{"id":"cf9d215d-9361-4005-815e-d5e1f67d6614","arxiv_id":"2601.07443","paper_version":2,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":5.0,"correctness_risk":"low","formal_verification":"none","parameter_count":5,"one_line_summary":"For a topological-insulator–ferromagnet junction with hexagonal warping, exceptional points form a hexagon whose positions and number can be tuned by an in-plane magnetic field.","lead":"This paper computes where exceptional points—places where two quantum states merge into one—appear at the junction between a topological insulator and a ferromagnet once the hexagonal distortion of the insulator surface is included. It finds six such points arranged in a hexagon, and shows that a magnetic field can move them or make four of them disappear.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Central EP prediction rests on the wide-band self-energy being purely imaginary, momentum-independent, and σz-only; a realistic FM lead may introduce off-diagonal or momentum-dependent imaginary components that alter the degeneracy conditions.","rationale":"The paper's algebra is correct under its stated Hamiltonian: the EP conditions are standard, the solutions reduce to the regular hexagon for Bx=0, and the annihilation at Bc is a straightforward consequence. The central risk is not internal inconsistency but physical modeling: all novelty (hexagonal EP arrangement, tunable annihilation) is a direct consequence of the specific form of Σ_L. Since the self-energy is not derived for the proposed heterostructure but borrowed from prior work, the prediction hangs on an assumption that may not be realized in a true TI–FM interface. This matches the reader's weakest assumption. The unphysical Γ=0, γ=1 parameter choice and the vague claim that warping acts 'through the eigenvectors' (which are actually independent of λ at the EPs) are secondary issues that do not break the central derivation. Therefore the CONDITIONAL verdict remains appropriate; no change is warranted beyond the conditions already stated.","tokens_in":7501,"tokens_out":9785,"duration_ms":98741,"concrete_test":"Using a tight-binding model for a semi-infinite FM lead with exchange splitting m and spin-conserving hopping t' to a Bi2Te3 surface, compute the retarded self-energy Σ_L(k, ω=0) at the EP momenta (|k| ≈ γ/α). Check whether Im Σ_L is (i) independent of k, (ii) diagonal in the spin basis with entries -Γ± (positive), and (iii) has no σx/σy components. If Σ_L has an imaginary σx component (e.g., for a lead magnetized along x), solve d_R^2 = d_I^2 and d_R·d_I = 0 with the corrected d_I; if the number or positions of EPs differ from the six-point formula in §III, the central claim is conditional on the model.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Equations (5)–(6) and the analytic EP locations follow only if d_I = (0, 0, -γ), i.e., if the lead self-energy is purely imaginary, momentum-independent, and diagonal in σz. The paper adopts Eq. (2) from Refs. [24,43–45] without deriving it for the Fe3GeTe2|Bi2Te3 interface proposed in Sec. VI. In a real ferromagnetic lead, the imaginary part of Σ_L can acquire off-diagonal spin components (e.g., if the lead magnetization is not exactly along z, or if interface hopping mixes spins) and can be momentum-dependent when the wide-band approximation fails. In that case d_R·d_I = -γλ(k_x^3 - 3k_x k_y^2) is replaced by a more general expression involving d_Ix and d_Iy; the factorization into three zero lines is lost, so the six-EP hexagonal pattern and the B_c = √(4/3)γ annihilation are not guaranteed. Additionally, the numerical choice Γ = 0, γ = 1 makes Γ_- = -1 in Eq. (2), implying a negative density of states for one spin channel—an unphysical parameter regime that cannot represent a passive lead. This weakens the 'realistic platform' claim even though it does not change the algebraic result.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies a two-band non-Hermitian Hamiltonian for a topological-insulator–ferromagnet junction, composed of the Fu-model surface state (linear Dirac term plus cubic hexagonal warping), a Zeeman field B_x, and a wide-band lead self-energy Σ_L = −iΓσ_0 − iγσ_z. The authors derive the non-Hermitian degeneracy conditions d_R^2 = d_I^2 and d_R·d_I = 0, obtaining exact exceptional-point locations that form a hexagon for λ ≠ 0 and are independent of the warping strength λ. They show that an in-plane magnetic field moves these EPs and that, above B_c = √(4/3)γ, four of the six EPs disappear, leaving two on the k_x = 0 line. Phase-rigidity plots corroborate the analytic results, and the λ = 0 limit is shown to produce an exceptional ring that fragments into six EPs for finite λ. The central claim is that TI–FM heterojunctions with hexagonally warped surface states provide a tunable platform for non-Hermitian physics.","tokens_in":7823,"tokens_out":7694,"duration_ms":85718,"significance":"If the model Hamiltonian is taken as given, the paper provides a clean, exactly