REVIEW 3 major objections 5 minor 29 references
Competing magnetic interactions in the antiferromagnetic topological insulator MnBi$_{2}$Te$_{4}$
T0 review · 3 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash
Pith's one-line read Inelastic neutron scattering on MnBi2Te4 finds intralayer frustration $|J_2/J_1| = 0.36$, placing the ferromagnetic triangular layers near the classical instability boundary at $1/3$ where helical and skyrmion phases compete.
desk verdict First INS spin-wave data on MnBi2Te4 make a credible case for frustrated intralayer exchange, but the headline |J2/J1|=0.36 rests on a uniform-broadening fit the paper itself contradicts. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The central object is a local-moment Heisenberg Hamiltonian on the stacked triangular lattice with intralayer nearest-neighbor exchange $J_1$, intralayer next-nearest-neighbor exchange $J_2$, interlayer exchange $J_c$, and single-ion anisotropy $D$ for the $S=5/2$ Mn moments. The load-bearing quantity is the dimensionless frustration ratio $|J_2/J_1|$ compared with the classical boundary $1/3$ between intralayer ferromagnetic and noncollinear order. The parameters are extracted by comparing powder-averaged linear spin-wave intensity calculations to the measured $S(Q,E)$, using separate fits for the low-energy gap edge (Model-c) and the full intralayer spectrum (Model-ab). Classical Monte Carlo simulations then map the fitted parameters into a field-temperature phase diagram that places the spiral and skyrmion phases nearby.
What would settle it
A single-crystal inelastic neutron scattering experiment that resolves the intralayer spin-wave branches along high-symmetry directions would settle it: if a direct fit gives $|J_2/J_1|$ clearly below $1/3$, or if the line shapes force a strongly momentum-dependent broadening instead of the uniform 0.85 meV assumed here, the proximity-to-instability claim fails.
Extended reading notes
Core claim
The paper's central claim is that the magnetic interactions in MnBi2Te4 form a frustrated $J_1$-$J_2$ triangular-lattice system rather than a simple nearest-neighbor ferromagnet. From the powder-averaged spin-wave spectrum the authors extract $SJ_1 = 0.28(2)$ meV and $SJ_2 = -0.10(2)$ meV, giving $|J_2/J_1| = 0.36$, while the low-energy gap edge and magnetization data yield an interlayer coupling $SJ_c$ near $-0.055$ to $-0.085$ meV and a single-ion anisotropy $SD$ near 0.12 meV. The excitations show a spin gap of about 0.5 meV and an intrinsic linewidth that requires convolution by roughly 0.85 meV, far beyond the instrumental resolution. The high-energy intralayer modes need an effective anisotropy closer to $SD = 0.55$ meV, a discrepancy the paper attributes to momentum-dependent broadening that powder data cannot resolve. The major finding is that $|J_2/J_1|$ sits close to the classical instability limit $1/3$ for ferromagnetic triangular layers, so the ground state is nearly degenerate with helical and multi-$q$ states; classical Monte Carlo on the fitted parameters finds vertical spiral, skyrmion, and up-up-down-down phases at nearby points in the field-frustration phase diagram, and DFT+U reproduces the measured exchange ratios only near $U = 2.7$ eV.
Load-bearing premise
The fitted ratio $|J_2/J_1|=0.36$ assumes a single Heisenberg model with a uniform, momentum-independent spin-wave lifetime broadening of 0.85 meV; if the broadening varies with momentum or additional couplings exist, the ratio shifts.
Editorial extensions
If this is right
- The measured frustration places MnBi2Te4 near the classical $|J_2/J_1|=1/3$ boundary, so small changes in exchange, anisotropy, or field could tip the triangular layers into a different magnetic state.
- The strong intrinsic line broadening seen in the spin-wave spectrum points to magnon coupling to other degrees of freedom, which single-crystal neutron studies could directly characterize.
- The DFT+U comparison fixes the effective correlation strength near $U=2.7$ eV, so the neutron data provide a quantitative anchor for electronic-structure calculations of this topological magnet.
- The nearby spiral, skyrmion, and up-up-down-down phases in the classical phase diagram give concrete search targets for chemical substitution, particularly in Sb-substituted compounds where the anisotropy is smaller.
