REVIEW 4 major objections 4 minor 5 references
Topological Braiding of Bloch Eigenmodes Protected by Non-Abelian Quaternion Invariants
T0 review · 4 major / 4 minor · reviewed 2026-08-06 · deepseek-v4-flash
Pith's one-line read Bloch eigenmodes in a gapped three-band system braid like strands, with each braid word fixed by a quaternion topological invariant.
desk verdict Nice mapping from Q8 charges to braid words, with a real acoustic experiment, but the 'protected minimum braid sequence' rests on an unproved mirror-symmetry step that the paper only imposes in the SI. 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 non-Abelian Berry phase (NABP), the matrix holonomy $W = \exp[\oint A_{3\times3}(q)\,dq]$ built from the non-Abelian Berry connection $A_{mn}(q) = \langle \psi_m(q)|\partial_q|\psi_n(q)\rangle$. PT symmetry makes the eigenvectors real and quantizes the rotation angle of $W$ to $0,\pm\pi,\pm2\pi$; lifting SO(3) to SU(2) identifies $W$ with an element of the quaternion group $Q_8$. The load-bearing identity is the square root of the holonomy: a hidden mirror symmetry forces the holonomy over each half of the Brillouin zone to be equal, so that $\exp(\pm\frac{\pi}{2}L_x)$, $\exp(\pm\frac{\pi}{2}L_y)$, and $\exp(\pm\frac{\pi}{2}L_z)$ are exactly the half-BZ holonomies, and these matrices coincide with the generators (or a conjugation of two generators) of the Artin braid group $B_3$. The squared half-BZ word reproduces the full-BZ NABP, which is why each quaternion invariant determines a topologically protected minimum braid sequence.
What would settle it
Compute the half-BZ non-Abelian Berry holonomies $W_- = \exp[\int_{-\pi}^{0} A(q)\,dq]$ and $W_+ = \exp[\int_{0}^{\pi} A(q)\,dq]$ for a PT-symmetric three-band model with the stated mirror symmetry; if $W_-$ and $W_+$ are not equal, or their product is not the predicted diagonal $\exp(\pm\pi L_{x,y,z})$ with the exact eigenvector exchange, the braid-word assignment fails. In the acoustic system, a direct check is whether the half-BZ phases $\Theta_n^\pm$ of Eq. (8) stay quantized to 0 or $\pi$ under perturbations that preserve the mirror symmetry.
Extended reading notes
Core claim
On the paper's own terms, the discovery is a correspondence between the quaternion group $Q_8 = \{\pm1, \pm i, \pm j, \pm k\}$ and Artin's braid group $B_3$: for every PT-symmetric, fully gapped three-band Hamiltonian, the non-Abelian Berry phase $W = \exp[\oint A(q)\,dq]$ equals one of the quaternion invariants, and after taking its square root the half-Brillouin-zone holonomy is exactly a braid generator. Explicitly, $\exp(\pm \frac{\pi}{2}L_x)$ maps to $b_{12}^{\pm 1}$, $\exp(\pm \frac{\pi}{2}L_z)$ maps to $b_{23}^{\pm 1}$, and $\exp(\pm \frac{\pi}{2}L_y)$ maps to $(b_{12}b_{23}b_{12}^{-1})^{\pm 1}$, so the full-BZ holonomy is the square of a minimum braid word such as $b_{23}b_{23}$ for the $k$-phase and $(b_{12}b_{23}b_{12}^{-1})^2$ for the $j$-phase. This braid word is a bulk property: the eigenmodes swap profiles without gap closing, without degeneracy, and without being edge states, yet the exchange is protected by the non-Abelian band topology. The paper further shows that the sign of the quaternion invariant is observable through half-Brillouin-zone geometric phases that quantize to 0 or $\pi$, and reports acoustic-cavity experiments in which the predicted braiding of eigenfunctions is directly measured.
Load-bearing premise
The load-bearing premise is that the hidden mirror symmetry makes the transport of the three wavefunctions over the first half of the Brillouin zone exactly the same as over the second half, so that the required quarter-turn rotation is completed precisely at the midpoint; if the two halves differ, the full round trip is a diagonal matrix that does not single out any braid word.
