Fusion Category Symmetries

• Fusion category symmetries are rigid semisimple monoidal categories with finite simple objects that encode both invertible and non-invertible topological defect lines.
• They underpin state-sum TQFTs and commuting-projector lattice Hamiltonians by mapping Hopf algebra data to concrete gapped phase realizations and boundary phenomena.
• The framework extends to classify both anomalous and non-anomalous symmetries, enabling explicit constructions of SPT phases and duality-induced edge modes.

Fusion category symmetries generalize finite group symmetries by encoding all finite–type topological defect lines—both invertible and non-invertible—via the structure of a rigid semisimple monoidal category with finitely many simple objects and fusion/rule data (fusion coefficients and associators). In 1+1 dimensions, these symmetries are realized in QFTs and lattice models as non-invertible symmetries, whose action on boundary conditions/gapped phases is controlled by the theory of module categories. The classification of gapped phases, construction of TQFTs and lattice Hamiltonians, and the anomaly structure of these symmetries now admit a complete description in terms of Hopf (and weak Hopf) algebra data, module categories, and categorical Morita theory. The framework extends to the description and realization of anomalous and non-anomalous fusion category symmetries, including SPT phases and their edge phenomena.

1. Algebraic Framework of Fusion Category Symmetries

A fusion category $\mathcal{C}$ over a field (characteristic zero) is defined as a rigid semisimple monoidal category with finitely many simple objects, simple tensor unit, and finite-dimensional Hom spaces. The data includes:

• Simple objects $x \in \mathrm{Ob}(\mathcal{C})$ representing topological defect lines.
• Fusion rules:

$$x \otimes y \cong \bigoplus_{z \in \mathrm{Irr}(\mathcal{C})} N_{xy}^z\, z, \qquad N_{xy}^z \in \mathbb{Z}_{\geq 0}$$

• Associators $$\alpha_{x,y,z} : (x \otimes y) \otimes z \to x \otimes (y \otimes z)$$, whose matrix elements in chosen bases are the $F$-symbols:

$$F^{abc}_d : \bigoplus_e \mathrm{Hom}(a \otimes b, e) \otimes \mathrm{Hom}(e \otimes c, d) \rightarrow \bigoplus_f \mathrm{Hom}(b \otimes c, f) \otimes \mathrm{Hom}(a \otimes f, d)$$

satisfying the pentagon relation, enforcing associativity up to coherent isomorphism.

Fusion category symmetries generalize group symmetries by allowing non-invertible objects and nontrivial self-fusion ($N_{xx}^y>0$, $x \neq y$). The trivial group case arises as $\mathcal{C} = \mathrm{Vec}_G$, while Tambara–Yamagami, Kac–Paljutkin, and other fusion categories arise when $N_{xy}^z$ and $x \in \mathrm{Ob}(\mathcal{C})$0-symbols implement duality or more exotic non-invertible symmetries (Inamura, 2021).

2. State-Sum TQFTs and Lattice Realizations

Any non-anomalous fusion category symmetry in 1+1d is realized via state-sum TQFT and commuting-projector lattice Hamiltonians. The central construction is as follows:

• Given a finite-dimensional semisimple Hopf algebra $x \in \mathrm{Ob}(\mathcal{C})$1, every non-anomalous $x \in \mathrm{Ob}(\mathcal{C})$2 is equivalent to $x \in \mathrm{Ob}(\mathcal{C})$3.
• Indecomposable gapped phases with $x \in \mathrm{Ob}(\mathcal{C})$4 symmetry correspond to $x \in \mathrm{Ob}(\mathcal{C})$5-simple left $x \in \mathrm{Ob}(\mathcal{C})$6-comodule algebras $x \in \mathrm{Ob}(\mathcal{C})$7:

