{"id":"24db9d69-351b-4620-a2c1-c51e5b64bdd1","arxiv_id":"2506.10615","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"In NiTe2, adding a Hubbard U of 5 eV to surface Ni atoms shifts the calculated SS2 Dirac crossing down by about 100 meV, matching the ARPES value of -1.42 eV.","lead":"This paper measures the electronic structure of the semimetal NiTe2 and shows that a discrepancy between theory and experiment in one surface state can be explained if the surface feels stronger electron-electron repulsion than the bulk. The finding suggests that surface-specific correlation effects matter for predicting topological surface states in this material family.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The SS2 quantitative match relies on a surface U = 5 eV that is fitted, not derived; the ab initio cRPA U = 2.61 eV for the monolayer is never tested, so 'only by incorporating surface correlation' is not quantitatively established.","rationale":"The central claim is that an accurate description of the topological surface state SS2 in NiTe2 is obtained only by incorporating surface electronic correlation. The quantitative evidence is the ~100 meV shift of the SS2 Dirac crossing in DFT+U with U = 5 eV applied to surface Ni atoms, which brings theory into agreement with ARPES (-1.42 eV). For this to be convincing, the value of U must be justified independently of the target experimental quantity. The paper provides cRPA values of 2.17 eV (bulk) and 2.61 eV (monolayer), showing a modest correlation enhancement at the surface, but the chosen U = 5 eV is nearly twice the monolayer value. Crucially, the authors explicitly state in the SM that they could not compute U for the actual slab surface due to high computational cost, so U = 5 eV is an adjustable parameter selected to match the experiment. The fact that applying the same U to all Ni atoms gives SS2 at -1.50 eV (an overcorrection) further shows that the surface-only U = 5 eV is finely tuned rather than derived. This makes the 'only by incorporating surface correlation' claim quantitatively unproven, although the direction of the shift is consistent with enhanced surface correlation. The proposed concrete test with U = 2.61 eV would show whether an ab initio motivated surface U reproduces the experimental crossing. If it does not, the paper should either derive a surface-specific U or present its conclusion as qualitative rather than quantitative. The Fermi-level alignment concern raised by the reader is also valid but secondary; even with perfect alignment, the U choice remains unresolved. Therefore the conditional verdict is appropriate and no change is needed.","tokens_in":23879,"tokens_out":5477,"duration_ms":64110,"concrete_test":"Compute the slab band structure with DFT+U using U = 2.61 eV (the cRPA monolayer value) applied only to surface Ni atoms, with all other parameters unchanged. If the SS2 crossing moves by less than ~70 meV and remains >50 meV above -1.42 eV, the match at U = 5 eV is due to an inflated U, and the conclusion that surface correlation alone resolves the discrepancy would need to be weakened.","verdict_should_be":"UNCHANGED","load_bearing_attack":"To support the claim that accurate surface states require surface electronic correlation, the paper compares the SS2 Dirac crossing at -1.42 eV (ARPES) with -1.30 eV (DFT) and shows that DFT+U with U = 5 eV on surface Ni atoms shifts the crossing by ~100 meV to match. The load-bearing assumption is that U = 5 eV is a legitimate surface correlation parameter. However, the paper's own cRPA estimate is 2.17 eV (bulk) and 2.61 eV (monolayer); U = 5 eV is roughly double the monolayer value. The SM states that computing U for surface and bulk Ni atoms in the 10-layer slab is 'forbiddingly expensive' and beyond their resources, so U = 5 eV is selected post hoc to reproduce the experimental position. This is not a prediction. Moreover, DFT+U with the same U = 5 eV applied to all Ni atoms gives SS2 at -1.50 eV (SM, Fig. S7e), so the surface-only U = 5 eV is precisely tuned. Without a demonstration that a physically derived U for the actual surface yields the same ~100 meV shift, the central claim is not quantitatively supported.