Noncommutative Triangular Prism

• Noncommutative triangular prism is a 4D operator subsystem in C*(Z3 * Z2) that models associativity constraints and noncommutative geometry.
• Its dilation properties, established by the Halmos–Mirman theorem, ensure a coherent extension of operator systems with group elements of orders 3 and 2.
• In fusion categories, triangular prism equations generalize the pentagon equations, providing a powerful localization tool for resolving associativity problems and categorification.

A noncommutative triangular prism is a foundational object in operator system theory, noncommutative convexity, and fusion category theory, serving as both a model for noncommutative geometry and a key combinatorial structure for describing associativity constraints in fusion categories. In operator algebra, it is realized as a four-dimensional operator subsystem $$\mathcal{S}_{\mathrm{prism}} \subset C^*(\mathbb{Z}_3 * \mathbb{Z}_2)$$ generated by canonical group elements from the free product of cyclic groups of orders 3 and 2. In the context of tensor categories, its associated combinatorial equations encode pivotal associativity data and generalize the classical pentagon equations to the non-symmetric (noncommutative) setting, providing powerful localization tools for rigidity and categorification problems.

1. Algebraic Definition and Operator System Structure

Let $\mathbb{Z}_3$ (generated by $w$, order 3) and $\mathbb{Z}_2$ (generated by $v$, order 2) be cyclic groups. The full group $C^*$-algebra of their free product $G = \mathbb{Z}_3 * \mathbb{Z}_2$ admits a canonical 4-dimensional operator subsystem

$$\mathcal{S}_{\mathrm{prism}} := \mathrm{span}\{1, w, w^2, v\} \subset C^*(G).$$

The canonical generators $u_1 = w$, $u_2 = w^2$, $u_3= v$ satisfy

$$w^3 = 1, \quad v^2 = 1, \quad w^* = w^2, \quad v^* = v.$$

Alternatively, selfadjoint coordinate generators are defined as $x = \mathrm{Re}\,w = (w + w^2)/2$, $y = \mathrm{Im}\,w = (w - w^2)/(2i)$, $z = v$, so that

$$\mathcal{S}_{\mathrm{prism}} = \mathrm{span}\{1, x, y, z\} \subset C^*(G).$$

The noncommutative triangular prism is identified as the joint matrix range of $(x, y, z)$ under completely positive unital maps, describing a noncommutative convex body in the sense of Arveson, dual to a noncommutative convex set in NCConv (Farenick et al., 23 Jan 2026).

2. Dilation Theory: The Halmos–Mirman Theorem for the Triangular Prism

For each $n \in \mathbb{N}$, the matrix range of the generating tuple is

$$W_n(\mathfrak{u}) = \left\{ (\varphi(x), \varphi(y), \varphi(z)) \in M_n(\mathbb{C})^3 : \varphi: \mathcal{S}_{\mathrm{prism}} \to M_n(\mathbb{C}) \text{ unital completely positive} \right\}.$$

The dilation theorem shows that for operators $T_1,T_2,T_3 \in B(H)$, the following are equivalent:

• $(T_1, T_2, T_3) \in W_n(\mathfrak{u})$,
• There exists a Hilbert space $K \supset H$, a unitary $U$ with $U^3 = 1$, a selfadjoint unitary $V$ with $V^2 = 1$, and an isometry $Z: H \to K$ such that

$$Z^* U Z = T_1 + i T_2,\quad Z^* U^2 Z = T_1 - i T_2,\quad Z^* V Z = T_3.$$

Block-matrix dilations for $U$ and $V$ are given explicitly via the Mirman dilation (for the normal operator $T_1 + i T_2$ to a unitary of order 3) and the Halmos dilation (for selfadjoint contraction $T_3$ to a symmetry), ensuring simultaneous dilation compatible with the group structure (Farenick et al., 23 Jan 2026). The proof utilizes the universal property of $C^*(\mathbb{Z}_3 * \mathbb{Z}_2)$ and Stinespring’s theorem to construct joint dilations.

3. Noncommutative Geometric Properties

Extreme Points and Representation Theory

Analysis of irreducible representations of $C^*(\mathbb{Z}_3 * \mathbb{Z}_2)$ reveals that for every $n$, there exists an irreducible $*$-representation $$\pi_n : C^*(\mathbb{Z}_3 * \mathbb{Z}_2) \to M_n(\mathbb{C})$$, producing a noncommutative matrix-extreme point at level $n$ for $P(3)^{\text{max}}$. Realizations can be constructed using finite groups such as $\mathrm{PSL}_2(\mathbb{F}_q)$ for suitable $q$. As $\mathbb{Z}_3 * \mathbb{Z}_2$ contains the free group $F_2$, it admits type II$_1$, type II$_\infty$, and type III factorial representations, giving rise to extreme points at level $\aleph_0$.

