{"id":"0e7412f9-3da2-4922-a906-17c420360f62","arxiv_id":"2412.16815","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"In the antiferromagnetic state of TbTi3Bi4, band structure reconstructions with wavevectors near (1/3, 0.28, 0) in reciprocal units are attributed to Fermi surface nesting of Tb 5dxz orbitals, providing a mechanism for the 1/3 magnetization plateau.","lead":"Researchers mapped the electronic bands of the magnetic metal TbTi3Bi4 and found that two distinct wavelike distortions appear exactly when it becomes antiferromagnetic. The patterns come from specific Tb orbitals and are linked to the material's unusual 1/3 magnetization step, showing how orbital shape can steer magnetism.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The orbital-selective nesting mechanism is load-bearing but rests on a 4f-less DFT model plus a visual nesting argument; no susceptibility calculation or 4f-inclusive check is provided.","rationale":"Read in good faith: the ARPES data show a reproducible, temperature-dependent reconstruction with onset near TN1, with q vectors matching NPD magnetic peaks and no STM charge order; those observations are strong. The contested part is not the existence of modulations but the mechanism: an AFM order driven by a 1/3 nesting instability of Tb 5dxz. The evidence for that mechanism is a nonmagnetic DFT calculation from which 4f electrons are excluded, plus visually nested Fermi-surface segments. This is partly an external-consensus issue, but the internal weak point is sharper: the calculation does not demonstrate an instability (no χ0(q)), and the orbital projection is computed in a model that omits the ordered moments. The paper itself flags the NPD structure as one plausible solution (Fig. 8), so the plateau model is correspondingly tentative. The proposed test would settle whether the nesting/5dxz claim has quantitative support. If it fails, the title-level 'orbital-selective ... driven by 1/3 nesting' should be softened to an observation of AFM-associated modulations; the central ARPES observation would still stand. The reader's conditional verdict already captures this fragility; my read does not move it to a different verdict.","tokens_in":17749,"tokens_out":6363,"duration_ms":62010,"concrete_test":"Recompute the nonmagnetic DFT band structure from Methods C and evaluate the static Lindhard susceptibility χ0(q) with the Wannier Hamiltonian for q in the (a*, b*) plane; check for local maxima at q1≈0.35a* and q2≈0.28b*. Then repeat the same χ0(q) and Tb-5dxz orbital-weight calculation with Tb 4f electrons treated as valence states (or with a DFT+U correction). If neither the 4f-free nor the 4f-inclusive χ0(q) peaks at the observed wavevectors, or if the 5dxz weight disappears once 4f states are included, the central 'nesting-driven orbital-selective AFM' claim is unsupported; if both persist, the concern is settled.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's core causal claim is not just that band folding appears below TN1—that is well supported—but that the folding is the fingerprint of an AFM order driven by a 1/3 nesting instability of Tb 5dxz orbitals. The load-bearing support for the orbital/nesting half is the nonmagnetic DFT in Methods C, which explicitly treats Tb with nine valence electrons and does not include 4f states. That calculation is used to (a) assign the reconstructed quasi-1D Fermi surface to Tb 5dxz, and (b) identify q1 and q2 as nesting vectors from parallel Fermi-surface segments. Neither step is sufficient: no q-resolved susceptibility (e.g., χ0(q)) is computed, so 'nesting instability' is inferred visually, and the 4f electrons—the same moments that order—are absent from the model that defines the 5d bands and their orbital content. If 4f–5d hybridization or exchange shifts the 5d bands, the orbital projection could change; the polarization-dependent ARPES in Appendix G tests only one β band, not the nesting susceptibility. The NPD model being explicitly non-unique (Fig. 8) is a secondary risk for the 3a plateau mechanism.