Reading between the lines
- The large gap between the low-energy anisotropy ($SD$ near 0.12 meV) and the high-energy effective value ($SD$ near 0.55 meV) suggests that the high-energy fit may be absorbing momentum-dependent broadening; resolving the linewidths on a single crystal would separate those effects.
- If the frustration ratio is really this close to 1/3, strain, surface termination, or small off-stoichiometry should also be able to tip the surface layers into noncollinear textures, which could connect the bulk frustration to surface-sensitive measurements.
- A direct extension would be to map the field-temperature phase diagram of Mn(Bi,Sb)2Te4, where the paper notes the anisotropy is smaller; skyrmion scattering should appear at accessible fields if the proximity argument is correct.
Signed reviews
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper reports powder inelastic neutron scattering (INS) measurements on the putative antiferromagnetic topological insulator MnBi2Te4. The authors model the spin dynamics with a Heisenberg Hamiltonian containing intralayer nearest- and next-nearest-neighbor exchanges J1 and J2, interlayer exchange Jc, and single-ion anisotropy D. Because one parameter set cannot simultaneously describe the low-energy gap edge and the high-energy intralayer modes, they construct two models: Model-c (low-energy gap edge, with resolution-limited broadening) and Model-ab (high-energy intralayer spectrum, with a uniform 0.85 meV Lorentzian broadening). The central claim is that |J2/J1| = 0.36, close to the classical instability limit |J2/J1| = 1/3 for ferromagnetic triangular layers. This is supported by DFT+U calculations at moderate U (about 2.7 eV) and by classical Monte Carlo simulations that show nearby spiral and skyrmion phases. The paper also emphasizes Ising-like anisotropy and metamagnetism.
Significance. If the extracted ratio |J2/J1| is robust, the result is significant: it would place MnBi2Te4 near a magnetic instability where small perturbations (chemical substitution, strain, field) could drive the system into helimagnetic or skyrmionic states, which would matter for quantum anomalous Hall and axion-insulator physics. The paper has clear strengths: it provides a quantitative two-model analysis of powder INS data, uses magnetization critical fields as independent constraints for the gap-edge model, performs DFT+U calculations over a range of U, and is unusually candid about inconsistencies between models. However, because the headline ratio rests entirely on a model whose broadening assumption the authors themselves acknowledge is inadequate, the numerical proximity to 1/3 is not yet established with the claimed precision.
major comments (3)
- [Model-ab fitting paragraph and Table I] The extraction of the headline ratio |J2/J1| = 0.36 depends exclusively on Model-ab, which convolutes calculated spectra with a single, Q-independent Lorentzian of FWHM = 0.85 meV. Because this width is larger than SJ1 = 0.28 meV, the individual spin-wave branches are not resolved in the powder-averaged data. The manuscript itself states that the gap-edge and high-energy modes cannot be modeled with a single set of Heisenberg parameters, and later concludes with 'clear evidence for strongly Q-dependent broadening.' Under these conditions the fitted (SJ1, SJ2) values are sensitive to the assumed line shape, energy-dependent background, and fitted Q range, so the numerical proximity to 1/3 is not secure. This is load-bearing because the paper's main claim is precisely that the system sits close to the classical instability limit. I request a robustness analysis: fits with Q- or energy-dependent broadening, alternative background models, and fits excluding the low-Q data that show the high-energy tail, to demonstrate that the ratio is stable or to quote a conservative range.
- [Table I; 'Overall, our findings' paragraph] The two independent parameter sets imply inconsistent single-ion anisotropies. Model-ab gives SD = 0.55 meV and an effective spin gap of 1.1 meV, while Model-c and the magnetization data give SD about 0.085–0.12 meV and the true gap of about 0.5 meV. The paper acknowledges that this discrepancy is unexplained. Nevertheless, the Monte Carlo phase diagram is computed at D/J1 = 0.4, obtained by combining SD from Model-c with SJ1 from Model-ab, and the text quotes 0.3 < D/J1 < 0.4 as an experimental finding. No single model in Table I yields D/J1 = 0.4; combining values from inconsistent models is an ad hoc procedure. The skyrmion phase in Fig. 3(a) appears only for |J2/J1| greater than about 0.5 at this anisotropy, whereas the fitted ratio is 0.36, so the proximity statement involves a substantial extrapolation. The authors should clearly label the Monte Carlo results as scenario calculations and present the phase diagram over the full range of D/J1 allowed by the data, including the Model-ab value D/J1 approximately equal to 2.
