Near-Group Fusion Algebras

• Near-group fusion algebras are fusion rings formed by extending an integral group ring with one non-invertible object, governed by the quadratic relation ρ² = ℓρ + Σg.
• They provide a prototype for fusion categories by revealing structural constraints and categorifiability conditions through controlled Frobenius–Perron dimensions and quadratic equations.
• These algebras underpin diverse applications in modular tensor categories, subfactor theory, and VOA realizations, with key examples like the Tambara–Yamagami case and Izumi-type systems.

Searching arXiv for recent and foundational papers on near-group fusion rings/categories and related VOA examples. Near-group fusion algebras are fusion rings obtained by adjoining a single non-invertible basis element to an integral group ring. In the standard notation $R(G,\ell)$, where $G$ is a finite group and $\ell\in \mathbb{Z}_{\ge 0}$, the distinguished basis is $G\cup\{\rho\}$ and the fusion rules are

$$g h = gh,\qquad g\rho=\rho g=\rho,\qquad \rho^2=\ell\rho+\sum_{g\in G} g.$$

Equivalent notation $K(G,\alpha)$ is also used, with the same ring-theoretic content and with $X^*=X$ for the non-invertible basis element. These rings are the simplest extensions of integral group rings by one non-invertible element, and they occupy a central place in the theory of fusion rings, fusion categories, modular tensor categories, and related VOA and subfactor constructions (Schopieray, 2022, Hannah et al., 2023, Dong et al., 2024).

1. Classical form and basic invariants

A fusion ring is a unital based $\mathbb{Z}_{\ge 0}$-ring of finite rank. Concretely, it is a finitely generated free $\mathbb{Z}$-module with a distinguished basis, nonnegative integer structure constants, and an anti-involution compatible with duality. The integral group ring $\mathbb{Z}G$ of a finite group $G$0 is the basic pointed example: its basis elements are all invertible and have Frobenius–Perron dimension $G$1. A near-group fusion ring is the corresponding “group plus one” extension, with basis $G$2, group multiplication on $G$3, trivial action of $G$4 on $G$5, and the single nontrivial rule $G$6 (Schopieray, 2022).

The Frobenius–Perron dimension of the distinguished non-invertible basis element is the positive root of

$G$7

Thus, if $G$8, then

$G$9

Near-group rings are therefore special cases of two-dimension fusion rings: all basis elements have Frobenius–Perron dimension in $\ell\in \mathbb{Z}_{\ge 0}$0 for some $\ell\in \mathbb{Z}_{\ge 0}$1 (Schopieray, 2022). In the language of fusion categories, a near-group fusion category is a fusion category whose Grothendieck ring is $\ell\in \mathbb{Z}_{\ge 0}$2; its simple objects are the invertible objects indexed by $\ell\in \mathbb{Z}_{\ge 0}$3 together with one distinguished non-invertible simple object.

The case $\ell\in \mathbb{Z}_{\ge 0}$4 is the Tambara–Yamagami case. Here

$\ell\in \mathbb{Z}_{\ge 0}$5

so the non-invertible object has Frobenius–Perron dimension $\ell\in \mathbb{Z}_{\ge 0}$6. Tambara–Yamagami categories are the most tractable non-pointed examples in this class and already exhibit many of the basic structural phenomena of near-group fusion algebras (Hannah et al., 2023).

A useful broader perspective is that near-group rings are exactly the near-integral fusion rings $\ell\in \mathbb{Z}_{\ge 0}$7 for which the maximal-rank integral subring is pointed. In that formulation the basis is $\ell\in \mathbb{Z}_{\ge 0}$8, the integral part is $\ell\in \mathbb{Z}_{\ge 0}$9, and the quadratic relation becomes $G\cup\{\rho\}$0. This recasts near-group rings as a pointed special case of a larger class characterized by a maximal-rank proper fusion subring (Dong et al., 2024).

The commutativity of a near-group fusion algebra is controlled entirely by the group part. The based ring $G\cup\{\rho\}$1 is commutative if and only if $G\cup\{\rho\}$2 is abelian; if $G\cup\{\rho\}$3 is non-abelian, then the group subring $G\cup\{\rho\}$4 is already noncommutative, so the full near-group ring is noncommutative as well (Izumi et al., 2019).

