{"id":"60516e1e-38ce-4bca-ad7d-3acb18e8717d","arxiv_id":"2411.12434","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"A centrosymmetric antiferromagnet with bulk magnetic octupoles acquires a multiferroic surface with net magnetization, electric dipole, and linear magnetoelectric response, demonstrated for FeF2.","lead":"Surfaces of a common type of antiferromagnet can develop net magnetism, electric polarization, and a linear magnetoelectric response all at once, even though the bulk material has none of these properties. This could make surface layers of such antiferromagnets useful for magnetoelectric devices and sensors.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The direct numerical evidence for the surface linear ME effect rests on DMCs extracted from a slab whose inversion symmetry was artificially broken by moving one F atom; no convergence or control study is provided, so the reported α_xz=α_yz may be an artifact.","rationale":"I agree with the reader's weakest-assumption identification: the artificial F-atom displacement in Supplement I is the most load-bearing concern. The DMC calculation is the only direct numerical computation of the linear ME response, and the paper itself states that the DMCs vanish in the undistorted slab. The ad hoc inversion breaking therefore changes the reference structure from the physical surface to a weakly polarized slab, and there is no evidence that the extracted DMCs are insensitive to the details of that distortion. This is a genuine correctness risk because the bulk quadratic ME effect could contaminate the tiny linear signal. The concern does not, however, undermine the symmetry-based prediction that the (110) surface belongs to a magnetic point group allowing α_xz=α_yz, nor does it invalidate the computed surface magnetization or electric dipole. Those results support a conditional acceptance provided the DMC procedure is validated. The reader's verdict of CONDITIONAL is therefore appropriate, and my stress-test does not change it.","tokens_in":11255,"tokens_out":13350,"duration_ms":148926,"concrete_test":"Repeat the DMC extraction of Fig. 3(b) using at least three amplitudes of the artificial F-atom displacement (e.g., 0.005, 0.01, 0.02, 0.05 Å), with the F atom placed on the top and bottom surfaces and displaced in both directions along [110], and extrapolate the extracted layer-resolved DMC to zero F-displacement. If the extrapolated values do not converge to a single nonzero limit, or if they depend on the chosen F atom or direction, the reported DMCs are artifacts of the artificial inversion breaking. As a cross-check, recompute the surface DMC in an asymmetric slab with only one physical (110) surface (the other side passivated with a nonmagnetic insulator) without any F displacement; agreement with the extrapolated values would indicate that the artificial distortion faithfully represents the surface's intrinsic electric field.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The most load-bearing weakness is the DMC calculation that provides the only explicit numerical evidence for a nonzero linear ME response. Supplement I states that the undistorted slab is inversion-symmetric, so the DMCs (and the linear ME response) are in principle expected to vanish, and that inversion was broken by 'slightly mov[ing] one F atom.' This ad hoc displacement is not a property of the physical (110) surface and is not shown to be a controlled perturbation: no amplitude, direction, or choice of F atom is varied, and no extrapolation to the undistorted limit is given. Because bulk FeF2 has a quadratic ME response, the internal electric field created by the displaced F atom can generate a contribution to the induced magnetic moment that scales like β E^2 and contaminates the linear DMC. Since the reported Z^m values are tiny (10^-4 μB/Å scale), this contamination could easily be of the same order. The central claim that the surface 'exhibits a linear magnetoelectric effect' therefore rests on an unvalidated numerical procedure; the symmetry analysis alone shows that α is allowed, not that it is nonzero. The surface magnetization and electric dipole calculations are not affected by this issue, and the conclusion that the surface is multiferroic in the broad coexistence sense retains some support, but the quantitative ME prediction needs independent confirmation.