{"id":"3d46ddee-0011-476d-98ca-1b8bb30c05d3","arxiv_id":"1908.10893","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"Weak localization in Li-doped epitaxial graphene separates intra- and intervalley scattering, showing intervalley rates grow as the square root of carrier density at low doping but exceed tight-binding predictions at high coverage.","lead":"This paper measures how lithium atoms on graphene surfaces change electron scattering, separating scattering within a valley from scattering between valleys. It explains most of the effect with known physics but finds a mismatch at high lithium coverage that points to band structure changes.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The high-doping discrepancy that motivates the 'beyond tight-binding' claim depends on the untested assumption that residual short-range disorder is unaffected by Li deposition; a desorption/re-dosing hysteresis check would settle it.","rationale":"The reader's weakest assumption is exactly the one I would flag: the Fig. 3 theory curves are generated from residual-disorder parameters fixed at Δn=0, and the entire interpretation of the high-density data as evidence for band-structure modifications depends on those parameters being independent of Li coverage. This assumption is not tested in the paper. The low-doping sqrt(n) agreement provides some support for a fixed short-range disorder background in that regime, but it does not constrain what happens above roughly 5×10^13 cm^-2, where the discrepancy appears. The paper's counterargument to new scatterers is based on calculated binding strengths for isolated Li adatoms, not on a measurement of the disorder background at high coverage. Moreover, the high-density points are one-sided (lower-bound) estimates, so the magnitude of the discrepancy is not directly measured. The proposed desorption/re-doping hysteresis check is a decisive experiment because it compares τ_i^-1 at the same carrier density and same disorder realization except for the deposition history; if the rate is history-dependent, the fixed-residual-disorder assumption fails and the central high-doping claim is not supported. If the rate retraces, the assumption survives and the discrepancy becomes a genuine indication of Li-induced band-structure effects. This does not change the overall CONDITIONAL verdict, but it sharpens the condition.","tokens_in":16976,"tokens_out":13666,"duration_ms":158546,"concrete_test":"On a sample such as SiC3 or SiC4, after the highest-coverage Li dose, anneal in small steps to desorb part of the Li and bring n back to a value previously measured during the deposition-up sweep; at each step remeasure the weak-localization curve and extract τ_i^-1 using the identical fitting procedure. Plot the desorption-down points on Fig. 3(a): if they retrace the deposition-up curve, the residual-disorder background is stable and the high-density deviation is a genuine Li/band effect; if they lie above the up-sweep curve at equal n, Li has irreversibly or history-dependently created additional intervalley scatterers, invalidating the fixed-Vres assumption and the TB comparison.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's theoretical comparison is anchored by the statement in the main text that Vres is 'explicitly not dependent on subsequent Li deposition,' so the residual-disorder density and strengths used for the curves in Fig. 3 are fixed at their Δn=0 values. The central high-coverage conclusion—that measured intervalley rates exceed tight-binding predictions and may indicate adatom-induced band-structure modifications—is only valid if this assumption holds. The paper rules out new short-range scatterers from Li by appeal to weak Li-graphene bonding, but that is a theoretical expectation, not an experimental control, and the summary itself lists alternative explanations (VHS position, resonant scattering, nonlocal screening, electron-electron interactions) that are not excluded by the data. In particular, the high-density points in Fig. 3(a) are lower bounds on τ_i^-1 (Supplement D), so the size of the claimed discrepancy is not actually measured; what is shown is that the lower bound already lies above the TB curve. A history-dependent test can separate the residual-disorder assumption from the band-structure interpretation: if the intervalley rate at a given carrier density is different when that density is approached by desorbing Li rather than by depositing it, the disorder background has been modified by Li and the comparison in Fig. 3(a) is not a clean test of band-structure effects.