REVIEW 3 major objections 5 minor 1 cited by
Structural disorder can make a topological insulator vanish and then reappear in amorphous quantum spin Hall systems.
Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →
T0 review · deepseek-v4-flash
2026-08-04 08:11 UTC pith:C2LFLVH4
load-bearing objection A careful numerical study of an amorphous BHZ model whose headline re-entrant N→T→N transition looks like a hard-cutoff artifact rather than generic amorphous physics. the 3 major comments →
Phase diagram of amorphous quantum spin Hall insulators
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
Core claim
For a finite-range amorphous BHZ model with hard cutoff R, the paper finds that increasing structural disorder (Gaussian site displacements) can drive nontrivial-to-trivial (N→T) and trivial-to-nontrivial (T→N) transitions, and in specific parameter regimes (notably R = 2.03, 3.03, 3.70, and partially R = 2.50), a re-entrant N→T→N cycle: a topologically nontrivial phase in the perfect lattice becomes trivial under weak disorder, then becomes nontrivial again at larger disorder. This is shown via a real-space Z2 marker, corroborated by bulk-boundary correspondence (edge-state localization) and Landauer-Büttiker conductance. The paper also finds that for a large cutoff approximating no cutoff
What carries the argument
The local real-space Z2 marker for Dirac-type Hamiltonians, expressed as C(r) = π tr_r W (Q P̂_x Q̂_y P − P̂_x Q̂_y P), where W is the product of gamma matrices not in the Hamiltonian, P is the projector onto occupied states, and Q = 1 − P. This marker serves as a disorder-compatible topological invariant that the paper uses to compute phase diagrams; edge-state spectra and two-terminal conductance provide independent checks.
Load-bearing premise
The load-bearing premise is that the hard cutoff R is a physically valid regularization of hopping, and that specific non-integer values like 2.03, 3.03, and 3.70 represent generic cases; the re-entrant behavior is absent in the no-cutoff limit (R = 24.03) and at exact integer R (where even infinitesimal disorder destroys the topological phase), so the effect hinges on this finite, non-integer cutoff.
What would settle it
A direct numerical check would be to compute the topological marker for a very large system with a smooth exponential cutoff (no hard truncation) at the same disorder strengths; if the re-entrant N→T→N signature disappears for all R, the central claim is an artifact of the hard cutoff. Alternatively, an experimental measurement of the Hall conductance in an amorphous HgTe/CdTe-like sample as a function of disorder strength (e.g., via ion irradiation) that shows only a single transition, not a re-entrant one, would contradict the prediction.
If this is right
- If re-entrant behavior is confirmed experimentally, amorphous samples of spin-orbit coupled materials could show a non-monotonic response to structural disorder, with topological edge channels reappearing at high disorder.\n
- The finding that short-ranged hopping models (small R) behave differently from the no-cutoff limit (large R) implies that the microscopic connectivity range is a key parameter controlling disorder-driven topological transitions.\n
- The paper's demonstration that integer cutoffs (e.g., R = 2.0) produce a spurious sudden loss of topology at infinitesimal disorder provides a caution for numerical studies of amorphous systems: the choice of cutoff must avoid coinciding with lattice neighbor distances.\n
- The re-entrant phase transitions, if robust, provide a concrete target for studying critical properties of disorder-driven topological transitions beyond the familiar Anderson-transition paradigm.\n
- The match between the real-space marker and conductance phase diagrams (with small finite-size discrepancies) supports the practical use of local markers for diagnosing topology in structural-disordered systems.
Where Pith is reading between the lines
- The re-entrant N→T→N cycle likely arises from a competition between two effects: weak disorder broadens and shifts the effective mass (favoring trivialization) while strong disorder induces a topological Anderson-like transition (favoring nontriviality); the paper does not identify the microscopic mechanism, but its data are consistent with such a picture.\n
- A testable extension would be to measure the longitudinal resistivity of an amorphous thin film as a function of annealing temperature (which controls structural disorder σ), predicting a non-monotonic signature: insulating at intermediate disorder but conducting at high disorder due to re-entrant edge states.\n
- If real amorphous materials have smooth hopping tails instead of a hard cutoff, the re-entrant behavior may be suppressed; the paper's Appendix C implies the effect is sensitive to the sharpness of connectivity truncation, suggesting that experimental realization would require systems with sharply decaying hoppings.
