mix two distinct resistance levels and a temperature-dependent spread.","Because larger graphene–substrate separation correlates with lower sheet resistance, decoupling schemes such as intercalation should improve local mobility, and the distance–resistance relation gives a quantitative target for such schemes.","The near-identical Dirac energies on S2 and S3 mean the resistance contrast is a mobility effect, so nanoscale mobility engineering must target the interfacial $6\\times6$ corrugation and its disorder rather than doping.","Rotating the current direction by 90° with respect to the substrate steps should channel current preferentially through low-resistance S2 terraces, creating intrinsic quasi-one-dimensional transport channels roughly ten Fermi wavelengths wide.","Because S2 and S3 differ in resistance and hence in Joule heating, adjacent terraces act as nanoscale heat sources and sinks, opening a route to study low-dimensional heat transport and thermoelectric behavior."],"supporting_citations":[{"why":"Supplies the STP extraction formula $\\rho_{\\mathrm{sheet}} = (dV_{\\mathrm{STP}}/dx)/j$ and the interpretation of measured step voltage drops.","marker":"[4]"},{"why":"Earlier atomic-scale STP transport study in epitaxial graphene that provides the observed spread in step and defect resistances the paper aims to explain.","marker":"[5]"},{"why":"Introduces the resistor-network current-density modeling and the residual-resistivity dipole concept used in the finite-element simulations.","marker":"[6]"},{"why":"Establishes the PASG method that yields large-area monolayer graphene without bilayer regions.","marker":"[9]"},{"why":"Documents the homogeneity, shallow step heights, and low resistance anisotropy of PASG graphene that justify the uniform-current approximation.","marker":"[28]"},{"why":"Theoretical basis for how localized scatterers produce spatial current and field variations, motivating the need to model the current density.","marker":"[29]"},{"why":"Explains the potential barrier at substrate steps by graphene detachment, used to interpret the additional voltage drop at each step.","marker":"[31]"},{"why":"Substrate remote-phonon scattering model used to assign the temperature dependence of $\\rho_{\\mathrm{el-phonon}}$.","marker":"[63]"}],"fun_headline_variants":["Nanoscale graphene resistance varies 270% with substrate stacking","Graphene's local resistance hinges on SiC stacking and gap","Substrate steps and spacing flip graphene sheet resistance","Graphene resistance changes up to 270% across SiC terraces"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole picture rests on the assumption that the local current density is the nearly uniform field produced by the simulation, and that the resistances assigned to single, double, and triple substrate steps are correct; if those inputs are wrong, every quoted sheet resistance shifts and the 270% spread may change or disappear.","fun_headline_variants_meta":{"raw":{"variants":["Nanoscale graphene resistance varies 270% with substrate stacking","Graphene's local resistance hinges on SiC stacking and gap","Substrate steps and spacing flip graphene sheet resistance","Graphene resistance changes up to 270% across SiC terraces"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000202,"raw_usage":{"total_tokens":1400,"prompt_tokens":983,"completion_tokens":417,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":599,"completion_tokens_details":{"reasoning_tokens":346}},"tokens_in":599,"tokens_out":417,"duration_ms":4906,"temperature":1.0,"reasoning_tokens":346,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T14:28:31.585647+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Simultaneously image the local current density and the electrochemical potential on the same PASG terraces, using a probe with current-density sensitivity, and test whether the potential slope divided by the true local current density still shows a ~270% terrace-to-terrace variation at 8 K.","supporting_citations":[{"cited_title":"A., Wenderoth, M., Electronic Transport Properties of 1D- Defects in Graphene and Other 2D-Systems, Ann","cited_arxiv_id":null,"evidence_quote":"Supplies the STP extraction formula $\\rho_{\\mathrm{sheet}} = (dV_{\\mathrm{STP}}/dx)/j$ and the interpretation of measured step voltage drops."},{"cited_title":"et al., Atomic scale transport in epitaxial graphene","cited_arxiv_id":null,"evidence_quote":"Earlier atomic-scale STP transport study in epitaxial graphene that provides the observed spread in step and defect resistances the paper aims to explain."},{"cited_title":"G., Schneider, M","cited_arxiv_id":null,"evidence_quote":"Introduces the resistor-network current-density modeling and the residual-resistivity dipole concept used in the finite-element simulations."},{"cited_title":"et al., Comeback of epitaxial graphene fo r electronics: large-area growth of bilayer-free graphene on SiC, 2D Mater","cited_arxiv_id":null,"evidence_quote":"Establishes the PASG method that yields large-area monolayer graphene without bilayer regions."},{"cited_title":"et al., Minimum Resistance Anisotropy of Epitaxial Graphene on SiC, ACS Appl","cited_arxiv_id":null,"evidence_quote":"Documents the homogeneity, shallow step heights, and low resistance anisotropy of PASG graphene that justify the uniform-current approximation."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Theoretical basis for how localized scatterers produce spatial current and field variations, motivating the need to model the current density."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Explains the potential barrier at substrate steps by graphene detachment, used to interpret the additional voltage drop at each step."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Substrate remote-phonon scattering model used to assign the temperature dependence of $\\rho_{\\mathrm{el-phonon}}$."}],"review_version":1}