{"id":"55ce34fa-45d1-49db-b725-bd8a95c7fc1c","arxiv_id":"2509.03322","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"A two-step MBE growth method creates a twisted Sb2Te3 homojunction whose Moiré pattern alters the topological surface state and adds new Landau-level features under magnetic field.","lead":"Researchers grew a thin film of the topological insulator Sb2Te3 in which the top layer is rotated relative to the layers beneath it, producing a Moiré pattern. They observed that this pattern changes the electronic states and produces new magnetic-field energy levels, offering a new platform for twistronics in topological materials.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The Moiré interpretation rests on a buried 7/8-QL twist that is inferred from surface topography and a period formula but never directly imaged; without cross-sectional confirmation, the electronic signatures are not uniquely attributable to a twisted interface.","rationale":"The paper's central contribution is a growth method that plausibly creates a twisted Sb2Te3 homojunction, and the STM data provide real supporting evidence: two rotation domains are resolved, the Moiré period (2.24 nm) matches the twist-angle formula for the measured 11.0° rotation, and position-dependent dI/dV spectra in the Supporting Information appear to rule out charged-defect-induced Landau-level splitting. The concern is not that these observations are fabricated or irrelevant; it is that the buried QL7/QL8 interface—the actual location of the twist—is never directly imaged or diffractively confirmed. Because the Dirac-point shift and the new Landau-level peaks are all attributed to this buried interface, the structural premise is load-bearing. The conditional verdict is appropriate: accept the growth method as a likely contribution, but do not treat the Moiré topological surface state as established until the interface stacking is directly verified. This matches the reader's weakest-assumption identification.","tokens_in":7986,"tokens_out":8653,"duration_ms":80062,"concrete_test":"Prepare a cross-sectional lamella from the sample in Fig. 2a across the Moiré region and perform atomic-resolution HAADF-STEM. From the atomic-column positions in the seventh and eighth QLs, directly extract the relative in-plane rotation at the buried interface. If the interface exhibits a uniform twist of 11.0° ± 1° over the imaged region, the structural premise is confirmed; if instead the twist is nonuniform, the interface is incoherent, or a different stacking is found, the central Moiré-surface-state interpretation is falsified. As a complementary check, record the Fourier transform of the STM topograph inside the Moiré region and require that both the atomic-lattice spot and the Moiré spot yield the same twist angle.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that a Moiré topological surface state is formed by a twisted topmost Sb2Te3 layer depends on a structural premise that is established only indirectly. The 11.0° relative rotation is measured from atomically resolved images of the two normal domains (Fig. 2b,c), and the observed 2.24 nm period is then matched to the formula λ_m = a/√2(1−cosθ) with a = 0.425 nm. This match is a necessary consistency check, but it is not a direct determination: the same surface period could in principle arise from a strain-modulated top layer, a reconstructed interface, or a domain boundary with a different local stacking, and no cross-sectional or diffraction data show the buried QL7/QL8 interface where the twist is supposed to live. Every electronic signature—the Dirac-point shift (85 to 121 meV in Fig. 2e,f) and the extra Landau-level peaks at B ≥ 6.3 T in Fig. 3c—is interpreted as a consequence of this buried twist. If the interface is not a single coherent 11° twist, those interpretations lose their foundation. The paper would be materially strengthened by direct visualization of the stacking across the interface.