betaFeedback

WorksMathematics

Grothendieck rings of complex fusion categories: rank 3

Abstract

Ostrik classified the pivotal fusion categories of rank three. We show that every complex fusion category of rank three admits a spherical structure and recover the five Grothendieck rings. Both steps use the mass at twist one in the Drinfeld center. For an irreducible cubic codegree polynomial, Galois averaging in the sphericalization determines the first six moments of the squared twist. For a spherical category, the first two moments give a single inequality that excludes all the remaining rings outside the list. The rank-specific descent theorem we use from the literature is Ostrik's theorem for a quadratic factor.

GPT-6.2 Astra assisted with the research. Complete human verification is pending, and the results should be treated with caution.

The classification and the proof

A complex fusion category C\mathcal{C} is a rigid, semisimple, C\mathbb{C}-linear tensor category with finite-dimensional morphism spaces, finitely many isomorphism classes of simple objects and simple unit. Its rank is the number of simple classes. Its Grothendieck ring K0(C)K_0(\mathcal{C}) records their tensor multiplicities and their distinguished basis. A pivotal structure is a monoidal isomorphism from the identity to the double dual. It is spherical when the two pivotal traces agree.

For nonnegative integers k,l,m,nk,l,m,n satisfying k2+l2=kn+lm+1k^2+l^2=kn+lm+1, denote by K(k,l,m,n)K(k,l,m,n) the based ring on 1,X,Y1,X,Y with

X2=1+mX+kY,Y2=1+lX+nY,XY=kX+lY.X^2 = 1 + mX + kY,\qquad Y^2 = 1 + lX + nY,\qquad XY = kX + lY.

The elementary derivation of this presentation appears in § 2.

Theorem 1.1. Up to based-ring isomorphism, the Grothendieck rings of complex fusion categories of rank three are exactly

Z[C3],K(1,0,0,0),K(1,0,1,0),K(1,0,2,0),K(1,1,1,0).\mathbb{Z}[C_3],\qquad K(1,0,0,0),\qquad K(1,0,1,0),\qquad K(1,0,2,0),\qquad K(1,1,1,0).

Every complex fusion category of rank three admits a spherical structure. Each of the five rings has a unitary realization.

The five rows of Table 1 specify unitary examples. A row can have several inequivalent associators.

Fusion ringA unitary realizationReference
Z[C3]\mathbb{Z}[C_3]Vec⁡C3\operatorname{Vec}_{C_3}[11 Theorem 1.1(i)]
K(1,0,0,0)K(1,0,0,0)An Ising category[12], [11 Thm. 1.1(iii)]
K(1,0,1,0)K(1,0,1,0)Rep⁡(S3)\operatorname{Rep}(S_3)[11 Theorem 1.1(iv)]
K(1,0,2,0)K(1,0,2,0)The E6E_6 even part of rank three[2, 5], [3 Sec. 1]
K(1,1,1,0)K(1,1,1,0)The even subcategory of SU(2)5SU(2)_5[9 Sec. 4]

Table 1. A unitary realization of each ring.

History. Two early sources of examples are Bion-Nadal’s construction of an E6E_6 subfactor [2] and the Tambara–Yamagami classification [12], which includes the Ising ring. Ostrik classified the rank-three categories admitting a ribbon structure [9]. Hagge and Hong classified the tensor structures on the E6E_6 ring by solving the pentagon equations [5]. Thornton proved that every near-group category, including the rank-three near-group family, admits a spherical structure [13 Theorem 1.3]. Ostrik subsequently classified all pivotal rank-three categories, up to tensor equivalence [11 Theorem 1.1]. His proof also gave spherical descent in the quadratic case and obstructions for individual cubic rings [11 Section 5]. We use his formal codegrees and central-induction results [10, 11] and the indicator formulas of Ng and Schauenburg [7, 8].