Editorial extensions
If this is right
- Each of the eight quaternion invariants corresponds to a specific minimal braid word in $B_3$; the $k$-phase gives $b_{23}^2$, the $i$-phase gives $b_{12}^2$, and the $j$-phase gives $(b_{12}b_{23}b_{12}^{-1})^2$.
- The braiding happens in the bulk of a gapped insulator, with no gap closing and no degenerate point, so it is accessible to ordinary non-interacting and classical wave systems.
- Because the braid word is fixed by the band topology, the exchange sequence survives local perturbations; disorder may distort the trajectories, but the final braid word remains the same.
- The sign of the quaternion invariant is physically readable: it decides in which half of the Brillouin zone the quantized $\pi$ geometric phase appears, which the full-BZ Zak phase alone misses.
- Relaxing the mirror symmetry enlarges the allowed braid sequences (e.g., $b_{12}^4 b_{23}^{-2}$), suggesting that braid sequences can refine non-Abelian topological classification into sub-genres.
Reading between the lines
- Beyond the paper: because the braid generators here are $3\times3$ orthogonal matrices, the same correspondence could serve as a classical or single-particle simulator of braid-group algebra, where the braided objects are bulk band indices rather than quasiparticles with anyonic statistics.
- Beyond the paper: the half-BZ geometric phases $\Theta_n^\pm$ are effectively mirror-symmetry-resolved holonomies, so a natural extension is to define analogous quantities on other symmetry-invariant submanifolds (e.g., $C_2$ or $C_3$ invariant planes) to diagnose braid words in higher-symmetry multiband models.
- Beyond the paper: the authors leave open exactly which braid word appears when mirror symmetry is broken but PT symmetry remains; a concrete test would be to identify the split point that makes the two factors equal and check whether arbitrary even-length words such as $b_{12}^4 b_{23}^{-2}$ are realized as claimed.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The manuscript proposes that the non-Abelian Berry phase (NABP) of a gapped, PT-symmetric three-band insulator can be decomposed into half-Brillouin-zone (BZ) rotations that act as generators of the braid group B3, and that each quaternion invariant in Q8 uniquely determines a 'minimum braid sequence' of the Bloch eigenmodes. The authors construct a one-dimensional tight-binding model with a hidden mirror symmetry, compute the NABP and define half-BZ geometric phases, and report acoustic experiments in coupled cavities with a synthetic dimension that realize braid sequences associated with the k, j, and -1 invariants.
Significance. If the central claim were established, the work would introduce a new class of topologically protected braiding in non-interacting systems, with implications for classical wave physics and potentially for single-particle quantum systems. The algebraic Table 1 relating SO(3) rotations to braid words is explicit, the proposed braid representation is concrete, and the acoustic experiment is a substantial proof-of-principle demonstration. However, the load-bearing step that mirror symmetry forces the half-BZ holonomy to be the square root of the full-BZ rotation is not proven, and several internal inconsistencies in the numerical example further undermine the conclusions.
major comments (4)
- [Topological braiding of Bloch modes, after Eq. (7)] The central step of the paper is the assertion that 'The mirror symmetry ensures that the NABPs accumulated over each half of the BZ are the same, i.e., it guarantees the ±π/2 rotation ... completes exactly in half BZ.' This is not a consequence of the symmetry. For a mirror symmetry R H(q) R^{-1}=H(-q), the half-BZ Wilson loops are related by U_- = R^T U_+^{-1} R (up to endpoint gauge choices), not by U_- = U_+. A concrete counterexample is R=diag(1,-1,1), W=exp(π L_z), and U_+ = exp(π n·L) with n=(1,1,0)/√2. Then U_+^2=I, U_+ exchanges modes 1 and 2, and R U_+^{-1} R U_+ = W, so the full-BZ NABP is the k-phase invariant, yet the half-BZ word is not b23. Thus the same quaternion invariant admits mirror-symmetric decompositions with different half-BZ braid words. SI §3 explicitly 'imposes' the equality of the half-BZ integrals rather than deriving it, and the Discussion concedes that without mirror symmetry the word changes. No argument is provided that mirror symmetry alone selects a unique word, so the claimed 'minimum braid sequence' is a model-dependent choice rather than a protected consequence of the Q8 charge.
