{"id":"ea07b747-008e-46c4-ae48-96fd02f30b84","arxiv_id":"2509.21877","paper_version":2,"verdict":"CONDITIONAL","confidence":"LOW","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"A first-principles DFT calculation predicts a tunnel magnetoresistance ratio above 1000 percent in Mn3Sn(01-10)/MgO(110)/Mn3Sn junctions with a 12-monolayer MgO barrier.","lead":"An ab initio calculation predicts that a sandwich of the antiferromagnet Mn3Sn, an MgO insulator, and another Mn3Sn layer shows a tunnel magnetoresistance ratio above 1000 percent when the two magnetic layers are switched between aligned and reversed states. The result matters because it suggests that antiferromagnets, which produce no stray magnetic fields, could still generate practically large readout signals in magnetic memory devices.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The >1000% TMR result depends on an untested coherent-interface premise: the 7.6% mismatch is imposed entirely on MgO, with no in-plane relaxation or dislocation/termination sensitivity check; relaxing this premise could change the MgO filtering and the TMR.","rationale":"The reader's weakest_assumption is exactly the premise I would flag: the coherent, defect-free interface with all 7.6% mismatch placed on MgO. My independent reading of the manuscript confirms this is the least-secure condition for the abstract claim. The symmetry analysis in Sec. III A (Eqs. (5)-(6)) is internally consistent, the Landauer-Büttiker/pwcond transport calculation is a standard tool, and the qualitative mechanism (Mn3Sn k-space spin splitting plus MgO filtering) is plausible and supported by the vacuum comparison in Appendix A. None of this, however, controls the quantitative headline: TMR >1000% at 12 ML is a ratio of two exponentially small conductances, and no k-point, thickness, strain-partition, or interface-termination convergence data are shown. The manuscript is also incomplete (placeholder figure captions, 'To be updated' markers), which further limits verification but is secondary to the scientific concern. Because the reader already conditioned acceptance on interface sensitivity, my pass does not change the verdict; it sharpens the required check.","tokens_in":17306,"tokens_out":5694,"duration_ms":56297,"concrete_test":"Recompute T_P, T_AP, and the k-parallel-resolved transmission for N_MgO(110)=12 in a near-commensurate supercell that distributes the 7.6% mismatch (e.g., 13 Mn3Sn in-plane cells matched to 14 MgO cells, residual strain <0.1%), with full relaxation of all atomic positions including in-plane coordinates; compare with the fully strained-MgO result. If the TMR ratio drops below ~1000% or the T_P-T_AP peak near k_parallel=0 disappears, the coherent-interface assumption is load-bearing for the central claim. A cheaper complementary check is to repeat the calculation with the strain placed on Mn3Sn instead of MgO and see whether the TMR magnitude survives.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central numerical claim is the zero-temperature TMR ratio ≳1000% at N_MgO(110)=12 (Sec. III B). For this number to be more than a property of the particular computational cell, the modeled interface must be representative. Sec. II A states that the ~7.6% lattice mismatch is accommodated by applying in-plane strain to MgO, and that only the interfacial distance is optimized over high-symmetry stackings. No calculation tests what happens if the strain is partitioned differently, if in-plane atomic positions relax, or if misfit dislocations or interface intermixing appear. The mechanism itself is explicitly attributed to MgO's k-parallel filtering (complex band structure, Fig. 7) and to the momentum-dependent Mn3Sn spin polarization. Both depend on the precise interface geometry: a 7.6% biaxially strained MgO(110) barrier has different decay rates and band alignment than a relaxed or dislocated barrier, and the interfacial density of states (Fig. 4) changes with termination and chemical bonding. Since the predicted magnitude above 1000% involves a ratio of two small tunneling conductances, it is especially sensitive to small changes in the minority (T_AP) channel. Thus the strongest claim is load-bearing on a structural premise that the paper asserts but does not validate. This is not an internal inconsistency; the calculation is a standard DFT/pwcond transport setup. But without an interface-sensitivity check, the headline number is not established.