- The distinction between R values close to versus far from lattice neighbor distances suggests that generic non-integer R values are not all equivalent, and future theories could classify amorphous topological phase diagrams by the set of included hopping shells.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper studies the effects of structural (positional) disorder on the two-dimensional Bernevig-Hughes-Zhang (BHZ) model with finite-range hopping, parameterized by a hard cutoff radius R and an exponentially decaying hopping amplitude. Using a real-space Z2 marker, the authors map phase diagrams as functions of disorder strength σ and mass parameter M for several R values. They report that, for certain R, a clean topological insulator first becomes trivial under weak disorder and then re-enters a topological phase at stronger disorder (N→T→N). They corroborate the bulk marker results with open-boundary edge-state densities and two-terminal conductance calculations. The paper also shows that for R=24.03 (the effectively no-cutoff limit) the topological phase remains stable for the entire disorder range studied, while for exact-integer R (e.g., R=2.0) even infinitesimal disorder destroys the topological phase due to a connectivity discontinuity (Appendix C).
Significance. If the re-entrant N→T→N sequence is a genuine property of amorphous topological insulators, it would be a nontrivial addition to the growing literature on disorder-driven topological transitions, with potential implications for amorphous-material experiments. The paper's methodology is solid: the real-space Z2 marker is an established tool, the clean-limit transition points are separately checked against hopping-distance estimates, and the marker results are cross-validated with edge-state and conductance calculations. The computations appear careful, and the paper contains no fitted parameters dressed as predictions. However, the headline re-entrant behavior is tied to a specific choice of hard cutoff: it appears only for R values slightly above lattice neighbor distances, and it is absent in the no-cutoff limit. Whether this is a generic amorphous-physics effect or an artifact of the hard cutoff remains unresolved, which limits the significance of the central claim in its current form.
major comments (3)
- [Sec. III (Figs. 2, 3) and Appendix C] The central claim of re-entrant N→T→N behavior is demonstrated only for R=2.03, 3.03, and 3.70, values that are 0.03 above the square-lattice neighbor distances 2, 3, and √13. Appendix C shows that for exact-integer R (R=2.0, 3.0) the same proximity to a neighbor distance causes a pathological sensitivity: infinitesimal disorder destroys the topological phase because pairs cross the cutoff. The re-entrant sequence for R=2.03 etc. is produced by the same cutoff-crossing mechanism at finite σ. In the no-cutoff limit (R=24.03, Fig. 3) no re-entrance is seen; the topological phase remains stable throughout. Thus the data as presented support a cutoff-induced connectivity effect rather than a generic property of amorphous hopping. The paper should either provide a physical argument for why hard truncation at these particular R values is representative, or demonstrate that re-entrance survives
- [Abstract and Sec. III (Fig. 2)] The abstract states as a general phenomenon that 'the system exhibits a re-entrant behaviour' without the caveat that this occurs only for R tuned to be slightly above a lattice neighbor distance. Given the paper's own Appendix C, this overstates the generality. The main text does acknowledge the distinction between R categories, but the abstract and the concluding paragraph should be revised to clearly state that re-entrance is a property of certain short-range hard-cutoff models, not of the amorphous BHZ model in general. Otherwise readers may take the headline as a statement about smooth, physically motivated hopping.
