REVIEW 3 major objections 4 minor 33 references
A Discrete Formulation of Second Stiefel-Whitney Class for Band Theory
T0 review · 3 major / 4 minor · reviewed 2026-08-11 · deepseek-v4-flash
Pith's one-line read The paper presents a fully discrete, gauge-fixing-free algorithm for computing the second Stiefel-Whitney class $w_2$ directly from Bloch states on a Brillouin-zone mesh, and verifies it on tight-binding models.
desk verdict A genuinely discrete, gauge-fixing-free w2 formula with real numerical checks, but the gauge-invariance proof is sketched and the -1 eigenvalue perturbation is not rigorously handled. 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 the SVD-normalized overlap matrix $w_{01} = L R^\top$ obtained from $\tilde w_{01} = \Phi_0^\top \Phi_1$, together with its explicit lift to the Pin$_+$ group. The lift is defined by decomposing $w = q_w R_1^{s_w}$, taking the principal logarithm of the $SO(r)$ part $q_w$ (assuming no eigenvalue $-1$), exponentiating the rotation parameters with the Clifford generators $\Sigma_{ij} = [\gamma_i,\gamma_j]/(-4)$, and multiplying by $\gamma_1$ when $s_w = 1$. This lift satisfies $u(w^\top) = u(w)^\dagger$, which makes the plaquette Wilson loop orientation-independent; the $\mathbb{Z}_2$ value $z_\square$ is read from whether the lifted loop is near $+1$ or $-1$. Gauge invariance follows because a gauge change shifts $z$ by a coboundary $(\delta\eta)_{012}$, which cancels when plaquette values are multiplied, and the Whitney sum formula follows from the anticommutation of lifts of odd orthogonal matrices.
What would settle it
Run the algorithm on the 4x4 model at $m=1$ with a mesh chosen so coarse that some plaquette Wilson loop is not close to $\pm1$, or so that some overlap matrix has a $-1$ eigenvalue in its $SO(r)$ part; if repeated random gauge transformations produce different values of $\prod_\square (-1)^{z_\square}$, the claimed gauge-invariant quantization is false. A cleaner test is to find any PT-symmetric configuration where two different infinitesimal gauge perturbations of the same overlap matrix yield different lifted signs, which would make the cocycle ill-defined.
Extended reading notes
Core claim
In a PT-symmetric band structure, the occupied Bloch states can be chosen real, giving a frame $\Phi_v$ at each mesh vertex $v$. For each link the paper forms $\tilde w_{01} = \Phi_0^\top \Phi_1$ and SVD-normalizes it to $w_{01} = L R^\top \in O(r)$, the discrete Wilson line. Each $w$ is lifted to $u(w) \in \mathrm{Pin}_+(r)$ by writing $w = q_w R_1^{s_w}$, diagonalizing the $SO(r)$ part $q_w$, exponentiating its logarithm with Clifford-algebra generators, and including a $\gamma_1^{s_w}$ factor when $\det w = -1$. On a plaquette $(0123)$ the lifted product $u(w_{01})u(w_{12})u(w_{23})u(w_{30})$ is close to either $+1$ or $-1$, and the bit $z_\square$ records which. The product $\prod_\square (-1)^{z_\square}$ over the entire torus is shown to be invariant under local $O(r)$ gauge transformations and under changes of lift, so it defines the $\mathbb{Z}_2$ number $\nu$ that pairs $w_2$ with the Brillouin-zone torus. The paper verifies this on the Hamiltonian $H_k = \sin k_x \, \sigma_x\otimes\sigma_0 + \sin k_y \, \sigma_y\otimes\sigma_y + (m-\cos k_x-\cos k_y)\,\sigma_z\otimes\sigma_0$, obtaining $\nu=1$ for $0<|m|<2$ and $\nu=0$ otherwise, and checks the Whitney sum formula on $H^{(x)}_k \oplus H^{(y)}_k$.
Load-bearing premise
The load-bearing premise is that the mesh is fine enough that every SVD-normalized overlap matrix has an $SO(r)$ part with no eigenvalue $-1$ and every plaquette Wilson loop sits clearly near $+1$ or $-1$; the paper removes the $-1$ cases by random gauge perturbations but does not rigorously prove the resulting $\mathbb{Z}_2$ class is unchanged by that removal.
