{"id":"84eddc56-b8c8-44db-861c-7fd6f0eff69c","arxiv_id":"1908.09309","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A crisscross antiferromagnetic lattice with 4'/m' symmetry realizes 2D and 3D higher-order topological insulators with quantized quadrupole moments, corner states, and chiral hinge states.","lead":"This paper constructs a 2D and a 3D tight-binding model of a crisscross antiferromagnet whose 4'/m' magnetic symmetries produce higher-order topological insulator phases. The 2D phase shows quantized electric and magnetic quadrupole moments with fractional corner charges, and the 3D phase hosts chiral hinge states and a predicted half-quantum spin-flop pumping effect.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The unmodified 3D model is not a strict second-order TI: the (001) surface hosts a PT-protected Dirac cone (Fig. 3b inset), so the claimed chiral hinge phase requires an additional S4-preserving Zeeman term not present in Eq. (6).","rationale":"The reader's weakest assumption concerned the stability of the assumed 4'/m' magnetic order, which is a legitimate model-input caveat but not the most load-bearing issue for the central claim. The more specific problem is that the 3D model, as written in Eq. (6), does not satisfy the defining boundary condition of a 3D second-order TI: all 2D surfaces should be gapped, yet the authors themselves show a gapless Dirac cone on the (001) surface when the z boundaries are open. The S4-based invariant in Eq. (7) is computed from bulk eigenvalues and does not guarantee gapped surfaces; the paper's own Fig. 3b inset demonstrates that PT symmetry forces a surface Dirac cone. The chiral hinge states shown in Fig. 3c appear in a rod geometry with periodic z, which removes the problematic surfaces. The discussion then invokes an external magnetic field to gap the surface Dirac cone, but that field is not part of the original Hamiltonian and breaks C4zT, modifying the magnetic point group. This is not an internal mathematical contradiction, but it makes the headline 3D HOTI claim overbroad. The 2D analysis, the charge quadrupole moment, and the corner-state calculation appear internally consistent, and I do not see a fatal error there. The concern therefore reinforces the reader's CONDITIONAL verdict rather than changing it; the paper would be strengthened by explicitly presenting the bare model as a parent phase whose strict HOTI boundary signature requires the S4-preserving Zeeman term.","tokens_in":7860,"tokens_out":20353,"duration_ms":224238,"concrete_test":"Exact-diagonalize Eq. (6) on a finite Lx x Ly x Lz hexahedron with open boundaries in all three directions and no Zeeman term, and plot the (001) surface spectral function. If a Dirac cone remains at the surface BZ center and connects the vertical hinge modes, the bare model is not a strict 3D second-order TI. Then add a small S4-preserving, C4zT-breaking sigma_z Zeeman term and verify that the surface gap opens and the four hinge modes form a single chiral loop; this would settle whether the central 3D HOTI claim requires an external field beyond the Hamiltonian as written.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The weakest point is the 3D claim. In the '3D Model' section the authors explicitly state that with open boundary in z the (001) surface 'is not gapped but exhibit a Dirac cone' (inset of Fig. 3b), protected by the combined PT symmetry, and that only after applying a magnetic field along z (which breaks C4zT but keeps S4) do 'connected hinge states in the hexahedron sample' appear. By the standard definition of a 3D second-order TI, all 2D surfaces must be gapped; here the z-normal surfaces are gapless for the Hamiltonian in Eq. (6). The S4 eigenvalue invariant v in Eq. (7) is a bulk criterion and does not by itself certify gapped surfaces. Thus the abstract's statement that the 3D stack 'possesses the HOTI phase holding chiral 1D metallic states on the hinge' is not true for the bare Hamiltonian; it requires an additional Zeeman term that changes the magnetic point group away from the 4'/m' symmetry used to construct the model. The paper discloses this limitation, but the strongest claim should be qualified as conditional on that extra term, and the v=1 phase designation needs a corresponding caveat.