{"id":"1cd3f4a2-b03b-492f-8695-a1c9ba31e749","arxiv_id":"1908.11354","paper_version":2,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Spin-orbit torques in a honeycomb antiferromagnet are computed numerically; they vanish in the symmetric model and become finite and anisotropic when sublattice symmetry is broken.","lead":"This paper uses quantum transport simulations to calculate spin-orbit torques in a model antiferromagnet on a honeycomb lattice. It finds that these torques vanish when the two sublattices are symmetric, and become finite and anisotropic when that symmetry is broken, which helps guide materials for antiferromagnetic memory devices.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Zero anti-damping torques in the symmetric model are established only within an assumed C∞v angular basis; the C3v lattice and transport setup may permit unmodeled harmonics.","rationale":"The reader's weakest_assumption correctly identifies the completeness of Eq. (25)/(38) as the central methodological risk, and I agree that the C∞v classification is not automatically a complete basis for the C3v lattice model. My read adds a second, closely related layer: the paper asserts, rather than derives, that exact sublattice symmetry forces B||=B⊥=0 even though the symmetry argument as written directly addresses δs− and not the δs+ components that carry the anti-damping torques. These two concerns are not fully independent, so I mark agreement as partial rather than full. I do not see a basis for rejecting the numerical results: the kwant scattering method is standard, the samples are large, the disorder averaging is repeated over many realizations, and the qualitative pattern—δs−=0 in the symmetric model and finite anisotropic anti-damping torques in the asymmetric model—is internally consistent and visually supported by the presented figures. The problem is the strength of the universal 'identically vanishing' claim relative to the evidence. A closed-form symmetry proof or a model-independent harmonic decomposition with error bars would settle it; without that, the correct editorial outcome is the reader's original CONDITIONAL verdict, so I recommend no change.","tokens_in":19414,"tokens_out":19146,"duration_ms":198082,"concrete_test":"Reanalyze the symmetric-model raw δs+ data with a complete C3v-invariant basis. Generate at least 200 Néel-vector orientations per Fermi energy and impurity concentration; include, besides the five C∞v vector forms in Eq. (25), the threefold harmonics allowed by C3v (e.g. terms built from the in-plane bond vectors e1,e2,e3 and the current direction, such as Σ_k (e_k·l)(e_k·[z×I]) and its cross product with l). Report the fit residuals and bootstrap error bars for the amplitudes of all C∞v-breaking forms, and the resulting change in aI and in the B||,B⊥ torques. If the C∞v-breaking amplitudes are below the statistical noise at both E=0.3w and E=0.05w, the ansatz is validated and the zero-torque claim survives; if they are on the order of the quoted coefficients, the claim is an artifact of the assumed angular basis.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim of Section VII is that the symmetric model has B||=B⊥=0 identically in all regimes, leaving only the Edelstein term (26)/(27) in the metal regime. The numerical fit uses Eq. (25) or (28), whose completeness is asserted from the C∞v symmetry of the low-energy effective model (23). The actual tight-binding model (1)-(6) has C3v point-group symmetry, and the two-terminal ribbon setup with zigzag edges and current along x has even lower symmetry. Section VI states that the exact sublattice symmetry 'also leads to' B||=B⊥=0, but no derivation is given; the symmetry argument for δs−=0 does not immediately constrain the uniformly polarized part δs+ that enters B||,B⊥. In particular, the B terms are odd under l→−l, while δs− is defined by A/B interchange; relating the two requires a concrete spin-rotation/sublattice operation that is not supplied. If additional C3v-allowed but C∞v-forbidden harmonics contribute to δs+, the least-squares fit to the five-vector ansatz can absorb them, biasing the extracted coefficients and concealing a finite anti-damping torque. With 30–80 disorder realizations and no reported error bars or released code, the 'identically vanishing' wording is stronger than what the numerical evidence alone establishes.