Fermionic-Membrane Toric Code

• Fermionic-Membrane Toric Code is a topological model in d≥7 that generalizes fermionic excitations to 3-brane membranes using a stabilizer Hamiltonian formulation.
• It employs flux-brane, electric-brane, and square stabilizers to enforce commuting-projector properties and a unique Z₂ fusion rule for membrane excitations.
• A specialized membrane-flip protocol detects fermionic self‐statistics through a Dijkgraaf–Witten twist, linking high-dimensional topology with quantum error correction.

The fermionic-membrane toric code is a family of exactly-solvable commuting-projector lattice models in $d \geq 7$ spatial dimensions, whose key feature is the realization of topological orders with extended (3-brane) excitations obeying nontrivial fermionic self-statistics. It generalizes the notion of fermionic quasi-particles ($p = 0$) in (2+1)D and fermionic loop excitations ($p = 1$) in (4+1)D to $p = 3$-dimensional membrane excitations, with explicit lattice realizations and an underlying cohomological classification. The Hamiltonian is formulated in the Pauli stabilizer formalism and encodes the unique nontrivial $\mathbb{Z}_2$-valued cohomology class in $H^{d+1}(B^3\mathbb{Z}_2, U(1))$, corresponding to a Dijkgraaf–Witten twist by $\frac12\mathrm{Sq}^4\mathrm{Sq}^1\,b_3$ and supporting a $\mathbb{Z}_2$ fusion rule and fermionic membrane statistics under the membrane-flip detection process (Feng et al., 31 Dec 2025).

1. Lattice Formulation and Stabilizer Structure

The fermionic-membrane toric code is defined on a $d$-dimensional cellulation (such as a simplicial or hypercubic lattice) of a spatial manifold $M_d$ for $d\geq7$. The local Hilbert space consists of qubits on each $(d-3)$-cell $c_{d-3}$, corresponding to the placement of a 3-form $\mathbb{Z}_2$ gauge field $B_{d-3}$ in cohomological terms. The Hamiltonian is an exactly solvable commuting-projector model with three types of stabilizers:

• Flux-brane stabilizers $G_{c_{d-4}}$ associated to $(d-4)$-cells:

$$G_{c_{d-4}} = X_{\delta c_{d-4}} \cdot \prod_{c'_{d-3}} Z_{c'_{d-3}}^{\int \delta c_{d-4} \cup_{d-6} c'_{d-3}}$$

where $X_{\delta c}$ applies $X$ to each $(d-3)$-cell in the coboundary $\delta c$, and the higher cup product $\cup_{d-6}$ ensures commutation structure required for fermionic statistics.

• Electric-brane stabilizers $B_{c_{d-2}}$ for each $(d-2)$-cell:

$B_{c_{d-2}} = Z_{\partial c_{d-2}}$

acting as products of $Z$ over the boundary $(d-3)$-cells.

• Square stabilizers $C_{c_{d-3}}$ enforcing $\mathbb{Z}_2$ structure on each $(d-3)$-cell:

$C_{c_{d-3}} = X_{c_{d-3}}^2$

These arise from the reduction of a parent $\mathbb{Z}_4$ structure, such that all local degrees of freedom are $\mathbb{Z}_2$ qubits.

All stabilizer terms commute due to the presence of the higher-cup product twist, and the full Hamiltonian sums these projector terms.

2. Membrane Excitations and Fusion Rules

Excitations of the code are supported on closed 3-dimensional membranes. A single membrane excitation is created by a local operator on a single $(d-3)$-cell $c$:

$$U(c) = X_c \prod_{c'} Z_{c'}^{\int c' \cup_{d-6} c}$$

This operator commutes with $G$ and $C$ except at the $(d-2)$-cells in the boundary of $c$, flipping the eigenvalues of the corresponding $B$ stabilizers, and hence creates a membrane defect in the dual lattice.

The fusion of two identical membrane excitations on the same support yields:

$(U(c))^2 = C_c \cdot (\text{products of }G)$

which acts as the identity in the ground-state sector, so membrane excitations obey a $\mathbb{Z}_2$ fusion rule.

3. Fermionic Statistics of 3-Brane Membranes

A defining aspect of this topological order is the fermionic self-statistics of the 3-brane excitations. This is captured by a closed membrane-flip process, an extension of the Kitaev 24-step loop-flip process for loops/strings, here realized for membranes on a small triangulated 3-sphere in the dual lattice. The detected phase is $\mu = -1$:

$$\mu = \prod’_{\{\sigma_4, \tau_4\}} [U_{(0\sigma_4)^*}, U_{(0\tau_4)^*}]^2$$

where the product is over unordered partitions of tetrahedra on the 3-sphere boundary, $U_{(0\sigma_4)^*}$ is a membrane operator dual to the 4-simplex labeled by $\{0\}\cup\sigma_4$, and $[\cdot,\cdot]$ indicates a group commutator. Cohomological evaluation using higher-cup identities shows that the net phase upon completing the membrane-flip is $-1$. This is the hallmark of fermionic 3-brane statistics (Feng et al., 31 Dec 2025).