- [Sec. IV (Fig. 4)] The conductance results, presented as a validation of the bulk marker, do not fully reproduce the re-entrant phase diagram. The text notes that for R=2.03 at M=4.15 and 4.20, the marker predicts a transition to a non-trivial state at large disorder but the conductance does not show it, and even at M=4.10 (clean-limit topological) the conductance is close to zero. The authors attribute this to finite size, which is plausible, but it means the re-entrant N→T→N sequence is experimentally/transport-validated only in a limited parameter region. Given the load-bearing nature of the re-entrance claim, this discrepancy should be addressed more quantitatively, e.g., by showing convergence with system size or by identifying the correlation-length limitation explicitly.
minor comments (5)
- [Appendix A] Typo: 'as ta function of the strength' should read 'as a function of the strength.'
- [References] Reference [39] contains the corrupted string 'C. BruHERE!ne' — this appears to be a LaTeX/encoding error and should be corrected to 'Brüne'.
- [Sec. IV] The software package is written as 'KW ANT'; it should be 'KWANT'.
- [Sec. III] The phrase 're-entrant phase transition is also observed only at R=2.50. In this case however, the structural disorder is not sufficient to drive the trivial phase into a topological one' is self-contradictory as written. Clarify what exactly is observed for R=2.50.
- [Sec. II] The physical interpretation of σ as proportional to temperature (σ² ∝ k_B T) is stated without a derivation or reference; since σ is just a parameter in the Gaussian displacement, this connection is optional but if kept it should be referenced or justified.
Circularity Check
No significant circularity: the phase diagram, edge-state signatures, and conductance are independent calculations on the same model, with no fitted parameter dressed as a prediction and no load-bearing self-citation.
full rationale
This is a direct numerical study, not a derivation that reduces to its inputs. The model Hamiltonian in Eqs. (2)-(3) is specified independently of the quantities to be predicted. The bulk topological phase is obtained by evaluating the real-space Z2 marker of Refs. [36,37] on the disordered Hamiltonian, and the phase boundaries are verified by independent probes: the clean-limit transition points (M≈2.64 for R=1.5 and M≈4.115 for R=2.0) are matched against the hopping-distance estimates in Eq. (10) and the surrounding text, and the disordered phase diagrams are corroborated by bulk-boundary edge-state spectra and Landauer conductance computed with KWANT (Figs. 4-5). No parameter is fitted to the re-entrant N→T→N sequence and then renamed as a prediction; the marker, the spectral gap, the edge-state localization, and the conductance are separate calculations on the same Hamiltonian. The references to prior work are standard external formulations (real-space markers, degree-theoretic invariants, the BHZ model, and transport formalism), and none of the load-bearing steps relies on a self-citation by the present authors. The sensitivity of the re-entrant behavior to the hard cutoff R (Appendix C and Fig. 3) is a physical-modeling caveat about the generality of the result, not evidence that an equation or fitted parameter is identical to an input. The model choices, such as the hard cutoff at specific R values, are ad hoc and raise selection/soundness concerns, but those are not circularity. Therefore no circular step is identified.
Axiom & Free-Parameter Ledger
free parameters (2)
- Hard cutoff radius R values =
1.70, 2.03, 2.50, 3.03, 3.38, 3.70, 24.03
- Number of disorder configurations averaged =
20
axioms (4)
- domain assumption The real-space Z2 marker of Chen (Eq. 9) correctly computes the Z2 index for disordered class AII Dirac-type Hamiltonians.
- domain assumption Gaussian structural displacement plus a hard hopping cutoff R is the right model of amorphization.
- standard math The BHZ model belongs to the Dirac-type class covered by the degree formula ν=(-1)^deg[n].
- domain assumption Time-reversal symmetry is preserved under structural disorder.
read the original abstract
In light of recent progress in the study of amorphous topological phases, we investigate the effects of structural disorder on the topological properties of a two-dimensional quantum spin Hall insulator modeled by the Bernevig-Hughes-Zhang Hamiltonian. Using a real-space formulation of the Z2 invariant for Dirac-type Hamiltonian, we map out the phase diagram as a function of disorder strength and the mass parameter. Our results reveal that under the influence of structural disorder, a system can either undergo a phase transition from a topologically non-trivial to a topologically trivial phase or from a trivial to non-trivial phase. Remarkably, in certain parameter regimes, the system exhibits a re-entrant behaviour: a topologically non-trivial phase in the perfect lattice undergoes a transition to a trivial state under the influence of weak disorder but re-emerges as the disorder strength is further increased. We corroborate these findings through analysis of the bulk-boundary correspondence and transport calculations.