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This manuscript studies current-induced spin-orbit torques (SOTs) in a two-dimensional honeycomb antiferromagnet with Rashba spin-orbit coupling and on-site disorder, using the kwant transport solver. The authors compute non-equilibrium spin densities on the two sublattices for many Néel-vector orientations and fit them to symmetry-motivated angular decompositions. In the symmetric model they conclude that the staggered polarization and all anti-damping torques vanish identically, leaving an isotropic Edelstein field-like torque in the metal regime and additional field-like harmonics in the half-metal regime. In an asymmetric model with s-d coupling on one sublattice only, they find finite anisotropic, disorder-dependent anti-damping torques. The final section uses these torques in Landau-Lifshitz-Gilbert simulations to illustrate pulsed-current Néel-vector switching.","tokens_in":19736,"tokens_out":10471,"duration_ms":103976,"significance":"If the central identity claims hold, the paper provides a useful microscopic benchmark for symmetry classifications of SOTs in antiferromagnets and identifies 2D confinement as a source of torque anisotropy. Its strengths are the transparent scattering-based methodology, the systematic angular sampling over 200 Néel orientations, the clear separation of metal and half-metal regimes, and the connection to experimentally relevant ferrimagnetic systems. The main limitation is that the zero-torque result is verified only within a five-vector ansatz whose completeness is asserted from an effective C∞v model while the actual tight-binding model has C3v symmetry; without either a symmetry proof or a completeness check, the word \"identically\" is stronger than the numerical and symmetry evidence presented.","major_comments":[{"comment":"The central claim that the symmetric model has B∥ = B⊥ = 0 identically is not derived. The stated sublattice symmetry Λx Heff[−l]Λx = Heff[l] is used to argue δs− = 0, but the anti-damping coefficients in Eq. (25) multiply vector forms built from δs+, and no operation is exhibited that forces those coefficients to zero. In addition, the completeness of the C∞v decomposition (25) is asserted rather than proved for the actual C3v lattice; the zigzag two-terminal ribbon has even lower symmetry. If C3v-allowed harmonics are present, a least-squares fit to the five-vector ansatz can absorb them, biasing the coefficients and concealing a finite anti-damping torque. Please provide an explicit symmetry derivation for B∥ = B⊥ = 0, or test the fit with an extended basis containing C3v-allowed terms and report the resulting coefficients.","section":"§VI–§VII, Eqs. (23)–(25)"},{"comment":"The \"identically vanishing\" statements are supported only by fits of disorder-averaged data, with 30–80 realizations and no reported error bars. Since the quantitative content of the paper is a set of fitted coefficients (aI, a′I, c, b⊥, b∥), the absence of confidence intervals makes it impossible to judge whether the zero values are exact to numerical precision or merely small. Please report standard errors or confidence intervals for the extracted coefficients, and state the residuals of the fits; for the asymmetric model, apply the same completeness check to Eq. (38), since the Fermi surfaces in Fig. 9 are warped by C3v effects and may require additional angular harmonics.","section":"§VII.A, Fig. 7 and §VIII, Fig. 11"},{"comment":"The prefactor in the linear-response formula for δs/δµ appears to be off by a factor of two relative to Eq. (13): from the zero-temperature limit of (13), δs/δµ = (1/4πℏ) Σα [...] = (1/(2h)) Σα [...], whereas Eq. (15) states 1/(4h). Since the normalized coefficients in Figs. 7 and 11 and the prefactor η in Eqs. (27) and (39) inherit this normalization, please clarify the definition of h and the summation convention, and correct the prefactor if needed.","section":"Eq. (15)"}],"minor_comments":[{"comment":"The caption of Fig. 12 states that the top and bottom panels show dynamics for the symmetric and asymmetric models, but the text does not state the initial Néel orientation or the duration of the current pulses; please add these parameters for reproducibility.","section":"Fig. 12 and §IX"},{"comment":"The notation σ in Eq. (12) is called a dimensionless 2D conductivity, but it is extracted from L⟨G⟩/W with a factor 2e²/h; please state the units explicitly (for example, in units of e²/h per square) to avoid confusion.","section":"Eq. (12)"},{"comment":"In Eq. (20), the torque terms contain (JA/ℏ)(l×s+ + m×s−) and (JA/ℏ)(m×s+ + l×s−); after substituting δs± in Eq. (21), the relation to the four torques T_l± and T_m± is clear, but the sign convention for δs− should be stated explicitly in the text.","section":"§V, Eq. (20)"}],"recommendation":"major_revision","confidential_remarks":"The central zero-torque claim is likely correct on physical grounds, but as written it is a strong identity claim resting on an unproven ansatz; the revision should either provide a rigorous symmetry derivation or soften the wording to numerical evidence within the C∞v decomposition. I also note that no code or data release is mentioned, which would strengthen reproducibility for a purely numerical paper."