solvable example in which hexagonal warping converts an exceptional ring into six isolated EPs and in which an in-plane field controls their positions and annihilation. The analytic solution is useful and the derivations are straightforward and correct: the listed EP coordinates satisfy the stated degeneracy conditions, and the phase-rigidity plots support the eigenvector coalescence. I found no parameter fitting or circularity; the EP positions follow from the model without adjustable parameters. The main value lies in the explicit analytic control of EP structure and in the proposal of a specific material platform. However, the transfer of the results to the proposed Fe₃GeTe₂|Bi₂Te₃ heterostructure depends on the validity of the wide-band, momentum-independent, σ_z-only self-energy, which is assumed rather than derived for that interface. The work is solid within its model but needs to address this gap before the stronger 'realistic platform' claim can be fully supported.","major_comments":[{"comment":"The entire degeneracy analysis rests on the self-energy form Σ_L = −iΓσ_0 − iγσ_z, which gives d_I = (0,0,−γ). This form is quoted from Refs. [24,43–45] but is not derived for the Fe₃GeTe₂|Bi₂Te₃ interface proposed in Sec. VI. If a realistic FM lead produces off-diagonal imaginary self-energy components (d_Ix, d_Iy ≠ 0) or momentum-dependent terms, Eq. (6) is modified and the factorization into the three lines k_x = 0, k_x = ±√3 k_y is lost. In that case the hexagonal six-EP pattern and the prediction B_c = √(4/3)γ are not guaranteed. Please either derive or estimate the neglected components for the proposed interface, or explicitly state the conditions under which the σ_z-only wide-band form applies and temper the platform claims accordingly.","section":"§II, Eq. (2); §III, Eqs. (5)–(6)"},{"comment":"The numerical calculations set Γ = 0, γ = 1. From Eq. (2), the spin-resolved broadening parameters are Γ_± = Γ ± γ, so this choice gives Γ_- = −1, i.e., a negative density of states for one spin channel of the passive lead. This is unphysical and weakens the experimental-relevance claim. Since Γ enters only through the common shift ϵ_0 = −iΓ, the degeneracy conditions and EP positions are independent of Γ. I recommend choosing Γ > γ > 0 (for example Γ = 1.5, γ = 1) or explicitly stating that Γ is set to zero for display only and has no effect on the EP locations.","section":"§III, parameter choice after Eq. (6)"}],"minor_comments":[{"comment":"The statement that 'four of the EPs annihilate at the critical field B_c' is imprecise. At B_c the square-root term vanishes, but the four off-axis EPs merge pairwise into two points at (±√3 B_c/(4α), B_c/(4α)); only for B > B_c do these two points disappear, leaving two EPs on the k_x = 0 line. Please clarify this in the text.","section":"§IV, critical-field discussion"},{"comment":"The abstract and outlook state that hexagonal warping plays a crucial role in 'determining the locations' of the EPs, but the derived locations are independent of λ. The warping determines whether the degeneracies are isolated EPs or an exceptional ring and fixes their hexagonal symmetry, not their coordinate values. The wording should be adjusted to avoid this mismatch.","section":"§VI and Abstract, wording"},{"comment":"The phase-rigidity definition r = ⟨Ψ_L|Ψ_R⟩/⟨Ψ_R|Ψ_R⟩ requires specification of the normalization convention for the left and right eigenvectors. In biorthogonal quantum mechanics one usually writes r = ⟨Ψ_L|Ψ_R⟩/√(⟨Ψ_L|Ψ_L⟩⟨Ψ_R|Ψ_R⟩) or states that the right eigenvectors are normalized and the left vectors are chosen accordingly.","section":"§III, Eq. (7)"},{"comment":"The Zeeman term B_x σ_x is included, but the manuscript does not comment on orbital effects of the in-plane magnetic field. For a strictly two-dimensional surface state these are usually negligible, but a brief statement would be helpful.","section":"§II, Eq. (4)"}],"recommendation":"major_revision","confidential_remarks":"The algebraic core is correct and the paper is clearly written. The main risk is the unexamined transfer of the wide-band σ_z-only self-energy to the proposed Fe₃GeTe₂|Bi₂Te₃ heterostructure; if this self-energy is not representative, the central six-EP prediction would not survive. The unphysical Γ = 0, γ = 1 parameter choice is easily fixed and does not affect the analytic results. I therefore see a major revision rather than rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a correct, modest analytic result — warping turns an exceptional ring into six field-tunable EPs — wrapped in a slightly too-grand 'realistic platform' package. The math is fine; the presentation needs honest parameter choices and a less mysterious account of how warping acts.