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports a first-principles study of the Mn3Sn(01-10)/MgO(110)/Mn3Sn magnetic tunnel junction. Using density functional theory with the Quantum ESPRESSO package and the pwcond Landauer-Buttiker transmission code, the authors compute zero-temperature tunneling conductances for parallel and antiparallel configurations of the noncollinear antiferromagnetic Mn3Sn electrodes. They find that the TMR ratio reaches about 1000% or more at 12 MgO monolayers, and they interpret this as arising from the momentum-dependent x-component of the Mn3Sn spin polarization combined with the k-parallel filtering of the MgO barrier, as supported by symmetry relations for the bulk spin polarization, momentum-resolved transmission maps, a vacuum-junction control calculation, and the complex band structure of MgO.","tokens_in":17597,"tokens_out":6758,"duration_ms":59693,"significance":"If the numerical result is robust, the paper provides a valuable ab initio prediction of a large TMR effect in an all-antiferromagnetic tunnel junction with a practical MgO barrier and in a crystallographic orientation relevant for spin-orbit-torque switching. The work is not circular: the TMR ratio is a direct transmission output, and the spin-polarization and complex-band analyses are independent interpretive checks. The symmetry argument leading to Eq. (5) is clean, and the vacuum-junction comparison in Appendix A is a useful control. The main weakness is that the central quantitative claim is presented without final figures, without convergence or uncertainty analysis, and without sensitivity tests for the assumed coherent-interface structure.","major_comments":[{"comment":"The headline claim that the TMR ratio reaches ≳1000% at N_MgO(110)=12 cannot be fully evaluated from the submitted manuscript because Fig. 5 is marked with the placeholder '!!To be updated!!!' and no numerical values for T_P, T_AP, or the TMR ratio are given in the text. Since the TMR ratio is a ratio of two exponentially small tunneling conductances, the authors should provide the final figure together with the raw T_P and T_AP values, a convergence check with respect to the 51x51 versus 101x101 k_parallel meshes, and a statement of the numerical uncertainty from the Gaussian smearing used in Eq. (4).","section":"Sec. III B, Fig. 5"},{"comment":"The interface model accommodates the ~7.6% lattice mismatch entirely by straining the MgO lattice, and only the interfacial distance is optimized over several high-symmetry stackings. No test addresses alternative strain partitioning, in-plane atomic relaxation, misfit dislocations, or different interface terminations. Because the predicted TMR ratio depends on the momentum-dependent filtering of MgO and on the interfacial Mn3Sn electronic structure, and because T_AP is a small conductance, these structural choices are load-bearing for the predicted ≳1000% effect; the authors should add explicit sensitivity calculations or provide a quantitative argument establishing that the chosen coherent-interface model is representative.","section":"Sec. II A"},{"comment":"The antiparallel configuration is constructed by doubling the scattering region and then cutting it in half before the transmission calculation, but the manuscript does not specify how the doubled cell is cut or verify that the Mn3Sn buffer layers are thick enough to decouple the two Mn3Sn/MgO interfaces. This matters because spurious coupling between the two interfaces in the doubled cell could affect the small T_AP conductance. The authors should describe the exact construction and report convergence with respect to the number of buffer layers.","section":"Sec. II B"},{"comment":"The self-consistent calculation constrains the initial Mn magnetic moments to 3.0 micro-Bohr magnetons with a 1.0 Ry penalty, and the paper states only that the inverse triangular structure is realized 'overall' with a 'tiny net magnetization'. Since the TMR mechanism is attributed to the specific magnetic configuration, the authors should show that the final results are insensitive to the constraint by comparing with an unconstrained calculation or by varying the constraint strength and initial moment directions.","section":"Sec. II A and Sec. II B"}],"minor_comments":[{"comment":"In the paragraph discussing Fig. 5, the text says 'the total transmission for the parallel configuration, T_P, and that for the parallel configuration, T_AP'; the second instance should read 'antiparallel configuration'.","section":"Sec. III B"},{"comment":"The numerical evaluation of Eq. (4) relies on a Gaussian smearing described only in footnote 93; please state the smearing width in the main text and