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper reports weak-localization magnetotransport measurements on Li-doped epitaxial graphene on SiC, from which the intravalley scattering rate τ_*^{-1} and intervalley scattering rate τ_i^{-1} are extracted separately as a function of the Li-induced carrier density Δn. The authors find that τ_i^{-1} grows approximately as √n for carrier densities below about 5×10^13 cm^-2, which they attribute to scattering off pre-existing short-range disorder whose rate is proportional to the graphene density of states. At higher densities, the extracted τ_i^{-1} (in several cases only a lower bound) lies above the predictions of both Dirac and tight-binding models computed with parameters fixed at Δn=0, and the authors suggest this may indicate adatom-induced band-structure modifications beyond the applied model, while explicitly listing alternative explanations. The paper includes detailed fits to the WL formula, a supplement documenting the error-bar procedure, and numerical tight-binding and Dirac calculations of scattering rates and mobility with no further free parameters after baseline fitting.","tokens_in":17265,"tokens_out":8236,"duration_ms":84872,"significance":"The paper's strongest contribution is the separate extraction of intra- and intervalley scattering channels in a controlled doping sequence, together with the demonstration that the intervalley rate follows a √n dependence consistent with density-of-states-enhanced scattering from residual short-range disorder. The comparison to theory is disciplined: the residual-disorder parameters are fixed at Δn=0, and the subsequent √n and tight-binding curves are genuine shape predictions rather than free fits. The careful reduced-χ2 analysis with an explicit error-bar definition is another strength, as is the transparent supplement that allows the fits to be reproduced. If the high-density discrepancy survives an experimental control for disorder modification by Li, the suggestion that alkali adatoms alter the graphene band structure beyond a rigid shift would be an important contribution to the adatom-graphene problem. However, one load-bearing assumption and the lower-bound character of the high-density points require attention before the central interpretation can be considered established.","major_comments":[{"comment":"The conclusion that the high-density deviation in Fig. 3(a) points to band-structure modifications beyond the tight-binding model rests on the assumption, stated as 'Since Vres is explicitly not dependent on subsequent Li deposition', that the residual short-range disorder (density nres and strengths V0,i, V0,*) is unaffected by Li deposition. The manuscript excludes new short-range scatterers only by arguing that Li-graphene bonding is weak, which is a theoretical expectation rather than an experimental control. A desorption/re-dosing hysteresis measurement would directly test this assumption: if τ_i^-1 at a given carrier density differs when that density is approached by desorbing Li rather than by depositing it, the disorder background has been irreversibly modified, and the Fig. 3(a) comparison is not a clean test of band-structure effects. Without such a control, the high-density discrepancy is equally consistent with Li-induced modification of the residual disorder, so this point is load-bearing for the central interpretation.","section":"Main text, paragraph following Eq. (2)"},{"comment":"At the highest carrier densities, the plotted values of τ_i^-1 are one-sided lower bounds: Supplement D explains that for these points the reduced chi-square flattens once Bi exceeds the accessible 100 mT range, so only a lower bound on Bi, and hence on τ_i^-1, can be determined. The main text should state this explicitly in the caption or body of Fig. 3(a), and the text 'the intervalley data lies well above the √n traces' should be rephrased as 'the lower bounds on the intervalley rate lie above the traces'. As it stands, the wording overstates the quantity actually measured, although the existence of a discrepancy in the sense of a lower bound above the theoretical curve is not in question.","section":"Fig. 3(a) and Supplement D"},{"comment":"The phrase 'leaving us with no free fitting parameters in our theory' is stronger than the procedure warrants. The values of V0,i and V0,* are fitted per sample to the Δn=0 intervalley rate and mobility, and nres is assumed to be 10^12 cm^-2 (Supplement I, Table I). What is correct and valuable is that, once those baseline fits are made, the doping dependence of the theoretical curves is predicted without additional fitting. The wording should say that rather than implying a fully parameter-free calculation, so that readers do not overestimate the independence of the comparison.","section":"Main text, paragraph after Eq. (2); Supplement I"}],"minor_comments":[{"comment":"The first sentence of Section F.1 contains a duplicated article: 'The the charged Li adatoms' should read 'The charged Li adatoms'.","section":"Supplement F.1"},{"comment":"The sentence 'All