Outline. Sections 2 and 3 relate the three formal codegrees to the mass at twist one in the center. Section 4 applies this relation to squared twists in the sphericalization. Six moments force the cubic ring and allow its spherical structure to descend. For a reducible codegree polynomial we invoke Ostrik’s quadratic descent or the canonical spherical structure in the weakly integral case. Section 5 then applies the same mass estimate to ordinary twists. A degree-two polynomial recovers both Ostrik’s structural reduction and his bound for near-group multiplicities [11 Theorems 3.3 and 4.1]. Each exclusion is independent of any bound on the structure constants.

The ring algebra and sphericalization

The two codegree coefficients

As in [11 Proposition 3.1], a rank-three fusion ring is either Z[C3]\mathbb{Z}[C_3], or all its basis elements are self-dual and it has the form K(k,l,m,n)K(k,l,m,n), with basis 1,X,Y1,X,Y and multiplication

X2=1+mX+kY,Y2=1+lX+nY,XY=kX+lY.(1)X^2=1+mX+kY,\qquad Y^2=1+lX+nY,\qquad XY=kX+lY. \tag*{(1)}

Here k,l,m,nk,l,m,n are nonnegative integers satisfying

k2+l2=kn+lm+1.(2)k^2+l^2=kn+lm+1. \tag*{(2)}

Here is a quick derivation. If X∗=Y≠XX^*=Y\ne X, the product X2X^2 has no unit constituent; writing d=FPdim⁡(X)=FPdim⁡(Y)d=\operatorname{FPdim}(X)=\operatorname{FPdim}(Y) gives d2=(a+b)dd^2=(a+b)d for nonnegative integers a,ba,b, so dd is an integer. The product XYXY has one unit constituent, hence d2=1+(c+e)dd^2=1+(c+e)d for integers c,e≥0c,e\ge0; therefore d∣1d\mid1 and d=1d=1, giving Z[C3]\mathbb{Z}[C_3]. Otherwise all three simples are self-dual. The involution on a based ring reverses products, so the ring is commutative. Frobenius reciprocity forces the coefficients in XYXY to be k,lk,l as displayed; the identity (X2)Y=X(XY)(X^2)Y=X(XY) gives (2), which also suffices for associativity.

The multiplication matrices in the ordered basis 1,X,Y1,X,Y are

NX=(0101mk0kl),NY=(0010kl1ln).N_X=\begin{pmatrix}0&1&0\\1&m&k\\0&k&l\end{pmatrix},\qquad N_Y=\begin{pmatrix}0&0&1\\0&k&l\\1&l&n\end{pmatrix}.

They are commuting real symmetric matrices, so their common characters χ1,χ2,χ3\chi_1,\chi_2,\chi_3 are real. Their formal codegrees are

fi=1+χi(X)2+χi(Y)2.f_i=1+\chi_i(X)^2+\chi_i(Y)^2.

Equivalently, they are the eigenvalues of multiplication by the Casimir element

Ω=1+X2+Y2=3+(l+m)X+(k+n)Y.\Omega=1+X^2+Y^2=3+(l+m)X+(k+n)Y.

Let τ\tau extract the coefficient of the unit. For the basis e0=1,e1=X,e2=Ye_0=1,e_1=X,e_2=Y we have τ(eiej)=δij\tau(e_i e_j)=\delta_{ij}. For multiplication by aa, the iith diagonal entry is τ(eiaei)\tau(e_i a e_i). Thus Tr⁡(a)=τ(aΩ)\operatorname{Tr}(a)=\tau(a\Omega) and

Tr⁡(a/Ω)=τ(a).(3)\operatorname{Tr}(a/\Omega)=\tau(a). \tag*{(3)}

Here Tr⁡\operatorname{Tr} is the regular trace of the rational ring, and Ω\Omega is invertible because its eigenvalues are positive. We shall use the same calculation in the signed ring. Taking a=1a=1 gives ∑i=131/fi=1\sum_{i=1}^{3}1/f_i=1. The codegrees are eigenvalues of an integral matrix. Consequently their characteristic polynomial has the form

p(t)=∏i=13(t−fi)=t3−bt2+ct−c.(4)p(t)=\prod_{i=1}^{3}(t-f_i)=t^3-bt^2+ct-c. \tag*{(4)}