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First, the punchline: this paper gives the first full numerical map of spin-orbit torques in a disordered honeycomb antiferromagnet with Rashba coupling, and its central qualitative finding is clean — in the sublattice-symmetric model the anti-damping torques vanish in every regime they tested, leaving only the Edelstein field-like torque in the metal regime. That goes beyond the ferromagnet analytics and is the thing to remember.\n\nWhat it does well: the kwant implementation is standard and the transport formalism is sound; the decomposition of the non-equilibrium spin density into symmetry-allowed vector forms is clearly laid out; they separate metal and half-metal regimes; and the disorder dependence of the asymmetric-model anti-damping torques makes physical sense and is a useful benchmark. The paper also honestly flags that the asymmetric model is closer to a ferrimagnet than an antiferromagnet.\n\nWhere it is soft: the claims of 'identically vanishing' B=0 are asserted in Sec. VI with, at most, a symmetry hint, not a derivation. The sublattice symmetry operation that forces δs−=0 does not by itself constrain the uniform component δs+ that enters the anti-damping torques. The fit ansatz (25)/(28) is motivated by C∞v symmetry of the low-energy model, but the full lattice is C3v and the two-terminal ribbon is lower still; if C3v-allowed extra harmonics exist, the least-squares fit could hide them. No error bars are given for 30–80 disorder realizations, and no code or data are released, so 'identically' is stronger than the numerical evidence. I also noticed a factor-of-two ambiguity in the spin density normalization (Eq. 15) that should be cleaned up — minor, but it matters for comparing coefficients.\n\nNone of this sinks the paper. The central result is probably right, and the paper is careful about what is and is not included. The missing derivation and the absence of error bars/code are addressable in revision.\n\nThis paper is for the antiferromagnetic spintronics subfield, especially people who work on symmetry classification of spin-orbit torques and on material/current-geometry choices for AFM memory devices. It deserves a serious referee: the numerical data are credible, the qualitative finding is new, and the weak spots are fixable. I would send it to peer review, and I would ask the authors to either prove the vanishing anti-damping claim or soften it to 'zero within numerical accuracy', and to release the code and data.","headline":"A clean numerical result — zero anti-damping torques in the symmetric Rashba honeycomb antiferromagnet — that would be more convincing if the 'identically vanishing' claims were derived or backed with error bars.","tokens_in":20233,"tokens_out":2313,"would_cite":false,"duration_ms":22242,"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 sublattice-symmetric Rashba honeycomb antiferromagnet exhibits no Néel and no anti-damping spin-orbit torques; only an isotropic Edelstein field-like torque remains, and broken sublattice symmetry restores finite anisotropic…","keywords":["spin-orbit torque","antiferromagnet","Rashba spin-orbit coupling","honeycomb lattice","Néel vector","Edelstein effect","anti-damping torque","sublattice symmetry"],"falsifier":"Compute the non-equilibrium spin density in the symmetric honeycomb model at a Fermi energy well above the Dirac point, where Fermi surfaces are strongly warped by the $C_{3v}$ lattice symmetry, and fit the response with basis functions that include threefold and sixfold angular harmonics. If any such harmonic, or any component of the torque along the anti-damping directions, is nonzero, then the identically vanishing anti-damping torque is an artifact of the symmetric $C_{\\infty v}$ ansatz rather than a property of the lattice model.","tokens_in":1955,"feed_emoji":"🧲","tokens_out":2781,"duration_ms":90236,"temperature":0.7,"pith_summary":"This