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents an ARPES and DFT study of the type-II Dirac semimetal NiTe2, focusing on the topological surface state SS2 that forms a Dirac-like crossing at the Γ point. The authors report that ARPES places this crossing at -1.42 eV, whereas their slab DFT calculation gives -1.30 eV, a discrepancy of about 100 meV. They show that applying DFT+U with U = 5 eV to the surface Ni atoms only shifts SS2 downward by roughly 100 meV, producing a one-to-one match with experiment, while leaving other surface states largely unchanged. They also argue that bulk electronic structure is well described by plain DFT, that bulk correlation is unimportant, and that Te vacancies produce an upward shift and therefore cannot explain the observed downward shift. The central claim is that an accurate description of the topological surface states requires incorporating surface electronic correlation.","tokens_in":24205,"tokens_out":2945,"duration_ms":37645,"significance":"If the central claim were quantitatively established, this would be a valuable result: it would identify a concrete case where surface-enhanced electron correlation shifts a topological surface state by a measurable amount, and it would provide a route to resolving a known discrepancy between theory and ARPES for NiTe2. The paper contains high-quality experimental data, a careful bulk DFT comparison, and an explicit attempt to rule out the vacancy alternative. However, the central quantitative conclusion currently rests on a Hubbard U chosen post hoc to match the experimental peak position, and the paper itself acknowledges that a first-principles surface U was not computed. The comparison between slab theory and experiment also lacks a discussion of Fermi-level alignment and slab-size convergence. These gaps mean the significance is conditional: the paper demonstrates a plausible mechanism, but it does not yet provide the parameter-free test that would justify the abstract's strong wording.","major_comments":[{"comment":"The surface Hubbard U = 5 eV is selected post hoc to reproduce the experimental SS2 position. The cRPA values quoted in the paper are 2.17 eV (bulk) and 2.61 eV (monolayer), and the Supplementary Material states that computing U for surface and bulk Ni atoms in the 10-layer slab is 'forbiddingly expensive' and beyond the authors' resources. Thus the ~100 meV shift is not a prediction from a derived parameter; it is a fit. The abstract's claim that an accurate description is obtained 'only by incorporating surface electronic correlation' is not quantitatively supported by this procedure. The authors should either compute a surface U from first principles (e.g., cRPA or DFPT for the slab, possibly with a reduced model), or explicitly treat U as an adjustable parameter and present the resulting uncertainty, including showing the SS2 position for U = 2.61 eV and for a range of U values.","section":"§3(d) and SM Sec. I"},{"comment":"The comparison between slab DFT and ARPES relies on an implicit alignment of the slab Fermi level with the experimental Fermi level, but the paper never states how the slab E_F is set (e.g., from the electron count, from a bulk reference, or from a band feature) nor what uncertainty this introduces. Slab eigenvalues can shift by tens of meV with slab thickness and vacuum size, and the 10-layer slab used here is not demonstrated to be converged. The 100 meV discrepancy that U = 5 eV is tuned to remove could be partly or wholly a slab finite-size artifact. Please provide the Fermi-level alignment procedure, a slab-thickness convergence test (or an estimate of its error), and a statement of the combined theoretical uncertainty on the SS2 crossing energy.","section":"§3(d) and Fig. 3"},{"comment":"The vacancy exclusion argument is based on calculations at 6% (VCA) and 11% (supercell) Te vacancy concentrations, which give upward shifts of ~200–230 meV for SS2, whereas the measured bulk vacancy concentration is ~1.5%. The paper does not demonstrate that the shift scales linearly or that the direction of the shift persists at 1.5%, so the extrapolation is an unsupported assumption. Since ruling out vacancies is part of the argument that correlation is the relevant mechanism, this gap should be addressed, for example by a lower-concentration supercell calculation (if feasible) or by an explicit discussion of the expected concentration dependence.","section":"SM Sec. VI"}],"minor_comments":[{"comment":"The title contains a typo: 'se mimetal' should be 'semimetal'.","section":"Title page"},{"comment":"The sentence 'The band structure of surface with Te vacancy matches very well with that of the pristine surface, except for a shift of all the bands related surface states towards EF suggesting that the vacancy induced effects are very similar to that observed using the VCA method' is a run-on; consider splitting it for clarity.","section":"SM Sec. VI"},{"comment":"'Nickle' should