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports the fabrication of a Moiré superlattice on the surface of the topological insulator Sb2Te3 using a two-step molecular beam epitaxy growth method. The authors argue that the topmost quintuple layer of one rotational domain overgrows a neighboring domain, creating a twisted homojunction whose surface exhibits a 2.24 nm Moiré period. Scanning tunneling microscopy and spectroscopy reveal a Dirac-point shift between the Moiré and normal regions, and Landau-level spectra measured under magnetic fields up to 10.5 T show additional peaks in the Moiré region. These data are modeled with a two-Dirac-cone Hamiltonian, yielding a satellite Dirac cone with an enhanced Landé g factor. The paper claims this is the first experimental realization of a Moiré topological surface state and suggests the growth method is broadly applicable to other van der Waals materials.","tokens_in":8238,"tokens_out":7518,"duration_ms":73145,"significance":"If the structural and electronic interpretations are correct, this work would be an important experimental milestone: it would demonstrate a route to Moiré engineering on topological insulator surfaces using epitaxial growth, avoiding the contamination and interface defects associated with manual stacking. The two-step growth method itself is plausible and potentially transferable, and the quantitative match between the measured Moiré period and the twist-angle formula is a genuine strength. The Landau-level data are extensive and the comparison between normal and Moiré regions is a useful approach. However, the central electronic-state claim relies on a five-parameter fit without reported uncertainties, and the structural identification of a buried twist is inferred rather than directly imaged. These limitations currently temper the impact of the result.","major_comments":[{"comment":"The claim that the Moiré pattern originates from a coherent 11.0° twist between the topmost quintuple layer and the underlying 7-QL film is inferred from atomic-resolution images of two normal domains and from the agreement of the measured 2.24 nm period with the twist-angle formula. The buried interface itself is never directly imaged, and the same surface period could in principle arise from strain modulation, an interfacial reconstruction, or a local stacking variation at the domain boundary. Because every electronic signature in the paper is interpreted through this structural premise, the authors should provide direct evidence for the twisted interface (e.g., cross-sectional STM or TEM, LEED, or atomically resolved imaging across the boundary that shows both lattices coexisting at the same location) or explicitly rule out alternative origins for the periodic modulation.","section":"Figure 2 and accompanying text (structural determination)"},{"comment":"The two-Dirac-cone model introduces five free parameters per Moiré region (ED1, vF1, ED2, vF2, and g), and the manuscript does not report the number of Landau-level peaks used, the fit uncertainties, or a quantitative comparison with the one-cone model. With five adjustable parameters, a good fit to a limited set of peaks is not a strong validation of the model. The statement that the one-cone model 'clearly fails' should be supported by a figure and goodness-of-fit metrics. Furthermore, the fitted g values vary from 1.06 to 8.50 meV/T with no monotonic trend versus the Moiré period, suggesting that this parameter may be absorbing unrelated physics rather than representing a robust Zeeman term.","section":"Section 4, Eqs. (2)-(4) and Table 1"},{"comment":"The authors explain the appearance of the new Landau-level features at B ≥ 6.3 T by the cyclotron orbit fitting within the 19 nm effective radius of the Moiré region. This finite-size threshold indicates that the observed peaks may be quantum-confinement states of a finite Moiré patch rather than intrinsic Landau levels of a periodic superlattice. The text should clarify whether the two-Dirac-cone model is intended to describe an infinite periodic Moiré system or a finite confined region, and if the latter, how the model accounts for the boundary conditions.","section":"Section 3, paragraph on the 6.3 T onset"},{"comment":"The reported Dirac-point shift from 85 to 121 meV is presented as evidence that the Moiré pattern modifies the electronic structure, but the ED1 values for the other Moiré regions in Table 1 are 153-160 meV, and no statistics or error bars are provided for any of the extracted energies. A shift of this magnitude could also result from local doping variations or tip-induced band bending. The authors should report measurements on multiple normal regions and compare the extracted ED1 and vF1 values with theoretical predictions for a twisted topological insulator surface (e.g., Refs. 14 and 15) to establish that the observed changes are attributable to the Moiré potential.","section":"Section 3, Fig. 2e-f and Table 1"}],"minor_comments":[{"comment":"The formula is written as λ_m = a/√2(1−cos θ), which is ambiguous and would give an incorrect value if read literally; it should be λ_m = a/√[2(1−cos θ)].","section":"Section 2, Moiré period formula"},{"comment":"The caption begins with 'Figure 3. Figure 3.'; the duplication should be removed.","section":"Figure 3 caption"},{"comment":"Reference 20 contains a typo in the journal name: 'Phys. Rev. ett.' should be 'Phys. Rev. Lett.'