We call the ring cubic when pp is irreducible over Q\mathbb{Q}. We will use the following rank-three identities:

b=9+(l+m)2+(k+n)2=7+3k2+3l2+m2+n2,(5)b=9+(l+m)^2+(k+n)^2=7+3k^2+3l^2+m^2+n^2, \tag*{(5)}
c=4b−9+(kl−mn)2+(2(k2−lm)−1)2.(6)c=4b-9+(kl-mn)^2+\left(2(k^2-lm)-1\right)^2. \tag*{(6)}

Both identities follow by expanding 3I+(l+m)NX+(k+n)NY3I+(l+m)N_X+(k+n)N_Y and using (2). The determinant identity is Ostrik’s identity [11 equation (12)], written as a sum of two squares. In particular,

c≥4b−8,(7)c\ge4b-8, \tag*{(7)}

because the last square in (6) is the square of an odd integer. Vieta’s identities also give

u:=∑i=13fi−2=1−2bc.(8)u:=\sum_{i=1}^{3}f_i^{-2}=1-\frac{2b}{c}. \tag*{(8)}

The signed quotient

Let SS be the sphericalization of a fusion category CC with self-dual rank-three ring. There is a tensor functor π:S→C\pi:S\to C. Each simple object VV of CC has two simple lifts V+,V−V_+,V_- with nonzero opposite dimensions. When VV is self-dual, its lifts are self-dual: the dual of V+V_+ is a lift of VV, and sphericality gives dim⁡(V+∗)=dim⁡(V+)\dim(V_+^*)=\dim(V_+), which rules out V−V_-. This observation is due to Thornton [13, proof of Lemma 2.3], as noted in [11 Section 5.1(a)]. Put ε=1−\varepsilon=1_-. Then

ε2≅1,dim⁡(ε)=−1,ε⊗V+≅V−.\varepsilon^2\cong1,\qquad\dim(\varepsilon)=-1,\qquad\varepsilon\otimes V_+\cong V_-.

The construction and these properties are given in [4 Remark 3.1 and Proposition 5.14] and [11 Section 5.1]. A useful realization is the equivariantization for the canonical C2C_2-action [1 Section 5.3].

All simples of SS are self-dual, so K0(S)K_0(S) is commutative. Its two quotients at ε=1\varepsilon=1 and ε=−1\varepsilon=-1 are respectively

R=K0(C),R~=K0(S)/(ε+1).R=K_0(C),\qquad\widetilde{R}=K_0(S)/(\varepsilon+1).

Choose lifts of X,YX,Y and denote their images in R~\widetilde{R} by X~,Y~\widetilde{X},\widetilde{Y}. Then R~=K(k~,l~,m~,n~)\widetilde{R}=K(\widetilde{k},\widetilde{l},\widetilde{m},\widetilde{n}), with the same multiplication formulas and associativity relation as above, but with integer, possibly negative, coefficients. For each corresponding pair of coefficients a,a~a,\widetilde{a},

∣a~∣≤a,a~≡a(mod2).(9)|\widetilde{a}|\le a,\qquad\widetilde{a}\equiv a\pmod{2}. \tag*{(9)}

Indeed, aa is the sum and a~\tilde{a} the difference of the two lift multiplicities. Changing the chosen lift changes the sign of the corresponding basis element.

The coefficient-of-1 functional τ\tau on R~\widetilde{R} satisfies

τ(X~iX~j)=δij,(X~0,X~1,X~2)=(1,X~,Y~).(10)\tau(\widetilde{X}_i\widetilde{X}_j)=\delta_{ij},\qquad(\widetilde{X}_0,\widetilde{X}_1,\widetilde{X}_2)=(1,\widetilde{X},\widetilde{Y}). \tag*{(10)}

For example, X+2X_+^2 contains 11 once and does not contain ε\varepsilon, since X+X_+ is self-dual and εX+≇X+\varepsilon X_+ \not\cong X_+. Thus taking the signed quotient preserves this pairing.