Editorial extensions
If this is right
- A band-structure code that outputs Bloch states on a finite $k$-mesh can evaluate $w_2$ directly, without constructing smooth gauges or counting Wilson-loop spectral crossings.
- The same plaquette construction works on any closed two-dimensional base manifold, so $w_2$ can be assigned in geometries beyond the torus whenever a real frame exists.
- The formula is manifestly $\mathbb{Z}_2$-quantized: each plaquette contributes exactly $0$ or $1$, so the result cannot drift continuously under small Hamiltonian perturbations.
- The lattice-level Whitney sum formula provides a practical consistency check and a way to compute $w_2$ of a direct-sum band group from its factors.
- The Villainized-field-theory interpretation gives a common language for the discrete invariant and for lattice constructions of the third integral Stiefel-Whitney class $W_3$ as a Pinc obstruction.
Reading between the lines
- The paper does not pursue this, but a numerical study that coarsens the mesh near a gap-closing point would turn "sufficiently fine" into a quantitative criterion for when SVD normalization and the $-1$-eigenvalue removal are reliable.
- The paper does not state this explicitly, but because the input is only the projection $P_v$ of the occupied bands, the same algorithm should apply to Wannier-interpolated or other projection-valued data sets without modification.
- The paper does not demonstrate it, but the lift protocol could be reused to compute $\mathbb{Z}_2$ invariants in magnetic space groups or superconducting classes where reality of the occupied frame is enforced by a different antiunitary symmetry.
- The paper does not carry this out, but a three-dimensional extension could pair $w_2$ on $k$-slices with $W_3$ on cubes to give a fully discrete $\mathbb{Z}_2$-monopole charge for nodal-line semimetals.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper proposes a fully discrete, gauge-fixing-free algorithm for computing the second Stiefel-Whitney class w2 of a real vector bundle over a closed manifold, motivated by band-theoretic applications in PT-symmetric systems. The method uses Bloch states sampled on a triangular or cubic lattice mesh, forms SVD-normalized overlap matrices on each link, lifts the resulting O(r) Wilson lines to Pin+(r) via a logarithmic protocol, and defines a Z2-valued 2-cocycle from the Pin+ Wilson loop on each plaquette. The authors claim the resulting Z2 number is manifestly quantized and independent of the gauge choices, and they demonstrate the algorithm on a known 4x4 tight-binding model, an embedded 8x8 model, and a Whitney-sum test. The paper also draws an analogy to lattice field theory through spinon decomposition and Villainization. The central claim is that the method provides a practical and rigorous numerical tool for w2 without global gauge fixing or continuum smoothness assumptions.
Significance. If the missing technical steps are supplied, this would be a valuable contribution: it offers a direct, parameter-free, and explicit algorithm for a topological invariant that is central to classifying higher-order topological insulators and nodal-line semimetals, and it avoids the spectral-crossing counting that can be numerically fragile. The construction is concrete and implementable, and the numerical checks against known values, including the Whitney sum formula, are appropriate and encouraging. The connection to lattice field theory is suggestive and may help unify viewpoints, although it is not the main result. The main strengths are that the algorithm has no fitted parameters, the quantization is built in by construction, and the tests are independent of the method's own output.
major comments (3)
- [Sec. II.B, Eq. (18)] The gauge-invariance of the lattice Z2 value rests on the identity u(w'_{01}) = (-1)^{eta_{01}} u(V_0)^dagger u(w_{01}) u(V_1), which is asserted without proof. This identity is not automatic for an arbitrary section of Pin+(r); it holds because both sides are lifts of the same O(r) element and Pin+(r) -> O(r) is a double cover, but the paper should state and prove this property, including its domain of validity, since the lift protocol fails when the SO part has a -1 eigenvalue. Without a derivation, the central invariance claim of the algorithm is not fully established.