be 'Nickel' in the figure caption.","section":"Fig. S1 caption"},{"comment":"'configiration' should be 'configuration'.","section":"SM Sec. I"},{"comment":"The label 'U = 0' in the first panel of Fig. 3(d) is slightly ambiguous because the panel refers to plain DFT; consider labeling it 'DFT (U = 0)' or 'PBE'.","section":"Fig. 3(d)"}],"recommendation":"major_revision","confidential_remarks":"The paper is potentially interesting, but the central quantitative claim is currently a fit rather than a falsifiable prediction. In my view, the authors need to either derive the surface U or substantially soften the abstract and conclusions. The Fermi-level alignment issue is another important gap that should be fixed before publication. I would not reject outright, because the experimental data and the qualitative trend of the U dependence are valuable and the manuscript's scope can accommodate the necessary revisions."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this paper has a solid experimental core and a plausible new idea, but the central claim—that surface electronic correlation is required to describe the SS2 surface state—is not quantitatively established. The match at U = 5 eV is fitted, not derived.\n\nWhat's new: the surface-restricted DFT+U calculation. Prior ARPES works (Refs 30, 32, 36) already had SS2 at -1.42 eV and DFT at -1.30 eV. The authors show that applying U to surface Ni atoms only shifts SS2 down by ~100 meV at U=5 eV, while leaving SS0/SS0' essentially unchanged, giving one-to-one agreement with their ARPES data. That's a clean demonstration that a surface-localized correlation term can do the job.\n\nWhat's good beyond that: the bulk comparison is careful—PBE matches the bulk valence band and FS; DFT+U and hybrid functionals all worsen the bulk description, so the bulk really does look weakly correlated. The vacancy check is a nice negative control: surface Te vacancies shift SS2 upward, opposite to experiment, so defects are unlikely to be the cause. The wavefunction analysis showing SS2 has substantial Ni 3d character, while SS0/SS0' do not, gives a physically sensible reason why correlation would selectively affect this state.\n\nThe soft spots are the load-bearing ones. U = 5 eV is not derived; their own cRPA gives 2.17 eV bulk and 2.61 eV monolayer, and the SM says computing the actual surface U in the slab is too expensive. So U = 5 eV is chosen to reproduce the measured position. That's fitting, not predicting. The SM also shows that applying U=5 eV to all Ni atoms pushes SS2 to -1.50 eV, overshooting, which means the surface-only restriction is doing real work and is precisely tuned. Two smaller issues: no error bars on the -1.42 eV crossing, and no statement of how the slab Fermi level was aligned, so the comparison rests on an unquantified few-tens-of-meV uncertainty.\n\nI don't think the paper's claim is wrong—the cRPA trend and the d-character argument point in the same direction. But the abstract's 'only by incorporating surface electronic correlation' is too strong for what is currently a parameter-scan demonstration.\n\nFor whom? Researchers working on NiTe2 and TMDC surface states, and people interested in surface-specific +U practice. It deserves a serious referee: an editor should send it out, and a referee should ask for the U=2.61 eV test, a derived surface U if feasible, the slab E_F reference, and quantification of the ARPES peak position. Without those, it's a plausible hypothesis rather than a demonstrated mechanism.","headline":"Plausible attribution of the SS2 mismatch to surface Ni 3d correlation, but the U=5 eV match is fitted rather than derived; worth refereeing, not desk-rejecting.","tokens_in":24752,"tokens_out":3823,"would_cite":false,"duration_ms":41130,"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":"The paper claims that topological surface states of NiTe2 are only matched to ARPES when surface electronic correlation is included, shifting the SS2 Dirac crossing by about 100 meV.","keywords":["type-II Dirac semimetal","NiTe2","topological surface states","surface electron correlation","DFT+U","ARPES","Hubbard U","transition metal dichalcogenides"],"falsifier":"Repeat the surface DFT+U calculation with $U$ set to the cRPA monolayer value of 2.61 eV in a slab that is converged with thickness: if the SS2 crossing still sits about 100 meV above the measured $-1.42$ eV, or if a measurement on a surface whose Ni $3d$ occupation is independently modified does not move the crossing as $U$ predicts, the surface-correlation explanation is