{"id":"49f2621d-91bb-434a-893f-58f252e9b066","arxiv_id":"1908.03082","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"MoTe2 hosts switchable polar domains and nanoscale 1T'/Td phase domain-wall superlattices with distinct interfacial conductance, potentially related to topological hinge states.","lead":"Using cryogenic electron microscopy and scanning tunneling microscopy, this paper reveals that the semimetal MoTe2 forms polar domains and nanoscale superlattices of phase domain walls at low temperature. The work shows that these domain walls have distinct electronic signatures and can be moved with an electron beam, pointing toward electrically controllable topological interfaces.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The polar-domain claim rests on DF-TEM contrast without direct polarization measurement; the paper's own note concedes atomic-resolution confirmation is missing, so conditional acceptance is appropriate.","rationale":"The paper is a careful experimental study with credible DFT support; the stacking model for 1T'/Td phases is internally consistent, and the cross-sectional HAADF-STEM image of a 1T' twin wall directly shows a Td unit. Credit is due for the in-situ videos, the reversibility arguments against beam damage, and the first-principles energy landscape linking Td and 1T' phases. However, the central novelty is the first observation of polar domains in Td-MoTe2, and the only direct evidence for polarity reversal is two-level DF-TEM contrast in the ab-plane view. The authors themselves state that the absolute polarization direction cannot be identified and that atomic-scale imaging is needed for confirmation. DF-TEM contrast can in principle arise from phase coexistence, diffraction conditions, thickness variations, or strain, and Supplementary Note 2 shows that 1T' remnants are present in the capped cross-section geometry. The phase-domain-wall and conducting-interface claims are built on this same assignment, which makes the missing atomic-resolution stacking determination the most load-bearing concern. The STM hinge-state interpretation is speculative but explicitly hedged, so it is not the primary weakness. The reader's conditional verdict is appropriate: the structural observations are plausible and well illustrated, but the polar-domain identification requires an independent structural check before the headline claim can be accepted as definitive. No stronger objection is warranted.","tokens_in":18285,"tokens_out":3987,"duration_ms":49332,"concrete_test":"Prepare a thin cross-section from the same crystal, identify a bright/dark ab-plane domain pair by DF-TEM at 80 K, then atomically resolve the same domain-wall region by HAADF-STEM or iDPC-STEM along [100]/[010] and determine the interlayer shear sequence (+M-P+ versus -M+P-) on both sides of the boundary; if the sequence does not flip across the contrast boundary, the polar-domain interpretation fails.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The load-bearing premise is that bright/dark ab-plane DF-TEM domains at 80 K are opposite polar variants Td-up and Td-down. The Methods section shows the contrast was obtained with g1+ = ±(1,2,1bar) DF spots, and the main text states that 'the absolute polarization direction cannot be identified in the ab-plane TEM view'; Supplementary Fig. 4 explicitly concedes that the Td-up/Td-down assignment 'needs further confirmation by measurements such as atomic-resolved STEM imaging.' The risk is not merely a labeling issue: because the same specimens retain 1T' phase under mechanical constraints (Supplementary Note 2), and DF-TEM contrast is sensitive to thickness, bending, and the local 1T'/Td phase fraction, the observed two-level contrast could reflect phase coexistence or mixed stacking rather than polarity reversal. The e-beam switching argument and the STM 'conducting interfacial state' interpretation both inherit this ambiguity. No atomic-resolution image of a nominal polar DW is provided, and the HAADF-STEM image in Fig. 2c only demonstrates a Td unit at a 1T' twin wall. Without an independent structural probe of the stacking sequence at a bright/dark boundary, the central claim of polar domains is underdetermined.