$x \in \mathrm{Ob}(\mathcal{C})$8

• The state-sum TQFT assigns to each triangulated surface $x \in \mathrm{Ob}(\mathcal{C})$9 a vector space built from $$x \otimes y \cong \bigoplus_{z \in \mathrm{Irr}(\mathcal{C})} N_{xy}^z\, z, \qquad N_{xy}^z \in \mathbb{Z}_{\geq 0}$$0 and its bimodules. Partition functions are expressed as weighted sums over edge/face labelings, implementing the categorical data of $$x \otimes y \cong \bigoplus_{z \in \mathrm{Irr}(\mathcal{C})} N_{xy}^z\, z, \qquad N_{xy}^z \in \mathbb{Z}_{\geq 0}$$1.
• The corresponding 1d lattice Hamiltonian is:

$$x \otimes y \cong \bigoplus_{z \in \mathrm{Irr}(\mathcal{C})} N_{xy}^z\, z, \qquad N_{xy}^z \in \mathbb{Z}_{\geq 0}$$2

where $$x \otimes y \cong \bigoplus_{z \in \mathrm{Irr}(\mathcal{C})} N_{xy}^z\, z, \qquad N_{xy}^z \in \mathbb{Z}_{\geq 0}$$3 are commuting projectors due to Frobenius separability and symmetry.

The global symmetry action on the lattice is implemented by the $$x \otimes y \cong \bigoplus_{z \in \mathrm{Irr}(\mathcal{C})} N_{xy}^z\, z, \qquad N_{xy}^z \in \mathbb{Z}_{\geq 0}$$4-comodule structure $$x \otimes y \cong \bigoplus_{z \in \mathrm{Irr}(\mathcal{C})} N_{xy}^z\, z, \qquad N_{xy}^z \in \mathbb{Z}_{\geq 0}$$5, giving explicit operators $$x \otimes y \cong \bigoplus_{z \in \mathrm{Irr}(\mathcal{C})} N_{xy}^z\, z, \qquad N_{xy}^z \in \mathbb{Z}_{\geq 0}$$6 whose commutation with $$x \otimes y \cong \bigoplus_{z \in \mathrm{Irr}(\mathcal{C})} N_{xy}^z\, z, \qquad N_{xy}^z \in \mathbb{Z}_{\geq 0}$$7 follows from Hopf algebra identities (Inamura, 2021).

3. Classification of Gapped Phases and Module Categories

Indecomposable module categories over $$x \otimes y \cong \bigoplus_{z \in \mathrm{Irr}(\mathcal{C})} N_{xy}^z\, z, \qquad N_{xy}^z \in \mathbb{Z}_{\geq 0}$$8 classify symmetry-respecting gapped phases; boundary conditions correspond to simple $$x \otimes y \cong \bigoplus_{z \in \mathrm{Irr}(\mathcal{C})} N_{xy}^z\, z, \qquad N_{xy}^z \in \mathbb{Z}_{\geq 0}$$9-modules $$\alpha_{x,y,z} : (x \otimes y) \otimes z \to x \otimes (y \otimes z)$$0: $$\alpha_{x,y,z} : (x \otimes y) \otimes z \to x \otimes (y \otimes z)$$1 where $$\alpha_{x,y,z} : (x \otimes y) \otimes z \to x \otimes (y \otimes z)$$2. The action of symmetry defects (simple bimodules $$\alpha_{x,y,z} : (x \otimes y) \otimes z \to x \otimes (y \otimes z)$$3) on boundaries is realized by tensor product over $$\alpha_{x,y,z} : (x \otimes y) \otimes z \to x \otimes (y \otimes z)$$4, inducing the module category structure: $$\alpha_{x,y,z} : (x \otimes y) \otimes z \to x \otimes (y \otimes z)$$5 This implements non-negative integer matrix representations (NIM-reps) of $$\alpha_{x,y,z} : (x \otimes y) \otimes z \to x \otimes (y \otimes z)$$6 on the vacua of the boundary system, providing a correspondence with the Ostrik classification of module categories and SPT phases (Inamura, 2021).

4. Anomaly Structure and Weak Hopf Generalization

A fusion category symmetry $$\alpha_{x,y,z} : (x \otimes y) \otimes z \to x \otimes (y \otimes z)$$7 is non-anomalous if and only if it admits a fiber functor (i.e., is Tannakian, equivalent to a representation category of a semisimple Hopf algebra). In the presence of anomaly, the appropriate categorical data generalizes to weak (pseudo-unitary connected) semisimple Hopf algebras: $$\alpha_{x,y,z} : (x \otimes y) \otimes z \to x \otimes (y \otimes z)$$8 State-sum constructions, commuting-projector Hamiltonians, and module-category boundary descriptions all extend to this context, with minimal modifications for the nontrivial comultiplication, antipode, and partial associativity of $$\alpha_{x,y,z} : (x \otimes y) \otimes z \to x \otimes (y \otimes z)$$9 (Inamura, 2021).