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript introduces a two-dimensional (2D) and a three-dimensional (3D) tight-binding model with a crisscross antiferromagnetic order on a square lattice, designed to respect the magnetic point group 4'/m' generated by C4zT and PT. In 2D, the authors find that for |λ1| < |λ3| the system is a second-order topological insulator (HOTI) with Wannier centers at the unit-cell corners, four zero-energy corner states carrying fractional charge e/2, a quantized charge quadrupole moment qxy = e/2, and a quantized magnetic quadrupole moment given by Eq. (5). The 3D model stacks such layers with kz-dependent hoppings, and the authors claim that when condition (8) holds, the system realizes a 3D HOTI with chiral hinge states, characterized by a Z2 invariant v defined through S4 eigenvalues. They also discuss axion-insulator-like transport, including half-quantized Hall conductance on side surfaces and half-quantum spin-flop pumping.","tokens_in":8035,"tokens_out":8596,"duration_ms":85231,"significance":"If the claims hold, this is a valuable minimal model of a magnetic HOTI with simultaneous quantized charge and magnetic quadrupole moments, and a concrete tight-binding setting for chiral hinge transport. The 2D part is well supported: the phase boundary follows from the analytic dispersion, the S4 eigenvalue counting is explicit, and the corner states and e/2 corner charges are confirmed by direct diagonalization of a 20×20 sample. The transport discussion is standard axion-insulator physics applied to the model. The main weakness is the overstatement of the 3D HOTI phase; the bare 3D model has a gapless (001) surface, so the chiral hinge phase is conditional on an additional Zeeman term not present in Eq. (6).","major_comments":[{"comment":"The abstract claims that the 3D system 'possesses the HOTI phase holding chiral 1D metallic states on the hinge,' but the bare Hamiltonian (6) is not a second-order topological insulator by the definition given in the introduction, which requires all (D−1)-dimensional boundaries to be gapped. The 3D Model section explicitly states that with open boundary in z the (001) surface 'is not gapped but exhibit a Dirac cone' (inset of Fig. 3b), protected by PT, and that only after applying a magnetic field along z, which breaks C4zT but preserves S4, do connected hinge states appear in a hexahedron sample. This is a load-bearing mismatch: the central 3D result is true only for a modified model not given in Eq. (6). Please qualify the claim, give the Zeeman term explicitly, and recompute (or state the invariance of) the v=1 invariant in that setting.","section":"Abstract and 3D Model (Eq. (6), Fig. 3b inset)"},{"comment":"The invariant v defined in Eq. (7) is computed from S4 eigenvalues of the 4'/m' Hamiltonian and is used to label the 'HOTI phase' via condition (8). However, a bulk S4 eigenvalue invariant by itself does not certify that all surfaces are gapped, and indeed the (001) surface is gapless here. The hinge-state calculation in Fig. 3c uses open boundaries only in x and y (with z periodic), so it does not probe the gapless z-normal surfaces. Please clarify what exactly v=1 classifies for the bare Hamiltonian, and how the classification changes after the S4-preserving Zeeman term is added: which high-symmetry points remain S4-invariant and whether the eigenvalue ratios entering Eq. (7) are unchanged.","section":"3D Model (Eq. (7), condition (8))"},{"comment":"The magnetic quadrupole moment tensor in Eq. (5) is a central advertised result, described as 'a unique feature compared with previous studies,' but the derivation is only sketched in one paragraph. Please show explicitly how the entries of the 3×3 matrix follow from the Wannier-center coordinates (0 or 1/2) and the local moment directions on the four sites, including the sign pattern, and state the units and factors clearly. Without this, a reader cannot verify the claimed quantization.","section":"2D Model (Eq. (5))"}],"minor_comments":[{"comment":"The title contains a typo: 'antif erromagnetic' should be 'antiferromagnetic'.","section":"Title"},{"comment":"The phrase '1/4 quantum magnetic quadrupole moment' in the introduction is inconsistent with the abstract's 'quantized magnetic quadrupole moment'; if the entries of Eq. (5) are meant, specify that the tensor components have magnitude g μB/4.","section":"Introduction, first paragraph"},{"comment":"Equation (4) is dimensionally ambiguous: ηM/ηΓ is a phase, so 'modulo 2' likely means taking the phase modulo 2π and dividing by π; please rewrite the formula as P^n_{x/y} = (e/2π) arg(η^n_M/η^n_Γ) or the equivalent used in Ref. [17].","section":"Eq. (4)"},{"comment":"The sentence 'Two Dirac cone come from up and down surface degenerate at Fermi level' should be reworded for grammar and to indicate which surfaces the two cones belong to.","section":"Fig. 3b inset text"},{"comment":"Reference [28] is cited as a proposed chiral HOTI (EuIn2As2); please check the journal/volume data and ensure the classification Z4 statement matches that reference.