- [Quantized half-BZ geometric phases and SI §2] The half-BZ geometric phases Θ_n^± defined in Eq. (8) are not independent invariants. The band indices m,n in Eq. (8) are chosen, as stated in SI §2, 'according to the braid sequence to satisfy the holonomy condition over half BZ.' Therefore the quantization of Θ_n^± to 0 or π, and the association of specific sign patterns with ±i, ±j, ±k, is a restatement of the assumed half-BZ braid word rather than an independent verification. The relative homotopy group π1(X,A)=Z2 guarantees only a ℤ2-valued invariant for the half-BZ loop; it does not select which pair of eigenmodes is compared or which square root of the full-BZ rotation is realized.
- [Discussion] The Discussion explicitly allows multiple braid words for the same quaternion invariant: relaxing mirror symmetry 'the array of allowed braid sequences is vastly increased,' and the k-phase can yield sequences such as b12^4 b23^{-2}. This contradicts the claim in Table 1 and the abstract that each quaternion invariant is robustly associated with a topologically protected minimum braid sequence. If the braid word can change without changing the Q8 charge, then the braid sequence is not an invariant of the non-Abelian band topology.
- [Topological braiding of Bloch modes, parameters after Eq. (7)] The parameters given for the k-phase in this section are inconsistent with the claimed eigenvector evolution. With v1=1, v2=0, v3=-1, ω1=ω2=0, ω3=-4, the Hamiltonian (6) at q=π is diag(-2,0,-2), which is gapless, contradicting the statement that the system is always gapped. At q=0 the Hamiltonian is not diagonal, so the eigenvectors are not the basis states (0,1,0)^T and (1,0,0)^T stated in the text. The parameter set in Table S3 for k (v1=0, v2=-1, v3=1, ω1=-4, ω3=0) is gapped but still does not give basis eigenvectors at q=0. The description of the braiding in Fig. 2(a) therefore does not match the stated model parameters; this must be corrected.
minor comments (4)
- [Table S1] In the row for -i, the entry 'exp(π L_z)' should be 'exp(-π L_x)'; exp(π L_z)=diag(-1,-1,1), which does not match the stated diag(1,-1,-1).
- [After Table 1] The sentence beginning 'To summarize the discussion so far...' is grammatically incomplete; it appears to lack a main clause.
- [Table 1] The third -1 row lists 'exp(2π M_a)' again; this is likely a typo for 'exp(-2π M_a)' or 'exp(2π M_b)' given the text around Eq. (5).
- [Eq. (4)] The notation identifies braid group generators with 3×3 rotation matrices; the paper should state explicitly that this is a finite-dimensional representation of B3, not an isomorphism, since the matrices satisfy b12^4 = b23^4 = 1.
Circularity Check
The 'topologically protected minimum braid sequence' reduces to a definition plus an imposed condition: braid generators are defined as π/2-rotation square roots (Eq. 4), the half-BZ equal-holonomy split is 'imposed' (SI §3) while asserted as mirror-guaranteed, and the half-BZ phases cited as the 'rigorous foundation' are defined from the assumed word.
-
self definitional
[Eq. (4) and Table 1]
"exp (±π/2 L_x) = (1 0 0; 0 0 ∓1; 0 ±1 0) → b12^{±1}, where b12^{±1} are two conjugate generators of the Artin's braiding group B3. In other words, the π/2 rotation maps one-to-one to a braiding operation."