paper asks which spin-orbit torques an electric current exerts on the staggered order of a two-dimensional honeycomb antiferromagnet with Rashba spin-orbit coupling. Using a microscopic tight-binding model with on-site disorder, the authors compute the non-equilibrium spin density that drives the torques. They establish that when the s-d exchange coupling is identical on the two sublattices, the staggered non-equilibrium polarization vanishes and all anti-damping torques vanish identically; in the metallic regime the only remaining torque is the isotropic Edelstein field-like torque. When the exchange coupling is placed on only one sublattice, finite, anisotropic, disorder-dependent anti-damping torques appear. The result matters because anti-damping torques and staggered Néel torques are the two mechanisms proposed for electrical switching of antiferromagnetic domains, and the paper shows both are entirely absent under exact sublattice symmetry.","feed_headline":"Sublattice symmetry kills anti-damping spin torques","feed_subtitle":"Exact sublattice symmetry leaves only the isotropic Edelstein torque; broken symmetry restores anisotropic anti-damping torques.","key_machinery":"The load-bearing object is the non-equilibrium spin-density response $\\delta s/\\delta\\mu$ computed from scattering states in a two-terminal geometry and decomposed onto the vector basis allowed by the low-energy $C_{\\infty v}$ symmetry. The argument is carried by the exact sublattice symmetry of the symmetric model: conjugation by $\\Lambda_x$ (exchange of sublattices) combined with $\\boldsymbol{\\ell}\\to -\\boldsymbol{\\ell}$ leaves the Hamiltonian invariant, which makes the staggered polarization $\\delta s_-$ vanish and forces the two anti-damping coefficients $B_\\perp$ and $B_\\parallel$ to zero. The remaining coefficients $a_I$, $a'_I$, $c$ are extracted numerically by fitting over many Néel-vector orientations; their dependence on disorder and on the polar angle $\\theta$ encodes the field-like torques. The same ansatz, applied to the asymmetric one-sublattice model with vector $n_A$, yields the four torque coefficients $a_I$, $b_\\perp$, $b_\\parallel$, $c$ that are the paper's main output.","core_discovery":"The central discovery is an exact symmetry-enforced vanishing of two of the three classes of spin-orbit torque in the symmetric model. With identical s-d exchange on A and B sublattices, the Hamiltonian obeys a sublattice symmetry $\\Lambda_x H[-\\boldsymbol{\\ell}]\\Lambda_x = H[\\boldsymbol{\\ell}]$; the paper argues and numerically confirms that this forces the non-equilibrium staggered spin density $\\delta s_-$ to zero and the anti-damping torque coefficients $B_\\perp = B_\\parallel = 0$ in all transport regimes. What remains in the metal regime is the single isotropic Edelstein term $\\delta s_+ \\propto \\hat{z}\\times j$, i.e. a purely field-like torque of the inverse spin-galvanic type. In the half-metal regime two additional field-like high-harmonic torques with coefficients $a'_I$ and $c$ appear, but still no anti-damping torque. If the sublattice symmetry is broken by coupling s-d exchange to only one sublattice, the torque acquires four angular coefficients; the anti-damping coefficients $b_\\perp$ and $b_\\parallel$ become finite and strongly depend on impurity concentration, while field-like coefficients remain largely disorder-insensitive.","pith_inferences":["If the $C_{\\infty v}$ ansatz is not complete, the reported zeros could fail at higher Fermi energies where Fermi-surface warping from the lattice $C_{3v}$ symmetry becomes significant; a direct check would be to fit scattering data including threefold and sixfold harmonics.","The symmetric-model result suggests a general design rule: any bipartite antiferromagnet with equal coupling on both sublattices and Rashba-type spin-orbit coupling should exhibit only field-like torques, so a measured anti-damping torque in such a system is a fingerprint of sublattice-symmetry breaking.","The disorder-dependence pattern (metal versus half-metal) could be tested in half-metallic antiferromagnets by comparing torque efficiencies in samples with different resistivities.","The switching mechanism via a transient fully magnetized state during an anti-damping-dominated pulse, which the paper notes resembles all-optical switching in GdFeCo, could be probed by time-resolved magneto-optical measurements on current pulses."],"forward_implications":["In a sublattice-symmetric Rashba honeycomb antiferromagnet, current-driven