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The 2D results appear sound and the paper is readable. The main issue is the qualification of the 3D phase: the authors themselves disclose the Dirac cone on the (001) surface and the need for a magnetic field along z, so the abstract and the v=1 'HOTI phase' label must be revised accordingly. I recommend major revision rather than rejection because the claims can be repaired by adding the Zeeman term explicitly and clarifying what the bulk invariant classifies. The novelty claim about the magnetic quadrupole moment would also benefit from a fuller derivation."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nTwo things to know. First, the 2D crisscross AFM model is a clean, minimal construction of a magnetic HOTI with a genuinely new feature: the quantized magnetic quadrupole moment. Second, the 3D chiral hinge phase advertised in the abstract is not realized by the bare Hamiltonian in Eq. (6); the (001) surface has a PT-protected Dirac cone, and the hinges only appear after adding a Zeeman field that breaks C4zT. The authors disclose this in the 3D section, but the abstract and the discussion do not carry the caveat.\n\nThe 2D part is solid. The Hamiltonian in Eq. (2) is fully determined by the 4'/m' constraints, the dispersion is explicit, and the S4 eigenvalue count for qxy is standard and correctly applied. The corner states and fractional e/2 charges in Fig. 2 back the claim. The MQM tensor in Eq. (5) follows from the pinned spin arrangement and the WC positions; I believe this is not present in earlier HOTI models, which usually rely on Mx/My or S4 charge quantization. The paper also clearly separates its Z2 classification from the Z4 case of EuIn2As2. Credit is due for the explicit quantities and for grounding the claims in the band representations.\n\nSoft spots, in order of size. The 3D claim is the biggest. With open boundaries in z, the (001) surface is gapless (Fig. 3b inset), so by the standard definition the system in Eq. (6) is not a 3D second-order TI. Adding a z-direction magnetic field gaps those surfaces but changes the magnetic point group away from the 4'/m' symmetry used to build the model. That is not a fatal flaw, but it means the abstract's 'possesses the HOTI phase' is an overstatement; it should say the hinge phase appears under an additional S4-preserving Zeeman term. A referee should request this qualification.\n\nSecond, the half-quantum spin-flop pumping is asserted in a short paragraph with no derivation. It may well follow from the axion action and the C4zT-pinned spin texture, but as written it is a suggestion, not a calculation. Either expand it or drop the strong claim.\n\nThird, the stability of the assumed 4'/m' magnetic order is never discussed. This is standard for model papers, so I call it minor.\n\nOverall, the paper is a useful contribution. The 2D model is well supported, the MQM is a new observable, and the 3D stack is a reasonable starting point for engineering chiral hinge states. I would send it to a serious referee and would cite it in work on magnetic HOTIs.","headline":"Solid 2D magnetic HOTI model with quantized MQM; the 3D chiral hinge claim needs a caveat about an extra Zeeman term that the abstract omits.","tokens_in":8639,"tokens_out":2936,"would_cite":true,"duration_ms":27849,"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":"The crisscross antiferromagnetic model realizes second-order topological insulator phases with quantized corner charges and chiral hinge states.","keywords":["higher-order topological insulator","antiferromagnetic topological insulator","magnetic point group","charge quadrupole moment","magnetic quadrupole moment","chiral hinge states","tight-binding model","Wannier centers"],"falsifier":"In the tight-binding model, the decisive calculation is to add a symmetry-breaking term that tilts the spins out of the $xy$-plane and track the four corner states: if they split or move off zero energy before the bulk gap closes, the claimed $\\mathbb{Z}_2$ protection fails. Experimentally, neutron diffraction or torque magnetometry showing ordered moments canted out of the plane would falsify the model's magnetic premise.","tokens_in":7599,"feed_emoji":"🧲","tokens_out":11447,"duration_ms":101329,"temperature":0.7,"pith_summary":"The paper introduces a two-dimensional tight-binding model of a crisscross antiferromagnet whose magnetic point group $4'/m'$ is generated by compositions of time reversal with inversion and with fourfold rotation. For intercell hopping stronger than intracell hopping ($|\\lambda_1|<|\\lambda_3|$), the model is claimed to be a second-order topological insulator: the bulk and