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports a combined ARPES, neutron powder diffraction, STM, magnetization, and DFT study of the kagome metal TbTi3Bi4. The authors identify two band modulations below the antiferromagnetic transition at TN1 ≈ 20.4 K: q1 ≈ 1/3 a* along the Tb zigzag-chain direction and q2 ≈ 0.28 b* perpendicular to it. They attribute these modulations to Fermi-surface nesting of quasi-1D Tb 5dxz orbitals, propose that the associated band folding produces a Dirac cone at the M1 point signaling a topological phase transition, and connect the resulting ~3a magnetic period to the observed 1/3 magnetization plateau. The experimental observations of band reconstruction, gap openings, and their temperature dependence are presented in detail and supported by NPD showing magnetic peaks at the same wavevector.","tokens_in":17983,"tokens_out":3729,"duration_ms":33275,"significance":"If the central claims hold, this work would provide a rare experimental example of orbital-selective band reconstruction driven by magnetism in a kagome metal, with implications for the LnTi3Bi4 family. The strengths of the paper are the multi-technique approach (VUV and soft-X-ray ARPES, NPD, STM, transport, magnetization), the clear temperature dependence showing that the band folding onsets at TN1, and the identification of two distinct modulation vectors consistent between ARPES and neutron diffraction. The STM exclusion of a charge-density-wave origin is a valuable control. The paper is, however, less convincing in three load-bearing areas: the orbital-selective nesting mechanism relies on a 4f-less DFT calculation, the topological interpretation is inferred from a single Dirac-cone-like crossing without symmetry or surface-state analysis, and the magnetic structure underlying the 1/3 plateau is admitted to be non-unique.","major_comments":[{"comment":"The assignment of the reconstructed quasi-1D Fermi surface to Tb 5dxz orbitals and the identification of q1 and q2 as nesting vectors rest entirely on DFT calculations that exclude Tb 4f electrons (Methods C: 'projector augmented wave potentials with nine valence electrons for Tb (f electrons are not considered)'). Since the 4f electrons are the ordered moments and can hybridize with the 5d states, the orbital projection and the Fermi-surface topology may change below TN1. The authors should provide a quantitative check of the nesting scenario, for example a q-resolved static susceptibility χ0(q) computed from the DFT band structure, or a 4f-inclusive calculation (DFT+U or open-core treatment) showing that the Tb 5dxz character and the nesting vectors are robust. Without such a check, the causal statement in the Discussion that 'the Tb 5dxz orbital plays a direct role in establishing the antiferromagnetic ground state' is not established; the ARPES data support band folding with q1 and q2, but the nesting-driven orbital-selective mechanism remains a plausible interpretation rather than a demonstrated one.","section":"Methods C and Fig. 7"},{"comment":"The claim of a 'symmetry-protected Dirac cone' and a topological phase transition is not supported by the evidence presented. The text states that 'the emergent gapless Dirac point at the BZ boundary must either be protected by magnetic symmetries or originate from surface states induced by bulk band inversions,' and then concludes that either scenario points to an underlying topological state. However, no magnetic-space-group analysis, no calculation of topological invariants, and no surface-state computation are provided to distinguish these alternatives or to establish protection. The abstract and conclusion nonetheless state the topological transition as a finding. The authors should either downgrade the topological claim to a more cautious statement—that a Dirac-cone-like crossing appears below TN1 with an as-yet-undetermined origin—or supply the missing symmetry and topological analysis.","section":"Results, Fig. 4 and Abstract"},{"comment":"The proposed magnetic structure with wavevector (0.35, 0.28, 0) is explicitly admitted to be 'one plausible solution derived from the NPD data; alternative magnetic configurations, e.g. spin helix, however, cannot be fully excluded.' Because the 1/3 magnetization plateau mechanism in Fig. 9(c) relies on a specific ~3a-period moment arrangement along the a-axis, the non-uniqueness of the Rietveld model is a load-bearing uncertainty for that mechanism. The authors should either present additional experimental constraints (e.g., single-crystal neutron diffraction, or field-dependent neutron data) that discriminate the proposed collinear-like structure from a spin helix, or frame the plateau mechanism more explicitly as a speculative proposal whose validity is contingent on the magnetic structure being confirmed.","section":"Fig. 8 and Discussion of the 1/3 plateau"}],"minor_comments":[{"comment":"In the abstract, the phrases 'Dirac cones only at the  point' and 