The general categorical input

For a spherical fusion category D\mathcal{D}, write F:Z(D)→DF: Z(\mathcal{D}) \to\mathcal{D} for the forgetful functor from its Drinfeld center and II for its right adjoint. If its Grothendieck ring is commutative, central induction gives a bijection between its ring characters χ\chi and the simple summands AχA_\chi of I(1)I(1), with

I(1)=⨁χAχ,dim⁡(Aχ)=dim⁡(D)fχ,θAχ=1.(11)I(1)=\bigoplus_\chi A_\chi,\qquad\dim(A_\chi)=\frac{\dim(\mathcal{D})}{f_\chi},\qquad\theta_{A_\chi}=1. \tag*{(11)}

These are the commutative-ring cases of Ostrik’s central-induction and trivial-twist theorems [11 Theorems 2.5 and 2.13]. Two consequences will be used repeatedly. First, dim⁡(D)/fχ\dim(\mathcal{D})/f_\chi is an algebraic integer, because it is an object dimension. Second, every fχf_\chi belongs to the field of simple object dimensions: the forgetful functor preserves spherical dimensions, so both the numerator and denominator in fχ=dim⁡(D)/dim⁡(F(Aχ))f_\chi=\dim(\mathcal{D})/\dim(F(A_\chi)) belong to that field. These are [11 Corollaries 2.14–2.15].

For a fusion category without a specified pivotal structure, every formal codegree still divides its global dimension [11 Corollary 2.14].

The dimension field in the cubic case

Lemma 2.1. Suppose that RR is cubic. The dimension field LL of SS is a totally real cyclic cubic field, and the dimension homomorphism identifies R~⊗Q\widetilde{R}\otimes\mathbb{Q} with LL. Set

x=dim⁡(X+),y=dim⁡(Y+),δ=1+x2+y2=dim⁡(C),x=\dim(X_+),\qquad y=\dim(Y_+),\qquad\delta=1+x^2+y^2=\dim(C),

so that dim⁡(S)=2δ\dim(S)=2\delta. Then

Tr⁡L/Q(1/δ)=1,Tr⁡L/Q(x/δ)=Tr⁡L/Q(y/δ)=0.(12)\operatorname{Tr}_{L/\mathbb{Q}}(1/\delta)=1,\qquad\operatorname{Tr}_{L/\mathbb{Q}}(x/\delta)=\operatorname{Tr}_{L/\mathbb{Q}}(y/\delta)=0. \tag*{(12)}

The formal codegrees of R~\widetilde{R} are the three conjugates gig_i of δ\delta, counted with multiplicity. Write

∏i(t−gi)=t3−b~t2+c~t−c~.\prod_i(t-g_i)=t^3-\widetilde{b}t^2+\widetilde{c}t-\widetilde{c}.

Then

0<b~≤b,c~c∈Z>0,0<b~c~≤bc.(13)0<\widetilde{b}\le b,\qquad\frac{\widetilde{c}}{c}\in\mathbb{Z}_{>0},\qquad0<\frac{\widetilde{b}}{\widetilde{c}}\le\frac{b}{c}. \tag*{(13)}

Moreover,

v:=∑igi−2=1−2b~c~.(14)v:=\sum_i g_i^{-2}=1-\frac{2\widetilde{b}}{\widetilde{c}}. \tag*{(14)}

Proof. The dimension homomorphism factors through R~⊗Q\widetilde{R}\otimes\mathbb{Q}, so its image is a field of degree at most three. The formal codegrees of SS are

2f1,2f2,2f3,2g1,2g2,2g3:(15)2f_1,2f_2,2f_3,2g_1,2g_2,2g_3: \tag*{(15)}

characters with ε=1\varepsilon= 1 come from RR, and those with ε=−1\varepsilon= -1 come from R~\widetilde{R}; the two lifts double the sum defining each codegree. Every formal codegree of a spherical category lies in its dimension field [11 Corollary 2.15]. Therefore LL contains all three roots of the irreducible cubic pp. It has degree exactly three and is its splitting field. Spherical dimensions are totally real [4 Corollary 2.10], so LL is totally real. The map R~⊗Q→L\widetilde{R} \otimes\mathbb{Q} \to L is now an isomorphism.