Figures
Forward citations
Cited by 1 Pith paper
-
Leveraging structural disorder to enhance topological phases
Structural disorder with site-separation penalties enhances 2D topological phases to strong disorder but harms 3D phases; spectral localizer using time-reversal symmetry enables Z2 diagnosis despite spin-frame scrambling.
Reference graph
Works this paper leans on
-
[1]
M. Z. Hasan and C. L. Kane, Rev. Mod. Phys.82, 3045 (2010)
2010
-
[2]
Qi and S.-C
X.-L. Qi and S.-C. Zhang, Rev. Mod. Phys.83, 1057 (2011)
2011
-
[3]
Thus,R=24.03 pro- vides a numerically accurate approximation toR→∞limit
Sincee 1−18 ≈10 −8, the hopping amplitudes beyond this range are negligible - on the order of machine precision for complex floating point calculations. Thus,R=24.03 pro- vides a numerically accurate approximation toR→∞limit. We notice that when the clean system is in topological phase (M<10.97), it remains stable for the entire range of disorder strength...
-
[4]
C.-K. Chiu, J. C. Y . Teo, A. P. Schnyder, and S. Ryu, Rev. Mod. Phys.88, 035005 (2016)
2016
-
[5]
A. P. Schnyder, S. Ryu, A. Furusaki, and A. W. W. Ludwig, Phys. Rev. B78, 195125 (2008)
2008
-
[6]
N.P.Mithcell, L.M.Nash, D.Hexner, A.M.Turner, and W.T.M.Irvine, Nature Physics14, 380 (2018)
2018
-
[7]
Li, R.-L
J. Li, R.-L. Chu, J. K. Jain, and S.-Q. Shen, Phys. Rev. Lett. 102, 136806 (2009)
2009
-
[8]
A. J. Ur ´ıa- ´Alvarez and J. J. Palacios, Amorphization-induced topological and insulator-metal transitions in bidimensional bixsb1−x alloys (2024), arXiv:2410.16034 [cond-mat.dis-nn]
Pith/arXiv arXiv 2024
-
[9]
Agarwala and V
A. Agarwala and V . B. Shenoy, Phys. Rev. Lett.118, 236402 (2017)
2017
-
[10]
Cheng, T
X. Cheng, T. Qu, L. Xiao, S. Jia, J. Chen, and L. Zhang, Phys. Rev. B108, L081110 (2023)
2023
-
[11]
Regis, V
V . Regis, V . Velasco, M. B. S. Neto, and C. Lewenkopf, Phys. Rev. B110, L161105 (2024)
2024
-
[12]
Agarwala, V
A. Agarwala, V . Juriˇci´c, and B. Roy, Phys. Rev. Res.2, 012067 (2020)
2020
-
[13]
Z. Zhang, P. Delplace, and R. Fleury, Sci- ence Advances9, 10.1126/sciadv.adg3186 (2023), https://www.science.org/doi/pdf/10.1126/sciadv.adg3186
-
[14]
A. G. Grushin and C. Repellin, Phys. Rev. Lett.130, 186702 (2023)
2023
-
[15]
Varjas, A
D. Varjas, A. Lau, K. P¨oyh¨onen, A. R. Akhmerov, D. I. Pikulin, and I. C. Fulga, Phys. Rev. Lett.123, 196401 (2019)
2019
-
[16]
Q. Marsal, D. Varjas, and A. G. Grushin, Proceedings of the National Academy of Sciences117, 30260 (2020), https://www.pnas.org/doi/pdf/10.1073/pnas.2007384117
-
[17]
P.Corbae, J.D.Hannukainen, Q.Marsal, D.M.-Segovia, and A.G.Grushin, EPL142, 16001 (2023)
2023
-
[18]
Y .-B. Yang, T. Qin, D.-L. Deng, L.-M. Duan, and Y . Xu, Phys. Rev. Lett.123, 076401 (2019)