The braid generator is defined to be exactly the square-root matrix: b12 := exp(±π/2 L_x) and, similarly, b23 := exp(±π/2 L_z). Since the full-BZ NABP for ±i is exp(±π L_x) = diag(1,−1,−1), the Table 1 entry 'i ↔ b12 b12' is the identity W = (√W)^2 restated in braid vocabulary; nothing in the topology selects this square-root factorization. The Discussion later concedes that the same k-phase admits other words (e.g., b12^4 b23^{-2}) when the equal-halves constraint is dropped, confirming that the 'minimum braid sequence' is a chosen square-root convention presented as an unveiled connection, not a derived topological consequence.
-
fitted input called prediction
[Main text after Eq. (7); SI §3]
"In the presence of mirror symmetry, we impose the condition exp[∫_{−π}^{0} A_{3×3}(q)dq] = exp[∫_{0}^{π} A_{3×3}(q)dq]. This constraint ensures that exp[∫_{−π}^{0} A_{3×3}(q)dq] ∈ G."
The central prediction — that each quaternion charge completes its ±π/2 rotation over half the BZ, yielding its 'protected minimum braid sequence' — is exactly this imposed condition, yet the main text states it as a theorem: 'The mirror symmetry ensures that the NABPs accumulated over each half of the BZ are the same... it guarantees the ±π/2 rotation.' The symmetry R H(q) R^{-1} = H(−q) only relates the half-BZ Wilson loops by inverse/conjugation, not by equality, and the Discussion concedes that without the condition the same k-phase can also give b12^4 b23^{-2}. The predicted half-BZ braid word (b23 for Q8 = k) is therefore the imposed square-root holonomy of Step 1 relabeled as a prediction; the output equals the input by construction.
1 more flagged steps
-
self definitional
[SI §2; main text 'Quantized half-BZ geometric phases']
"The indices are chosen according to the braid sequence to satisfy the holonomy condition over half BZ. ... This establishes a rigorous foundation for decomposing NABPs into sequences of braiding operations."
The half-BZ phases Θ_n^± are constructed by selecting the comparison pair (band indices m, n in ⟨ψ_n(−π)|ψ_m(0)⟩ and ⟨ψ_m(0)|ψ_n(π)⟩) 'according to the braid sequence,' so the braid word is an input to the definition of the phases. These same phases are then invoked as the 'rigorous foundation for decomposing NABPs into sequences of braiding operations.' The foundation presupposes the decomposition it purports to justify: the ℤ2 quantization from π1(X, A) = ℤ2 only quantizes the phase after the comparison pair has been fixed by the assumed word.
full rationale
The quaternion classification is externally grounded (refs. 22, 23 are not self-citations; ref. 33 independently proves the SO(3)-braid decompositions), and the paper's self-citations (refs. 12, 44, 45 sharing author G. Ma) concern the acoustic coupling scheme and are not load-bearing. The circularity sits in the central claim that each quaternion invariant is 'robustly associated with a topologically protected minimum braid sequence.' Three reduction steps: (1) Eq. (4) defines the braid generators as the π/2 rotations, so Table 1's mapping from the full-BZ NABP to a braid word is the trivial identity W = (√W)^2 restated in braid notation. (2) The step that selects a definite half-BZ word — mirror symmetry making the two half-BZ holonomies equal so the ±π/2 rotation 'completes exactly in half BZ' — is written in SI §3 as 'we impose the condition,' not derived; the stated symmetry R H(q) R^{-1} = H(−q) relates the half-BZ loops by inverse/conjugation, and the Discussion concedes that without the constraint the same k-phase can be decomposed as b12^4 b23^{-2}. The predicted half-BZ word is therefore the imposed holonomy relabeled. (3) The half-BZ geometric phases offered as the 'rigorous foundation' are defined with indices chosen 'according to the braid sequence,' presupposing the word they are said to justify. The acoustic experiment demonstrates the profile exchanges for models engineered so the half-BZ holonomy equals the imposed square root, and the displayed eigenfunctions come from least-squares fitting (Methods), so it inherits the assumption rather than independently confirming robustness. Central predictions reduce by construction; score 6.