Apply the regular-trace calculation (3) to the basis 1,x,y1,x,y, whose pairing is (10) and whose Casimir element is δ\delta. This gives (12). The embeddings of LL are precisely the three characters of the signed algebra, and its Casimir element has image δ\delta; hence its codegrees are gi=σi(δ)>0g_i = \sigma_i(\delta) > 0. Their reciprocal sum is one, giving the displayed polynomial and (14).

Formula (5), valid also for the signed coefficients, gives

b~=7+3k~2+3l~2+m~2+n~2≤b.\widetilde{b} = 7 + 3\widetilde{k}^{2} + 3\widetilde{l}^{2} + \widetilde{m}^{2} + \widetilde{n}^{2} \le b.

Finally, formal codegrees divide the global dimension even without a pivotal structure [11 Corollary 2.14]. Thus δ/f1\delta/f_1 is an algebraic integer in LL. Taking its field norm yields

NL/Q(δ/f1)=c~c∈Z>0.N_{L/\mathbb{Q}}(\delta/f_1) = \frac{\widetilde{c}}{c} \in\mathbb{Z}_{>0}.

The last inequality in (13) follows.

Central induction and twist moments

Let F:Z(D)→DF : Z(\mathcal{D}) \to\mathcal{D} be the forgetful functor for a spherical fusion category D\mathcal{D}, and let I:D→Z(D)I : \mathcal{D} \to Z(\mathcal{D}) be its right adjoint. We use the induction formula [11 Proposition 2.4]

FI(V)≅⨁W∈Irr⁡DW⊗V⊗W∗.(16)FI(V) \cong\bigoplus_{W \in\operatorname{Irr}\mathcal{D}} W \otimes V \otimes W^{*}. \tag*{(16)}

Together with adjunction, this identifies [I(V):A]=[F(A):V][I(V) : A] = [F(A) : V] and dim⁡CEnd⁡(I(V))=[FI(V):V]\dim_{\mathbb{C}} \operatorname{End}(I(V)) = [FI(V) : V]. Write d=dim⁡(D)d = \dim(\mathcal{D}), and for A∈Irr⁡Z(D)A \in\operatorname{Irr} Z(\mathcal{D}) put aA=dim⁡(A)a_A = \dim(A) and let θA\theta_A be its twist. The numbers aAa_A are real, and the twists are roots of unity. We use the identities

∑AaA2=d2.(17)\sum_A a_A^2 = d^2. \tag*{(17)}
∑AaA2θAn=d∑V∈Irr⁡Ddim⁡(V)νn(V)(n≥1).(18)\sum_A a_A^2\theta_A^n = d\sum_{V \in\operatorname{Irr}\mathcal{D}} \dim(V)\nu_n(V) \qquad(n \ge1). \tag*{(18)}

where νn\nu_n is the nnth Frobenius–Schur indicator for the specified pivotal structure. To see (18), multiply the Ng–Schauenburg identity

tr⁡(θn∣I(V))=dνn(V)\operatorname{tr}(\theta^n|_{I(V)}) = d\nu_n(V)

by dim⁡(V)\dim(V) and sum over VV. Adjunction and sphericality of the forgetful functor give aA=∑V[I(V):A]dim⁡(V)a_A = \sum_V [I(V) : A]\dim(V), yielding the formula. The indicator identity is [8 Theorem 4.1], and (17) is Müger’s dimension formula for the center [6]. Also,

νn(V)∈Q(ζn).(19)\nu_n(V) \in\mathbb{Q}(\zeta_n). \tag*{(19)}

because the cyclic operator defining the indicator has nnth power equal to the identity [7 Theorem 5.1]. The trace identity uses sphericality. The cyclotomicity assertion holds in any pivotal category.