2019
-
[19]
Cassella, P
G. Cassella, P. d’Ornellas, T. Hodson, W. M. H. Natori, and J. Knolle, Nature Communications14, 6663 (2023)
2023
-
[20]
M. Costa, G. R. Schleder, M. Buongiorno Nardelli, C. Lewenkopf, and A. Fazzio, Nano Let- ters19, 8941 (2019), pMID: 31679336, https://doi.org/10.1021/acs.nanolett.9b03881
-
[21]
Mansha and Y
S. Mansha and Y . D. Chong, Phys. Rev. B96, 121405 (2017)
2017
-
[22]
S. Kim, A. Agarwala, and D. Chowdhury, Phys. Rev. Lett.130, 026202 (2023)
2023
-
[23]
Spring, A
H. Spring, A. R. Akhmerov, and D. Varjas, SciPost Phys.11, 022 (2021)
2021
-
[24]
M. F. Mart ´ınez, L. Jezequel, J. H. Bardarson, T. K. Kvorning, and J. D. Hannukainen, A one-particle density matrix frame- work for mode-shell correspondence: Characterizing topol- ogy in amorphous higher-order topological insulators (2025), arXiv:2509.03632 [cond-mat.mes-hall]
Pith/arXiv arXiv 2025
-
[25]
Manna, S
S. Manna, S. K. Das, and B. Roy, Phys. Rev. B109, 174512 (2024)
2024
-
[26]
L. Jezequel, J. H. Bardarson, and A. G. Grushin, Explicit equiv- alence between the spectral localizer and local chern and wind- ing markers (2025), arXiv:2508.00214 [cond-mat.mes-hall]
Pith/arXiv arXiv 2025
-
[27]
A. K. Ghosh, Q. Marsal, and A. M. Black-Schaffer, Laser- induced topological phases in monolayer amorphous carbon (2025), arXiv:2508.09571 [cond-mat.mes-hall]. 9
arXiv 2025
-
[28]
S. T. Ciocys, Q. Marsal, P. Corbae, D. Varjas, E. Kennedy, M. Scott, F. Hellman, A. G. Grushin, and A. Lanzara, Nature Communications15, 8141 (2024)
2024
-
[29]
Corbae, S
P. Corbae, S. Ciocys, and D. Varjas, Nat. Mat.22, 200 (2023)
2023
-
[30]
Rechtsman, A
M. Rechtsman, A. Szameit, F. Dreisow, M. Heinrich, R. Keil, S. Nolte, and M. Segev, Phys. Rev. Lett.106, 193904 (2011)
2011
-
[31]
Zhou, G.-G
P. Zhou, G.-G. Liu, X. Ren, Y . Yang, H. Xue, L. Bi, L. Deng, Y . Chong, and B. Zhang, Light: Science & Applications9, 133 (2020)
2020
-
[32]
C. W. Groth, M. Wimmer, A. R. Akhmerov, J. Tworzydło, and C. W. J. Beenakker, Phys. Rev. Lett.103, 196805 (2009)
2009
-
[33]
St ¨utzer, Y
S. St ¨utzer, Y . Plotnik, Y . Lumer, P. Titum, N. H. Lindner, M. Segev, M. C. Rechtsman, and A. Szameit, Nature560, 461 (2018)
2018
-
[34]
J. Song, H. Liu, H. Jiang, Q.-f. Sun, and X. C. Xie, Phys. Rev. B85, 195125 (2012)
2012
-
[35]
- a prototypical model of time-reversal symmetric, two- dimensional topological insulator, with extended hoppings. To continuously tune the system from a crystalline to an amor- phous configuration, we introduce structural disorder via ran- dom displacements of the lattice sites, drawn from a normal distribution. We diagnose the topological nature of the ...