Duality, Exactness, and Lifting

The operator system $\mathcal{S}_{\mathrm{prism}}$ is not exact, following from the non-exactness of $C^*(\mathbb{Z}_3 * \mathbb{Z}_2)$ which contains $F_2$. However, $\mathcal{S}_{\mathrm{prism}}$ is OMAX, possessing the lifting property. Its dual operator system is

$$\mathcal{S}_{\mathrm{prism}}^\delta \cong \{ (z_1, z_2, z_3, z_4, z_5) \in \mathbb{C}^5 : z_1+z_2+z_3 = z_4+z_5 \},$$

completely order isomorphic to an OMIN system encoding the classical triangular prism as $\mathrm{conv}\{1, \omega, \omega^2\} \times [-1,1]$.

Tensor Products and Complete Positivity

Tensor product structures for $\mathcal{S}_{\mathrm{prism}}$ display strong noncommutative behavior: $$\mathcal{S}_{\mathrm{prism}} \otimes_{\min} \mathcal{S}_{\mathrm{prism}} \neq \mathcal{S}_{\mathrm{prism}} \otimes_{c} \mathcal{S}_{\mathrm{prism}} \neq \mathcal{S}_{\mathrm{prism}} \otimes_{\max} \mathcal{S}_{\mathrm{prism}},$$ with automatic complete positivity of all positive maps into or out of $\mathcal{S}_{\mathrm{prism}}$ (Farenick et al., 23 Jan 2026).

4. Triangular Prism Equations in Fusion Categories

For a pivotal fusion category $\mathcal{C}$, the triangular prism equations (TPE) encode associativity data for nine objects $X_1, \dots, X_9$ and six morphisms $\alpha_1, \dots, \alpha_6$ in appropriate Hom spaces. The general TPE is

$$\sum_{\beta_0 \in B_0} T(\rho^{-2}(\alpha_2),\rho(\alpha_3),\rho^{-1}(\alpha_1),\beta_0)\,T(\rho^{-1}(\alpha_5),\beta_0',\rho(\alpha_4),\rho^{-1}(\alpha_6)) = \cdots,$$

where the right side involves summation over auxiliary bases and simple objects, with each $T(-,-,-,-)$ a tetrahedral string diagram valuation. The TPE uses the automorphisms $\rho$ (cyclic "third-leg rotation") and $\sigma$ (pivotal shift), allowing for bookkeeping of tensor product associators without relying on commutativity or symmetry (Liu et al., 2022).

5. Relationship with the Pentagon Equations

In the spherical case (when all duals and the pivotal structure square to the identity), the TPE specialize to the classical Mac Lane pentagon equations for associators, up to an explicit change of basis. This is formalized via a $4\times 4$ matrix $M_{i_1,i_2,i_3}$ constructed from tetrahedron invariants, so that

$$(\mu_1,\mu_2,\mu_3,\mu_4) \mapsto (\alpha_1,\alpha_2,\alpha_3,\alpha_4) = M_{i_1,i_2,i_3} \cdot (\mu_1, \mu_2, \mu_3, \mu_4)^T,$$

and the TPE becomes the PE. This demonstrates that, in the presence of a symmetric (spherical) structure, noncommutative data reduces to classical coherence equations (Liu et al., 2022).

6. Localization and Applications in Categorification

The triangular prism equations permit a localization strategy for the analysis of large, overdetermined systems of associativity constraints. The general method involves selecting a minimal set of variables (F-symbols or tetrahedron invariants), extracting a subset of equations, computing a Gröbner basis, and recursively reducing complexity. This approach is especially effective for ruling out possible fusion ring categorifications, with variable elimination and sequential subsystem analysis drastically shrinking the candidate solution space.

Applications include the resolution of the second Frobenius–Schur indicator conjecture (e.g., showing $\nu_2(Y^*) = 1$ in pivotal fusion categories whenever $\mathrm{Hom}(1 \to X^* X \to Y)$ is odd-dimensional) and the full classification of non-pointed, integral, unitary, 1-Frobenius simple fusion categories up to rank $8$ and Frobenius–Perron dimension $20,\!000$. The only resulting categories in this range are representation rings $\mathrm{Rep}(\mathrm{PSL}_2(q))$ with $q \in \{4,5,7,8,9,11\}$ a prime power (Liu et al., 2022).

7. Categorical and Noncommutative Features

The prism equations do not assume commutativity or braiding. The orientation, duality, and order of the tensor factors are essential, with all operations sensitive to the lack of symmetric monoidal structure. The automorphisms $\rho$ and $\sigma$ are instrumental in encoding noncommutative associator manipulation, and even in the categorical Grothendieck ring, localization arguments utilize noncommutative multiplication. This consolidates the TPE as a unifying categorical tool for both noncommutative tensor categories and their classical commutative limits.

---

References:

• "Representations of noncommutative cubes and prisms" (Farenick et al., 23 Jan 2026)
• "Triangular Prism Equations and Categorification" (Liu et al., 2022)