'leading to the emergence of Dirac cones' contain missing symbols (M1 or similar); these should be completed before publication.","section":"Abstract"},{"comment":"The text says TbTi3Bi4 'crystallizes in a centrosymmetric cubic structure (space group Fmmm, No. 69)'; Fmmm is an orthorhombic space group, not cubic. Please correct this.","section":"Methods A"},{"comment":"The caption refers to 'the  pockets' and 'the  point' with missing symbols; the momentum labels (M1, M2, or similar) should be specified for clarity.","section":"Fig. 5 caption"},{"comment":"The quantity 'Peak 3' is introduced in the text before its definition in the caption of Fig. 4(g); consider defining it in the main text where it is first mentioned.","section":"Fig. 4 and Appendix D"},{"comment":"The manuscript cites ref. [17] (arXiv:2405.16831) for recent single-crystal neutron diffraction results showing anisotropic moments along a and b; given the reliance on that result for the plateau mechanism, the authors should state explicitly which quantities are taken from that work and which are measured here.","section":"Introduction and Discussion"}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Read the TbTi3Bi4 paper. My take: the core experimental observation is solid and worth publishing, but the strong interpretive claims—orbital-selective nesting-driven AFM and a topological transition—outrun the evidence.\n\nWhat's new: first ARPES of this compound through TN1. The dual band modulations (q1 ~ 1/3 a*, q2 ~ 0.28 b*) are clearly resolved, temperature-dependent, and track the AFM transition. Gap sizes (77.6 meV along a*, 27 meV along b*) and the NPD wavevector (0.35, 0.28, 0) match the band folding. STM rules out a CDW at that wavevector. Combined with the Rietveld refinement showing magnetic peaks below TN1, it's a coherent package and a genuine advance for the LnTi3Bi4 family.\n\nSoft spots are where they go beyond observation. The mechanism—nesting instability of Tb 5dxz—rests on DFT with 4f electrons excluded, plus a visual identification of parallel Fermi-surface segments. There is no q-resolved susceptibility (e.g., χ0(q)) and no calculation including 4f to check whether hybridization shifts the 5d bands. The polarization-dependent ARPES tests only one β band. So the orbital-selectivity story is plausible but under-supported. The NPD model is explicitly non-unique—the authors say a spin helix can't be excluded—which doesn't affect the band reconstruction but weakens the specific 3a real-space picture used for the 1/3 plateau mechanism. The Dirac cone at M1 and 'topological phase transition' is also overstated: they admit the crossing could be surface states or protected by magnetic symmetry, and the conclusion says 'hinting at.' If they want that claim, they need a symmetry analysis or spin-resolved/inversion evidence. The 1/3 plateau mechanism is post-hoc consistency, but they flag it as a proposal with future neutron tests.\n\nCircularity isn't a problem: q vectors and gaps are measured, not fitted. This is an experimental paper, and those measurements are the strength.\n\nWho's this for? The kagome/correlated-magnetism community will want it. ARPES people will appreciate the data quality. I'd send it to review, but I'd ask for toned-down topology claims and, ideally, a χ0(q) calculation or at least a frank limitation statement about the 4f-less DFT. It deserves a serious referee, but the revision should be substantive, not cosmetic.","headline":"Solid ARPES/NPD observation of dual band modulations at the AFM transition, but the orbital-selective nesting mechanism and topological claims go beyond the evidence.","tokens_in":18618,"tokens_out":4308,"would_cite":true,"duration_ms":34440,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["71.18.+y","75.25.-j","79.60.-i"],"model":"deepseek-v4-flash","headline":"Antiferromagnetic order in TbTi3Bi4 is driven by nesting of Tb $5d_{xz}$ Fermi surfaces, producing Dirac cones at M1.","keywords":["kagome metal","orbital selectivity","antiferromagnetism","Fermi surface nesting","ARPES","TbTi3Bi4","Dirac cone","magnetization plateau"],"falsifier":"Track the magnetic peaks at $(q_1, q_2, 0)$ with neutron diffraction on a field-oriented single crystal while sweeping the field through the 1/3 magnetization plateau: if the $b$-axis ($q_2$) modulation does not collapse as the plateau forms, the proposed spin-flip mechanism is wrong.","tokens_in":2002,"feed_emoji":"🧲","tokens_out":8367,"duration_ms":120180,"temperature":0.7,"pith_summary":"This