Pith/arXiv arXiv 2025
-
[36]
C. P. Orth, T. Sekera, C. Bruder, and T. L. Schmidt, Scientific Reports6, 24007 (2016)
2016
-
[37]
Bernevig, T.Hughes, and S.Cheng.Zhang, Nat
B. Bernevig, T.Hughes, and S.Cheng.Zhang, Nat. Mat.314, 1757 (2006)
2006
-
[38]
Chen, Phys
W. Chen, Phys. Rev. B107, 045111 (2023)
2023
-
[39]
L. A. Oliveira and W. Chen, Phys. Rev. B109, 094202 (2024)
2024
-
[40]
B.A.Bernevig, T.L.Hughes, and S.- C.Zhang, Science314, 1757 (2006), https://www.science.org/doi/pdf/10.1126/science.1133734
-
[41]
Konig, S
M. Konig, S. Wiedmann, C. BruHERE!ne, A. Roth, H. Buh- mann, L. W. Molenkamp, X.-L. Qi, and S.-C. Zhang, Science 318, 766-770 (2007)
2007
-
[42]
Chen, Phys
W. Chen, Phys. Rev. B101, 195120 (2020)
2020
-
[43]
C. Wang, T. Cheng, Z. Liu, F. Liu, and H. Huang, Phys. Rev. Lett.128, 056401 (2022)
2022
-
[44]
Franca and A
S. Franca and A. G. Grushin, Phys. Rev. Mater.8, L021201 (2024)
2024
-
[45]
Bernevig and T
B. Bernevig and T. Hughes,Topological insulators and topolog- ical superconductors(Princeton Uni. Press, Princeton, 2013)
2013
-
[46]
Fu and C
L. Fu and C. L. Kane, Phys. Rev. B76, 045302 (2007)
2007
-
[47]
Kitaev, Annals of Physics321, 2 (2006), january Special Issue
A. Kitaev, Annals of Physics321, 2 (2006), january Special Issue
2006
-
[48]
Bianco and R
R. Bianco and R. Resta, Phys. Rev. B84, 241106 (2011)
2011
-
[49]
d’Ornellas, R
P. d’Ornellas, R. Barnett, and D. K. K. Lee, Phys. Rev. B106, 155124 (2022)
2022
-
[50]
Sykes and R
J. Sykes and R. Barnett, Materials for Quantum Technology2, 025005 (2022)
2022
-
[51]
Sykes and R
J. Sykes and R. Barnett, Phys. Rev. B103, 155134 (2021)
2021
-
[52]
J. D. Hannukainen, M. F. Mart ´ınez, J. H. Bardarson, and T. K. Kvorning, Phys. Rev. Lett.129, 277601 (2022)
2022
-
[53]
R. Fan, P. Zhang, and Y . Gu, SciPost Phys.15, 249 (2023)
2023
-
[54]
von Gersdorff, S
G. von Gersdorff, S. Panahiyan, and W. Chen, Phys. Rev. B103, 245146 (2021)
2021
-
[55]
Corbae, S
P. Corbae, S. Ciocys, D. Varjas, E. Kennedy, S. Zeltmann, M. Molina-Ruiz, S. M. Griffin, C. Jozwiak, Z. Chen, L.-W. Wang, A. M. Minor, M. Scott, A. G. Grushin, A. Lanzara, and F. Hellman, Nature Materials22, 200 (2023)
2023
-
[56]
Landauer, IBM Journal of Research and Development1, 223 (1957)
R. Landauer, IBM Journal of Research and Development1, 223 (1957)
1957
-
[57]
M.Buttiker, Phys. Rev. B38, 9375 (1988)
1988
-
[58]
C. W. Groth, M. Wimmer, A. R. Akhmerov, and X. Waintal, New Journal of Physics16, 063065 (2014)
2014
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.