edges stay gapped, while each corner of a finite square hosts a zero-energy state carrying charge $e/2$, equivalently a charge quadrupole moment $q_{xy}=e/2$. An additional feature is a quantized magnetic quadrupole moment, which earlier charge-quadrupole models do not possess. Stacking such layers along $z$ with interlayer hoppings chosen so that the $k_z=0$ and $k_z=\\pi$ planes are in different 2D phases gives a three-dimensional second-order topological insulator with chiral hinge states and a $\\mathbb{Z}_2$ invariant; the paper further predicts half-quantum spin-flop pumping in response to a $z$-directed electric field.","feed_headline":"Antiferromagnet lattice hosts e/2 corner charges and hinge currents","feed_subtitle":"A 4'/m' crisscross AFM model becomes a second-order topological insulator with quantized magnetic quadrupole effects.","key_machinery":"The machinery is a four-site square-lattice tight-binding Hamiltonian whose spin configuration and hopping phases are fixed by the magnetic point group $4'/m'$; the two independent real parameters $\\lambda_1$ and $\\lambda_3$ control intra- and intercell hopping, and the dispersion $E(\\mathbf{k})=\\pm 2\\sqrt{\\lambda_1^2+\\lambda_3^2+\\lambda_1\\lambda_3(\\cos k_x+\\cos k_y)}$ has doubly degenerate bands throughout the Brillouin zone. The classification runs through the unitary rotoinversion $\\mathcal{S}_4=\\mathcal{P}\\mathcal{C}_{4z}$: its eigenvalues at $\\Gamma$ and $M$ fix each occupied band's Wannier-center coordinate via $P_{x/y}=\\frac e2(\\eta_M/\\eta_\\Gamma \\bmod 2)$, so the transition at $|\\lambda_1|=|\\lambda_3|$ is exactly the Wannier centers moving from the cell center to its corners. The same $\\mathcal{S}_4$ eigenvalue data at the four $\\mathcal{S}_4$-invariant momenta $\\Gamma$, $M$, $Z$, and $R$ enters the $\\mathbb{Z}_2$ invariant $(-1)^v=\\xi_R\\xi_M/\\xi_Z\\xi_\\Gamma$ for the 3D stack, and the mirror-time-reversal symmetry $\\mathcal{M}_z\\mathcal{T}$ confines all moments to the $xy$-plane to produce the quantized magnetic quadrupole moment.","core_discovery":"The central discovery is that a single symmetry-enforced crisscross antiferromagnetic lattice realizes higher-order topological insulating phases in both two and three dimensions. In 2D, the $4'/m'$ symmetries force the Wannier centers of the occupied bands to sit at the unit-cell center in the trivial phase and at the unit-cell corners in the nontrivial phase; the corner phase has $q_{xy}=e/2$, four corner states each with fractional charge $e/2$, and the quantized magnetic quadrupole tensor $\\varrho_{ij}$ of Eq. (5). In 3D, the paper shows that placing the $k_z=0$ and $k_z=\\pi$ planes in opposite 2D phases produces chiral one-dimensional hinge states connecting valence and conduction bands, with a $\\mathbb{Z}_2$ invariant $v$ determined by $\\mathcal{S}_4$ eigenvalues at $\\Gamma$, $M$, $Z$, and $R$. Side surfaces remain insulating but carry massive Dirac cones of opposite mass on neighboring faces, which is exactly what forces the hinge modes to exist; the resulting axion-type response gives half-quantum surface Hall conductances and the proposed spin-flop pumping.","pith_inferences":["The paper leaves implicit that any material with the same magnetic point group and a band inversion at an $\\mathcal{S}_4$-invariant momentum should show identical corner charges and magnetic-quadrupole quantization, regardless of the microscopic hopping details.","A direct test of the model's rigidity is to add a small canting angle $\\delta$ that tilts the moments out of the $xy$-plane; if the four corner states split before the bulk gap closes, the topological phase is destroyed by spin fluctuations at finite temperature, which the paper does not analyze.","The same Wannier-center bookkeeping likely extends to other magnetic point groups combining time reversal with fourfold rotation, in which case a quantized magnetic quadrupole moment would be a general feature of such antiferromagnetic higher-order topological insulators rather than a crisscross-lattice speciality."],"forward_implications":["In the 2D nontrivial phase, cutting a finite square leaves four zero-energy corner states that share two electrons at half filling, so each corner carries an $e/2$ charge exponentially localized at the corner.","Because the magnetic quadrupole moment is quantized whenever the Wannier centers sit at the cell corners, the corner phase is not only a charge quadrupole insulator but also a magnetic quadrupole insulator, a combination absent from prior charge-only quadrupole models.","The 3D model is a chiral second-order topological insulator whenever condition (8) holds, and its $\\mathbb{Z}_2$ invariant