2. Structural constraints and categorifiability

Near-group fusion algebras are easy to write down, but categorifiability is highly constrained. For general two-dimension fusion rings with quadratic relation

$G\cup\{\rho\}$5

one basic dichotomy is that if $G\cup\{\rho\}$6, then $G\cup\{\rho\}$7, while $G\cup\{\rho\}$8 forces $G\cup\{\rho\}$9 irrational. Applied to near-group rings, this yields: if $$g h = gh,\qquad g\rho=\rho g=\rho,\qquad \rho^2=\ell\rho+\sum_{g\in G} g.$$0, then $$g h = gh,\qquad g\rho=\rho g=\rho,\qquad \rho^2=\ell\rho+\sum_{g\in G} g.$$1 is irrational. Moreover, if a categorifiable near-group ring has irrational Frobenius–Perron dimension, then $$g h = gh,\qquad g\rho=\rho g=\rho,\qquad \rho^2=\ell\rho+\sum_{g\in G} g.$$2 (Schopieray, 2022).

This divisibility condition is one of the main structural restrictions. For near-group rings $$g h = gh,\qquad g\rho=\rho g=\rho,\qquad \rho^2=\ell\rho+\sum_{g\in G} g.$$3, irrational categorifiable cases must have

$$g h = gh,\qquad g\rho=\rho g=\rho,\qquad \rho^2=\ell\rho+\sum_{g\in G} g.$$4

The paper on categorification of integral group rings extended by one dimension develops this in the broader setting of two-dimension and two-orbit fusion rings, using pseudounitarity, formal codegrees, induction to the Drinfeld center, Galois action, and number-theoretic bounds on sums of roots of unity (Schopieray, 2022).

A complete classification is obtained there for elementary abelian $$g h = gh,\qquad g\rho=\rho g=\rho,\qquad \rho^2=\ell\rho+\sum_{g\in G} g.$$5-groups. If $$g h = gh,\qquad g\rho=\rho g=\rho,\qquad \rho^2=\ell\rho+\sum_{g\in G} g.$$6, then $$g h = gh,\qquad g\rho=\rho g=\rho,\qquad \rho^2=\ell\rho+\sum_{g\in G} g.$$7 is categorifiable if and only if

$$g h = gh,\qquad g\rho=\rho g=\rho,\qquad \rho^2=\ell\rho+\sum_{g\in G} g.$$8

Equivalently, for $$g h = gh,\qquad g\rho=\rho g=\rho,\qquad \rho^2=\ell\rho+\sum_{g\in G} g.$$9, the only categorifiable near-group ring is the Tambara–Yamagami case $K(G,\alpha)$0. This gives an infinite family in which the near-group ansatz exists ring-theoretically but almost never admits a fusion-category realization (Schopieray, 2022).

An older classification for abelian $K(G,\alpha)$1 in the $K(G,\alpha)$2-setting imposes a different but related dichotomy. For a near-group $K(G,\alpha)$3-category of type $K(G,\alpha)$4 with $K(G,\alpha)$5, either

$K(G,\alpha)$6

Evans and Gannon call such systems of type $K(G,\alpha)$7. In the case $K(G,\alpha)$8, $K(G,\alpha)$9 must be cyclic, and for most $X^*=X$0 there is exactly one $X^*=X$1-category of type $X^*=X$2, namely $X^*=X$3 with $X^*=X$4, except for $X^*=X$5, which have additional exceptional categories (Evans et al., 2012).

The same work studies the second class $X^*=X$6, especially $X^*=X$7, and constructs over 40 new finite depth subfactors with Jones index ranging from around 6.85 to around 14.93. In that regime, the near-group fusion algebra is no longer a sporadic phenomenon but part of a substantial infinite landscape tied to $X^*=X$8 principal graphs and Izumi-type systems (Evans et al., 2012).

3. Centers, modular data, and Frobenius–Schur indicators

The Drinfeld center of a near-group fusion category is a modular tensor category, and its modular data often admit unexpectedly explicit descriptions. For near-group $X^*=X$9-categories of type $\mathbb{Z}_{\ge 0}$0, Evans and Gannon compute the tube algebra and the corresponding doubles, obtaining modular data relevant to conformal field theory. For type $\mathbb{Z}_{\ge 0}$1 with $\mathbb{Z}_{\ge 0}$2 of odd order, they conjecture and verify in many cases that the modular data of the double are controlled by quadratic forms on $\mathbb{Z}_{\ge 0}$3 and on a second group $\mathbb{Z}_{\ge 0}$4 of order $\mathbb{Z}_{\ge 0}$5 (Evans et al., 2012).