parameters enterning the matrix element' contains a typo; 'enterning' should be 'entering'.","section":"Supplement I"},{"comment":"The phrase 'the extracted τi was indistinguishable from zero' is confusing when read against the axes of Fig. 2(c) and Fig. 3(a), which plot τ_i^{-1}; consider rewording to 'the intervalley time τ_i was too short to resolve, so only a lower bound on τ_i^{-1} could be determined'.","section":"Supplement D and main text after Eq. (1)"},{"comment":"Reference 17 is listed as 'Phys. Rev. XXX XXX, XXX (2019), arXiv:1904.08191'; if a published version now exists, the citation should be updated to the final journal reference.","section":"Reference 17"}],"recommendation":"major_revision","confidential_remarks":"The DFT parameters used in the 'no free parameters' comparison (ZLi = 0.9, d = 1.78 Å) are taken from Ref. 17, a preprint co-authored by one of the present authors (K.K.). This is not circular because those parameters are not fitted to the transport data, but the referee should be aware that the theoretical input and the experiment share an author; an explicit acknowledgement of this in the paper would be appropriate. The high-density interpretation is already hedged in the abstract, but the hysteresis control suggested in Major Comment 1 would substantially strengthen the paper's central claim."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nWhat you should know: this is a careful weak-localization study that, for the first time, separates intra- and intervalley scattering rates in alkali-doped graphene. The low-doping result—intervalley rate growing as sqrt(n), tied to rising density of states amplifying scattering off pre-existing short-range disorder—is solid. The fits have clear chi-square minima, the sqrt(n) dependence is a genuine shape prediction after the Δn=0 baseline is fixed, and the inverse mobility comparison in Fig. 3(b) works without free parameters. That is a real contribution.\n\nThe soft spot is the high-doping claim. The paper wants to say measured intervalley rates exceed tight-binding predictions, implying adatom-induced band-structure modifications. Two problems. First, the highest-density points are only lower bounds on tau_i^-1 (Supplement D); what is actually shown is that the lower bound already lies above the TB curve. That is suggestive, not a measurement of the discrepancy's size. Second, the whole comparison rests on the assumption that residual short-range disorder (density and strengths) is unaffected by Li deposition. The paper states this explicitly and argues Li bonds weakly; that's a theoretical expectation, not a control. Lithium could screen or modify defects, or create new short-range scatterers, and any of those would shift the predicted curves. A desorption/re-dosing run—approach the same n by removing Li instead of adding it—would test this directly. The authors have the stage for it; they just didn't do it.\n\nThe DFT Coulomb parameters (ZLi=0.9, d=1.78 Å) come from a co-authored paper. That is a self-citation, but not a fatal one: the mobility match gives independent support to that input, and the low-doping intervalley story doesn't depend on it.\n\nBottom line: the low-doping story is credible and should be cited. The high-doping 'beyond tight-binding' conclusion is a reasonable hypothesis, but it is conditional. A referee should ask for the hysteresis control or at least a much sharper discussion of the disorder-background assumption. I would send this to review, not desk reject. It deserves referee time, and a knowledgeable referee can probably extract the solid core from the speculative tail.","headline":"A careful WL study that cleanly separates intra- and intervalley scattering in Li-doped graphene; the low-doping sqrt(n) result is solid, but the high-doping 'beyond tight-binding' claim rests on an untested disorder-background assumption and lower-bound data points.","tokens_in":17824,"tokens_out":2373,"would_cite":true,"duration_ms":25612,"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":"Weak localization measurements separately quantify intra- and intervalley scattering in Li-doped graphene, revealing a high-doping discrepancy with tight-binding theory.","keywords":["weak localization","lithium adatoms","intervalley scattering","intravalley scattering","epitaxial graphene","tight-binding model","carrier density scaling","magnetoresistance"],"falsifier":"Prepare two graphene samples with very different initial defect densities, measure their intervalley rates as lithium is added, and check whether the low-density $\\sqrt n$ portions scale with the initial defect density as the residual-disorder explanation predicts. If the high-density excess above the tight-binding curve appears at the same carrier density regardless of initial defects, the fixed-disorder picture is