distinguishes it from the $\\mathbb{Z}_4$ classification of $\\mathrm{EuIn}_2\\mathrm{As}_2$.","An electric field along $y$ induces opposite half-quantum Hall currents on the two $x$-normal side surfaces, connected by the surface states on the top and bottom faces, a manifestation of the topological magnetoelectric effect with axion angle $\\theta=\\pi$.","An electric field along $z$ pumps charge from two diagonal hinges to the other two; because the hinge spins are pinned in the $xy$-plane by $\\mathcal{M}_z\\mathcal{T}$, the spin direction flips on each side surface, giving the half-quantum spin-flop pumping signature."],"supporting_citations":[{"why":"Defines the quantized charge quadrupole moment and its relation to corner states in 2D second-order topological insulators; the paper's $q_{xy}$ formula is taken from here.","marker":"[13]"},{"why":"Sets up higher-order bulk-boundary correspondence and the chiral-versus-helical hinge-state classification, and supplies the $\\mathcal{S}_4$ polarization formula used for Wannier centers.","marker":"[17]"},{"why":"Provides an earlier chiral higher-order topological insulator built from $\\mathcal{S}_4$ rotoinversion; the paper contrasts its new magnetic quadrupole with this construction.","marker":"[24]"},{"why":"Another $\\mathcal{S}_4$-symmetric chiral higher-order topological insulator model that the paper compares with; also used for the behavior of connected hinge states when a Zeeman field preserves $\\mathcal{S}_4$.","marker":"[26]"},{"why":"Supplies the $\\mathrm{EuIn}_2\\mathrm{As}_2$ example whose $\\mathbb{Z}_4$ classification is contrasted with the $\\mathbb{Z}_2$ invariant proposed here.","marker":"[28]"},{"why":"Gives the half-quantum Hall conductance $e^2/2h\\,\\mathrm{sgn}(m)$ of a massive Dirac fermion, which underlies the predicted surface transport.","marker":"[31]"},{"why":"Provides the $\\theta=\\pi$ axion electrodynamics that the paper uses to interpret the side-surface magnetoelectric response.","marker":"[33]"}],"fun_headline_variants":["Crisscross AFM hosts higher-order topology with e/2 corners","AFM crisscross lattice gives fractional corner charges and hinge modes","Higher-order topological insulator from crisscross antiferromagnetism","e/2 corners and hinge currents from crisscross AFM model","Crisscross AFM: quantized magnetic quadrupole and hinge states"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"Everything rests on the magnetic moments staying exactly in the $xy$-plane in the crisscross pattern of Fig. 1; any canting, domain formation, or extra Heisenberg term that breaks the combined mirror-time-reversal symmetry $\\mathcal{M}_z\\mathcal{T}$ removes the quantization of the Wannier centers and with it the charge and magnetic quadrupole moments.","fun_headline_variants_meta":{"raw":{"variants":["Crisscross AFM hosts higher-order topology with e/2 corners","AFM crisscross lattice gives fractional corner charges and hinge modes","Higher-order topological insulator from crisscross antiferromagnetism","e/2 corners and hinge currents from crisscross AFM model","Crisscross AFM: quantized magnetic quadrupole and hinge states"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000579,"raw_usage":{"total_tokens":2737,"prompt_tokens":959,"completion_tokens":1778,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":575,"completion_tokens_details":{"reasoning_tokens":1684}},"tokens_in":575,"tokens_out":1778,"duration_ms":13535,"temperature":1.0,"reasoning_tokens":1684,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T11:15:55.038096+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"In the tight-binding model, the decisive calculation is to add a symmetry-breaking term that tilts the spins out of the $xy$-plane and track the four corner states: if they split or move off zero energy before the bulk gap closes, the claimed $\\mathbb{Z}_2$ protection fails. Experimentally, neutron diffraction or torque magnetometry showing ordered moments canted out of the plane would falsify the model's magnetic premise.","supporting_citations":[{"cited_title":"van Miert and C","cited_arxiv_id":null,"evidence_quote":"Supplies the $\\mathrm{EuIn}_2\\mathrm{As}_2$ example whose $\\mathbb{Z}_4$ classification is contrasted with the $\\mathbb{Z}_2$ invariant proposed here."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the half-quantum Hall conductance $e^2/2h\\,\\mathrm{sgn}(m)$ of a massive Dirac fermion, which underlies the predicted surface transport."},{"cited_title":"Okuma, M","cited_arxiv_id":null,"evidence_quote":"Provides the $\\theta=\\pi$ axion electrodynamics that the paper uses to interpret the side-surface magnetoelectric response."}],"review_version":1}