- [Monte Carlo simulation paragraph and Fig. 3] The classical Monte Carlo phase diagram is computed using parameters obtained from the same INS data, so Fig. 3 is not an independent confirmation of skyrmion proximity; it is a scenario calculation whose input parameters are the same uncertain quantities that are being tested. This is not a statistical circularity in the strict sense, but the text's claim that skyrmion phases 'could be accessed in chemically tuned compounds' overstates the predictive content, especially given that the model places the skyrmion phase outside the fitted |J2/J1| range at D/J1 = 0.4. I recommend explicit language stating that the skyrmion and spiral phases occur for parameter ratios beyond those inferred from the current fits, and that the quoted proximity to the 1/3 instability is therefore a motivation for further study rather than a quantitative prediction.
minor comments (5)
- [Abstract and Introduction] There are several typographical errors: 'teteradymite' should be 'tetradymite' (appearing twice), and 'Componds' should be 'Compounds' in the final section before the acknowledgments.
- [References] Reference [17] appears only as 'Supplementary Material reference'; the authors should include a complete citation to the supplementary material.
- [Fig. 1(c)] The comparison of Model-ab to the data at low Q is described as 'less satisfactory,' but the text does not quantify this disagreement or explain whether the fitting Q range (0.8–1.9 Å−1) was chosen to exclude the problematic low-Q region. A brief statement on how the Q range was selected would help the reader assess the fit's sensitivity.
- [Eq. (1)] For clarity, the sign convention and the parameterization in terms of S times the coupling constants should be stated explicitly in the text near Eq. (1); the table lists SJ1, SJ2, SJc, and SD, but the conversion from these products to the Hamiltonian parameters J1, J2, Jc, and D is not spelled out.
- [DFT+U discussion] The sentence 'Recent first-principles electronic structure calculations with U = 5 eV predict that |J2/J1| < 0.03' would be easier to check if the authors explicitly contrasted this with their own DFT+U result at 5 eV in Table I, which gives |J2/J1| approximately 0.07; the difference likely arises from details of the method, and a one-sentence explanation would prevent confusion.
Circularity Check
No significant circularity: the central J2/J1 ratio is an experimentally fitted quantity, and the DFT+U section is a consistency comparison rather than the source of the experimental result.
full rationale
The paper's main claim, |J2/J1| = 0.36, is obtained by fitting powder-averaged inelastic neutron scattering data to a Heisenberg model (Eq. 1) with parameters listed in Table I. The exchange parameters are not defined in terms of the ratio they produce; the ratio is a derived property of independently varied fit parameters. The 'major finding' is therefore an experimental inference, not a self-referential construction. The paper explicitly separates Model-c and Model-ab because 'the sharpness of intralayer modes (at the gap edge) and the broad, high energy interlayer modes cannot be consistently modeled with a single set of Heisenberg parameters.' This is an honest limitation about model consistency and Q-dependent broadening, not a circular step. The DFT+U discussion notes that 'the value of U is critical' and that ratios at U approximately 2.7 eV are 'consistent with the INS data.' Even though U is not determined from first principles independently, the experimental J2/J1 value does not come from the DFT calculation; the DFT comparison is presented as a consistency check, and the paper states the DFT exchange values are generally larger than experiment. The classical Monte Carlo phase diagram is computed from the experimentally extracted parameter range and explicitly places skyrmion phases at |J2/J1| >= 0.5, 'slightly larger frustration ratio than our INS data suggests,' so it is an extrapolation rather than a circular prediction of the measured ratio. The main self-citations, e.g., Ref. [19] for the powder-averaged spin-wave calculation procedure and Ref. [10] for sample characterization, are methodological or background citations and are not load-bearing in the derivation of J2/J1. Overall, the paper's central quantitative result is a fit to data, not a consequence of a self-cited uniqueness theorem or a parameter renamed as a prediction, so no circularity score above 1 is warranted.