confirmed; if the excess instead tracks the amount of lithium deposited, new scatterers or adatom-induced band changes are responsible.","tokens_in":16768,"feed_emoji":"🧲","tokens_out":8812,"duration_ms":81564,"temperature":0.7,"pith_summary":"The paper tries to establish that weak-localization magnetoresistance can separate the two scattering channels in graphene that ordinary conductivity blurs together, and that lithium adatoms affect both channels in measurable, partly predictable ways. In epitaxial graphene, adding Li increases the intravalley scattering rate as expected for Coulomb scattering off charged adatoms, and it also increases the intervalley rate in a way that at low doping follows $\\tau_i^{-1}\\propto\\sqrt{n}$, the signature of extra carriers making pre-existing short-range defects more effective through the graphene density of states. A parameter-free tight-binding calculation that includes trigonal warping and band nonlinearity reproduces the mobility and the low-density intervalley data, but above $n\\approx 5\\times10^{13}\\,\\mathrm{cm}^{-2}$ the measured intervalley rate rises well above the calculation. The authors take that residual discrepancy as evidence that lithium at high coverage modifies the graphene band structure beyond simple Dirac-cone physics, consistent with photoemission reports; a sympathetic reader would care because it shows transport and photoemission can be connected through quantitatively separated scattering rates.","feed_headline":"Li adatoms unmask graphene's hidden intervalley scattering","feed_subtitle":"Magnetoresistance fits show intervalley scattering grows as the square root of carrier density, then outruns tight-binding theory","key_machinery":"The central machinery is the graphene weak-localization formula $\\Delta\\sigma(B_\\perp)=\\frac{e^2}{\\pi h}[F(\\tau_B^{-1}/\\tau_\\phi^{-1}) - F(\\tau_B^{-1}/(\\tau_\\phi^{-1}+2\\tau_i^{-1})) - 2F(\\tau_B^{-1}/(\\tau_\\phi^{-1}+\\tau_*^{-1}+\\tau_i^{-1}))]$, where $F(z)=\\ln z+\\psi(z^{-1}+\\tfrac12)$ and $\\tau_B^{-1}=4eDB_\\perp/\\hbar$; it converts magnetic-field scans into separate dephasing, intravalley, and intervalley rates. The interpretive counterpart is the Born-approximation scattering rate $\\tau_\\alpha^{-1}(\\varepsilon_k)=\\frac{2\\pi}{\\hbar}n_\\alpha\\int \\frac{dk'}{(2\\pi)^2}|V^\\alpha_{kk'}|^2\\delta(\\varepsilon_k-\\varepsilon_{k'})$ evaluated on nearest-neighbor tight-binding bands, which supplies the parameter-free comparison via the residual-disorder strength $V_{\\rm res}$ and the DFT-derived Coulomb potential of Li. The physical idea carrying the argument is that lithium doping changes only the carrier density, so the intervalley scattering from pre-existing defects scales with the graphene density of states $D(E_F)\\propto\\sqrt n$; any deviation from that scaling at high density is read as band-structure modification.","core_discovery":"The central discovery is that intervalley scattering in Li-doped epitaxial graphene is enhanced far more than a long-range Coulomb scatterer should cause, and that the enhancement splits into a low-density part that a fixed population of short-range defects can explain and a high-density part that it cannot. Fitting the weak-localization magnetoconductance to the standard graphene formula yields $\\tau_i^{-1}$ that grows as $\\sqrt n$ up to about $5\\times10^{13}\\,\\mathrm{cm}^{-2}$, exactly the density-of-states scaling expected when doping merely raises the Fermi energy while the density and strength of residual short-range scatterers stay fixed. The same fixed-defect assumption, evaluated on nearest-neighbor tight-binding bands, reproduces the measured mobility without free parameters. Above that density, the measured intervalley rate exceeds both the Dirac-model and tight-binding predictions, so the paper concludes that high-coverage Li induces band-structure modifications, such as a shifted van Hove singularity or an impurity band, that the model does not include.","pith_inferences":["A natural test of the shifted-van-Hove explanation would be a continuous density sweep through $n\\sim10^{14}\\,\\mathrm{cm}^{-2}$ to look for a kink in $\\tau_i^{-1}$ when the Fermi level crosses the saddle point.","Comparing lithium with heavier alkali adatoms such as potassium or cesium would separate charge-transfer effects from adatom-specific band-structure effects, because all alkalis dope similarly but perturb the Dirac bands differently.","If the high-density excess is due to an impurity band, then its contribution to intervalley scattering should depend on temperature in a way that elastic scattering off fixed defects does not; low-temperature weak-localization measurements across a wider temperature range could reveal that."],"forward_implications":["Weak localization can resolve short-range intervalley scattering