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript argues that the (110) and (1-10) surfaces of rutile FeF2, a centrosymmetric collinear compensated antiferromagnet with ferroically ordered bulk magnetic octupoles, are multiferroic: they exhibit a net magnetization, a net electric dipole moment, and a linear magnetoelectric response, all in the absence of spin-orbit coupling. Using DFT+U calculations and a multipole decomposition of the magnetization density, the authors compute layer-resolved surface magnetization, electric dipoles, and magnetoelectric multipoles. They further compute dynamical magnetic charges (DMCs) to claim explicit confirmation of the surface linear ME effect, and they propose a bulk-boundary correspondence linking the surface ME response to the bulk magnetic octupoles.","tokens_in":11580,"tokens_out":10239,"duration_ms":91079,"significance":"If the predictions are correct, this is a valuable extension of surface magnetism to the emerging class of altermagnetic materials, and the identification of a surface linear ME effect without spin-orbit coupling would be notable. The paper's strengths include a clear symmetry-based framework, explicit first-principles calculations with provided computational parameters, layer-resolved predictions that are falsifiable by surface magnetometry, and a bulk-boundary correspondence that goes beyond the specific example. The multipole decomposition is used consistently with the authors' prior work, and the calculations for surface magnetization and electric dipole are straightforward. However, the quantitative confirmation of the linear ME effect via DMCs is not yet convincing, and the multiferroic classification is not fully justified.","major_comments":[{"comment":"The DMC calculation that is presented as confirmation of the surface linear ME effect is not a controlled perturbation. As stated in Supplement I, the undistorted slab is inversion-symmetric, so the linear DMC is expected to vanish by symmetry, and the authors break inversion by slightly moving one F atom. No details are given for this displacement (amplitude, direction, or which F atom is moved), no convergence with respect to the displacement is reported, and no extrapolation to the zero-displacement limit is provided. Because the bulk has a quadratic ME response, the internal electric field generated by the displaced F atom can induce a contribution proportional to βE^2 that contaminates the linear term; with the reported DMC values on the order of 10^-4 μB/Å, such contamination could easily be comparable in magnitude. The claim in the main text that the linear response 'confirms the emergence of surface ME response' is therefore not supported by the presented data. A control study varying the artificial F displacement, and ideally comparing different choices of the displaced atom, is required.","section":"Supplement I and Fig. 3"},{"comment":"The computed DMC is ∂m_z/∂u_Fe,[110], i.e., the magnetic response to a displacement of Fe atoms along the surface normal [110], while the claimed linear ME tensor components are α_xz and α_yz. The mapping between the displacement along [110] and the electric-field components E_x and E_y is not explained. Since the linear ME tensor is not necessarily symmetric, the response to a [110] field need not equal α_xz or α_yz. The paper should state explicitly which DMC component is computed, how it relates to the reported α components, and why displacing only Fe atoms, and not F atoms, captures the relevant lattice-mediated response.","section":"Main text, 'Next, we confirm...' paragraph and Fig. 3"},{"comment":"The derivation of the bulk-boundary correspondence assumes that the intrinsic electric field at the (110) surface has the form E=(E_x,E_y,0) with E_x=E_y, and E_x=-E_y for the (1-10) surface. This assumption is never derived from the surface structure or computed from first principles. It is load-bearing because it is used to obtain the relations M^surf_xz = ± M^surf_yz and to connect the surface ME response to the bulk octupole components. Please provide a justification for this form of the intrinsic field, or test it explicitly in the slab calculations.","section":"Bulk-boundary correspondence, Eqs. (1)-(3)"},{"comment":"The classification of the surface as multiferroic rests on the coexistence of a net magnetization and a net electric dipole moment at the surface. However, the paper's own definition of an ME multiferroic requires two primary ferroic orders. Neither the surface magnetization nor the surface electric dipole is demonstrated to be switchable, and the electric dipole is a boundary property rather than a bulk ferroelectric order. Please clarify whether the