{"id":"e15e2999-cb18-439d-8670-b067578201c6","arxiv_id":"1908.04322","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"DFT calculations predict that CrI3 proximity induces a 40 meV exchange bias in MnBi2Te4 films, enabling zero-field QAH states with Chern numbers 1 and 3.","lead":"This computational paper predicts that placing a magnetic insulator, CrI3, on MnBi2Te4 thin films creates a strong magnetic pinning effect of about 40 meV, which could stabilize the zero-field quantum anomalous Hall effect. The work also shows that an electric field could drive the system into higher Chern number states.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"High-CN gated QAH claim rests on an unverified isolated-slab approximation; full heterostructure calculations under field are not shown.","rationale":"The reader's conditional verdict already captures the main methodological caveats, including the isolated-slab treatment of the gated high-CN states, the unnamed vdW functional, and the lack of Hubbard U. I agree that these warrant conditional acceptance rather than rejection. My focus is on the gated high-CN claim, because it is part of the abstract's headline and is the one piece explicitly computed only on isolated MnBi2Te4 slabs. The zero-field QAH claim for 3 and 5 SL heterostructures is directly supported by explicit DFT band structures, SOC-evolution tests, and Wilson-loop calculations, so it is less vulnerable. The exchange-bias magnitude of 40 meV is supported by multiple stacking and supercell tests, though the fixed ~7% strain and GGA functional remain quantitative uncertainties. The concrete test proposed here would settle whether the high-CN gated state survives in the full heterostructure; until that is run, the electrically tunable high-CN conclusion should be regarded as conditional.","tokens_in":11829,"tokens_out":4507,"duration_ms":50033,"concrete_test":"Compute the electronic structure and Wilson-loop Chern number of the full 3-SL MnBi2Te4/CrI3 heterostructure (same sqrt3 x sqrt3 / 1 x 1 model) under out-of-plane electric fields epsilon = 0, 0.004, 0.005, 0.015 V/A, using the paper's GGA+vdW method. If CN=3 at 0.015 V/A and no Cr-eg band crosses the Fermi level or the relevant gap, the high-CN claim is supported. If CN differs or Cr-eg bands intervene, the isolated-slab approximation fails and the high-CN conclusion requires revision.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The abstract claims that high Chern number (CN=3) QAH can be reached by electric gating, but the supporting calculation is performed only on isolated MnBi2Te4 slabs, not on the MnBi2Te4/CrI3 heterostructure. The section 'Electrically tunable high-Chern-number QAH effect' states: 'we only consider isolated MnBi2Te4 models when applying an electric field,' justified by the weak perturbation of CrI3 on the low-energy MnBi2Te4 bands. However, the same paper shows that Cr-eg bands shift under an electric field (Fig. 1b) and, for 1-4 SL films, Cr-eg bands form the lowest conduction bands. The two topological transitions for 3 SL, at epsilon = 0.002 V/A (Gamma) and 0.0141 V/A (Gamma-K), are derived from the isolated slab (Fig. 3). If, in the actual heterostructure, Cr-eg states cross the Fermi level or the relevant band gap at these fields, or if the band inversion is modified by the CrI3 layer, then the CN=3 state and the 'electrically tunable zero-field QAH' conclusion do not follow. This is an explicitly stated limitation rather than a hidden flaw, but it is load-bearing for a headline claim and is not tested anywhere in the manuscript.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript proposes using a monolayer of the ferromagnetic insulator CrI3 as a magnetic proximity layer on MnBi2Te4 thin films to restore the out-of-plane surface magnetism needed for the quantum anomalous Hall (QAH) effect. First-principles DFT calculations on MnBi2Te4/CrI3 heterostructures find an interfacial ferromagnetic exchange coupling of about 40 meV, attributed to the long tails of Cr-eg Wannier orbitals hybridizing with Te-p states. The authors show that the heterostructure remains insulating, that 3- and 5-SL MnBi2Te4/CrI3 host a QAH state with Chern number 1, and that a CrI3/MnBi2Te4/CrI3 sandwich can drive even-SL films into a QAH state with Chern number -1. They further claim that an applied electric field can produce a high-Chern-number QAH state with CN=3, based on calculations performed on isolated MnBi2Te4 slabs. The central exchange-bias and zero-field QAH results are supported by calculations on the full heterostructure for