paper argues that the antiferromagnetic state of the kagome metal TbTi3Bi4 is not a uniform band reconstruction but an orbital-selective one: the Tb $5d_{xz}$ orbitals, concentrated on quasi-1D Fermi surfaces along the zigzag chains, nest with wavevector $q_1 \\approx a^*/3$ and a weaker $q_2 \\approx 0.28 b^*$, and the resulting magnetic order folds the electronic bands in a momentum-dependent way. The authors establish this by combining temperature-dependent ARPES, neutron powder diffraction, STM, and orbital-resolved DFT, showing that the band folding appears only below the Néel temperature $T_{N1} \\approx 20.4$ K and only on selected Fermi-surface sheets. The same wavevector appears in the refined magnetic structure, and no charge order is seen at that wavevector, so the modulation is magnetic in origin. If correct, the finding connects Fermi-surface nesting to a specific orbital and explains the 1/3 fractional magnetization plateau as a consequence of the approximate $3a$ period of this magnetic order. This matters because it demonstrates an orbital-specific control knob for magnetism in the LnTi3Bi4 family.","feed_headline":"Kagome metal's magnetic order pinned to one orbital","feed_subtitle":"Band folding at 1/3 a* and 0.28 b* creates Dirac cones and explains the 1/3 magnetization plateau.","key_machinery":"The central object is the quasi-one-dimensional Fermi surface of the intercalated Tb zigzag chains, specifically its Tb $5d_{xz}$ orbital weight, which the calculations show is concentrated on the $k_z = \\pi$ sheets parallel to the $a^*$ direction. Its nesting vector $q_1 \\approx a^*/3$ (with a secondary $q_2 \\approx 0.28 b^*$) matches the magnetic wavevector $(0.35, 0.28, 0)$ refined from neutron data. This nesting is proposed to enhance the RKKY exchange between Tb $4f$ moments, and the folded-band analysis of the ARPES data carries the argument: the emergent bands at M1 and the gapped pockets are reproduced by folding the normal-state bands by $q_1$ and $q_2$, connecting the observed electronic reconstruction directly to the magnetic periodicity.","core_discovery":"Below $T_{N1} \\approx 20.4$ K, TbTi3Bi4 enters an antiferromagnetic state whose electronic structure shows two coexisting band modulations: a strong folding along $a^*$ with $q_1 \\approx a^*/3$, seen as a gap of roughly 43 meV on the outer pocket at M1 and a 77.6 meV gap on the quasi-1D Fermi surface, and a weaker folding along $b^*$ with $q_2 \\approx 0.28 b^*$, with a gap near 27 meV. The folded bands live mainly in the $k_z = \\pi$ plane, exactly where the band-structure calculation places a well-nested quasi-1D Fermi surface dominated by the Tb $5d_{xz}$ orbital. Neutron powder diffraction finds magnetic peaks at the same $(q_1, q_2, 0)$ wavevector below $T_{N1}$, while STM sees no charge order at that wavevector, so the authors conclude the modulations are the band-structure signature of the antiferromagnetic order itself. The momentum-dependent folding produces symmetry-protected Dirac cones at the M1 point, and the approximate $3a$ period along the zigzag chains provides a natural microscopic picture for the 1/3 magnetization plateau: an intralayer spin flip at the first metamagnetic field produces the plateau, and a second flip saturates the magnetization. The paper's central claim is that this entire chain is driven by nesting of the intercalated Tb $5d_{xz}$ orbitals, making orbital selectivity the organizing principle of the magnetic ground state.","pith_inferences":["Because the calculation that identifies the Tb $5d_{xz}$ nesting excludes the Tb $4f$ electrons, the paper implicitly assumes the itinerant $d$-orbital physics is decoupled enough from the localized $f$ moments to set the ordering wavevector; a resonant ARPES experiment at the Tb M edge could test this directly.","The $q_2$ modulation is presented as the weaker, secondary instability; an interesting extension is whether doping or strain that shifts the $b^*$ nesting condition would suppress the 1/3 plateau while leaving the $a^*$ order intact.","The same orbital-selective nesting mechanism may apply to other intercalated kagome metals with zigzag chains, suggesting a design rule: align the chain direction with the dominant nesting vector to select the magnetic order.","The proposed spin-flip sequence for the plateau is a minimal model; field-dependent neutron diffraction on a single crystal could distinguish it from the alternative spin-helix configuration the authors say cannot be