Assumptions & free parameters
free parameters (2)
- per-phase tight-binding parameters (omega1, omega2, omega3, v1, v2, v3, w1, w2, u) =
listed in SI Table S3 for theory and Table S4 for COMSOL
- experimental fit parameters (omega_i, t_i, gamma, G0) =
ranges given in SI Tables S5-S7
assumptions (6)
- domain assumption Three-band PT-symmetric gapped Hamiltonians H = R E R^T have topological space X = O(3)/(Z2 x Z2 x Z2) with pi1(X) = Q8, and NABPs map to quaternion invariants.
- domain assumption Mirror symmetry at q = 0 and pi ensures half-BZ NABPs are equal and the relative homotopy group pi1(X, A) = Z2 quantizes half-BZ geometric phases.
- standard math All homotopy loops on SO(3) can be decomposed into 48 braid sequences generated by b12 and b23 with relations b12 b23 b12 = b23 b12 b23 and b12^8 = b23^8 = 1.
- ad hoc to paper The 3x3 matrices b12 = exp(pi/2 L_x) and b23 = exp(pi/2 L_z) represent braid generators acting on the Bloch eigenmode frame.
- domain assumption PT-symmetry imposes that the rotation angle over the full BZ takes only the quantized values 0, +/-pi, and +/-2pi.
- domain assumption The acoustic cavity system with two coupling tubes realizes the tight-binding Hamiltonian Eq. (6) with tunable hopping signs.
invented entities (1)
-
Half-BZ quantized geometric phases Theta_n^+ and Theta_n^-
Cite this review
Pith. "Pith review of Topological Braiding of Bloch Eigenmodes Protected by Non-Abelian Quaternion Invariants." pith.science (2026). https://pith.science/paper/JU5H2XAL
@misc{pith2026250701809,
author = {Pith},
title = {Pith review of: Topological Braiding of Bloch Eigenmodes Protected by Non-Abelian Quaternion Invariants},
year = {2026},
howpublished = {\url{https://pith.science/paper/JU5H2XAL}},
note = {Machine review of arXiv:2507.01809}
}
read the original abstract
Braiding has attracted significant attention in physics because of its important role in describing the fundamental exchange of particles. Infusing the braiding with topological protection will make it robust against imperfections and perturbations, but such topological braiding is believed to be possible only in interacting quantum systems, e.g., topological superconductors. Here, we propose and demonstrate a new strategy of topological braiding that emerges from non-Abelian topological insulators, a class of recently discovered multi-band topological phase. We unveil a mathematical connection between braiding and non-Abelian quaternion invariants, by which Bloch eigenmodes under parallel transport produce braid sequences protected by the non-Abelian band topology. The braiding is also associated with geometric phases quantized over half the Brillouin zone. This new type of non-Abelian topological braiding is experimentally realized in acoustic systems with periodic synthetic dimensions. The results show that the principle discovered here is a new strategy towards topological braiding and can be extended for other types of classical waves and non-interacting quantum systems.
Reference graph
Works this paper leans on
-
[1]
Wu, Q., Soluyanov, A. A. & Bzdušek, T. Non -Abelian band topology in noninteracting metals. Science 365, 1273–1277 (2019)
work page 2019
-
[2]
Sun, X.-Q., Zhang, S. -C. & Bzdušek, T. Conversion Rules for Weyl Points and Nodal Lines in Topological Media. Phys. Rev. Lett. 121, 106402 (2018)
work page 2018
-
[3]
The Fundamental Group of $SO(n)$ Via Quotients of Braid Groups
Hajdini, I. & Stoytchev, O. The Fundamental Group of $SO(n)$ Via Quotients of Braid Groups. Preprint at https://doi.org/10.48550/arXiv.1607.05876 (2016)
work page Pith review arXiv doi:10.48550/arxiv.1607.05876 2016
-
[4]
Jiang, T. et al. Four-band non-Abelian topological insulator and its experimenta l realization. Nat Commun 12, 6471 (2021)
work page 2021
-
[5]
Chen, Z.-G., Wang, L., Zhang, G. & Ma, G. Chiral Symmetry Breaking of Tight -Binding Models in Coupled Acoustic-Cavity Systems. Phys. Rev. Applied 14, 024023 (2020)
work page 2020
Reviewed August 6, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.