multiple stackings and supercells, whereas the gated high-CN claim rests on an explicitly stated but untested isolated-slab approximation.","tokens_in":12127,"tokens_out":6986,"duration_ms":64735,"significance":"If the central predictions hold, the work provides a concrete route to zero-field QAH effect in an intrinsic magnetic topological insulator by solving the surface-magnetism problem that currently limits MnBi2Te4 films. The reported exchange energy of ~40 meV is large compared with the relevant magnetic ordering temperatures, and the orbital-overlap mechanism (Cr-eg tails penetrating through the van der Waals gap) is a physically appealing and testable explanation. The paper also delivers a symmetry-based scenario for electric-field-induced Chern-number jumps of 3, which is of independent interest. Strength of the manuscript includes explicit testing of multiple stacking geometries and supercell sizes, SOC-scaling analysis showing topological phase transitions, and Wilson-loop Berry-phase calculations for the Chern numbers in the heterostructure.","major_comments":[{"comment":"The CN=3 claim is computed for isolated MnBi2Te4 slabs, not for the MnBi2Te4/CrI3 heterostructure. The text states explicitly that \"we only consider isolated MnBi2Te4 models when applying an electric field,\" yet the abstract and conclusion present the high-CN state as a property of the proposed heterostructure platform. Since Fig. 1(b) shows that an electric field shifts the Cr-eg bands and Fig. 2 shows Cr-eg bands forming the lowest conduction bands for 1-4 SL films, a full heterostructure calculation under field is needed to verify that the Cr-eg states do not cross the relevant band gap and that the topological transitions occur at the quoted fields. Without such a calculation, the headline claim of \"electrically tunable zero-field QAH effect\" is not supported.","section":"Electrically tunable high-Chern-number QAH effect (Fig. 3)"},{"comment":"The symmetry argument for a Chern-number jump by 3 relies on the C3 and combined time-reversal-mirror (TM) symmetries of the isolated MnBi2Te4 slab. The actual CrI3/MnBi2Te4 heterostructure, especially under an electric field, may not preserve the TM symmetry because the magnetic CrI3 layer breaks time reversal and the interface has reduced symmetry. If the TM symmetry is broken, the three gapless points along the Gamma-K lines are no longer symmetry-protected, and the CN=3 plateau may be absent. The manuscript does not analyze the symmetry of the full heterostructure under an electric field, so the relevance of the isolated-slab result to the heterostructure remains unestablished.","section":"Electrically tunable high-Chern-number QAH effect (symmetry argument)"},{"comment":"The methods section does not identify the van der Waals functional used in the VASP calculations, referring only to \"van der Waals interactions in DFT calculations\" with a citation to the VASP code. Since the central quantitative result is the ~40 meV exchange coupling and the interlayer distance directly controls the Cr-eg/Te-p overlap, the choice of vdW functional (e.g., DFT-D3, optB88, vdW-DF) could affect the reported energies. In addition, the methods describe only Wannier functions of bulk MnBi2Te4 and do not explain how the heterostructure band structures and Wilson-loop Berry phases (including CrI3 states) were computed. This hampers reproducibility of both the exchange energy and the Chern numbers.","section":"Appendix 1: Calculation Methods"}],"minor_comments":[{"comment":"The term \"exchange bias\" is used for a large interfacial exchange coupling that pins the adjacent Mn layer's magnetization. In the strict magnetism literature, exchange bias denotes a shift of the hysteresis loop, which is not directly computed here. Consider using \"interfacial exchange coupling\" or \"proximity exchange\" to avoid ambiguity.","section":"Introduction and throughout"},{"comment":"The statement that the QAH gaps are 49 and 14 meV for 3 and 5 SLs should clarify that these are the gaps of the MnBi2Te4 bands (as listed in Table S1), not necessarily the fundamental charge gap of the heterostructure, because Cr-eg bands appear as the lowest conduction bands for thin films. The Chern number is evaluated for the full occupied manifold, so the distinction matters for interpreting the gap values.","section":"QAH effect (band gaps)"},{"comment":"The caption states the transitions occur at epsilon = 0.002 V/A and 0.0141 V/A, while the subpanels are labeled epsilon = 0.004 (CN=0) and 0.015 (CN=3). This is consistent with the