Assumptions & free parameters
free parameters (6)
- SJ1 (nearest neighbor intralayer exchange) =
0.28(2) meV (Model-ab)
- SJ2 (next-nearest neighbor intralayer exchange) =
-0.10(2) meV (Model-ab)
- SJc (interlayer exchange) =
-0.055 meV (Model-c INS); -0.085(5) meV (magnetization)
- SD (single-ion anisotropy) =
0.12 meV (Model-c); 0.55(5) meV (Model-ab)
- U (DFT+U Hubbard parameter) =
~2.7 eV
- Broadening FWHM for Model-ab =
0.85 meV
assumptions (4)
- domain assumption The spin dynamics are described by the local-moment Heisenberg Hamiltonian of Eq. (1) with bilinear J1, J2, Jc and single-ion anisotropy D; no higher-order or anisotropic exchange terms are included.
- standard math Linear spin-wave theory and powder averaging provide a valid approximation to the observed INS cross-section.
- domain assumption Classical Monte Carlo at temperature T=0.08 J1 captures the low-temperature phase diagram of the quantum S=5/2 Heisenberg model.
- domain assumption DFT+U with the Dudarev formulation and PBE functional gives reliable exchange parameters when U is set to an appropriate value.
Cite this review
Pith. "Pith review of Competing magnetic interactions in the antiferromagnetic topological insulator MnBi$_{2}$Te$_{4}$." pith.science (2026). https://pith.science/paper/5IFVXJQ5
@misc{pith2026190802332,
author = {Pith},
title = {Pith review of: Competing magnetic interactions in the antiferromagnetic topological insulator MnBi$_2$Te$_4$},
year = {2026},
howpublished = {\url{https://pith.science/paper/5IFVXJQ5}},
note = {Machine review of arXiv:1908.02332}
}
abstract
The antiferromagnetic (AF) compound MnBi$_{2}$Te$_{4}$ is suggested to be the first realization of an antiferromagnetic (AF) topological insulator. Here we report on inelastic neutron scattering studies of the magnetic interactions in MnBi$_{2}$Te$_{4}$ that possess ferromagnetic (FM) triangular layers with AF interlayer coupling. The spin waves display a large spin gap and pairwise exchange interactions within the triangular layer are frustrated due to large next-nearest neighbor AF exchange. The degree of frustration suggests proximity to a variety of magnetic phases, potentially including skyrmion phases, that could be accessed in chemically tuned compounds or upon the application of symmetry-breaking fields.
Figures
Figures from the paper (1 more)
Reference graph
Works this paper leans on
-
[1]
Tokura, K
Y. Tokura, K. Yasuda, and A. Tsukazaki, Nat. Rev. Phys. 1, 126 (2019)
2019
-
[2]
(Q = 0.7 ˚A−1) in Fig. 2(c1) indicates a spin gap with an onset of ∆ ≈ 0.5 meV consistent with sizable uniaxial magnetic anisotropy. The quantitative details of the magnetic interactions become more apparent based on fitting the data to a local-moment Heisenberg model, H =−J1 ∑ ⟨ij⟩|| Si·Sj−J2 ∑ ⟨⟨ij⟩⟩|| Si·Sj−Jc ∑ ⟨ij⟩⊥ Si·Sj−D ∑ i S2 i,z (1) where J1 and...
