Spin Bott Index in Quantum Spin Hall Systems

• Spin Bott index is a real-space topological invariant that decomposes the occupied subspace into effective spin sectors to replace Brillouin-zone-based Z₂ diagnostics.
• It applies to finite, disordered, and non-periodic systems—including quasicrystals and amorphous materials—where traditional momentum-space methods fail.
• The method utilizes projected position operators and unitary stabilization via singular value decomposition to robustly quantify topological phase transitions.

The spin Bott index is a real-space topological invariant used primarily to identify time-reversal-invariant quantum spin Hall phases in finite, disordered, amorphous, quasicrystalline, and other non-periodic systems. It is constructed by decomposing the occupied subspace into two effective spin sectors and taking half the difference of their Bott indices, thereby replacing Brillouin-zone-based $\mathbb Z_2$ diagnostics when translational symmetry is absent; in a distinct literature on real Bott towers, the same phrase is not standard, but the explicit Bott-matrix formula for the spin obstruction $w_2$ plays the closest analogous role (Huang et al., 2018, Huang et al., 2018, DSouza et al., 2016).

1. Conceptual scope and terminological usage

In condensed-matter theory, the spin Bott index was introduced because the standard $\mathbb Z_2$ invariant for quantum spin Hall insulators is ordinarily defined from Bloch bands, Berry-curvature constructions, hybrid Wannier charge centers, parity eigenvalues, or Pfaffians, all of which rely on translational symmetry and a Brillouin zone. In quasicrystals and other non-periodic systems, Bloch momentum $k$ is not a good quantum number, Brillouin-zone formulas fail, and the relevant electronic states may be critical rather than Bloch-like. The spin Bott index addresses this by working directly in real space on finite samples (Huang et al., 2018, Huang et al., 2018).

The construction generalizes the ordinary Bott index, which had already been used for Chern topology in nonperiodic systems. The ordinary Bott index measures the noncommutativity of projected position operators, while the spin Bott index introduces a spin-sector decomposition of the occupied subspace so that opposite spin sectors can carry opposite topological winding, as expected in a quantum spin Hall phase. In that sense it is a real-space analogue of spin Chern or $\mathbb Z_2$ diagnostics rather than a net Chern marker (Huang et al., 2018).

A separate mathematical usage arises in the topology of real Bott towers and generalized real Bott manifolds. The relevant papers do not define a named invariant called “spin Bott index”; instead, they derive necessary-and-sufficient spin criteria from explicit formulas for the second Stiefel–Whitney class $w_2$, expressed in terms of the Bott matrix or associated column vectors. A plausible implication is that the phrase can be used informally for these parity obstructions, but this is an interpretive rather than standard terminological usage (DSouza et al., 2016, İlhan et al., 2021).

2. Real-space construction for quantum spin Hall phases

The starting point is the projector onto the occupied subspace,

$$P=\sum_i^{N_{\mathrm{occ}}} |\psi_i\rangle\langle\psi_i|.$$

From it one defines the projected spin operator

$$P_z=P\hat s_z P,\qquad \hat s_z=\frac{\hbar}{2}\sigma_z.$$

If $[H,\hat s_z]=0$, the spectrum of $P_z$ consists of the two values $w_2$0. When spin is not strictly conserved, the eigenvalues spread, but the construction remains valid provided the spectrum still splits into two groups separated by zero; this separation is the spin gap of $w_2$1 (Huang et al., 2018, Huang et al., 2018).

Let

$w_2$2

The positive- and negative-eigenvalue subspaces define the effective spin-sector projectors

$w_2$3

For each sector one then forms projected position operators using rescaled coordinates $w_2$4,

$w_2$5

The complementary terms $w_2$6 are included to make the matrices closer to unitary and improve numerical stability (Huang et al., 2018).

Because the finite-sample projected operators are not exactly unitary, the standard implementation uses a singular value decomposition. For any projected operator $w_2$7,

$w_2$8

Applying this to $w_2$9 and $\mathbb Z_2$0 gives stabilized unitary matrices $\mathbb Z_2$1 and $\mathbb Z_2$2. The sector Bott indices are then

$\mathbb Z_2$3

and the spin Bott index is

$\mathbb Z_2$4

In the condensed-matter usage, $\mathbb Z_2$5 indicates a quantum spin Hall phase and $\mathbb Z_2$6 a trivial insulator. The construction requires both a bulk energy gap and an open spin gap of $\mathbb Z_2$7; if spin mixing closes the spin gap, the decomposition into $\mathbb Z_2$8 and $\mathbb Z_2$9 ceases to be well defined (Huang et al., 2018, Huang et al., 2018).

3. Bott-index algebra, logarithmic formulation, and stability

The spin Bott index inherits its basic algebraic structure from the ordinary Bott index. For a pair of unitary operators $k$0 satisfying the spectral condition $k$1, equivalently $k$2, the Bott index is defined by

$k$3

with $k$4 taken as the principal logarithm. In the infinite-dimensional formulation, additional trace-class conditions such as $k$5 and $k$6 ensure that the logarithm is trace class and that the index is well defined (Toniolo, 2021).