Obstructions to anomaly-freeness are realized concretely as the nonexistence of "magnetic" Lagrangian algebras in the Drinfeld center $F$0 or, equivalently, the failure to construct a single-vacuum module category (Zhang et al., 2023).

5. Tetrahedral Symmetry and Structural Constraints

The $F$1-symbols ($F$2-symbols) in any fusion category satisfy full tetrahedral $F$3 symmetries under minimal hypotheses: no pivotal or unitarity structure is required, as the tetrahedral group acts on the set of associator transformations by permuting the tensor factors and dualities. Explicitly,

$F$4

and similar relations for all permutations, as established in (Fuchs et al., 2021). This symmetry is critical in both categorical computations and in the geometric interpretation of state-sum (TQFT) amplitudes, where the $F$5-symbol corresponds to tetrahedron weights that are invariant under apex relabelling.

6. Examples: Group and Duality Symmetries, SPT Edge Modes

For $F$6 with $F$7 finite, phases are labelled by subgroups $F$8 and cocycle data $F$9. The algebra $$F^{abc}_d : \bigoplus_e \mathrm{Hom}(a \otimes b, e) \otimes \mathrm{Hom}(e \otimes c, d) \rightarrow \bigoplus_f \mathrm{Hom}(b \otimes c, f) \otimes \mathrm{Hom}(a \otimes f, d)$$0 and its boundary module structure encode SPT and Higgs phases, boundary ground state degeneracies, and edge anomalies (in SPTs, nontrivial edge degeneracy arises from the projective module structure of $$F^{abc}_d : \bigoplus_e \mathrm{Hom}(a \otimes b, e) \otimes \mathrm{Hom}(e \otimes c, d) \rightarrow \bigoplus_f \mathrm{Hom}(b \otimes c, f) \otimes \mathrm{Hom}(a \otimes f, d)$$1 under $$F^{abc}_d : \bigoplus_e \mathrm{Hom}(a \otimes b, e) \otimes \mathrm{Hom}(e \otimes c, d) \rightarrow \bigoplus_f \mathrm{Hom}(b \otimes c, f) \otimes \mathrm{Hom}(a \otimes f, d)$$2). In the case of maximal Higgsing, the dual- or self-dual Ising fixed point can be constructed explicitly.

Kramers–Wannier-like dualities (Tambara–Yamagami categories) admit gapped symmetric phases only when the fiber functor (i.e., bosonization) exists. The obstruction is diagnosed via fusion rules and the existence (or not) of the requisite module category structure (Inamura, 2021, Inamura, 2021, Zhang et al., 2023).

7. Edge Modes and Physical Realizations

On an open lattice with $$F^{abc}_d : \bigoplus_e \mathrm{Hom}(a \otimes b, e) \otimes \mathrm{Hom}(e \otimes c, d) \rightarrow \bigoplus_f \mathrm{Hom}(b \otimes c, f) \otimes \mathrm{Hom}(a \otimes f, d)$$3-localized Hilbert spaces, unique simple $$F^{abc}_d : \bigoplus_e \mathrm{Hom}(a \otimes b, e) \otimes \mathrm{Hom}(e \otimes c, d) \rightarrow \bigoplus_f \mathrm{Hom}(b \otimes c, f) \otimes \mathrm{Hom}(a \otimes f, d)$$4 implies unique ground state, but in the SPT case, the edge hosts a degeneracy equal to $$F^{abc}_d : \bigoplus_e \mathrm{Hom}(a \otimes b, e) \otimes \mathrm{Hom}(e \otimes c, d) \rightarrow \bigoplus_f \mathrm{Hom}(b \otimes c, f) \otimes \mathrm{Hom}(a \otimes f, d)$$5 at each boundary. This degeneracy is "topologically protected," realizing the expected SPT edge physics.

The construction provides an explicit algebraic and lattice realization for any non-anomalous fusion category symmetry, including non-invertible and duality-induced symmetries (Inamura, 2021).

---

References:

• "On lattice models of gapped phases with fusion category symmetries" (Inamura, 2021)
• "Tetrahedral symmetry of 6j-symbols in fusion categories" (Fuchs et al., 2021)
• "Anomalies of (1+1)D categorical symmetries" (Zhang et al., 2023)
• "Topological field theories and symmetry protected topological phases with fusion category symmetries" (Inamura, 2021)