Both estimates below use the following observation. If μ\mu is a finite positive measure on the unit circle T\mathbb{T} and qq is a polynomial, then

∣q(1)∣2μ({1})≤∫T∣q(z)∣2 dμ(z).(20)|q(1)|^2\mu(\{1\}) \le\int_{\mathbb{T}} |q(z)|^2\,d\mu(z). \tag*{(20)}

The right side depends only on the moments through degree deg⁡q\deg q.

The central sign. The equivariantization model of SS gives a central copy of Rep⁡(C2)\operatorname{Rep}(\mathbb{C}_2) whose sign representation lifts ε\varepsilon [1 Definition 5.13 and Corollary 5.15]. Its underlying fusion subcategory has the usual tensor product of representations, with pivotal map on the sign equal to minus its usual pivotal map because dim⁡(ε)=−1\dim(\varepsilon)=-1. For even nn, cyclic rotation on the line Hom⁡(1,ε⊗n)\operatorname{Hom}(1,\varepsilon^{\otimes n}) is the ordinary rotation multiplied by this inverse pivotal scalar. Thus

ν2j(ε)=−1(j≥1).(21)\nu_{2j}(\varepsilon)=-1 \qquad(j\geq1). \tag*{(21)}

Lemma 3.1 (A forced squared-twist sector). In the cubic situation of Lemma 2.1, the simple constituents of I(1)I(1) and I(ε)I(\varepsilon) are twelve distinct simple objects of Z(S)Z(S). Their squared twists are all one, and the sum of their squared dimensions is

2δ2(u+v).(22)2\delta^2(u+v). \tag*{(22)}

The same statement holds after every embedding of LL.

Proof. Since K0(S)K_0(S) is commutative, I(1)I(1) is multiplicity-free. The formal-codegree theorem [11 Theorem 2.13] and (15) give its six simple constituent dimensions as

δf1, δf2, δf3, δg1, δg2, δg3.\frac{\delta}{f_1},\ \frac{\delta}{f_2},\ \frac{\delta}{f_3},\ \frac{\delta}{g_1},\ \frac{\delta}{g_2},\ \frac{\delta}{g_3}.

They are positive under every real embedding and their squared sum is δ2(u+v)\delta^2(u+v). Their twists are one [11 Theorem 2.5].

Write ε^\widehat{\varepsilon} for the canonical central lift of ε\varepsilon. The projection formula gives I(ε)≅ε^⊗I(1)I(\varepsilon)\cong\widehat{\varepsilon}\otimes I(1). Thus its six constituents have the opposite, negative dimensions, and are disjoint from those of I(1)I(1). The indicator identity gives tr⁡(θ2∣I(ε))=−2δ=dim⁡I(ε)\operatorname{tr}(\theta^2|_{I(\varepsilon)})=-2\delta=\dim I(\varepsilon). All six constituent dimensions are negative, so their absolute values sum to 2δ2\delta. Equality in the triangle inequality forces each squared twist to equal one. Their squared dimensions give the second copy of δ2(u+v)\delta^2(u+v) in (22).

The cubic case

The following estimate is the step that treats an arbitrary categorification of a cubic ring. It will force b≤16b\leq16, so its final arithmetic consequence can be read off without a search.

Proposition 4.1 (Cubic codegree bound). If K(k,l,m,n)K(k,l,m,n) is cubic and admits a complex fusion categorification, then

c≤72b.(23)c\leq\frac{7}{2}b. \tag*{(23)}

Proof. Use the notation of Lemmas 2.1–3.1, and put

M=Q(ζ120).M=\mathbb{Q}(\zeta_{120}).

The fields LL and MM are linearly disjoint: L/QL/\mathbb{Q} is Galois of degree 33, whereas [M:Q]=φ(120)=32[M:\mathbb{Q}]=\varphi(120)=32. Choose three automorphisms σ1,σ2,σ3\sigma_1,\sigma_2,\sigma_3 of Q‾\overline{\mathbb{Q}} restricting to the three embeddings of LL and fixing MM pointwise. Fix these choices for the rest of the proof. In particular, they fix every indicator of each of the orders

2, 4, 6, 8, 10, 12,2,\ 4,\ 6,\ 8,\ 10,\ 12,

by (19). The same choices work for all six orders, since each divides 120120. We apply these automorphisms only to algebraic scalar identities; no choice of a field of definition for the associators is needed.