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This manuscript reports an experimental and theoretical study of domain walls in semimetallic MoTe2. Using in-situ cryogenic dark-field TEM in cross-section and ab-plane geometries, HAADF-STEM, and low-temperature STM, together with DFT calculations, the authors claim the first observation of polar domains of the low-temperature Td Weyl phase (Td↑ and Td↓), abundant nanoscale 1T'/Td phase domain walls forming superlattice-like arrays along the c axis, reversible electron-beam manipulation of polar domains, and distinct local conductance at polar and phase domain walls. The DFT part identifies a centrosymmetric T0 reference structure with two instabilities, maps out an energy landscape with four minima (Td↑, Td↓, 1T'-I, 1T'-II), and proposes a low-energy path for Td↑↔Td↓ switching through a 1T' intermediate. The paper explicitly notes that the absolute polarization direction is not determined in the ab-plane TEM view and that the Td↑/Td↓ assignment needs further confirmation.","tokens_in":18653,"tokens_out":5407,"duration_ms":55478,"significance":"The topic is timely and the potential payoff is high: if the polar and phase domain structures are confirmed, MoTe2 becomes a controllable platform for engineering topological interfacial states, with implications for polar metals and Weyl semimetals. The paper has notable strengths: the DFT calculations are genuinely first-principles, with the Hubbard U and exchange J taken from prior literature rather than fitted to the observed domain patterns; the energy landscape is internally consistent; and the combination of DF-TEM, atomic-resolution HAADF-STEM, and STM provides multiple complementary observations. The group-theoretical family tree linking T0, 1T', and Td phases is elegant and gives testable structural predictions. The main weakness is that the central polar-domain assignment rests on DF-TEM contrast alone, with the absolute polarization direction unmeasured and an explicit admission in the Supplementary Material that atomic-resolution confirmation is still required.","major_comments":[{"comment":"The central claim that the bright and dark ab-plane DF-TEM contrasts correspond to opposite polar variants Td↑ and Td↓ is underdetermined. The main text states that 'the absolute polarization direction cannot be identified in the ab-plane TEM view' and the assignment is made 'for the sake of simplicity'; Supplementary Fig. 4 explicitly says the assignment 'needs further confirmation by measurements such as atomic-resolved STEM imaging.' Because the DF-TEM contrast was obtained with g1± = ±(1,2,1bar) spots (Methods), it is sensitive not only to polarity but also to specimen thickness, bending, and the local 1T'/Td phase fraction; the manuscript itself invokes mechanical constraints to explain retained 1T' phase in the capped cross-section specimens (Supplementary Note 2). No atomic-resolution image of a nominal polar DW is provided. The authors should provide an independent structural probe—for example, atomic-scale STEM across a bright/dark boundary or a direct measurement of the polarization direction—before the existence of polar domains can be considered established.","section":"In-plane view of polar domains and domain walls / Methods / Supplementary Fig. 4"},{"comment":"The STM evidence for 'conducting interfacial states' does not yet support the hinge-state interpretation. The measurements were performed on Fe-doped MoTe2 with ~1.06% Fe impurities, and the manuscript notes that the quasiparticle interference patterns are dominated by Fe-dopant scattering (Supplementary Fig. 6). The first-type DWs show enhanced conductance in the empty state, but this is a local spectroscopic feature that could arise from topography, tip effects, or strain; no transport measurement or direct edge/hinge-state signature is presented. The text appropriately hedges this as a 'possible signature,' but the abstract and title elevate the conducting interfacial states to a headline result. The authors should either present direct evidence for a conducting hinge mode (e.g., spatially resolved spectroscopy with polarization dependence and comparison to theory) or clearly relegate the hinge-state assignment to a conjecture.","section":"Phase domain wall conductance / Fig. 5"},{"comment":"The identification of the superlattice-like modulations as −(Td)m(1T')n− stackings is plausible but not uniquely determined by the data. The thinnest periodicity of ~4 nm is interpreted as six layers, but Fig. 3c lists three possible six-layer configurations; the choice of −(Td)4(1T')2− as 'most likely' is based on energetic and lattice-mismatch reasoning rather than direct atomic-resolution imaging of the modulation. The paper also notes that the periodicity changes within a single twin domain, which is taken as evidence for opposite