.","section":"References"},{"comment":"The g-factor values are given in meV/T; the authors should state the conversion to a dimensionless g factor or define an effective g value to make the magnitude of the enhancement more transparent.","section":"Table 1"},{"comment":"The claim that the two-step growth method 'can be widely applied to other van der Waals materials' is not supported by any demonstration in this manuscript; it should be framed as a proposal rather than an established feature.","section":"Section 5 (summary)"},{"comment":"The statement that 'the height of 1 QL of Sb2Te3 in the Moiré region is slightly higher than that of the normal region' is not quantified; a line-profile height comparison should be included in the Supporting Information.","section":"Section 3, height measurement"}],"recommendation":"major_revision","confidential_remarks":"This is a potentially impactful result from an experienced group, and the growth method is novel. However, the structural identification of the buried twist and the statistical justification of the two-Dirac-cone fit are load-bearing and need to be substantially strengthened. The paper would be suitable for publication after major revision, provided the authors add direct or at least more conclusive evidence for the twisted interface and a more rigorous analysis of the Landau-level fits, including uncertainties and comparison with alternative models."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nThis paper is worth your time. It reports the first controlled Moiré superlattice on a topological insulator surface, and the most solid part is the growth method. The two-step MBE approach—low-temperature growth to lock in rotational domains, then higher-temperature growth to make the top layer overgrow the domain boundary—is simple, clever, and apparently works. The STM evidence is internally consistent: two adjacent normal domains differ by 11.0 ± 0.3°, and the Moiré period (2.24 nm) matches the twist formula with the known Sb2Te3 lattice constant. The Dirac-point shift between normal and Moiré regions is a clean zero-field spectroscopic difference, and the Landau-level data are high quality. The extra peaks in the Moiré region at B ≥ 6.3 T are a real effect, and the authors do a decent job ruling out charged defects with position-dependent spectra.\n\nThe soft spots are in the interpretation, not the fabrication. The twist is inferred from surface atomic images and the Moiré period, not from direct cross-sectional imaging of the buried 7/8-QL interface. That is a fair request, though not a fatal flaw: a strain-modulated top layer or a different local stacking could in principle produce the same surface period, but the coincidence with the measured domain rotation makes the twist interpretation the natural one. A cross-section or diffraction measurement would settle it.\n\nMore concerning is the electronic-state claim. The two-Dirac-cone model fits the Landau fans with five free parameters per Moiré region, and the extracted g factor varies by roughly a factor of eight across the four regions (1.06 to 8.50 meV/T) with no error bars. The satellite cone is never observed directly—no QPI, no ARPES—and the paper's phrase 'established Moiré topological surface state' is stronger than the data justify. The fit is suggestive, not conclusive. I would want uncertainty quantification, a microscopic justification for the g factor, and an independent probe of the second cone before believing that part.\n\nBottom line: the growth method is the lasting contribution and will be useful to the twistronics community. The electronic interpretation needs more work and more careful language. This deserves a serious referee, but the electronic claims should not survive in their current form.