\n\nThe genuinely new piece is the combination of Fu's hexagonal warping with the standard wide-band self-energy for a TI–FM junction. The degeneracy conditions are derived cleanly, the EP locations are correct, and the phase-rigidity plots match the analytic points. Solving Eq. (6) for the cubic term gives three lines in k space; intersecting those with the circle from Eq. (5) gives exactly the six points stated, and the Bc = 4/3 γ critical field follows. That is a solid, parameter-free result within the model. The λ=0 exceptional-ring limit and the λ≠0 hexagonal pattern are a nice pedagogical demonstration.\n\nSoft spots, in order of size. First, the numerical choice Γ=0, γ=1 is physically impossible: it makes Γ− = −1, a negative spin-channel density of states. The result itself does not depend on that choice, but the authors keep it throughout the figures and then call the setup 'realistic.' They should pick a physical regime, e.g., Γ > γ > 0, and show the same plots. Second, the sentence that warping acts 'through the eigenvectors' is not explained, and at every EP the cubic term vanishes, so the warping term's explicit contribution to d_R is zero. What warping does is enforce the d_R·d_I = 0 condition on a set of rays, which turns the ring into isolated points. That is a geometric statement about the degeneracy condition, not about eigenvectors, and should be rewritten. Third, the wide-band self-energy is borrowed from Refs. [24,43–45] without a derivation for the specific Fe3GeTe2|Bi2Te3 interface; if the real interface creates momentum-dependent or off-diagonal imaginary self-energy components, the factorization into rays and circle breaks. That is a scope limitation, not a fatal flaw, but the paper should state it explicitly rather than implying the platform is ready-made.\n\nThe paper deserves a serious referee. It is not a landmark, but it is a clean analytic contribution to the non-Hermitian junction niche, and it can be fixed with modest revision: physical parameters, a clarified role for warping, and a more careful experimental framing.","headline":"Six warping-tuned EPs with field annihilation: correct analytic model, overplayed 'realistic platform' — worth a referee after honest parameter fixes.","tokens_in":8332,"tokens_out":5908,"would_cite":true,"duration_ms":57315,"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":"A topological insulator–ferromagnet junction with hexagonal warping hosts six exceptional points — located exactly by a simple formula — that an in-plane magnetic field can move and annihilate; above a critical field only two survive.","keywords":["exceptional points","non-Hermitian Hamiltonian","topological insulator","hexagonal warping","ferromagnet junction","magnetic-field tuning","phase rigidity","Bi2Te3"],"falsifier":"Build a tight-binding or density-functional model of a concrete topological-insulator–ferromagnet junction (for example Bi2Te3 with a ferromagnetic layer), compute the full interface self-energy without the wide-band approximation, and test whether the degeneracy conditions γ² = (−αky+Bx)² + (αkx)² + λ²(kx³−3kxky²)² and λγ(kx³−3kxky²) = 0 still force the six formula points. Any momentum dependence or imaginary σx/σy component in the self-energy would shift or erase the analytic exceptional-point locations; alternatively, a measurement that maps the phase-rigidity hexagon in momentum space woul","tokens_in":7367,"feed_emoji":"🧲","tokens_out":9483,"duration_ms":88489,"temperature":0.7,"pith_summary":"This paper sets out to show that the interface between a topological insulator with a hexagonally warped surface state and a ferromagnetic lead is a controllable source of exceptional points, the non-Hermitian degeneracies where eigenvalues and eigenvectors coalesce at once. Modeling the ferromagnet as a wide-band reservoir gives the junction a constant imaginary self-energy, and the paper solves the resulting non-Hermitian Hamiltonian analytically. It finds six exceptional points whose locations form a hexagon and are given by closed-form expressions in the Fermi velocity, the imaginary spin splitting, and the applied magnetic field. The warping term does not set the positions, but it is what breaks the continuous exceptional ring of the unwarped problem into six discrete points. If the picture holds, bismuth-telluride–ferromagnet heterostructures become a tunable laboratory for non-Hermitian physics, with an external field controlling how many exceptional points are present.","feed_headline":"Magnetic field erases four of six exceptional points","feed_subtitle":"Hexagonal warping arranges the degeneracies; an in-plane magnetic field leaves just two tunable exceptional points.","key_machinery":"The argument is carried by expressing the effective non-Hermitian junction Hamiltonian as H_eff = ε0 + d·σ, with the complex vector d = dR + i dI. For this junction dR = (−αky + Bx, αkx, λ(kx³ − 3kxky²)) and dI = (0, 0, −γ). Non-Hermitian degeneracies occur precisely when dR² = dI² and dR·dI = 0; these two simultaneous conditions force the real and imaginary parts of both eigenvalues to coincide, and solving them yields the exceptional-point locations. The phase rigidity r = ⟨ΨL|ΨR⟩/⟨ΨR|ΨR⟩, which vanishes at an exceptional point, supplies the