4. Cohomological and TQFT Classification

The fermionic-membrane toric code realizes a 3-form $\mathbb{Z}_2$ gauge theory in $(d+1)$ spacetime dimensions with a nontrivial Dijkgraaf–Witten cocycle twist. The relevant cohomology class is

$$\frac12 \mathrm{Sq}^4 \mathrm{Sq}^1 b_3 \in H^{d+1}(B^3\mathbb{Z}_2, U(1)) \cong \mathbb{Z}_2$$

on oriented manifolds equivalently expressed as $\pi w_5 \cup b_3$, where $w_5$ is the fifth Stiefel–Whitney class and $b_3$ is the 3-form gauge field. This twist enforces that the minimal membrane excitation carries fermionic statistics.

In the Dijkgraaf–Witten framework, this code is the unique nontrivial 3-form $\mathbb{Z}_2$ gauge theory in $(d+1)$ dimensions admitting such fermionic 3-brane excitations. As such, it sits naturally within the periodic table of extended fermionic statistics: with $p=0$ for fermionic particles (twist $\mathrm{Sq}^2$), $p=1$ for fermionic loops ($\mathrm{Sq}^2 \mathrm{Sq}^1$), $p=3$ for membranes, and analogously for higher $p$ (Feng et al., 31 Dec 2025).

5. Generalizations and Relation to Other Fermionic Codes

The construction extends to arbitrarily high spatial dimension $d \geq 7$ using the same principles: qubits placed on $(d-3)$-cells, electric and flux stabilizers on boundaries and coboundaries, and a higher-cup twist for fermionic statistics. The resulting codes form a hierarchy where "fermionic-membrane" code refers to the member with 3-brane excitations. In $8+1$ dimensions, condensing e$^2$m$^2$ volumes in a $\mathbb{Z}_4$ volume-only toric code leads to a fermionic-volume toric code, whose elementary excitation is a fermionic 5-volume brane, detected by a 48-step volume-flip process.

This classification fits into the conjectured periodic table of fermionic extended excitations: | Excitation dimension $p$ | Detecting cohomology twist | Model name | |-------------------------|----------------------------|----------------------------| | 0 (particle) | $\mathrm{Sq}^2$ | Fermionic toric code | | 1 (loop) | $\mathrm{Sq}^2 \mathrm{Sq}^1$ | Fermionic-loop toric code | | 3 (membrane) | $\mathrm{Sq}^4 \mathrm{Sq}^1$ | Fermionic-membrane toric code | | 5 (volume) | $\mathrm{Sq}^6 \mathrm{Sq}^1$ | Fermionic-volume toric code |

This suggests that for each odd $p$, one can realize explicit stabilizer Hamiltonians whose minimal $p$-brane excitation exhibits fermionic self-statistics detected via a generalized flip process.

6. Detection Protocols and Physical Realization

The fermionic nature of the membrane excitations is revealed by a specific membrane-flip protocol, realized via a product of commutators of membrane operators associated with the boundaries of higher-dimensional simplices. Such a protocol ensures that the resulting phase under a full flip is $-1$. The commuting-projector lattice construction is directly compatible with Pauli stabilizer codes, and all local projectors can be implemented using tensor-network representations advantageous for studying topological orders and quantum error-correcting codes in high dimensions.

A plausible implication is that, given the unique cohomological anomaly, these codes can only be realized intrinsically in dimensions $d \geq 7$, or as boundary theories of suitable SPT bulk phases in $d+1$ dimensions.

7. Context and Classification in Topological Phases

The fermionic-membrane toric code, and its relatives, contribute to the systematic lattice realization and diagnosis of higher-form topological orders with nontrivial generalized statistics. Cohomological twists classified by higher iterates of Steenrod squares (e.g., $\mathrm{Sq}^4 \mathrm{Sq}^1$) define a full set of such phases, distinct from ordinary bosonic higher-form gauge theories. These models provide explicit instances crucial for analyzing anomalies, boundary-bulk correspondence, and the algebraic theory of braided fusion $n$-categories (Feng et al., 31 Dec 2025).

The approach unifies and extends previous works on fermionic loop statistics in four and higher dimensions, as in the Walker–Wang models of fermionic-loop toric codes and the associated gravitational/cobordism anomalies in $(4+1)$D and beyond (Fidkowski et al., 2021).