note that the same smearing is used in the numerator and denominator.","section":"Sec. II C, Eq. (4)"},{"comment":"The schematic in Fig. 1(d) uses the labels 'Parallel MgO' and 'Antiparallel MgO', but the caption does not explain that these refer to the relative orientation of the cluster octupole moments of the two Mn3Sn electrodes; please clarify.","section":"Fig. 1"},{"comment":"The phrase 'momenum dependence' in the first paragraph of Appendix A should be corrected to 'momentum dependence'.","section":"Appendix A"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is not in publishable form because of the placeholder markers and the missing final Fig. 5 data, in addition to the interface-sensitivity and convergence issues described in the main report. I would advise the editor to request the final figures and the additional calculations before any further decision; the scientific approach is standard and the predicted effect, if confirmed, would be of interest to the mesoscopic magnetism and spintronics community."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a potentially interesting prediction—>1000% TMR in an all-antiferromagnetic Mn3Sn/MgO/Mn3Sn junction with a practical MgO barrier—but the manuscript is not ready for referees. It is an unfinished draft with 'To be updated!!!' placeholders in key figures, and the headline number rests on an untested assumption about how the 7.6% lattice mismatch is accommodated.\n\nWhat's new: earlier Mn3Sn-based TMR studies used vacuum, very thin barriers, or hBN. This is the first calculation I know of for the Mn3Sn(01-10)/MgO(110) orientation with a 12-monolayer MgO barrier, which is the geometry relevant to the SOT switching experiments. The mechanism analysis is a genuine step beyond just reporting a number: the symmetry argument for why the x-component of the spin polarization survives the k_z sum, the k_parallel-resolved transmission maps, and the MgO complex-band decay rates all hang together coherently. The vacuum-junction comparison (Appendix A) is a useful control to show the MgO filtering changes the k_parallel distribution.\n\nThe central calculation itself is standard DFT + Landauer-Büttiker, and the TMR is a direct output, not a fit. No circularity problem.\n\nSoft spots, in order of severity. First, the manuscript is incomplete: Figs. 2, 5, and 7 carry 'To be updated!!!' markers. No serious journal should send that to referees. Second, the headline >1000% number has no convergence analysis—no k_parallel mesh convergence, no smearing-width dependence, no sensitivity of T_AP, which is the small number in the denominator. TMR ratios computed as ratios of two tiny tunneling conductances are notoriously sensitive. Third, the interface model: the 7.6% mismatch is fully imposed on MgO, and only the interfacial distance is relaxed for a few high-symmetry stackings. The authors do not test in-plane relaxation, different strain partitioning, misfit dislocations, or chemical intermixing. The mechanism depends on MgO's k_parallel filtering, which will change under biaxial strain, so this is a real gap, not a footnote concern. The AP-state construction (doubling the cell, then cutting) is a bit tersely described but seems procedurally okay.\n\nWho this is for: people working on Mn3Sn spintronics and antiferromagnetic MTJs. If the calculation survives a completed draft with convergence checks and interface sensitivity tests, the prediction would be a useful benchmark. My recommendation: don't desk-reject the science, but send the authors back for a completed manuscript. The current version is not refereeable. If they address the placeholders and add at least a strain-partitioning or interface-relaxation test, I'd be happy to referee the revised version.","headline":"Promising but unfinished ab initio TMR prediction for Mn3Sn/MgO/Mn3Sn that needs completion and interface-sensitivity checks before it can be taken seriously.","tokens_in":18204,"tokens_out":2536,"would_cite":false,"duration_ms":21590,"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":"The paper predicts that an all-antiferromagnetic Mn3Sn/MgO/Mn3Sn tunnel junction reaches a tunnel magnetoresistance ratio above 1000 percent, driven by the momentum-dependent spin polarization of Mn3Sn and MgO's filtering of tunneling…","keywords":["tunnel magnetoresistance","antiferromagnetic tunnel junction","Mn3Sn","noncollinear antiferromagnet","MgO barrier","spin polarization","first-principles transport","cluster octupole"],"falsifier":"Grow an epitaxial $\\mathrm{Mn_3Sn}(01\\bar{1}0)/\\mathrm{MgO}(110)/\\mathrm{Mn_3Sn}$ stack with about 12 MgO monolayers, switch the cluster octupole moments between the parallel and antiparallel configurations, and measure the resistance-area product and TMR at low temperature; a TMR much below 1000 percent, or an RA far outside the predicted range of about 1 to 10 kOhm-micrometer-squared, would contradict the coherent-interface calculation.","tokens_in":17101,"feed_emoji":"🧲","tokens_out":9976,"duration_ms":83436,"temperature":0.7,"pith_summary":"This paper predicts that a magnetic tunnel junction built entirely from the antiferromagnet Mn3Sn, separated by a MgO barrier about twelve monolayers thick, can reach a tunnel magnetoresistance ratio above 1000 percent. The result matters because antiferromagnets normally have no net magnetization to act as a readable magnetic state; Mn3Sn's inverse 120-degree spin order breaks time-reversal symmetry and creates a momentum-space spin polarization even though the total moment vanishes. The prediction comes from first-principles electronic-structure and Landauer-Buettiker transport calculations, which trace the effect to the x-component of the spin polarization surviving the momentum sum, together with MgO's filtering of tunneling electrons near the zone center. If correct, all-antiferromagnetic junctions could offer the same readout contrast as ferromagnetic MgO junctions while keeping antiferromagnetic advantages such as the absence of stray fields.","feed_headline":"Antiferromagnetic tunnel junction predicted to exceed 1000% TMR","feed_subtitle":"At 12 MgO monolayers, the junction's resistance changes by an order of magnitude between magnetic states.","key_machinery":"The load-bearing objects are the momentum-resolved spin polarization of Mn3Sn and the complex band structure of the MgO barrier. From a Wannier-based tight-binding model, the paper computes $s_n(\\mathbf{k})=\\langle\\psi_{n,\\mathbf{k}}|\\mathbf{s}|\\psi_{n,\\mathbf{k}}\\rangle$ and its symmetry relations, then projects it onto the two-dimensional Brillouin zone as $p(k_\\parallel)$; the relations $s_x(k_x,k_y,k_z)=s_x(k_x,k_y,-k_z)$ and $s_z(k_x,k_y,k_z)=-s_z(k_x,k_y,-k_z)$ select the $x$-channel as the only one that survives the $k_z$ sum. On the barrier side, the lowest imaginary part of the complex wavevector, $\\kappa_{z0}(k_\\parallel)$, gives the dominant exponential decay of tunneling states; it is small near $k_\\parallel \\sim 0$, matching the region where $|p(k_\\parallel)|$ is large. The tunneling conductance is assembled by the Landauer-Buettiker formula from momentum-resolved transmissions $T(k_\\parallel)$ calculated with the scattering-state method.","core_discovery":"On the paper's own terms, the central claim is that the all-antiferromagnetic junction $\\mathrm{Mn_3Sn}(01\\bar{1}0)/\\mathrm{MgO}(110)/\\mathrm{Mn_3Sn}$ shows a finite TMR ratio reaching $\\gtrsim 1000\\%$, with the reported value at $N_{\\mathrm{MgO}(110)}=12$ monolayers and with the transport genuinely in the tunneling regime. The mechanism combines two ingredients. First, the inverse $120^{\\circ}$ antiferromagnetic order of Mn3Sn gives a momentum-dependent spin polarization whose $s_x$ component is even under $k_z\\to -k_z$ while $s_z$ is odd, so after summing over $k_z$ only the $x$-component, aligned with the easy axis of the cluster octupole moment, contributes to the difference between the parallel and antiparallel configurations. Second, the MgO(110) barrier filters the tunneling electrons: its smallest complex-band decay rate $\\kappa_{z0}(k_\\parallel)$ is lowest near $k_\\parallel \\sim 0$, and the momentum-resolved transmission difference $T_P(k_\\parallel)-T_{AP}(k_\\parallel)$ is largest there. A comparison with a vacuum-barrier junction shows that MgO changes the momentum pattern rather than merely acting as a spacer.","pith_inferences":["The symmetry-based argument implies the effect is structural rather than accidental, so moderate changes in MgO thickness or small lattice relaxation may shift the magnitude without destroying the mechanism.","The same even/odd symmetry selection could be applied to other noncollinear antiferromagnets, so a bulk Fermi-surface calculation alone might predict which electrode/barrier combinations are worth full junction simulations.","The ideal-interface, zero-temperature calculation likely overestimates the room-temperature ratio; phonon and interface disorder scatterings will broaden the k-parallel selection, and quantifying that reduction is a natural next test."],"forward_implications":["An Mn3Sn electrode can work with a reasonably thick MgO barrier