\n\nBest,\n[You]","headline":"A clever two-step MBE growth method likely creates a twisted Sb2Te3 homojunction with a real Moiré pattern, but the electronic-state evidence is overinterpreted.","tokens_in":8877,"tokens_out":4008,"would_cite":true,"duration_ms":36622,"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":"A two-step MBE growth produces a Moiré superlattice on the Sb2Te3 topological surface state, evidenced by a shifted Dirac point and Landau levels that require a second Dirac cone.","keywords":["Moiré superlattice","topological insulator","Sb2Te3","molecular beam epitaxy","two-step growth","Landau levels","Dirac cone","twistronics"],"falsifier":"Measure the twist angle of the buried interface directly — for example by electron diffraction or cross-sectional transmission electron microscopy on identically grown films — and confirm both the 11.0° rotation and the 2.24 nm period. If the buried interface is not a coherent twist, or if the extra Landau-level peaks do not follow the two-Dirac-cone fan when the period is known independently, the central claim fails.","tokens_in":7761,"feed_emoji":"🌀","tokens_out":11727,"duration_ms":96970,"temperature":0.7,"pith_summary":"This paper is trying to establish that a Moiré superlattice — a periodic interference pattern created by stacking two atomic lattices with a small twist — can be grown directly on the surface of a topological insulator, something theory predicts but experiments had not realized. The route is a two-step molecular beam epitaxy growth: a low-temperature step forms Sb2Te3 films with rotational domains, and a high-temperature step lets the top quintuple layer overgrow the domain boundary with an ~11° twist, producing a 2.24 nm Moiré period. The evidence is spectroscopic as well as structural: the surface Dirac point shifts from 85 to 121 meV in the Moiré region, and under magnetic fields new Landau-level peaks appear above 6.3 T that a single Dirac cone cannot explain. A two-Dirac-cone model, with the second cone carrying an enhanced Zeeman term, fits the fans for four Moiré periods. If the interpretation is right, topological-insulator twistronics becomes accessible by growth rather than by manual stacking, and the method should transfer to other van der Waals materials.","feed_headline":"Moiré pattern forms on a topological surface state","feed_subtitle":"Twisted Sb2Te3 top layer shifts the Dirac point and adds a satellite cone to the Landau fan.","key_machinery":"The structural machinery is the Moiré interference itself, quantified by the twist-angle relation $\\lambda_m = a/\\sqrt{2(1-\\cos\\theta)}$, which converts the measured 11.0° rotation of the top quintuple layer into a predicted 2.21 nm period that matches the observed 2.24 nm. The electronic machinery is a two-Dirac-cone model. The first cone is the ordinary topological surface state with Hamiltonian $H_1 = v_1(\\sigma_x\\pi_y - \\sigma_y\\pi_x)$ and Landau levels $E_{n,1}=E_{D1}+\\mathrm{sgn}(n)v_{F1}\\sqrt{2e\\hbar|n|B}$. The second cone adds a Zeeman term, $H_2 = v_2(\\sigma_x\\pi_y - \\sigma_y\\pi_x) + gB\\sigma_z$, giving $E_{n,2}=E_{D2}+\\mathrm{sgn}(n)\\sqrt{2v_{F2}^2e\\hbar|n|B + g^2B^2}$ and a zeroth level $E_{0,2}=E_{D2}-gB$; it is this Zeeman term that produces the extra Landau-level features and the nonzero slope of the zeroth level versus field. The model does the work of explaining why one Dirac cone cannot account for the measured fans, and it yields the fitted Fermi velocities, Dirac-point offsets, and the enhanced $g$ factor.","core_discovery":"The central claim is that a twisted top layer of Sb2Te3 produces a genuine Moiré topological surface state with measurable electronic consequences. In the grown films, two rotational domains differ by 11.0° ± 0.3°, and the eighth quintuple layer extends over the seventh to create a Moiré pattern whose 2.24 nm period matches the twist-angle formula. Scanning tunneling spectroscopy shows the Dirac point moves from 85 to 121 meV on the Moiré region. Under magnetic fields up to 10.5 T, the Moiré region develops Landau-level peaks absent on pristine Sb2Te3, beginning at 6.3 T where the cyclotron orbit first fits inside the ~19 nm Moiré patch; position-dependent spectra rule out charged-defect splitting. The fans from four different Moiré periods (2.08, 2.17, 2.24, 2.41 nm) are fit by a two-Dirac-cone model — one cone for the original surface state and one satellite cone with a magnetic-field-dependent Zeeman energy — while the one-cone model fails, and the fitted g factor of the satellite cone is roughly an order of magnitude larger than that of the original cone.","pith_inferences":["A testable extension follows from the size argument: if the 6.3 T onset is really a cyclotron-orbit effect, growing Moiré domains of different radii should move the onset as $B \\propto 1/r^2$; if it does not, the onset has a different origin.","An independent structural probe of the buried twist interface would separate the Moiré interpretation from a strain or polycrystalline-overlayer explanation, since all current angle evidence comes from top-surface atomic images.","The two-Dirac-cone fit treats the second cone as an independent object, but