excluded."],"forward_implications":["The same $(1/3, 0.28, 0)$ wavevector governs both the magnetic order and the band folding, so the antiferromagnetic state is electronically driven rather than purely a local-moment effect.","The Dirac cones at M1 below $T_{N1}$ imply a magnetic-order-induced topological transition in the antiferromagnetic state.","The 1/3 magnetization plateau follows from the approximate $3a$ magnetic period along the zigzag chains, with an intralayer spin flip at the first metamagnetic field and a second flip at higher field.","In the LnTi3Bi4 family, the presence or absence of a well-nested quasi-1D Fermi surface predicts which compounds show the 1/3 plateau, consistent with the contrast between GdTi3Bi4 and EuTi3Bi4.","Tuning the Fermi surface by gating or doping should change the nesting vector and could stabilize other fractional magnetization plateaus."],"supporting_citations":[{"why":"Establishes the 1/3 magnetization plateau and zigzag-chain crystal context in TbTi3Bi4 that the paper's mechanism aims to explain.","marker":"[15]"},{"why":"Maps the magnetic landscape and transitions of TbTi3Bi4, providing the baseline the ARPES and neutron data are compared against.","marker":"[16]"},{"why":"Supplies the single-crystal neutron diffractometry showing a-axis moments roughly three times the b-axis moments, used to support the anisotropic gap analysis.","marker":"[17]"},{"why":"Gives the twofold van Hove singularity and charge-order framework in kagome superconductors, the comparison that motivates the imperfect-nesting argument against a charge density wave.","marker":"[25]"},{"why":"Shows tunable magnetism and band structure across the RETi3Bi4 family, including which compounds have the nested quasi-1D Fermi surface (Gd versus Eu).","marker":"[29]"},{"why":"Provides the symmetry representations used to argue that no symmetry-enforced degeneracy exists at L1 or M1, so the observed Dirac cone requires magnetic symmetry or a topological origin.","marker":"[31]"},{"why":"Supplies the framework of $4f$\\textendash$5d$ indirect exchange in rare-earth metals that the paper invokes for orbital-selective magnetism.","marker":"[36]"},{"why":"Gives the orbital-dependent exchange mechanism stabilizing a helical Q structure in Gd compounds, cited as parallel support for orbital-selective exchange.","marker":"[38]"}],"fun_headline_variants":["Orbital-selective band modulations reveal hidden Dirac cones in kagome metal","Unidirectional band folding ties antiferromagnetism to orbital nesting","Kagome metal's 1/3 magnetization plateau traced to orbital-specific nesting","Single orbital steers antiferromagnetic order in TbTi3Bi4","Dirac cones and a fractional plateau from orbital-selective magnetism"],"cache_read_input_tokens":20736,"weakest_assumption_plain":"The argument depends on two fragile premises: that a band-structure calculation which leaves out Tb's localized $4f$ electrons still gets the orbital character and nesting right, and that the proposed magnetic structure\\textemdash which the authors concede is one plausible solution, not excluding a spin helix\\textemdash is the true one.","fun_headline_variants_meta":{"raw":{"variants":["Orbital-selective band modulations reveal hidden Dirac cones in kagome metal","Unidirectional band folding ties antiferromagnetism to orbital nesting","Kagome metal's 1/3 magnetization plateau traced to orbital-specific nesting","Single orbital steers antiferromagnetic order in TbTi3Bi4","Dirac cones and a fractional plateau from orbital-selective magnetism"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00025,"raw_usage":{"total_tokens":1679,"prompt_tokens":1195,"completion_tokens":484,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":811,"completion_tokens_details":{"reasoning_tokens":401}},"tokens_in":811,"tokens_out":484,"duration_ms":4717,"temperature":1.0,"reasoning_tokens":401,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T10:14:40.341238+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Track the magnetic peaks at $(q_1, q_2, 0)$ with neutron diffraction on a field-oriented single crystal while sweeping the field through the 1/3 magnetization plateau: if the $b$-axis ($q_2$) modulation does not collapse as the plateau forms, the proposed spin-flip mechanism is wrong.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Shows tunable magnetism and band structure across the RETi3Bi4 family, including which compounds have the nested quasi-1D Fermi surface (Gd versus Eu)."}],"review_version":1}