transitions being below the labeled values, but the relation should be stated explicitly to avoid confusion.","section":"Fig. 3 caption"},{"comment":"The sentence \"For instance, epsilon = 0.03 V/A induces a TPT with three gapless points along the Gamma-K lines in the 2-SL-thick MnBi2Te4 film at epsilon = 0.023 V/A\" is confusingly worded; it should be rephrased to indicate that the transition occurs at 0.023 V/A and that the CN=3 state is realized at 0.03 V/A.","section":"Sandwiched MnBi2Te4 structures"},{"comment":"The kinetic energy cutoff of 270 eV is rather low for first-row elements and heavy elements; a convergence test with respect to the cutoff would strengthen confidence in the quantitative energy differences.","section":"Appendix 1"}],"recommendation":"major_revision","confidential_remarks":"The core exchange-bias and zero-field QAH findings in the full heterostructure are solid and likely of high interest. The gated high-CN (CN=3) claim, however, is presented as a central result in the abstract while being computed only for isolated MnBi2Te4 slabs; the symmetry-based argument may also not survive the inclusion of CrI3. I recommend requesting either a full heterostructure calculation under electric field or a revision that explicitly reframes the high-CN result as a property of isolated MnBi2Te4 films under gating, not of the proposed heterostructure device. Given the prominence of the CN=3 claim in the abstract and conclusion, this is a load-bearing issue. The vdW functional specification is also a reproducibility concern for a computational letter."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Read the MnBi2Te4/CrI3 paper. My take: the exchange-bias result is more solid than I expected, and the even-SL sandwich twist is the real selling point. The high-CN gating claim is the soft spot, and the authors admit it.\n\nWhat's new: the ~40 meV FM coupling between CrI3 and the adjacent MnBi2Te4 layer, explained by Cr-eg Wannier tails reaching across the vdW gap, is a concrete mechanism. They test two stackings, two supercells, with and without SOC, and get the same ~40 meV. That is reproducible DFT evidence, not a fitted parameter. The other new thing: sandwiching an even-SL MnBi2Te4 film between two CrI3 layers flips the top SL's moment and gives CN=-1, extending QAH to even SLs. That's a clean, testable prediction.\n\nWhere it wobbles: the abstract's headline claim of 'electrically tunable high-Chern-number QAH' rests on field calculations done only on isolated MnBi2Te4 slabs, not on the heterostructure. The paper says this explicitly, so it is not a hidden error. But it is load-bearing: they show Cr-eg states shift under field and sit at the conduction band edge for 1-4 SLs, so the isolated-slab model could easily miss a crossing or a change in band inversion in the real interface. The symmetry argument that CN jumps by 3 at generic k-points is fine, but the actual transition fields for the heterostructure remain unverified. That makes the CN=3 part a proposal within a proposal, not a result on the full system.\n\nMinor issues: the vdW functional is not identified, no Hubbard U is used for Mn/Cr d electrons, and the small model stretches CrI3 by 7%. The multi-supercell check mitigates the strain concern, and GGA without U is standard for this class of proposal, so I don't weight those heavily.\n\nWho this is for: anyone working on MBT thin films or magnetic proximity in vdW heterostructures. It deserves a serious referee; the exchange-bias mechanism and even-SL sandwich prediction are worth testing experimentally. The gated CN=3 calculation should be repeated on the full interface before publication, or explicitly reframed as a prediction for MnBi2Te4 under field with CrI3 assumed passive.\n\nI'd send it to review, with a request for the missing heterostructure field calculation.","headline":"Strong exchange-bias result, one clean new prediction (even-SL sandwich), and one untested isolated-slab claim in the gated CN=3 part.","tokens_in":12620,"tokens_out":1738,"would_cite":true,"duration_ms":17099,"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":"Stacking CrI3 on MnBi2Te4 creates a 40 meV exchange bias that pins surface spins, restoring the magnetic gap and enabling zero-field QAH states.","keywords":["exchange bias","quantum anomalous Hall effect","MnBi2Te4","CrI3","magnetic proximity","Chern number","van der Waals heterostructure","electric-field gating"],"falsifier":"Measure the zero-field Hall resistance of a 3-SL MnBi2Te4 film capped with monolayer CrI3 below 10 K. Seeing no quantized