a Moiré band with mini-Brillouin-zone effects could produce fan deviations at low Landau index; comparing the four fitted $g$ factors against twist angle would test whether the Zeeman term is really Moiré-induced.","If the $g$ factor is Moiré-induced, its value should depend on period or twist angle; the four fitted values (1.06, 2.23, 4.25, 8.50 meV/T) vary widely, so a systematic twist-angle dependence would be a direct way to check the mechanism."],"forward_implications":["The two-step growth method, being an in-vacuum growth route rather than manual stacking, can be applied to other van der Waals materials to build Moiré superlattices without contamination or interface defects.","On the topological surface state, the Moiré pattern acts as a band-engineering tool: it shifts the Dirac point and creates a satellite Dirac cone whose Landau quantization is visible in tunneling spectra.","The enhanced g factor of the satellite cone, about an order of magnitude larger than the intrinsic cone's, makes the Moiré region a candidate platform for spin manipulation of topological surface states.","Because the onset of the new Landau-level peaks is tied to the cyclotron orbit fitting in the ~19 nm Moiré patch, larger Moiré domains should show the extra features at lower magnetic fields.","With future doping, the Moiré topological surface state should be able to access the correlated and superconducting phases predicted for this system."],"supporting_citations":[{"why":"Supplies the twist-angle/Moiré-period relation and the Moiré band framework used to predict a 2.21 nm period.","marker":"1"},{"why":"Predicts that a Moiré potential on a topological insulator renormalizes the Fermi velocity and creates satellite Dirac cones, the theory the two-cone fit tests.","marker":"14"},{"why":"Predicts Moiré surface states on topological insulators and enhanced superconductivity, motivating the search for the state realized here.","marker":"15"},{"why":"Theoretical treatment of Moiré Landau fans, providing the expectation of unusual Landau quantization that motivates the magnetic-field study.","marker":"16"},{"why":"Provides the two-step MBE growth precedent for lateral overgrowth that the present method adapts to twisted Sb2Te3.","marker":"19"},{"why":"Supplies the normal-region Landau-level baseline and the critical thickness for Sb2Te3 topological surface states.","marker":"21"},{"why":"Shows a Zeeman effect in topological surface states and separates it from charged-defect Landau-level splitting, supporting the g-factor interpretation.","marker":"25"},{"why":"Supplies the cyclotron-radius formula used to set the 6.3 T onset field for the new Landau-level peaks.","marker":"27"},{"why":"Provides the induced-Zeeman-effect model for Moiré surface states from which the second-cone Hamiltonian with g factor is taken.","marker":"28"}],"fun_headline_variants":["Twisted Top Layer Creates Moiré on Sb2Te3","New Landau Levels from Moiré on Sb2Te3","Moiré Topological Surface State: Fabricated and Probed","Twist Induced Moiré on Topological Insulator"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the topmost layer is a coherent, uniformly twisted Sb2Te3 layer rather than a strained or polycrystalline overlayer; the 11.0° twist angle is inferred from atomically resolved top-surface images and the matching Moiré period, with no independent cross-sectional or diffraction confirmation of the buried interface.","fun_headline_variants_meta":{"raw":{"variants":["Twisted Top Layer Creates Moiré on Sb2Te3","New Landau Levels from Moiré on Sb2Te3","Moiré Topological Surface State: Fabricated and Probed","Twist Induced Moiré on Topological Insulator"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000846,"raw_usage":{"total_tokens":3680,"prompt_tokens":942,"completion_tokens":2738,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":558,"completion_tokens_details":{"reasoning_tokens":2676}},"tokens_in":558,"tokens_out":2738,"duration_ms":20057,"temperature":1.0,"reasoning_tokens":2676,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T16:30:58.878281+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the twist angle of the buried interface directly — for example by electron diffraction or cross-sectional transmission electron microscopy on identically grown films — and confirm both the 11.0° rotation and the 2.24 nm period. If the buried interface is not a coherent twist, or if the extra Landau-level peaks do not follow the two-Dirac-cone fan when the period is known independently, the central claim fails.","supporting_citations":[],"review_version":2}