- [Sec. II.B, after Eq. (7)] The handling of overlap matrices whose SO part has a -1 eigenvalue is a load-bearing point for the algorithm's well-definedness. The paper states that such an occurrence is nongeneric and can be removed by random gauge transformations, but it does not prove that the set of gauge choices for which any link matrix has a -1 eigenvalue has measure zero, nor does it explicitly argue that two different good perturbations produce the same Z2 value. The latter would follow from Eq. (18) once proved, but the logical connection is not made. Since the algorithm is undefined on the bad set, the 'manifestly quantized' claim requires a precise statement of the measure-zero condition and the invariance on the good set.
- [Sec. III] The numerical tests reported in Eqs. (28) and (31) confirm the expected values, but essential reproducibility parameters are missing: the mesh size L, the number of random gauge samples, and the threshold used in the proximity tests such as 'W_box ~ 1' versus 'W_box ~ -1' are not reported. Without these, the reader cannot assess the robustness of the quantization or reproduce the results. For a paper whose central deliverable is a numerical algorithm, this is a necessary detail.
minor comments (4)
- [Eq. (14) and neighboring text] The symbol W012 is used for both the O(r) Wilson loop in Eq. (13) and its Pin+(r) lift in Eq. (14); please use distinct notations (for example, W012 and tilde{W}_{012}) to avoid ambiguity.
- [Sec. II.B, paragraph after Eq. (7)] The phrase 'randomness of the overlap matrix can be ensured by applying random gauge transformations' is vague; a precise probability statement about the codimension of the -1 eigenvalue locus in the gauge group would be more useful.
- [Ref. [21]] The phrase 'see the Supplementary Material of Ref. [21]' is not specific; please provide a section number or a more detailed pointer to the review of Wilson-loop approaches.
- [Abstract] The abstract's claim of a 'fully discrete' method is slightly overstated because the algorithm requires a sufficiently fine mesh and a generic gauge choice; consider adding qualifiers such as 'for sufficiently fine meshes'.
Circularity Check
No circularity: the discrete w2 algorithm is defined directly from Bloch-state overlap matrices and checked against independent known invariants.
full rationale
The construction in Sec. II.B defines w01 = L R^T via SVD normalization of the overlap Phi0^T Phi1 (Eq. 4), lifts each O(r) matrix to Pin+ by the explicit protocol in Eqs. (6)-(10), forms the Pin+ Wilson loop W012 in Eq. (14), and assigns the Z2 plaquette value z012 by comparison with +/-1 (Eq. 15); the invariant is the sum over plaquettes. No parameter is fitted to the target value, and the claimed outputs are verified against independent known values: the 4x4 model with known w2 (Eq. 28) and the Whitney sum formula for the direct-sum 2x2 models (Eq. 31). The only author self-citations (Refs. [13], [17], [22]) appear in introductory context or in the field-theory rationale of Sec. IV; the central algorithm does not depend on them. The gauge-invariance relation Eq. (18) is a standard lifting-covariance statement following from the fact that each O(r) element has exactly two Pin+ lifts differing by -1; it is a proof obligation, not an identification of input with output. The unproven assumption about avoiding -1 eigenvalues of the SO(r) part is a well-definedness and robustness concern, not a circularity.
Assumptions & free parameters
assumptions (3)
- domain assumption The Brillouin-zone mesh is sufficiently fine that the O(r) Wilson loop on each plaquette is close to the identity, making the Pin+ loop close to ±1.
- domain assumption The SO(r) part of the SVD-normalized overlap matrix generically has no -1 eigenvalue, so the logarithm in Eq. (7) is well defined.
- domain assumption The original O(N) bundle over the BZ is flat and topologically trivial, so that the gauge theory treatment applies.
Cite this review
Pith. "Pith review of A Discrete Formulation of Second Stiefel-Whitney Class for Band Theory." pith.science (2026). https://pith.science/paper/SHC4J7RI
@misc{pith2026241218796,
author = {Pith},
title = {Pith review of: A Discrete Formulation of Second Stiefel-Whitney Class for Band Theory},
year = {2026},
howpublished = {\url{https://pith.science/paper/SHC4J7RI}},
note = {Machine review of arXiv:2412.18796}
}
abstract
Topological invariants in band theory are often formulated assuming that Bloch wave functions are smoothly defined over the Brillouin zone (BZ). However, first-principles band calculations typically provide Bloch states only at discrete points in the BZ, rendering standard continuum-based approaches inapplicable. In this work, we focus on the second Stiefel-Whitney class $w_2$, a key $\mathbb{Z}_2$ topological invariant under PT symmetry that characterizes various higher-order topological insulators and nodal-line semimetals. We develop a fully discrete, gauge-fixing-free formula for $w_2$ which depends solely on the Bloch states sampled at discrete BZ points. Furthermore, we clarify how our discrete construction connects to lattice field theory, providing a unifying perspective that benefits both high-energy and condensed matter approaches.