contradicted.","tokens_in":23684,"feed_emoji":"🔬","tokens_out":6512,"duration_ms":68442,"temperature":0.7,"pith_summary":"This paper tries to establish that the missing ingredient in theoretical descriptions of the topological surface states of the type-II Dirac semimetal NiTe$_2$ is electron correlation localized at the surface, not in the bulk. Angle-resolved photoemission spectroscopy places the Dirac-like conical crossing of the surface state SS2 at about $-1.42$ eV below the Fermi level, while a plain DFT slab calculation puts it near $-1.30$ eV. Adding a Hubbard $U=5$ eV interaction to the $3d$ orbitals of surface Ni atoms only shifts the crossing by roughly 100 meV and reproduces the ARPES spectra one-to-one, while leaving the bulk bands and the near-$\\bar{\\Gamma}$ surface states SS0 and SS0$'$ essentially unchanged. A sympathetic reader would care because this separates surface from bulk correlation and says that band-topology calculations for this material class need to treat the surface as a distinct correlated subsystem.","feed_headline":"Surface nickel repulsion shifts NiTe2's Dirac crossing 100 meV","feed_subtitle":"Adding a Hubbard U only to surface Ni atoms maps the ARPES Dirac crossing exactly, while plain DFT sits 100 meV off.","key_machinery":"The load-bearing object is the topological surface state SS2, a Dirac-like conical crossing at the $\\bar{\\Gamma}$ point about 1.4 eV below the Fermi level, formed inside a parity-inverted gap from strongly hybridized Ni $3d$ and Te $5p$ orbitals. The argument runs on a slab calculation that switches on an on-site Hubbard $U$ only for the $3d$ electrons of the surface Ni atoms, leaving bulk Ni and all Te atoms at their DFT level. The paper demonstrates that increasing this surface $U$ moves SS2 monotonically downward while leaving the surface states SS0 and SS0$'$ near the Fermi level essentially fixed, which is what isolates surface correlation as the mechanism rather than a global band shift.","core_discovery":"The central claim is that an accurate description of topological surface states in NiTe$_2$ requires surface electronic correlation, and that DFT+U applied to the surface Ni atoms alone provides it. The bulk electronic structure, including the type-II Dirac point at about 55 meV below the Fermi level along $\\Gamma$--A, is already well captured by plain DFT; the paper shows that DFT+U, SCAN, and HSE06 all shift the bulk Ni $3d$ states away from the energy where they are observed, confirming that bulk correlation is weak. The decisive evidence is the surface state SS2, a Dirac-like conical crossing inside a parity-inverted gap of hybridized Ni $3d$--Te $5p$ character: ARPES places the crossing at $-1.42$ eV, plain DFT slab calculations at $-1.30$ eV, and DFT+U with $U=5$ eV on surface Ni atoms shifts it by about 100 meV into one-to-one agreement, while neither chalcogen vacancies nor bulk $U$ reproduce the experimental shift.","pith_inferences":["The paper's quantitative case leans on a surface $U=5$ eV, nearly twice the cRPA monolayer value of 2.61 eV that it cites; if the mechanism is correct, real surfaces screen far more weakly than a free monolayer, a claim that could be tested by comparing surfaces prepared under different conditions.","A natural extension the authors do not pursue is to check whether the crossing position tracks the surface Ni $3d$ occupation predicted by the $U$-dependent calculation, which would directly tie the shift to charge redistribution rather than to a one-electron potential shift.","The same surface-state sensitivity to $U$ suggests that intercalation or strain, both of which alter surface screening, should move SS2 measurably; existing strain studies could be reanalyzed for this correlation effect.","The Fermi-level alignment between the 10-layer slab and the ARPES data is assumed rather than demonstrated; a slab-thickness convergence study of the SS2 crossing position would separate finite-size errors from the correlation shift the paper attributes to $U$."],"forward_implications":["If the claim is right, theory–experiment comparisons for topological surface states in the 1T-MX$_2$ family must treat the surface as a separately correlated subsystem instead of applying a single bulk correlation scheme.","The energy of the SS2 crossing becomes a quantitative probe of the effective surface Hubbard $U$, since it moves monotonically with $U$ while other surface features do not.","Predictions of surface-state dispersion and topology that rely on plain DFT slab or Green's-function calculations will be systematically off in energy for states