plateau at $h/e^2$ with vanishing longitudinal resistance, or no surface magnetic gap in spin-resolved photoemission, would contradict the central prediction; likewise, observing no exchange-bias shift of the magnetization loop would rule out the 40 meV coupling.","tokens_in":2233,"feed_emoji":"🧲","tokens_out":4131,"duration_ms":125969,"temperature":0.7,"pith_summary":"The paper claims that stacking a monolayer of the ferromagnetic insulator CrI3 on a thin film of the layered antiferromagnet MnBi2Te4 creates a roughly 40 meV ferromagnetic exchange bias at the van der Waals interface, strong enough to pin the surface Mn moments out of plane. This matters because pure MnBi2Te4 films show gapless surface states, so the predicted zero-field quantum anomalous Hall effect has only been observed under fields above 6 T. With the CrI3 cap, the paper predicts that 3- and 5-septuple-layer films become zero-field QAH insulators with Chern number $C=1$, that vertical electric fields can switch the system through $C=0$ to $C=3$, and that a CrI3/MnBi2Te4/CrI3 sandwich turns even-layer films into $C=-1$ QAH insulators. The microscopic origin is traced to long tails of Cr-$e_g$ orbitals reaching across the van der Waals gap and hybridizing with Te $p$ orbitals, so the exchange coupling is both large and short-ranged.","feed_headline":"40 meV exchange bias opens zero-field QAH in MnBi2Te4/CrI3","feed_subtitle":"A CrI3 cap pins the topological magnet's surface spins, predicting quantized Hall states without an external field.","key_machinery":"The load-bearing object is the Cr-$e_g$ orbital tail. In the interface model, the Cr d-electron $e_g$-character localized orbital extends beyond the van der Waals gap and overlaps strongly with Te $p$ orbitals on the MnBi2Te4 side, forming a long exchange pathway from Cr-$e_g$ through I, Te, Bi, Te to Mn-$t_{2g}$ states. This orbital tail is what rationalizes a 40 meV FM coupling that is insensitive to stacking and spin-orbit coupling. The other pieces are standard topological-band tools: Berry-phase invariants give the Chern numbers, and symmetry arguments ($C_3$ rotation plus combined time-reversal/mirror symmetry $\\mathcal{T}M$) dictate that gap-closing transitions at generic points occur in groups of three or six, explaining how the Chern number can jump by 3. The electric-field calculations use isolated MnBi2Te4 slabs, relying on the fact that CrI3 modifies the low-energy band structure only weakly.","core_discovery":"In the authors' own terms: the MnBi2Te4/CrI3 interface is ferromagnetic, with an energy difference between FM and AFM alignments of about 40 meV that barely changes with stacking, with or without spin-orbit coupling, and persists in a larger supercell. CrI3 couples only to the neighboring MnBi2Te4 septuple layer, leaving the AFM order between MnBi2Te4 layers intact, and its magnitude exceeds the 3-4 meV MnBi2Te4 interlayer coupling and the roughly 10 meV CrI3 interlayer coupling. Because the coupling is so strong relative to the magnetic ordering temperatures, the surface moments of MnBi2Te4 are pinned out of plane, restoring the magnetic surface gap that ARPES experiments found missing. The resulting heterostructure remains an insulator with no charge transfer, the MnBi2Te4 band topology is preserved, and Berry-phase calculations give $C=1$ for 3- and 5-SL films (gaps of 49 and 14 meV). Applying a vertical electric field to the isolated MnBi2Te4 film drives gap closings at $\\Gamma$ and along $\\Gamma$--$K$, producing first a trivial $C=0$ state and then a high-Chern-number $C=3$ QAH state; in the symmetric sandwich, an even 4-SL film becomes a $C=-1$ QAH insulator with a 34 meV gap.","pith_inferences":["A cheap first experimental check is magnetometry: if the 40 meV bias is real, the heterostructure's magnetic hysteresis loop should shift below the CrI3 Curie temperature, before any transport measurement is attempted.","Inserting a monolayer hexagonal boron nitride spacer between CrI3 and MnBi2Te4 should exponentially suppress the Cr-$e_g$/Te-$p$ overlap; measuring the coupling decay with spacer thickness would separate the orbital-tail mechanism from strain-induced effects.","The symmetry argument that gap closings occur in triples along $\\Gamma$--$K$ is not specific to MnBi2Te4; any threefold-symmetric magnetic topological film under a perpendicular field should show Chern-number jumps of $\\pm 3$, so the high-Chern-number route could be tested in other van der Waals magnets.","If the strained CrI3 monolayer turns out to lose its out-of-plane ferromagnetism, the same