Figures
Reference graph
Works this paper leans on
-
[1]
We show why the eigenvalues of qw can be assumed not to include −1 for our purpose
= u(w01)† (12) holds. We show why the eigenvalues of qw can be assumed not to include −1 for our purpose. The eigenvalues of an SO(r) matrix w ∈ SO(r) may have a pair of eigenval- ues {−1, −1}. 5 The pair {−1, −1} splits into pairs of complex eigenvalues {eiθ, e−iθ} under infinitesimal per- turbations. Therefore, if w ∈ SO(r) is random, the ab- sence of −...
-
[2]
of w′ 01 is independently defined from u(w01) via (10). However, it satisfies the following re- lation using the lift u(Vv) of the gauge transformation matrices: u(w′
-
[3]
= (−1)η01 u(V0)†u(w01)u(V1), (18) where η01 ∈ {0, 1} is some 1-cochain. Thus, the trans- formation of the Z2-valued z012 under the gauge trans- formation can be written as z′ 012 := z012 + (δη)012 mod 2, (19) (δη)012 := η12 + η02 + η01 mod 2. (20) The change (δη)012 is canceled between adjacent trian- gles, proving the gauge invariance of (−1)ν(Y2). Simi-...
-
[4]
C.-K. Chiu, J. C. Y . Teo, A. P. Schnyder, and S. Ryu, Clas- sification of topological quantum matter with symmetries, Rev. Mod. Phys. 88, 035005 (2016)
work page 2016
-
[5]
G. E. V olovik, The universe in a helium droplet, V ol. 117 (OUP Oxford, 2003)
work page 2003
-
[6]
M. Z. Hasan and C. L. Kane, Colloquium: Topological insulators, Rev. Mod. Phys. 82, 3045 (2010)
2010
-
[7]
Qi and S.-C
X.-L. Qi and S.-C. Zhang, Topological insulators and su- perconductors, Rev. Mod. Phys. 83, 1057 (2011)
2011
- [8]
Show all 33 references
-
[9]
D. J. Thouless, M. Kohmoto, M. P. Nightingale, and M. den Nijs, Quantized hall conductance in a two- dimensional periodic potential, Phys. Rev. Lett. 49, 405 (1982)
1982
-
[10]
C. L. Kane and E. J. Mele, Z2 topological order and the quantum spin hall effect, Phys. Rev. Lett. 95, 146802 (2005)
2005
-
[11]
Fu and C
L. Fu and C. L. Kane, Time reversal polarization and a Z2 adiabatic spin pump, Phys. Rev. B 74, 195312 (2006)
2006
-
[12]
R. Yu, X. L. Qi, A. Bernevig, Z. Fang, and X. Dai, Equiv- alent expression of 𭟋2 topological invariant for band insu- lators using the non-abelian berry connection, Phys. Rev. B 84, 075119 (2011)
2011
-
[13]
A. P. Schnyder, S. Ryu, A. Furusaki, and A. W. W. Lud- wig, Classification of topological insulators and super- conductors in three spatial dimensions, Phys. Rev. B 78, 195125 (2008)
2008
-
[14]
Fukui, Y
T. Fukui, Y . Hatsugai, and H. Suzuki, Chern num- bers in discretized brillouin zone: Efficient method of computing (spin) hall conductances, Journal of the Physical Society of Japan 74, 1674 (2005), https://doi.org/10.1143/JPSJ.74.1674
2005 doi
-
[15]
A. A. Soluyanov and D. Vanderbilt, Wannier representa- tion of 𭟋2 topological insulators, Phys. Rev. B83, 035108 (2011)
2011
-
[16]
Morikawa and H
O. Morikawa and H. Suzuki, Winding number on 3d lat- tice (2024), arXiv:2412.03888 [hep-lat]
2024 arXiv
-
[17]
Shiozaki, A discrete formulation for three-dimensional winding number (2024), arXiv:2403.05291 [cond- mat.mes-hall]