The index is integer-valued. Writing the eigenvalues of the commutator unitary as $k$7, $k$8, one obtains

$k$9

hence

$\mathbb Z_2$0

For finite periodic lattices, the Bott index can be constructed from projected coordinate unitaries, and rigorous results identify it with the Chern number: $\mathbb Z_2$1 This relation explains why a spin-resolved difference of Bott indices can reproduce the quantum spin Hall classification in crystalline and disordered settings (Toniolo, 2017, Toniolo, 2021).

Homotopy invariance is a central structural property. Under continuous unitary deformations that preserve the spectral admissibility condition, the Bott index remains constant. Time-dependent analyses of projected Bott invariants on large finite tori show that, for local bounded Hamiltonians that are initially gapped, the Bott index of the time-evolved Fermi projection is constant in the thermodynamic limit, and any change at finite size is a finite-size effect whose characteristic time diverges with system size. This suggests that spin-resolved Bott constructions inherit stability under local driving so long as the occupied projector remains well defined and the spin-gap condition persists (Toniolo, 2017).

4. Diagnostic content, model systems, and observable signatures

The spin Bott index was benchmarked on the Kane–Mele model, where it agrees with the standard $\mathbb Z_2$2 invariant from hybrid Wannier charge center flow. Reported representative values include $\mathbb Z_2$3 in a quantum spin Hall phase and $\mathbb Z_2$4 in a trivial insulator. The same framework remains applicable when Rashba coupling breaks exact $\mathbb Z_2$5 conservation, provided the spin gap of $\mathbb Z_2$6 stays open (Huang et al., 2018).

In the disorder Kane–Mele model,

$\mathbb Z_2$7

the spin Bott index captures a disorder-driven transition consistent with the topological Anderson insulator scenario and previous conductance results. In the reported numerics, the transition occurs around $\mathbb Z_2$8. This establishes the invariant as a disorder-compatible real-space diagnostic rather than a purely crystalline one (Huang et al., 2018).

The Penrose-type quasicrystal lattice provides the canonical non-periodic example. There the absence of a periodic unit cell and Brillouin zone makes conventional band-topology tools unavailable, but direct real-space computation yields $\mathbb Z_2$9 in one study and $w_2$0 as the hallmark value in another. The corresponding phase is supported by a bulk gap under periodic boundary conditions, mid-gap boundary states under open boundary conditions, edge localization of the in-gap wavefunctions, and a quantized two-terminal conductance plateau

$w_2$1

The edge states are reported to be robust against boundary-shape changes, which is particularly significant for quasicrystalline geometry (Huang et al., 2018, Huang et al., 2018).

A parameter-space phase diagram was constructed in terms of the on-site $w_2$2-$w_2$3 offset $w_2$4, the spin-orbit coupling strength $w_2$5, and the bonding-length scale $w_2$6. The three identified regimes are normal insulator, quantum spin Hall insulator with $w_2$7, and weak metal. In the $w_2$8-dependent analysis, the spin Bott index switches to $w_2$9 at a critical value around

$$P=\sum_i^{N_{\mathrm{occ}}} |\psi_i\rangle\langle\psi_i|.$$0

after which a later gap closing leads to the weak-metal regime. The reported interpretation is that increasing hybridization drives an $$P=\sum_i^{N_{\mathrm{occ}}} |\psi_i\rangle\langle\psi_i|.$$1-$$P=\sum_i^{N_{\mathrm{occ}}} |\psi_i\rangle\langle\psi_i|.$$2 band inversion, spin-orbit coupling reopens the gap into a topological phase, and sufficiently strong hopping eventually overwhelms that gap (Huang et al., 2018).

5. Higher-order superconductors and related real-space Bott developments

The Bott-index framework has been extended from quantum spin Hall insulators to time-reversal-invariant higher-order topological superconductors hosting Majorana Kramers pairs. In that setting an ordinary chiral Bott index vanishes by time-reversal symmetry, because the relevant eigenphases occur in $$P=\sum_i^{N_{\mathrm{occ}}} |\psi_i\rangle\langle\psi_i|.$$3 pairs. The remedy is a spin-resolved Bott construction that separates the two Kramers-related sectors and thereby detects corner topology in real space under open boundary conditions and for arbitrary sample shapes (Luo et al., 2024).