switching of the Néel vector cannot proceed through the two torques previously thought responsible; only the field-like Edelstein torque acts, which is generally too weak to switch.","The metal-regime spin-orbit torque in such a symmetric antiferromagnet is fully described by one angle-independent coefficient $a_I$, independent of impurity concentration, making it a direct analogue of the Edelstein effect.","Breaking sublattice symmetry creates anti-damping torques whose anisotropy is controlled by the Néel-vector orientation: in-plane orientations maximize them in the metal regime.","Anti-damping torques scale oppositely with disorder in the two regimes — suppressed in the metal regime, enhanced in the half-metal regime — so impurity engineering can tune them.","In the asymmetric model, the staggered or Néel torque is no longer a useful notion; dynamics is governed by the single-lattice torque $T^A$, and the model applies to ferrimagnets like GdFeCo."],"supporting_citations":[{"why":"Provides the tight-binding scattering method used to compute transmission coefficients and local non-equilibrium spin densities in disordered samples.","marker":"51"},{"why":"Defines the Kane-Mele honeycomb-lattice tight-binding model that the Rashba and s-d exchange terms extend.","marker":"52,53"},{"why":"Supplies the analytic Rashba ferromagnet result with white-noise disorder that the symmetric metal regime reproduces.","marker":"55"},{"why":"Relates non-equilibrium spin density to spin-orbit torques and Gilbert damping, motivating the decomposition into torque coefficients.","marker":"56"},{"why":"Supply the $C_{\\infty v}$ symmetry classification of spin-density vector forms on which the fitting ansatz is based.","marker":"58,59"},{"why":"Identifies the inverse spin-galvanic Edelstein effect that matches the metal-regime torque of the symmetric model.","marker":"60"},{"why":"Supply the magnetic-point-group symmetry classification of antiferromagnetic spin-orbit torques that the numerical decomposition confirms.","marker":"37,43"}],"fun_headline_variants":["Symmetry kills anti-damping torques in antiferromagnet","Broken sublattice symmetry revives spin-orbit torques","Exact symmetry leaves only Edelstein torque","Honeycomb antiferromagnet: symmetry gates spin torques","Anti-damping torques vanish under sublattice symmetry"],"cache_read_input_tokens":22400,"weakest_assumption_plain":"The paper's conclusions rely on the assumption that the non-equilibrium spin density contains only the vector harmonics allowed by the low-energy $C_{\\infty v}$ symmetry; if additional angular harmonics allowed by the full $C_{3v}$ lattice symmetry are present, the fitted zero anti-damping coefficients would be incomplete and could be nonzero.","fun_headline_variants_meta":{"raw":{"variants":["Symmetry kills anti-damping torques in antiferromagnet","Broken sublattice symmetry revives spin-orbit torques","Exact symmetry leaves only Edelstein torque","Honeycomb antiferromagnet: symmetry gates spin torques","Anti-damping torques vanish under sublattice symmetry"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000224,"raw_usage":{"total_tokens":1496,"prompt_tokens":1017,"completion_tokens":479,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":633,"completion_tokens_details":{"reasoning_tokens":410}},"tokens_in":633,"tokens_out":479,"duration_ms":4530,"temperature":1.0,"reasoning_tokens":410,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T10:17:21.848864+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the non-equilibrium spin density in the symmetric honeycomb model at a Fermi energy well above the Dirac point, where Fermi surfaces are strongly warped by the $C_{3v}$ lattice symmetry, and fit the response with basis functions that include threefold and sixfold angular harmonics. If any such harmonic, or any component of the torque along the anti-damping directions, is nonzero, then the identically vanishing anti-damping torque is an artifact of the symmetric $C_{\\infty v}$ ansatz rather than a property of the lattice model.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the analytic Rashba ferromagnet result with white-noise disorder that the symmetric metal regime reproduces."},{"cited_title":"Anisotropy of spin-transfer torques and Gilbert damping induced by Rashba coupling","cited_arxiv_id":"1907.02041","evidence_quote":"Relates non-equilibrium spin density to spin-orbit torques and Gilbert damping, motivating the decomposition into torque coefficients."}],"review_version":1}