On the unit circle T\mathbb{T}, define the finite positive measure

μ=∑i=13∑A∈Irr⁡Z(S)σi(aA)24σi(δ)2Dirac⁡σi(θA)2.(24)\mu=\sum_{i=1}^{3}\sum_{A\in\operatorname{Irr} Z(S)}\frac{\sigma_i(a_A)^2}{4\sigma_i(\delta)^2}\operatorname{Dirac}_{\sigma_i(\theta_A)^2}. \tag*{(24)}

Here Dirac⁡z\operatorname{Dirac}_z is unit point mass at zz. Every central dimension belongs to LL, since the forgetful functor is spherical. Its conjugates are real, and σi(δ)>0\sigma_i(\delta)>0. Thus the coefficients in (24) are nonnegative. Twists and their conjugates are roots of unity, so the support is contained in T\mathbb{T}. Equation (17) gives

μ(T)=3.(25)\mu(\mathbb{T})=3. \tag*{(25)}

For 1≤j≤61\le j\le6, apply σi\sigma_i to (18), with n=2jn=2j, and divide by 4σi(δ)24\sigma_i(\delta)^2. Since the indicators are fixed, summing over ii gives

∫Tzj dμ(z)=12∑V∈Irr⁡Sν2j(V)Tr⁡L/Q(dim⁡(V)δ)=12(ν2j(1)−ν2j(ε))=1.(26)\begin{aligned} \int_{\mathbb{T}}z^j\,d\mu(z)=\frac{1}{2}\sum_{V\in\operatorname{Irr} S}\nu_{2j}(V)\operatorname{Tr}_{L/\mathbb{Q}}\left(\frac{\dim(V)}{\delta}\right) \\ &=\frac{1}{2}\left(\nu_{2j}(1)-\nu_{2j}(\varepsilon)\right)=1. \tag*{(26)} \end{aligned}

For the second equality, the four non-unit lifts have dimensions ±x,±y\pm x,\pm y and vanish under the weighted trace by (12); the traces for 11 and ε\varepsilon are 11 and −1-1. The last equality uses (21) and νn(1)=1\nu_n(1)=1. This step does not require knowing any indicator of X±X_\pm or Y±Y_\pm.

Lemma 3.1 forces the lower bound

μ({1})≥32(u+v).(27)\mu(\{1\})\ge\frac{3}{2}(u+v). \tag*{(27)}

Indeed, in each of the three summands defining μ\mu, those twelve objects contribute (u+v)/2(u+v)/2 at one. The sums u,vu,v are rational and hence independent of the embedding.

Apply (20) to the Fejér polynomial

P(z)=∣1+z+⋯+z6∣2=7+∑j=16(7−j)(zj+z−j)(z∈T).P(z)=\lvert1+z+\cdots+z^6\rvert^2=7+\sum_{j=1}^{6}(7-j)(z^j+z^{-j})\qquad(z\in\mathbb{T}).

Its value at one is 4949. Since the negative moments are the complex conjugates of the positive ones, (25) and (26) imply

49μ({1})≤∫TP(z) dμ(z)=7⋅3+2∑j=16(7−j)=63.49\mu(\{1\})\le\int_{\mathbb{T}}P(z)\,d\mu(z)=7\cdot3+2\sum_{j=1}^{6}(7-j)=63.

Combining this with (27) yields

u+v≤67.u+v\le\frac{6}{7}.