that conductivity measurements mask behind dominant Coulomb scattering, so the technique is a practical probe of adatom-induced disorder channels in graphene.","The observed $\\tau_i^{-1}\\propto\\sqrt{n}$ behavior confirms that residual short-range disorder in epitaxial graphene is pre-existing and is not created or modified by low-coverage lithium deposition.","Above $n\\approx5\\times10^{13}\\,\\mathrm{cm}^{-2}$, the intervalley rate exceeds tight-binding predictions even with trigonal warping, so high-density transport data carry information about band-structure changes that photoemission has also observed.","Because the mobility calculation matches experiment without free parameters while the intervalley rate does not, the combination of zero-field conductivity and weak localization is what exposes the missing physics."],"supporting_citations":[{"why":"It provides the graphene weak-localization magnetoconductance formula from which the intra- and intervalley rates are extracted.","marker":"[26]"},{"why":"It supplies the DFT-based parameters for the Li impurity potential and the impurity-band picture used in the tight-binding calculations.","marker":"[17]"},{"why":"It documents the sample stage, annealing procedures, and lithium dosing behavior that the present measurements build on.","marker":"[24]"},{"why":"It reports ARPES evidence of Li-induced band-structure modification and superconductivity, the comparison point for the high-density discrepancy.","marker":"[14]"},{"why":"It establishes Coulomb scattering off charged adatoms as the dominant mobility-limiting mechanism in alkali-doped graphene.","marker":"[9]"},{"why":"It identifies the atomic-scale defects that act as residual short-range disorder in epitaxial graphene on SiC.","marker":"[32]"}],"fun_headline_variants":["Li doping reveals extra intervalley scattering in graphene","Intervalley scattering in Li-doped graphene outruns theory","High Li coverage bends graphene's scattering rules","Graphene's intervalley scattering spikes beyond tight-binding model","Li adatoms boost intervalley scattering more than expected"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole comparison rests on the assumption that lithium deposition changes only the carrier density and leaves the number, positions, and scattering strengths of the pre-existing short-range defects untouched; if lithium creates new scatterers, screens the old ones, or reshapes the bands, the predicted $\\sqrt n$ scaling and the size of the high-density discrepancy would both change.","fun_headline_variants_meta":{"raw":{"variants":["Li doping reveals extra intervalley scattering in graphene","Intervalley scattering in Li-doped graphene outruns theory","High Li coverage bends graphene's scattering rules","Graphene's intervalley scattering spikes beyond tight-binding model","Li adatoms boost intervalley scattering more than expected"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000604,"raw_usage":{"total_tokens":2810,"prompt_tokens":930,"completion_tokens":1880,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":546,"completion_tokens_details":{"reasoning_tokens":1803}},"tokens_in":546,"tokens_out":1880,"duration_ms":12327,"temperature":1.0,"reasoning_tokens":1803,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T10:31:10.007290+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Prepare two graphene samples with very different initial defect densities, measure their intervalley rates as lithium is added, and check whether the low-density $\\sqrt n$ portions scale with the initial defect density as the residual-disorder explanation predicts. If the high-density excess above the tight-binding curve appears at the same carrier density regardless of initial defects, the fixed-disorder picture is confirmed; if the excess instead tracks the amount of lithium deposited, new scatterers or adatom-induced band changes are responsible.","supporting_citations":[{"cited_title":"McCann , author K","cited_arxiv_id":null,"evidence_quote":"It provides the graphene weak-localization magnetoconductance formula from which the intra- and intervalley rates are extracted."},{"cited_title":"Signatures of adatom effects in the quasiparticle spectrum of Li-doped graphene","cited_arxiv_id":"1904.08191","evidence_quote":"It supplies the DFT-based parameters for the Li impurity potential and the impurity-band picture used in the tight-binding calculations."},{"cited_title":"Khademi , author E","cited_arxiv_id":null,"evidence_quote":"It documents the sample stage, annealing procedures, and lithium dosing behavior that the present measurements build on."},{"cited_title":"Mallet , author I","cited_arxiv_id":null,"evidence_quote":"It identifies the atomic-scale defects that act as residual short-range disorder in epitaxial graphene on SiC."}],"review_version":1}