term 'multiferroic' is used in a weaker sense, namely coexistence of magnetic and polar responses, and discuss the implications for the central claim.","section":"Abstract and 'Emergent surface multiferroicity' section"}],"minor_comments":[{"comment":"The notation for the octupole components O_zyx and O_zxy appears before the definitions of the local and global octupole tensors; consider defining these components explicitly in the text to avoid confusion.","section":"Eq. (1)"},{"comment":"The statement that the non-relativistic surface linear ME effect is 'general to all d-wave NRSS materials' goes beyond the demonstrated example; please soften this claim or provide additional symmetry arguments that justify the generalization.","section":"Bulk-boundary correspondence section"},{"comment":"The dependence of the results on the Hubbard U (6 eV) and Hund's exchange J (0.95 eV) parameters is not tested; a brief statement about the robustness of the qualitative conclusions to these choices would be helpful.","section":"Supplement I"},{"comment":"In Fig. 3(a), the data points are described as 'shown as squares' in the caption, but the symbol is not identified in the main text; this is a minor editorial point.","section":"Fig. 3"}],"recommendation":"major_revision","confidential_remarks":"The paper is from a leading group and the physics is potentially important. The main issue is the DMC calculation: the authors need either to replace it with a controlled study (e.g., varying the artificial F displacement and extrapolating to zero) or to clearly demote it from 'confirmation' to a qualitative check. The multiferroic classification also needs careful wording. I believe the manuscript can be revised to address these issues without changing the core message."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: the genuinely new result is the coexistence of a surface electric dipole and a linear magnetoelectric response alongside the already-predicted surface magnetization, in a compensated antiferromagnet without spin-orbit coupling. The connection to bulk magnetic octupoles, via a bulk-boundary correspondence that should hold for all d-wave altermagnets, is a real conceptual advance. The symmetry analysis is clean, and the DFT+U layer-resolved plots for the surface dipole and magnetization are consistent with the magnetic point group m'm2'. If the linear ME response holds up, this is an important result.\n\nThe soft spot is exactly where the reader's report puts it, and the paper's own Supplement I concedes it: the undistorted slab is inversion-symmetric, so the DMCs should vanish in principle. To access them, the authors slightly move one F atom to break inversion. There is no control study, no variation of displacement amplitude or direction, and no extrapolation to the undistorted limit. Since bulk FeF2 has a quadratic ME response, the internal field from that displaced F atom can generate a magnetic response that scales like beta E^2 and contaminate the supposed linear term. The reported Z^m values are on the order of 10^-4 muB/Å, so contamination is a real possibility, not a remote worry. That means the quantitative alpha_xz = alpha_yz claim is not yet earned.\n\nWhat is earned: the symmetry argument that a linear surface ME response is allowed and would have that sign structure, and the surface magnetization and electric dipole calculations, which do not depend on the distorted slab at all. So the central 'multiferroic surface' conclusion, in the broad coexistence sense, rests on solid footing. Minor quibble: calling a static surface dipole 'multiferroic' is loose, but that is standard usage in this area. The reliance on the authors' own multipole formalism is a bit self-referential, but the surface results are computed from first principles, not derived only from the bulk octupoles, so that is acceptable.