Td domains, but this inference relies on the same unconfirmed Td↑/Td↓ assignment. An atomic-resolution cross-section image that resolves the stacking sequence across one short-period modulation would remove the ambiguity.","section":"Cross-sectional view of abundant phase domain walls / Fig. 3"}],"minor_comments":[{"comment":"The phrase 'low-temperature(T)' should read 'low-temperature Td' or 'low-temperature (Td)' for clarity.","section":"Abstract"},{"comment":"'combing first-principle calculations' should be 'combining first-principles calculations'.","section":"Results"},{"comment":"The manuscript uses inconsistent spacing in 'e- beam' and 'e-beam'; please standardize.","section":"Throughout"},{"comment":"The sentence 'JEOL-2010F field-emission TEM quipped with a low-T sample stage' contains a typo; 'equipped' is intended.","section":"Methods"},{"comment":"The note that the Td↑/Td↓ assignments are 'only for simplicity and needs further confirmation' is important enough to be stated in the main text rather than only in the Supplementary Material.","section":"Supplementary Fig. 4"},{"comment":"'Figures 4c show four TEM snapshots' should be 'Figure 4c shows four TEM snapshots'.","section":"Fig. 4c caption"},{"comment":"The statement that Weyl points in Td↑ and Td↓ are 'topologically identical' is imprecise; the authors should clarify that the Weyl points have the same locations but opposite chirality.","section":"Discussion"}],"recommendation":"major_revision","confidential_remarks":"The paper is within the journal's scope and the experimental work is substantial. The main concern is the polar-domain assignment: the authors should either provide an independent structural measurement of the polarization direction or atomic-resolution imaging of a nominal polar DW. The STM hinge-state claim should be either strengthened or clearly softened in the abstract and title."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is the first real-space observation of nanoscale 1T'/Td phase boundaries and e-beam-switchable domains in MoTe2, and it deserves a serious referee. The main caveat is honest, and the authors state it themselves: the bright/dark DF-TEM domains are assigned Td↑/Td↓ without measuring the absolute polarization direction.\n\nWhat is genuinely new: the P/M-plus/minus stacking notation, the T0 parent structure, the observation that 1T' twin walls already contain Td units at room temperature, the quasi-periodic 1T'/Td superlattice-like arrays in capped specimens at 80 K, and reversible e-beam domain-wall motion. The DFT energy landscape is a real addition: four minima plus a low-energy path through a 1T' intermediate. The U/J values are taken from earlier work, not fitted to the domain patterns, so circularity is not an issue.\n\nWhat is softer: the central claim that the two DF-TEM contrasts are opposite polar variants rests on indirect evidence. The authors note the absolute polarization direction cannot be identified from the ab-plane view and that atomic-resolved STEM confirmation is needed. The stress-test worry about phase coexistence or thickness contrast is reasonable but not fatal; the contrast reversal between g1+ and g1- imaging and the absence of domains at 300 K strengthen the inversion-related interpretation. Still, no atomic-scale image of a nominal polar DW is shown, so the polar label is a plausible interpretation rather than a direct measurement. The STM 'conducting hinge state' is properly labeled possible; the conductance enhancements are local but not a smoking gun. QPI shows no resolvable difference between domains, which limits the STM-based domain assignment.\n\nThis is a paper for people working on MoTe2, polar metals, or domain walls in topological materials. It is well illustrated, transparent about its caveats, and the microscopy looks carefully done. A referee should push for a direct probe of the stacking sequence across a nominal polar DW and for a clearer separation between hinge-state language and the conductance data, but the core observations deserve publication.","headline":"First real-space images of 1T'/Td domain walls and e-beam-switchable domains in MoTe2, with the polar-domain assignment an honest inference rather than a direct measurement.","tokens_in":19072,"tokens_out":3325,"would_cite":true,"duration_ms":32453,"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":"MoTe2 domain walls move under an electron beam and carry conducting states.","keywords":["MoTe2","Weyl semimetal","polar domain walls","phase domain walls","transmission