Parafermion Algebras in CFT and VOAs

• Parafermion algebras are operator structures that generalize fermionic and bosonic systems, playing a pivotal role in conformal field theory and vertex operator algebras.
• They are constructed via coset methods in affine Kac–Moody and lattice VOAs, yielding detailed modular data and fusion categories with practical computational applications.
• The framework unifies classic Green-type Lie algebras with modern parafermion vertex operator algebras, underpinning both rational and logarithmic conformal field theories.

Parafermion algebras constitute a foundational class of operator algebras that generalize both Clifford (fermion) and boson algebras, and occupy a central role in conformal field theory (CFT), the representation theory of vertex operator algebras (VOAs), topological phases of matter, and quantum computation. In mathematics, "parafermion algebra" refers, depending on context, to either the classic Green-type parafermion Lie algebra or, more prevalently in contemporary research, to the family of parafermion vertex operator algebras defined via coset (commutant) constructions in affine Kac–Moody and lattice VOAs. These structures encode chiral symmetry algebras for a broad class of rational and logarithmic CFTs and underlie the modular data of many fusion categories.

1. Classic Parafermionic Lie Algebras

The original parafermionic algebra, due to H.S. Green, is the universal enveloping algebra $\mathscr{P}_n$ generated by creation and annihilation operators $\{a_i^\dagger,\, a^j\}$ ($i,j=1,\dots,n$) subject to cubic exchange relations: $$\begin{aligned} &[[a_i^\dagger,a^j],a_k^\dagger] = 2\delta^j_k a_i^\dagger, \quad [[a_i^\dagger,a^j],a^k] = -2\delta_i^k a^j, \ &[[a_i^\dagger,a_j^\dagger],a_k^\dagger]=0, \quad [[a^i,a^j],a^k]=0, \end{aligned}$$ where $[\cdot,\cdot]$ is the Lie bracket, and $[\![x, y]\!] := xy + yx$. This algebra is isomorphic to the simple Lie algebra $\mathfrak{so}(2n+1,\mathbb{C})$ under an explicit identification of generators. Its Fock space representations $\mathscr{V}_n(p)$, for integer $p\ge 0$, are induced by enforcing exclusion principles analogous to finite-dimensional representations of $U(\mathfrak{so}(2n+1))$ and possess a rich cohomological structure computed via Kostant’s theorem, relating their algebraic invariants to combinatorics of self-conjugate partitions and Schur polynomial identities (Popov, 2014).

2. Parafermion Vertex Operator Algebras: Construction

Modern developments focus on parafermion vertex operator algebras $K(\mathfrak{g},k)$—hereafter abbreviated as PVOAs—defined as cosets (commutants) of Heisenberg VOAs within simple affine Kac–Moody VOAs at positive integral level $k$. Precisely, let $\mathfrak{g}$ be a simple Lie algebra of rank $\ell$, $Q$ its root lattice, $Q_L$ the long root lattice, $h^\vee$ the dual Coxeter number. For level $k\in\mathbb{Z}_{>0}$, the construction is: $$K(\mathfrak{g},k) := \mathrm{Com}_{L_{\mathfrak{g}}(k,0)} M_{\mathfrak{h}}(k),$$ where $L_{\mathfrak{g}}(k,0)$ is the simple affine VOA and $M_{\mathfrak{h}}(k)$ the rank-$\ell$ free-boson (Heisenberg) VOA (Dong et al., 2018, Dong et al., 2014).

This commutant is a simple, CFT-type, $C_2$-cofinite and rational VOA with central charge

$$c_{K(\mathfrak{g},k)} = \frac{k \cdot \dim\mathfrak{g}}{k + h^\vee} - \ell.$$

The canonical decomposition of $L_{\mathfrak{g}}(k,0)$ as a module for $M_{\mathfrak{h}}(k) \otimes K(\mathfrak{g},k)$ underpins the structure theory: $$L_{\mathfrak{g}}(k,\Lambda) \cong \bigoplus_{\mu \in (\Lambda + Q)/kQ_L} M_{\mathfrak{h}}(k, \mu) \otimes M_{\Lambda, \mu},$$ where $M_{\Lambda, \mu}$ are precisely the irreducible $K(\mathfrak{g},k)$-modules (Dong et al., 2018).

3. Module Classification, Fusion Rules, and Modular Data

The irreducible modules of $K(\mathfrak{g},k)$ are parameterized by pairs $(\Lambda,i)$, where $\Lambda\in P_k$ (level-$k$ dominant weights