For conserved $$P=\sum_i^{N_{\mathrm{occ}}} |\psi_i\rangle\langle\psi_i|.$$4, the paper introduces

$$P=\sum_i^{N_{\mathrm{occ}}} |\psi_i\rangle\langle\psi_i|.$$5

and also an equivalent spin-twisted form

$$P=\sum_i^{N_{\mathrm{occ}}} |\psi_i\rangle\langle\psi_i|.$$6

When spin is conserved, the invariant decomposes as

$$P=\sum_i^{N_{\mathrm{occ}}} |\psi_i\rangle\langle\psi_i|.$$7

When $$P=\sum_i^{N_{\mathrm{occ}}} |\psi_i\rangle\langle\psi_i|.$$8 is broken, the formulation based on the spin-projected operator

$$P=\sum_i^{N_{\mathrm{occ}}} |\psi_i\rangle\langle\psi_i|.$$9

remains valid as long as the spin-projection gap of $$P_z=P\hat s_z P,\qquad \hat s_z=\frac{\hbar}{2}\sigma_z.$$0 stays open. The resulting series of spin Bott indices does not merely distinguish topological from trivial phases; it encodes which corners host Majorana Kramers pairs and tracks their redistribution under boundary cleavage (Luo et al., 2024).

The same real-space Bott philosophy has also been broadened in directions adjacent to, but distinct from, the spin Bott index. The Bott metric uses the same plaquette operator as the Bott index and extracts the real part rather than the imaginary part of the trace-log,

$$P_z=P\hat s_z P,\qquad \hat s_z=\frac{\hbar}{2}\sigma_z.$$1

so that topology is encoded in the phase of the loop while quantum geometry is encoded in its contraction. That framework is compatible with spinful Hilbert spaces because internal labels may include spin, although no separate spin Bott metric is derived there (Chatterjee et al., 6 Apr 2026).

A further neighboring development is the extension of Bott-index methods to topological pumping in quasiperiodic systems. That work does not define a spin Bott index explicitly, but it states that the method can be used to study topological spin pumping in spinful electron systems or in topological magnets. This suggests a natural route from static spin Bott diagnostics to time-dependent spin-resolved transport settings (Yoshii et al., 2021).

6. Spin obstruction in real and generalized Bott manifolds

In the topology of real Bott towers, the phrase “spin Bott index” is best understood cautiously. The relevant invariant is not named that way in the literature; rather, the spin obstruction is exactly the second Stiefel–Whitney class $$P_z=P\hat s_z P,\qquad \hat s_z=\frac{\hbar}{2}\sigma_z.$$2, computed explicitly from the Bott matrix. A real Bott tower $$P_z=P\hat s_z P,\qquad \hat s_z=\frac{\hbar}{2}\sigma_z.$$3 is encoded by an upper triangular Bott matrix

$$P_z=P\hat s_z P,\qquad \hat s_z=\frac{\hbar}{2}\sigma_z.$$4

and the cohomology ring is

$$P_z=P\hat s_z P,\qquad \hat s_z=\frac{\hbar}{2}\sigma_z.$$5

The total Stiefel–Whitney class is

$$P_z=P\hat s_z P,\qquad \hat s_z=\frac{\hbar}{2}\sigma_z.$$6

and the orientability criterion is

$$P_z=P\hat s_z P,\qquad \hat s_z=\frac{\hbar}{2}\sigma_z.$$7

Under this orientability hypothesis, Theorem 4.10 gives the spin condition: $$P_z=P\hat s_z P,\qquad \hat s_z=\frac{\hbar}{2}\sigma_z.$$8 for all $$P_z=P\hat s_z P,\qquad \hat s_z=\frac{\hbar}{2}\sigma_z.$$9. Equivalently, $[H,\hat s_z]=0$0 is spin if and only if $[H,\hat s_z]=0$1. The paper also gives the explicit expansion

$[H,\hat s_z]=0$2

This is the precise mod-$[H,\hat s_z]=0$3 parity obstruction that one could reasonably interpret as an index-like spin criterion, although the paper does not assign it the name “spin Bott index” (DSouza et al., 2016).

The same theme persists for generalized real Bott manifolds. There the spin structure is characterized in terms of column vectors of the associated matrix $[H,\hat s_z]=0$4, and the principal theorem gives necessary-and-sufficient parity conditions involving self-dot products and pairwise dot products of the columns. In the special case $[H,\hat s_z]=0$5, the criterion simplifies to

$[H,\hat s_z]=0$6

The paper also reformulates the obstruction in terms of acyclic $[H,\hat s_z]=0$7-weighted digraphs, where indegrees, common in-neighbor sums $[H,\hat s_z]=0$8, and edge-weight self-dot-products encode the same information. Here again, the operative object is not a named index but a combinatorial parity obstruction equivalent to $[H,\hat s_z]=0$9 under orientability (İlhan et al., 2021).

Taken together, these two literatures use “spin Bott” language in fundamentally different ways. In non-periodic quantum systems, the spin Bott index is an established real-space invariant built from projected spin sectors and logarithms of commutator-like products. In Bott-manifold topology, the closest precise analogue is the explicit Bott-matrix or column-vector formula for the spin obstruction $P_z$0, which detects when the manifold admits a spin structure (Huang et al., 2018, DSouza et al., 2016).