electron microscopy","scanning tunneling microscopy","higher-order topology","in-situ cryo-TEM"],"falsifier":"Atomic-resolution HAADF-STEM across one of the sharp bright/dark boundaries would settle the matter: if the Mo zigzag chains and Te environment do not reverse their ±c displacement direction across the wall, the Td↑/Td↓ assignment is wrong. Alternatively, if a free-standing, uncapped crystal shows no 1T'/Td coexistence on cooling, the superlattice interpretation of the capped cross-section images would be an artifact of mechanical constraint.","tokens_in":18114,"feed_emoji":"🔬","tokens_out":4989,"duration_ms":48276,"temperature":0.7,"pith_summary":"This paper claims that the layered semimetal MoTe2, in its low-temperature polar Td phase, forms two opposite-polarization variants—Td↑ and Td↓—separated by domain walls, and that these walls coexist with abundant 1T'/Td phase boundaries that stack into nanoscale superlattice-like arrays along the c axis. It further claims that focused electron-beam irradiation at 80 K reversibly drives the polar walls, and that scanning tunneling spectroscopy sees enhanced conductance at the phase walls, a possible signature of conducting hinge states. If true, the significance is that domain walls in a polar Weyl semimetal become self-assembled, addressable interfaces for topological electronic states rather than unwanted defects.","feed_headline":"Polar walls in MoTe2 switch under an electron beam","feed_subtitle":"Cryo-TEM reveals Td↑/Td↓ polar domains and 1T'/Td phase walls, with conductance hints at the walls.","key_machinery":"The argument is carried by a compact stacking notation: each Te–Mo–Te layer is marked P or M for the two orientations of the Te octahedral distortion around the Mo zigzag chains, and each interlayer shear is marked + or − for the direction of Te–Te displacement. The 1T' and Td polymorphs are then just different sequences—1T'-I = +M+P+, 1T'-II = −M−P−, Td↑ = +M−P+, Td↓ = −M+P−—and a domain wall is a place where the sequence changes. This notation makes the microscopic relationship between the phases explicit, and together with the DFT potential-energy landscape (a high-symmetry T0 parent with two instabilities) it explains why polar switching prefers to pass through the nonpolar 1T' phase and why 1T' twin walls carry Td layers.","core_discovery":"The central discovery is the real-space coexistence and control of three kinds of interfaces in one crystal: polar domain walls between Td↑ and Td↓, phase domain walls between Td and the nonpolar higher-order topological 1T' phase, and twin walls of the 1T' phase that themselves contain Td-character layers. In cross-section dark-field TEM, the 1T' twins appear as micrometer-scale stripes, but on cooling the 1T' phase nucleates thin Td platelets that arrange quasi-periodically—the smallest observed period is about 4 nm, six layers—so the crystal becomes a natural superlattice of topological and non-topological units. In the ab-plane view, bright and dark domains at 80 K are assigned to Td↑ and Td↓, and the same walls move reversibly when the electron beam is focused on them. First-principles calculations show all four phases descend from a common centrosymmetric orthorhombic parent T0, with double-well instabilities in the interlayer shear coordinate and the monoclinic angle, and the lowest-energy path between the two polar states runs through the 1T' phase. STM on Fe-doped crystals finds that 1T'/Td walls show orientation-dependent enhanced conductance, which the authors interpret as a possible conducting hinge state.","pith_inferences":["If the e-beam switching works by the DFT-identified low-energy path through the intermediate 1T' phase, then ultrafast optical or electrical excitation that transiently stabilizes the 1T' shear should also toggle the polar state—an extension the paper does not demonstrate.","The same P/M and +/− stacking logic should apply to other layered polar Weyl semimetals such as WTe2, where analogous polar walls and phase walls may exist and could be searched for with the same dark-field TEM contrast.","The assignment of Td↑/Td↓ from bright/dark contrast could be tested directly by atomic-resolution STEM across a single wall; if confirmed, the same contrast could serve as a quick nondestructive readout of polarization in other van der Waals polar metals."],"forward_implications":["Polar walls in Td-MoTe2 are switchable at 80 K with a focused electron beam, with no apparent Te loss or damage, so the domains can be written and erased in a