-
[3]
J. Zhang, C.-Z. Chang, P. Tang, Z. Zhang, X. Feng, K. Li, L.-l. Wang, X. Chen, C. Liu, W. Duan et al. Science 339, 1582 (2013)
work page 2013
- [4]
-
[5]
Chang, W
C.-Z. Chang, W. Zhao, D. Y. Kim, H. Zhang, B. A. Assaf, D. Heiman, S.-C. Zhang, C. Liu, M. H. W. Chan, and J. S. Moodera, Nat. Mater. 14, 473 (2015)
2015
-
[6]
M. M. Otrokov, T. V. Menshchikova, M. G. Vergniory, I. P. Rusinov, A. Y. Vyazovskaya, M. K. Yu, G. Bihlmayer, A. Ernst, P. M. Echenique, A. Arnau et al., 2D Materials 4, 025082 (2017)
work page 2017
- [7]
-
[8]
M. M. Otrokov, I. I. Klimovskikh, H. Bentmann, A. Zeugner, Z. S. Aliev, S. Gass, A. U. B. Wolter, A. V. Ko- roleva, D. Estyunin, A. M. Shikinet al., arXiv:1809.07389 (2018)
arXiv 2018
Show all 29 references
-
[9]
S.-H. Lee, Y. Zhu, Y. Wang, L. Miao, T. Pillsbury, S. Kempinger, D. Graf, N. Alem, C. Chang, N. Samarth, and Z. Mao, arXiv:1812.00339 (2018)
2018 arXiv
-
[10]
M. M. Otrokov, I. P. Rusinov, M. Blanco-Rey, M. Hoff- mann, A. Y. Vyazovskaya, S. V. Eremeev, A. Ernst, P. M. Echenique, A. Arnau, and E. V. Chulkov, Phys. Rev. Lett. 122, 107202 (2019)
2019
-
[11]
J.-Q. Yan, Q. Zhang, T. Heitmann, Z. Huang, K. Y. Chen, J. G. Cheng, W. Wu, D. Vaknin, B. C. Sales, and R. J. McQueeney Phys. Rev. Mater. 3, 064202 (2019)
2019
-
[12]
R. S. K. Mong, A. M. Essin, and J. E. Moore, Phys. Rev. B 81, 245209 (2010)
2010
-
[13]
Y. Gong, J. Guo, J. Li, K. Zhu, M. Liao, X. Liu, Q. Zhang, L. Gu, L. Tang, X. Feng et al. , Chin. Phys. Lett. 36, 076801 (2019)
2019
-
[14]
Y. Deng, Y. Yu, M.-Z. Shi, J. Wang, X.-H. Chen, and Y. Zhang, arXiv:1904.11468 (2019)
2019 arXiv
-
[15]
Tanaka and N
Y. Tanaka and N. Uryu, Progr. Theor. Phys. 55, 1356 (1976)
1976
-
[16]
Murao, F
K. Murao, F. Matsubara, and T. Kudo, J. Phys. Soc. Jpn. 65, 1399 (1996)
1996
-
[17]
A. O. Leonov and M. Mostovoy, Nat. Commun. 6, 8275 (2015)
2015
-
[18]
Supplementary Material reference
-
[19]
Vaknin, D
D. Vaknin, D. M. Pajerowski, D. L. Schlagel, K. W. Dennis, and R. J. McQueeney, Phys. Rev. B 99, 220404 (2019)
2019
-
[20]
R. J. McQueeney, J. Q. Yan, S. Chang, and J. Ma, Phys. Rev. B 78, 184417 (2008)
2008
-
[21]
Szuszkiewicz, E
W. Szuszkiewicz, E. Dynowska, B. Witkowska, and B. Hennion, Phys. Rev. B 73, 104403 (2006)
2006
-
[22]
J.-Q. Yan, S. Okamoto, M. A. McGuire, A. F. May, R. J. McQueeney, and B. C. Sales et al. , arXiv:1905.00400 (2019)
2019 arXiv
-
[23]
Xiang, C
H. Xiang, C. Lee, H.-J. Koo, X. Gong, and M.-H. Whangbo, Dalton Trans. 42, 823 (2013)
2013
-
[24]
S. L. Dudarev, G. A. Botton, S. Y. Savrasov, C. J. Humphreys, and A. P. Sutton, Phys. Rev. B. 57, 1505 (1998)
1998
-
[25]
A. M. McGuire, Crystals 7, 121 (2017)
2017
-
[26]
R. J. Birgeneau, W. B. Yelon, E. Cohen, and J. Makovsky, Phys. Rev. B 5, 2607 (1972)
1972
-
[27]
W. B. Yelon, Oelete, and C. Vettier. J.Phys. C: Solid State Phys. 8, 2760 (1975)
1975
-
[28]
Y.-J.Hao, P. Liu, Y. Feng, X.-M. Ma, E. F. Schwier, M. Arita, S. Kumar, C. Hu, R. Lu, M. Zeng et al. , arXiv:1907.03722 (2019)
2019 arXiv
-
[29]
Swatek, Y
P. Swatek, Y. Wu, L. L. Wang, K. Lee, B. Schrunk, J.-Q. Yan, and A. Kaminski, arXiv:1907.09596 (2019)
2019 arXiv
Reviewed August 14, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.