with strong $3d$ character, even when their qualitative topology is correct.","Chalcogen vacancies can be ruled out as the origin of the experimental shift, because the paper finds they move SS2 toward the Fermi level by about 200 meV, opposite to the observed discrepancy.","Bulk correlation treatments (DFT+U, SCAN, HSE06) are all inconsistent with the measured bulk valence band, so any future many-body description of NiTe$_2$ must be surface-specific rather than applied to the whole crystal."],"supporting_citations":[{"why":"Provides the earlier observation of bulk states and spin-polarized topological surface states in NiTe2, giving the SS2 crossing context that this paper extends.","marker":"[30]"},{"why":"Documents the same ~100 meV discrepancy between the ARPES SS2 crossing and DFT calculations, the baseline this paper sets out to explain.","marker":"[32]"},{"why":"Shows Green's-function surface band structure is insensitive to strain in NiTe2, motivating the slab DFT+U approach used here.","marker":"[34]"},{"why":"Supplies the slab-configuration DFT calculation that gives better SS2 agreement in Pd/Pt-based systems, the methodological precedent for the slab approach.","marker":"[35]"},{"why":"Reports the ARPES-measured SS2 crossing at about -1.42 eV in NiTe2, one of the experimental anchors for the energy comparison.","marker":"[36]"},{"why":"Provides the cRPA screened Hubbard U values for bulk (2.17 eV) and monolayer (2.61 eV) NiTe2 used to argue surface correlation is enhanced.","marker":"[19]"},{"why":"Gives GW-level many-body results showing a type-II Dirac cone is preserved in NiTe2 but not PtSe2/PtTe2, framing the correlation question for this material.","marker":"[23]"}],"fun_headline_variants":["Surface Hubbard U pins NiTe2 Dirac crossing to ARPES","NiTe2 surface states demand correlation to match ARPES","100 meV shift: surface U fixes NiTe2 topological states","Surface-only DFT+U resolves NiTe2 Dirac surface states","NiTe2: strong surface correlation explains ARPES"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The comparison hinges on aligning the slab Fermi level with the experimental Fermi level to within about 50 meV, and on accepting the surface $U=5$ eV, which is roughly twice the cRPA monolayer value, as physically justified rather than a parameter tuned to match the measurement.","fun_headline_variants_meta":{"raw":{"variants":["Surface Hubbard U pins NiTe2 Dirac crossing to ARPES","NiTe2 surface states demand correlation to match ARPES","100 meV shift: surface U fixes NiTe2 topological states","Surface-only DFT+U resolves NiTe2 Dirac surface states","NiTe2: strong surface correlation explains ARPES"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000648,"raw_usage":{"total_tokens":2981,"prompt_tokens":958,"completion_tokens":2023,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":574,"completion_tokens_details":{"reasoning_tokens":1936}},"tokens_in":574,"tokens_out":2023,"duration_ms":19288,"temperature":1.0,"reasoning_tokens":1936,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T04:22:12.184162+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Repeat the surface DFT+U calculation with $U$ set to the cRPA monolayer value of 2.61 eV in a slab that is converged with thickness: if the SS2 crossing still sits about 100 meV above the measured $-1.42$ eV, or if a measurement on a surface whose Ni $3d$ occupation is independently modified does not move the crossing as $U$ predicts, the surface-correlation explanation is contradicted.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the earlier observation of bulk states and spin-polarized topological surface states in NiTe2, giving the SS2 crossing context that this paper extends."},{"cited_title":"Zheng, R","cited_arxiv_id":null,"evidence_quote":"Shows Green's-function surface band structure is insensitive to strain in NiTe2, motivating the slab DFT+U approach used here."},{"cited_title":"Settembri, F","cited_arxiv_id":null,"evidence_quote":"Supplies the slab-configuration DFT calculation that gives better SS2 agreement in Pd/Pt-based systems, the methodological precedent for the slab approach."},{"cited_title":"Sharma, A","cited_arxiv_id":null,"evidence_quote":"Provides the cRPA screened Hubbard U values for bulk (2.17 eV) and monolayer (2.61 eV) NiTe2 used to argue surface correlation is enhanced."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives GW-level many-body results showing a type-II Dirac cone is preserved in NiTe2 but not PtSe2/PtTe2, framing the correlation question for this material."}],"review_version":1}