numerical check that the eigenvectors also coalesce.","core_discovery":"The central result is an analytic description of the exceptional points. With lead self-energy ΣL = −iΓ σ0 − iγ σz and Zeeman field Bx σx, the eigenvalues coalesce when γ² = (−αky+Bx)² + (αkx)² + λ²(kx³−3kxky²)² and λγ(kx³−3kxky²)=0. Solving these gives six k-space points: (0,(Bx±γ)/α) and (±√3(Bx±√(4γ²−3Bx²))/(4α),(Bx±√(4γ²−3Bx²))/(4α)). The locations do not depend on the warping strength λ, yet λ fragments the exceptional ring found at λ=0 into six discrete exceptional points with hexagonal symmetry. A magnetic field moves the points, and above Bc=√(4/3)γ four annihilate, leaving two on the kx=0 line. Phase-rigidity plots confirm that eigenvalues and eigenvectors coalesce.","pith_inferences":["The four-point annihilation at the critical field resembles a pair-annihilation event in the complex-energy landscape; if the exceptional points carry integer winding charges, the two survivors may be constrained by a charge-conservation rule that the paper does not derive.","The paper treats the magnetic field purely as a Zeeman term and the self-energy as momentum-independent; a tight-binding or first-principles version of the same junction is the natural next check, since orbital coupling and momentum-dependent lead self-energies would modify the two degeneracy equations.","The same two-equation construction could be exported to other surface-state symmetries — trigonal warping, strained surfaces, or p-wave magnets — to predict when exceptional points replace exceptional rings.","If the field-tunable annihilation survives in a real material, the sharp sensitivity of the remaining exceptional points could make the junction a magnetic-field-actuated sensing element; that device-level possibility goes beyond what the paper claims."],"forward_implications":["In the linear-plus-warping surface-state junction with a ferromagnetic lead, exceptional points appear at six explicitly known momenta arranged in a hexagon, so no numerical search is needed to locate them.","An in-plane magnetic field moves all six exceptional points and can switch their number: above Bc = √(4/3)γ four annihilate and two remain, giving field control over non-Hermitian degeneracy.","Hexagonal warping acts as a symmetry selector: at λ = 0 the system has a continuous exceptional ring, while any finite λ fragments it into exactly six exceptional points.","Because the exceptional-point positions do not depend on λ but their pattern does, the prediction is insensitive to the warping strength; the phase-rigidity contours merely sharpen as λ increases.","Junctions made from a material such as Bi2Te3, with its naturally warped surface states and existing ferromagnet heterostructure recipes, offer a concrete place to look for these effects."],"fun_headline_variants":["Hexagonal warping splits exceptional ring into six points","Magnetic field leaves just two exceptional points","Warping and field tune exceptional point count","From six to two: field controls exceptional points","Hexagonal warping shapes exceptional points in heterojunctions"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The analytic six-point pattern rests on the wide-band approximation in which the ferromagnetic lead contributes a momentum-independent, purely σz imaginary self-energy and the magnetic field enters only as a Zeeman term; if real interfaces add momentum-dependent or σx/σy imaginary self-energy pieces, or if orbital effects matter, the degeneracy equations and the hexagonal pattern no longer follow.","fun_headline_variants_meta":{"raw":{"variants":["Hexagonal warping splits exceptional ring into six points","Magnetic field leaves just two exceptional points","Warping and field tune exceptional point count","From six to two: field controls exceptional points","Hexagonal warping shapes exceptional points in heterojunctions"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000708,"raw_usage":{"total_tokens":3018,"prompt_tokens":727,"completion_tokens":2291,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":471,"completion_tokens_details":{"reasoning_tokens":2228}},"tokens_in":471,"tokens_out":2291,"duration_ms":17636,"temperature":1.0,"reasoning_tokens":2228,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-04T06:21:23.529506+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Build a tight-binding or density-functional model of a concrete topological-insulator–ferromagnet junction (for example Bi2Te3 with a ferromagnetic layer), compute the full interface self-energy without the wide-band approximation, and test whether the degeneracy conditions γ² = (−αky+Bx)² + (αkx)² + λ²(kx³−3kxky²)² and λγ(kx³−3kxky²) = 0 still force the six formula points. Any momentum dependence or imaginary σx/σy component in the self-energy would shift or erase the analytic exceptional-point locations; alternatively, a measurement that maps the phase-rigidity hexagon in momentum space woul","supporting_citations":[],"review_version":1}