semimetal.","Because Td↑ and Td↓ are related by inversion, their Weyl points sit at the same momenta with opposite chirality; a Td↑/Td↓ wall projects opposite Fermi arcs and is a natural platform for studying Fermi-arc reconstruction.","Phase walls between the Weyl Td phase and the higher-order topological 1T' phase show enhanced, orientation-dependent tunneling conductance, consistent with conducting hinge states at the walls.","The quasi-periodic 1T'/Td stacking at low temperature is a natural superlattice whose period (as small as 4 nm) can be read directly from TEM contrast, giving a real-space handle on topological phase coexistence."],"supporting_citations":[{"why":"Predict the type-II Weyl semimetal phase in Td-MoTe2 and report its electronic signature, establishing why the polar phase is topologically nontrivial.","marker":"16,17"},{"why":"Show that 1T'-MoTe2 is a higher-order topological insulator with protected hinge modes, making the 1T'/Td wall an interface between topologically distinct phases.","marker":"35-37"},{"why":"Characterize Fermi arcs and the type-II Weyl character of MoTe2, providing the band-structure context for interpreting domain-wall conductance.","marker":"40,41"},{"why":"Demonstrates ultrafast interlayer-slide switching in the same material, supporting the idea that interlayer displacement is the key order parameter.","marker":"31"},{"why":"Discusses polar metals and explains why itinerant carriers do not necessarily forbid polar domains, underpinning the possibility of Td↑/Td↓ domains in a semimetal.","marker":"22"},{"why":"Report ferroelectric-like behavior and polar switching in MoTe2, supporting the polar nature of the Td phase.","marker":"38,39"},{"why":"Group-subgroup tables used to establish the common T0 parent and the Pm subgroup linking 1T' and Td.","marker":"57"},{"why":"Shows electrostatic-doping-driven phase transitions in MoTe2, an alternative switching pathway that the authors contrast with their e-beam mechanism.","marker":"45"}],"fun_headline_variants":["Electron beam flips polar domains in Weyl semimetal MoTe2","Polar and phase walls in MoTe2 offer conductive interfaces","MoTe2 makes natural superlattice of polar and phase walls","Cryo-TEM spots switchable domain walls in MoTe2"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The bright/dark contrast in the in-plane images is assigned to Td↑ versus Td↓ domains, but the absolute polarization direction is never measured; if the contrast came from some other structural difference, the central polar-domain claim would weaken.","fun_headline_variants_meta":{"raw":{"variants":["Electron beam flips polar domains in Weyl semimetal MoTe2","Polar and phase walls in MoTe2 offer conductive interfaces","MoTe2 makes natural superlattice of polar and phase walls","Cryo-TEM spots switchable domain walls in MoTe2"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00119,"raw_usage":{"total_tokens":4929,"prompt_tokens":981,"completion_tokens":3948,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":597,"completion_tokens_details":{"reasoning_tokens":3871}},"tokens_in":597,"tokens_out":3948,"duration_ms":28316,"temperature":1.0,"reasoning_tokens":3871,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T14:24:22.022243+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Atomic-resolution HAADF-STEM across one of the sharp bright/dark boundaries would settle the matter: if the Mo zigzag chains and Te environment do not reverse their ±c displacement direction across the wall, the Td↑/Td↓ assignment is wrong. Alternatively, if a free-standing, uncapped crystal shows no 1T'/Td coexistence on cooling, the superlattice interpretation of the capped cross-section images would be an artifact of mechanical constraint.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Demonstrates ultrafast interlayer-slide switching in the same material, supporting the idea that interlayer displacement is the key order parameter."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Discusses polar metals and explains why itinerant carriers do not necessarily forbid polar domains, underpinning the possibility of Td↑/Td↓ domains in a semimetal."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Group-subgroup tables used to establish the common T0 parent and the Pm subgroup linking 1T' and Td."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Shows electrostatic-doping-driven phase transitions in MoTe2, an alternative switching pathway that the authors contrast with their e-beam mechanism."}],"review_version":1}