\n\nBottom line: I would not accept the linear ME response as numerically established, but I would absolutely send this to a serious referee. The symmetry argument and the surface dipole/magnetization results deserve publication, and the DMC issue is fixable with a proper control calculation. Anyone working on altermagnetism, surface magnetism, or magnetoelectric effects should engage with this paper.","headline":"A real step beyond the same group's surface-magnetization paper, but the only direct numerical evidence for the linear magnetoelectric effect has a hole the authors themselves admit; still worth a serious referee.","tokens_in":12067,"tokens_out":2416,"would_cite":true,"duration_ms":23860,"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":"A centrosymmetric, collinear, compensated antiferromagnet with ferroically ordered bulk magnetic octupoles develops a multiferroic surface—net magnetization, electric polarization, and a linear magnetoelectric response—even without…","keywords":["multiferroics","magnetoelectric effect","antiferromagnets","altermagnetism","magnetic octupoles","surface magnetization","FeF2","bulk-boundary correspondence"],"falsifier":"Measure the (110) surface of FeF2 with nitrogen-vacancy magnetometry or surface magneto-optical Kerr effect while sweeping an electric field along [110]; the predicted change in the [001] magnetization should be linear in the field and obey $\\alpha_{xz}=\\alpha_{yz}$. Observing no such linear change, or a different sign pattern between magnetic domains, would falsify the claim.","tokens_in":11058,"feed_emoji":"🧲","tokens_out":10437,"duration_ms":83824,"temperature":0.7,"pith_summary":"This paper establishes that a centrosymmetric, collinear, compensated antiferromagnet whose bulk hosts ferroically ordered magnetic octupoles develops a multiferroic surface: the surface acquires a net magnetization, a net electric dipole moment, and a linear magnetoelectric response, none of which exist in the bulk. Using FeF2 as a concrete example, first-principles calculations show the (110) and (1̄10) surfaces behave this way even in the absence of spin-orbit coupling, because the relevant multipoles are non-relativistic in origin. The result matters because it offers a route to magnetoelectric multiferroic behavior at the surfaces of ordinary antiferromagnets, and ties the surface response to the bulk octupolar order through a bulk–boundary correspondence.","feed_headline":"Antiferromagnet surfaces turn multiferroic with no spin-orbit coupling","feed_subtitle":"An antiferromagnet with no bulk multiferroicity becomes multiferroic at its surface, as predicted for FeF2.","key_machinery":"The load-bearing object is the magnetic octupole $O_{ijk}=\\int \\mu_i(\\mathbf{r}) r_j r_k\\, d^3r$, a rank-3, inversion-symmetric, time-reversal-broken multipole whose ferroic ordering defines d-wave altermagnetism. The argument runs: bulk ferroic octupoles couple to the surface's intrinsic electric field, producing surface magnetoelectric multipoles $M_{ij}=\\int r_i\\mu_j(\\mathbf{r})\\,d^3r$, which in turn set the non-zero components of the surface linear magnetoelectric tensor ($\\alpha_{xz}$ and $\\alpha_{yz}$). The surface linear response is confirmed computationally through layer-resolved dynamical magnetic charges—derivatives of the total magnetic moment with respect to Fe displacements—which are largest at the surface and vanish in bulk-like layers.","core_discovery":"On the paper's own terms, the central discovery is that bulk magnetic octupoles—rank-3 magnetic multipoles that break time reversal while preserving inversion—act as the source of surface multiferroicity. When a surface cuts the crystal, its intrinsic electric field acts on these bulk octupoles to generate surface magnetoelectric multipoles, a net magnetization, and an electric dipole moment. For FeF2 the calculations predict the (110) surface to have a linear magnetoelectric response with $\\alpha_{xz}=\\alpha_{yz}$ and the (1̄10) surface with $\\alpha_{xz}=-\\alpha_{yz}$; the net magnetization lies along [00̄1] for (110) and [001] for (1̄10), while the electric dipole points along the surface normal. Because the octupolar order exists without spin-orbit interaction in d-wave spin-split antiferromagnets (altermagnets), the surface multiferroicity is predicted to be general to that class of materials.","pith_inferences":["If this bulk–boundary correspondence is generic, surfaces and interfaces of altermagnets could become a systematic platform for multiferroic and magnetoelectric engineering without heavy elements, a device-oriented consequence the paper does not discuss.","The same mechanism may generate higher-order surface multipoles, such as surface magnetic toroidal moments, with experimental consequences the paper leaves implicit.","Because the dynamical magnetic charge calculation breaks inversion by a single artificial fluorine shift, a complementary approach—an asymmetric slab or surface passivation—would test how much of the predicted response is intrinsic rather than an artifact of that shift.","The predicted linear magnetoelectric response is lattice-mediated only; an independent electronic contribution could be probed by measuring