K. Shiozaki, A discrete formulation for three-dimensional winding number (2024), arXiv:2403.05291 [cond- mat.mes-hall]
2024 arXiv
-
[18]
Alexandradinata and B
A. Alexandradinata and B. A. Bernevig, Berry-phase de- scription of topological crystalline insulators, Phys. Rev. B 93, 205104 (2016)
2016
-
[19]
Tanizaki, A
Y . Tanizaki, A. Tomiya, and H. Watanabe, Lattice gra- dient flows (de-)stabilizing topological sectors (2024), arXiv:2411.14812 [hep-lat]
2024 arXiv
-
[20]
the most probable interpolation
∼ (−1)z′ 012 1 and u(w01)u(w12)u(w20) ∼ (−1)z012 1, we finally ob- tain ˜W012 ∼ (−1)z012+z′ 012+(s′∪s)012 . (23) Taking the cohomology classes, the right-hand side is nothing but the relation (2). D. Third integral SW class Finally, the 3rd integral SW class[δZ/2] is obtained ...
2024
-
[21]
Ono and K
S. Ono and K. Shiozaki, Towards complete charac- terization of topological insulators and superconduc- tors: A systematic construction of topological invari- ants based on atiyah-hirzebruch spectral sequence (2024), arXiv:2311.15814 [cond-mat.mes-hall]
2024 arXiv
-
[22]
C. Fang, H. Weng, X. Dai, and Z. Fang, Topological nodal line semimetals*, Chinese Physics B 25, 117106 (2016)
2016
-
[23]
J. Ahn, S. Park, D. Kim, Y . Kim, and B.-J. Yang, Stiefel–whitney classes and topological phases in band theory, Chinese Physics B 28, 117101 (2019)
2019
-
[24]
Davoyan, W
Z. Davoyan, W. J. Jankowski, A. Bouhon, and R.- J. Slager, Three-dimensional PT -symmetric topological phases with a pontryagin index, Phys. Rev. B109, 165125 (2024)
2024
-
[25]
J. Ahn, D. Kim, Y . Kim, and B.-J. Yang, Band topology and linking structure of nodal line semimetals with Z2 monopole charges, Phys. Rev. Lett. 121, 106403 (2018)
2018
-
[26]
Chen, Instanton Density Operator in Lattice QCD from Higher Category Theory (2024), arXiv:2406.06673 [hep-lat]
J.-Y . Chen, Instanton Density Operator in Lattice QCD from Higher Category Theory (2024), arXiv:2406.06673 [hep-lat]
2024
-
[27]
Sachdev and R
S. Sachdev and R. Jalabert, Effective lattice models for two-dimensional quantum antiferromagnets, Modern Physics Letters B 4, 1043 (1990)
1990
-
[28]
Rabinovici and S
E. Rabinovici and S. Samuel, The CP N −1 Model: A Strong Coupling Lattice Approach, Phys. Lett. B101, 323 (1981)
1981
-
[29]
Di Vecchia, A
P. Di Vecchia, A. Holtkamp, R. Musto, F. Nicodemi, and R. Pettorino, Lattice CP N −1 Models and Their Large N Behavior, Nucl. Phys. B 190, 719 (1981)
1981
-
[30]
M. B. Einhorn and R. Savit, Topological Excitations in the Abelian Higgs Model, Phys. Rev. D 17, 2583 (1978)
1978
-
[31]
Berg and M
B. Berg and M. Luscher, Definition and Statistical Dis- tributions of a Topological Number in the Lattice O(3) Sigma Model, Nucl. Phys. B 190, 412 (1981)
1981
-
[32]
G. W. Moore, A Comment On Berry Connections (2017), arXiv:1706.01149 [hep-th]
2017 arXiv
-
[33]
Mack and V
G. Mack and V . B. Petkova, Z2 Monopoles in the Standard SU(2) Lattice Gauge Theory Model, Z. Phys. C 12, 177 (1982)
1982
Reviewed August 11, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.