A concrete modern example is the near-group fusion category $\mathbb{Z}_{\ge 0}$6 of type $\mathbb{Z}_{\ge 0}$7. Its simple objects are the invertible objects of $\mathbb{Z}_{\ge 0}$8 together with a unique non-invertible object $\mathbb{Z}_{\ge 0}$9, and

$\mathbb{Z}$0

Yu proves that its center is

$\mathbb{Z}$1

and also constructs two non-trivial faithful extensions of $\mathbb{Z}$2 explicitly, whose Drinfeld centers can also be obtained from representation categories of quantum groups at root of unity (Yu, 2023). This is a particularly transparent instance in which a non-pointed near-group category has a center decomposing into a pointed factor and a quantum-group factor.

Higher Frobenius–Schur indicators supply another layer of structure. For near-group categories with $\mathbb{Z}$3, indicators of the non-invertible simple object can be written in terms of $\mathbb{Z}$4, the number of $\mathbb{Z}$5-th roots of the identity in $\mathbb{Z}$6, together with small correction terms coming from the center. Tucker proves that the near-group fusion rings $\mathbb{Z}$7 exhibit Frobenius–Schur indicator rigidity: all categorifications are distinguished by their Frobenius–Schur indicators (Tucker, 2015).

For near-group categories with $\mathbb{Z}$8 and $\mathbb{Z}$9 odd, the same paper derives indicator formulas controlled by quadratic Gauss sums associated to a metric group $\mathbb{Z}G$0 and a second metric group $\mathbb{Z}G$1 of order $\mathbb{Z}G$2. In all known examples with $\mathbb{Z}G$3, the non-invertible object has Frobenius–Schur indicators given by quadratic Gauss sums. This ties the center, modular data, and arithmetic of quadratic forms directly to the intrinsic fusion algebra (Tucker, 2015).

Minimal modular extensions sharpen this picture. Schopieray studies weakly integral braided fusion categories with elementary fusion rules and classifies near-group braided fusion categories satisfying the minimal modular extension conjecture. A braided near-group fusion category possesses a minimal modular extension if and only if it is braided equivalent to one of four types: $\mathbb{Z}G$4 or its reverse, a braided Tambara–Yamagami category $\mathbb{Z}G$5, a symmetrically braided near-group fusion category, or one of seven nonsymmetrically braided near-group fusion categories described explicitly in the paper (Schopieray, 2021). The remaining Tambara–Yamagami braided fusion categories then yield arbitrarily large families of braided fusion categories with identical fusion rules violating the minimal modular extension conjecture.

4. Generalized near-group and near-integral frameworks

The classical near-group condition allows exactly one non-invertible simple object. A systematic generalization keeps the group action but permits multiple non-invertibles, provided the group of invertible objects acts transitively on them. In Dong’s formulation, a generalized near-group fusion category is a fusion category $\mathbb{Z}G$6 such that $\mathbb{Z}G$7 acts transitively on the set of non-isomorphic non-invertible simple objects. If $\mathbb{Z}G$8 is a chosen non-invertible simple object and $\mathbb{Z}G$9 its stabilizer, then

$G$00

and the category is said to be of type $G$01 (Dong, 2019).

The classical near-group case is recovered when there is exactly one non-invertible simple object: then $G$02, and the defining rule reduces to

$G$03

which matches the usual near-group formula $G$04 (Dong, 2019). The generalized Tambara–Yamagami case corresponds to $G$05, so all products of non-invertibles decompose purely into invertibles.

This broader framework leads to a structural classification in the slightly degenerate braided setting. Slightly degenerate generalized near-group fusion categories are exactly one of the following: $G$06, where $G$07 is an Ising category and $G$08 is slightly degenerate pointed; $G$09, where $G$10 is a Yang–Lee category; $G$11, where $G$12 is a slightly degenerate fusion category of the form $G$13 and $G$14 is non-degenerate pointed; or a prime category generated by a $G$15-dimensional simple object (Dong, 2019). A companion paper develops the same theory further, including an explicit example arising from an extension of a Fibonacci category and a classification of braided generalized Tambara–Yamagami categories of dimension $G$16 (Dong et al., 2020).

A complementary unification comes from near-integral fusion. There a near-integral fusion ring is one with a proper fusion subring of maximal rank. Such a ring is determined by an integral fusion subring $G$17 and one extra self-dual basis element $G$18, with

$G$19

Near-group fusion rings are exactly the case $G$20. The character-theoretic description is especially sharp: a fusion ring is near-integral if and only if there exists a ring homomorphism whose kernel is a maximal-rank proper fusion subring, and near-group rings are the pointed instances of this phenomenon (Dong et al., 2024).