the response at frequencies above optical phonons."],"forward_implications":["The (110) surface of FeF2 will show a linear magnetoelectric response $\\alpha_{xz}=\\alpha_{yz}$, the (1̄10) surface $\\alpha_{xz}=-\\alpha_{yz}$, and both will switch sign with the magnetic domain.","All d-wave spin-split antiferromagnets (altermagnets) with ferroic magnetic octupoles are expected to exhibit a nonrelativistic surface linear magnetoelectric effect and surface multiferroicity, without spin-orbit coupling.","The sign of the surface magnetization and of the magnetoelectric multipoles flips when the magnetic domain is reversed, while the electric dipole direction stays fixed, giving a simple domain readout.","The bulk–boundary correspondence explains which surface orientations show the effect: the orientation of the intrinsic surface electric field relative to the bulk octupole components selects the surface magnetization and magnetoelectric tensor components."],"supporting_citations":[{"why":"Earlier work classified surface magnetization in antiferromagnets; this paper builds on it to add polarization and the linear magnetoelectric response.","marker":"[4]"},{"why":"Supplies the irreducible tensor decomposition of magnetic octupoles and the quadratic magnetoelectric tensor used to identify the ordering in FeF2.","marker":"[10]"},{"why":"Establishes that d-wave altermagnets host ferroically ordered magnetic octupoles, the bulk property that the surface mechanism relies on.","marker":"[11]"},{"why":"Provides the atomic-site multipole extraction method used to compute the magnetic octupole and magnetoelectric multipole components.","marker":"[26]"},{"why":"Defines dynamical magnetic charges, the method used to compute the lattice-mediated surface linear magnetoelectric response.","marker":"[30]"},{"why":"Earlier calculation of uncompensated magnetization at the FeF2 (110) surface, whose result the present surface magnetization agrees with.","marker":"[29]"}],"fun_headline_variants":["Surface turns antiferromagnet into multiferroic","Multiferroicity emerges at antiferromagnet surface","Octupoles drive surface multiferroicity","Antiferromagnet surface gets magnetoelectric order","Surface multiferroicity without spin-orbit coupling"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The calculation that produces the surface magnetoelectric response relies on artificially shifting a single fluorine atom to break the slab's inversion symmetry, and the physical surface response is assumed to be the same as the response to that artificial shift.","fun_headline_variants_meta":{"raw":{"variants":["Surface turns antiferromagnet into multiferroic","Multiferroicity emerges at antiferromagnet surface","Octupoles drive surface multiferroicity","Antiferromagnet surface gets magnetoelectric order","Surface multiferroicity without spin-orbit coupling"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000431,"raw_usage":{"total_tokens":2172,"prompt_tokens":892,"completion_tokens":1280,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":508,"completion_tokens_details":{"reasoning_tokens":1204}},"tokens_in":508,"tokens_out":1280,"duration_ms":9577,"temperature":1.0,"reasoning_tokens":1204,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T17:32:11.784296+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the (110) surface of FeF2 with nitrogen-vacancy magnetometry or surface magneto-optical Kerr effect while sweeping an electric field along [110]; the predicted change in the [001] magnetization should be linear in the field and obey $\\alpha_{xz}=\\alpha_{yz}$. Observing no such linear change, or a different sign pattern between magnetic domains, would falsify the claim.","supporting_citations":[{"cited_title":"Urru and N","cited_arxiv_id":null,"evidence_quote":"Supplies the irreducible tensor decomposition of magnetic octupoles and the quadratic magnetoelectric tensor used to identify the ordering in FeF2."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the atomic-site multipole extraction method used to compute the magnetic octupole and magnetoelectric multipole components."},{"cited_title":"Ye and D","cited_arxiv_id":null,"evidence_quote":"Defines dynamical magnetic charges, the method used to compute the lattice-mediated surface linear magnetoelectric response."},{"cited_title":"Munoz, A","cited_arxiv_id":null,"evidence_quote":"Earlier calculation of uncompensated magnetization at the FeF2 (110) surface, whose result the present surface magnetization agrees with."}],"review_version":1}