rather than only with idealized barrier-free junctions.","The experimentally relevant orientation for spin-orbit torque switching, Mn3Sn(01-10), is the same geometry that shows the large TMR, so electrical write and magnetoresistive read are compatible in one junction.","The TMR ratio is governed by the x-component of the spin polarization along the easy axis, so reading the octupole orientation through resistance is a viable readout.","The exponential decay of transmission with barrier thickness and the small Fermi-level local density of states inside MgO confirm that the calculation describes tunneling rather than pinhole-like conduction.","The MgO barrier is not inert: the vacuum-barrier comparison shows that its k-parallel filtering shapes the momentum-resolved transmission pattern."],"supporting_citations":[{"why":"shows that noncollinear antiferromagnets can carry spin-polarized current, the premise for using Mn3Sn as an electrode.","marker":"[17]"},{"why":"proposes tunnel magnetoresistance in noncollinear antiferromagnetic junctions and supplies the theoretical framework this paper extends.","marker":"[21]"},{"why":"reports the octupole-driven magnetoresistance in Mn3Sn junctions, the experimental motivation and benchmark for the predicted effect.","marker":"[22]"},{"why":"provides the coherent spin-dependent tunneling picture for MgO barriers that the k-parallel filtering analysis builds on.","marker":"[4]"},{"why":"defines the k-parallel-projected effective spin polarization for noncollinear magnets used in Eq. (4).","marker":"[35]"},{"why":"identifies the Mn3Sn(01-10) orientation as favorable for spin-orbit torque switching, motivating the junction geometry studied here.","marker":"[74]"},{"why":"characterizes the cluster octupole order that defines the parallel and antiparallel magnetic configurations.","marker":"[76]"}],"fun_headline_variants":["Antiferromagnetic MTJ predicted to exceed 1000% TMR","No net spin but 1000% TMR in Mn3Sn/MgO/Mn3Sn junction","MgO barrier key to 1000% TMR in all-antiferromagnetic junction","TMR >1000% predicted in Mn3Sn/MgO/Mn3Sn tunnel junction","Spin-split Mn3Sn yields 1000% TMR in antiferromagnetic junction"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The calculation assumes that the Mn3Sn/MgO interface is coherent and defect-free, with the 7.6 percent lattice mismatch absorbed entirely by straining the MgO lattice; if real interfaces relax through dislocations, intermixing, or different atomic terminations, the predicted above-1000-percent ratio could change substantially.","fun_headline_variants_meta":{"raw":{"variants":["Antiferromagnetic MTJ predicted to exceed 1000% TMR","No net spin but 1000% TMR in Mn3Sn/MgO/Mn3Sn junction","MgO barrier key to 1000% TMR in all-antiferromagnetic junction","TMR >1000% predicted in Mn3Sn/MgO/Mn3Sn tunnel junction","Spin-split Mn3Sn yields 1000% TMR in antiferromagnetic junction"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00149,"raw_usage":{"total_tokens":6051,"prompt_tokens":1083,"completion_tokens":4968,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":699,"completion_tokens_details":{"reasoning_tokens":4851}},"tokens_in":699,"tokens_out":4968,"duration_ms":29751,"temperature":1.0,"reasoning_tokens":4851,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T15:45:43.013344+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Grow an epitaxial $\\mathrm{Mn_3Sn}(01\\bar{1}0)/\\mathrm{MgO}(110)/\\mathrm{Mn_3Sn}$ stack with about 12 MgO monolayers, switch the cluster octupole moments between the parallel and antiparallel configurations, and measure the resistance-area product and TMR at low temperature; a TMR much below 1000 percent, or an RA far outside the predicted range of about 1 to 10 kOhm-micrometer-squared, would contradict the coherent-interface calculation.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"proposes tunnel magnetoresistance in noncollinear antiferromagnetic junctions and supplies the theoretical framework this paper extends."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"reports the octupole-driven magnetoresistance in Mn3Sn junctions, the experimental motivation and benchmark for the predicted effect."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"identifies the Mn3Sn(01-10) orientation as favorable for spin-orbit torque switching, motivating the junction geometry studied here."},{"cited_title":"Suzuki, T","cited_arxiv_id":null,"evidence_quote":"characterizes the cluster octupole order that defines the parallel and antiparallel magnetic configurations."}],"review_version":1}