5. NIM-representations, algebra objects, and noncommutative realizations

The near-group relation $G$21 is also an efficient starting point for the study of module categories and algebra objects. In the NIM-rep approach, a near-group fusion ring $G$22 acts on a $G$23-module with basis $G$24, and the action of the non-invertible element $G$25 is encoded by a symmetric matrix $G$26, while the $G$27-action decomposes $G$28 into $G$29-orbits governed by subgroups $G$30. The basic constraints are

$G$31

where $G$32 (Hannah et al., 2023).

This matrix equation completely controls the low-orbit cases. For one orbit, NIM-reps are parametrized by pairs $G$33 satisfying

$G$34

with divisibility and positivity constraints. For two orbits, they are parametrized by tuples $G$35 satisfying

$G$36

$G$37

and

$G$38

a perfect square (Hannah et al., 2023). In the Tambara–Yamagami case $G$39, every irreducible NIM-rep has at most two group orbits.

Admissible NIM-reps reconstruct algebra objects. In the one-orbit case, the associated algebra object has the form

$G$40

In the two-orbit admissible case $G$41, the algebra object is purely group-like: $G$42 This recovers known classifications of algebra objects in near-group, especially Tambara–Yamagami, categories from a fusion-ring and NIM-rep viewpoint (Hannah et al., 2023).

The noncommutative side is far more rigid. For a non-abelian finite group $G$43, a unitary near-group fusion category with Grothendieck ring $G$44 exists if and only if $G$45 is an extra-special $G$46-group. In that case $G$47, $G$48, and $G$49. Moreover, for each extra-special $G$50-group $G$51 there are exactly three inequivalent unitary near-group categories, distinguished by the third Frobenius–Schur indicator $G$52 (Izumi et al., 2019).

Izumi and Tucker give a purely algebraic construction of all these noncommutative near-group fusion categories from pointed categories categorically Morita equivalent to them. Their categories are of the form

$G$53

with $G$54 built from $G$55, $G$56, and $G$57 inflated from a generator of $G$58. The group of invertible objects in $G$59 is the central product of $G$60 copies of $G$61 when $G$62 is even and the central product of one copy of $G$63 and $G$64 copies of $G$65 when $G$66 is odd, while the unique non-invertible simple object satisfies

$G$67

This realizes the full noncommutative near-group classification group-theoretically (Izumi et al., 2019).

6. Pointed limits, VOA realizations, and adjacent phenomena

Near-group fusion algebras are best understood against the backdrop of pointed fusion rings. A pointed fusion ring is a fusion ring in which every simple basis element is invertible, equivalently an integral group ring $G$68. Several papers in the area make this contrast explicit: near-group rings are “near” the pointed case because they add exactly one non-invertible simple object to a group ring, while preserving a large invertible sector (Schopieray, 2022).

A clean VOA example of the pointed limit is the even part $G$69 of the symplectic fermion vertex operator superalgebra. Its simple modules are

$G$70

and their fusion algebra is isomorphic to the group algebra of the Klein four group: $G$71 All simple modules are invertible, so this is not a genuine near-group fusion algebra but rather a pointed baseline from the VOA side (Abe et al., 2011). This example is useful precisely because it isolates the group-like sector without any extra non-invertible simple object.

A more general pointed-orbifold criterion appears for cyclic orbifolds of lattice VOAs. If $G$72 is an even positive-definite lattice and $G$73, then the orbifold VOA $G$74 has group-like fusion if and only if $G$75 acts trivially on the discriminant group $G$76, equivalently

$G$77

When this holds and $G$78 is fixed point free, the representation category is a pointed modular tensor category completely determined by its quadratic space structure (Lam, 2018). This again identifies a supply of explicitly computable group-like fusion rings that can serve as the invertible sector in broader near-group constructions.

A recent physics-oriented development studies near-group fusion algebras as commutative non-invertible fusion algebras of type $G$79, where $G$80 is finite Abelian and

$G$81

The paper emphasizes Fibonacci, Ising, and $G$82 fusion rules as concrete instances, develops a spurion-labeling scheme for coupling constants directly at the level of the near-group fusion algebra, and interprets radiative violations of tree-level non-invertible selection rules as “loop-induced groupification” (Suzuki et al., 20 Aug 2025). This does not alter the algebraic definition, but it shows that near-group fusion algebras now function as organizing principles well beyond their original categorical setting.

Taken together, these developments place near-group fusion algebras at a nexus of pointed fusion theory, categorifiability, module-category reconstruction, Drinfeld-center computations, subfactor and quantum-group realizations, and recent applications of non-invertible fusion rules. The pointed case supplies the ambient metric-group data, while the single non-invertible generator encodes the first genuinely non-pointed deformation of the group-ring paradigm.