exchange-bias function might be achievable with a different magnetic insulator or with CrI3 under less strain, which would make the prediction robust but shift the material of choice."],"forward_implications":["A monolayer CrI3 cap should make 3- and 5-SL MnBi2Te4 films zero-field QAH insulators with Chern number $C=1$ and gaps of 49 and 14 meV, respectively.","A vertical electric field around 0.015 V/Å on a 3-SL film should drive a topological transition into a $C=3$ QAH state; fields near 0.03 V/Å can put a 2-SL film into $C=3$.","Sandwiching an even-SL MnBi2Te4 film between two CrI3 layers should force all septuple-layer magnetizations into a net out-of-plane arrangement, giving a $C=-1$ QAH insulator, for example a 34 meV gap for 4 SL.","Because the exchange bias is stronger than MnBi2Te4 interlayer coupling but does not transfer charge, the same capping approach should work for pinning surface order of bulk MnBi2Te4, enabling ARPES observation of the axion-insulator surface gap.","Other magnetic insulators with out-of-plane magnetization could replace CrI3 and produce the same exchange bias, as long as their $e_g$-like orbitals couple across the van der Waals gap."],"supporting_citations":[{"why":"Documents the gapless surface states of MnBi2Te4 films, the problem the heterostructure is designed to solve.","marker":"[1–4]"},{"why":"Shows quantized Hall resistance only above 6 T in MnBi2Te4 films, the baseline the zero-field prediction must beat.","marker":"[5–7]"},{"why":"Predicted the intrinsic QAH state in MnBi2Te4 films, the starting point that motivates the surface-magnetism fix.","marker":"[24]"},{"why":"Independent theoretical prediction of MnBi2Te4 QAH and surface magnetism, used to identify odd-SL films as QAH insulators.","marker":"[30]"},{"why":"Establishes monolayer CrI3 as a ferromagnetic insulator with out-of-plane magnetization, the exchange-bias material chosen.","marker":"[38]"},{"why":"Provides the bulk CrI3 Curie temperature (61 K) used to argue the 40 meV bias exceeds magnetic ordering energies.","marker":"[37]"},{"why":"Gives the CrI3 interlayer coupling (about 10 meV), the comparison showing the interface bias is much stronger.","marker":"[44]"},{"why":"Supplies hybrid-functional corrections placing the Cr-$e_g$ states about 1.5 eV above Cr-$t_{2g}$, supporting the no-charge-transfer band alignment.","marker":"[46]"}],"fun_headline_variants":["CrI3 cap pins surface spins, restoring zero-field QAH in MnBi2Te4","40 meV exchange bias opens QAH in MnBi2Te4/CrI3 without a field","Magnetic proximity gives MnBi2Te4 a 40 meV gap for QAH","Electric gates tune MnBi2Te4/CrI3 to high Chern number QAH","Strong CrI3 coupling restores zero-field QAH in MnBi2Te4 films"],"cache_read_input_tokens":14720,"weakest_assumption_plain":"The calculation stretches the CrI3 monolayer by about 7% to match the MnBi2Te4 lattice and assumes the density-functional model captures the strained interface; if that strain eliminates CrI3's out-of-plane ferromagnetism, or if the isolated-slab electric-field model misses CrI3's response to gating, the 40 meV bias and the predicted topological phase boundaries do not follow.","fun_headline_variants_meta":{"raw":{"variants":["CrI3 cap pins surface spins, restoring zero-field QAH in MnBi2Te4","40 meV exchange bias opens QAH in MnBi2Te4/CrI3 without a field","Magnetic proximity gives MnBi2Te4 a 40 meV gap for QAH","Electric gates tune MnBi2Te4/CrI3 to high Chern number QAH","Strong CrI3 coupling restores zero-field QAH in MnBi2Te4 films"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000267,"raw_usage":{"total_tokens":1678,"prompt_tokens":1071,"completion_tokens":607,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":687,"completion_tokens_details":{"reasoning_tokens":487}},"tokens_in":687,"tokens_out":607,"duration_ms":6010,"temperature":1.0,"reasoning_tokens":487,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:45:34.473730+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the zero-field Hall resistance of a 3-SL MnBi2Te4 film capped with monolayer CrI3 below 10 K. Seeing no quantized plateau at $h/e^2$ with vanishing longitudinal resistance, or no surface magnetic gap in spin-resolved photoemission, would contradict the central prediction; likewise, observing no exchange-bias shift of the magnetization loop would rule out the 40 meV coupling.","supporting_citations":[{"cited_title":"Sivadas, S","cited_arxiv_id":null,"evidence_quote":"Gives the CrI3 interlayer coupling (about 10 meV), the comparison showing the interface bias is much stronger."}],"review_version":1}