Substitute (8) and (14) to obtain

bc+b~c~≥47.(28)\frac{b}{c}+\frac{\widetilde{b}}{\widetilde{c}}\ge\frac{4}{7}. \tag*{(28)}

But (13) bounds the second ratio by the first. Therefore b/c≥2/7b/c\ge2/7, which is (23). □\square

Remark 4.2 (The role of the six indicators). The moment argument uses only the fields containing the indicators, not their values, signs or sizes. Its atom bound is sharp for the stated moment information: the measure with mass 9/79/7 at 11 and mass 2/72/7 at each nontrivial seventh root has mass 33 and first six moments 11. The choice of order twelve is also natural for field disjointness. At order fourteen, Q(ζ14)\mathbb{Q}(\zeta_{14}) already contains a cubic field, so the degree argument used above cannot be continued uniformly.

The only cubic ring and its spherical structure

Proposition 4.3. A cubic rank-three fusion ring admits a complex fusion categorification only if it is isomorphic to K(1,1,1,0)K(1,1,1,0). Every categorification of this ring admits a spherical structure.

Proof. Combine (7) and Proposition 4.1:

4b−8≤c≤72b,henceb≤16.4b-8\le c\le\frac{7}{2}b,\qquad\text{hence}\qquad b\le16.

If one of k,lk,l is zero, interchange X,YX,Y so that l=0l=0. Equation (2) becomes k(k−n)=1k(k-n)=1, giving k=1,n=0k=1,n=0. The character X↦0X\mapsto0, Y↦−1Y\mapsto-1 then has formal codegree 22, contrary to irreducibility of pp. Thus k,l≥1k,l\ge1.

By (5),

3(k2+l2)+m2+n2≤9.3(k^2+l^2)+m^2+n^2\le9.

This forces k=l=1k=l=1. Associativity now gives m+n=1m+n=1. The two resulting choices are interchanged by exchanging XX and YY, so the ring is K(1,1,1,0)K(1,1,1,0).

It remains to descend sphericality to the given categorification, rather than merely exhibit a different spherical categorification. For K(1,1,1,0)K(1,1,1,0), (9) gives

k~,l~,m~∈{1,−1},n~=0.\tilde{k},\tilde{l},\tilde{m}\in\{1,-1\},\qquad\tilde{n}=0.

Signed associativity implies l~m~=1\tilde{l}\tilde{m}=1. Replacing X~\widetilde{X} by l~X~\tilde{l}\widetilde{X} and Y~\widetilde{Y} by k~Y~\tilde{k}\widetilde{Y} makes all four signed parameters equal to the original ones. For these choices of lifts, each signed multiplicity equals the corresponding unsigned multiplicity. Since these are respectively the difference and sum of the two lift multiplicities, the unwanted lift has multiplicity zero in every product. The three chosen lifts therefore form a fusion subcategory of SS.

The restriction of π\pi to this subcategory is a tensor equivalence with CC: it is bijective on simple objects and fully faithful. Transporting its spherical structure gives a spherical structure on CC. This is also the section criterion of [11 Section 5.1(c)].

Spherical descent in the remaining cases

Theorem 4.4. Every complex fusion category of rank three admits a spherical structure.

Proof. The pointed case is weakly integral. For a self-dual ring, the Frobenius–Perron dimension of the category is a formal codegree. If pp splits over Q\mathbb{Q}, this dimension is an integer and again the category is weakly integral. The canonical spherical structure exists in both cases [4 Propositions 8.23–8.24]. If pp has an irreducible quadratic factor, Ostrik’s descent theorem [11 Theorem 5.1] gives a spherical structure on the original category, with no pivotality hypothesis. Proposition 4.3 handles the irreducible cubic.

A two-moment test for the five rings

Let K(k,l,m,n)K(k,l,m,n) be the ring of a spherical rank-three category. Put x=FPdim⁡(X)x=\operatorname{FPdim}(X), y=FPdim⁡(Y)y=\operatorname{FPdim}(Y) and D=1+x2+y2D=1+x^2+y^2. We may use these positive dimensions in the twist identities, as in [11 Lemma 3.4]. Indeed, if DD is an integer, use the canonical positive spherical structure [4 Propositions 8.23–8.24]. Otherwise the spherical dimension field contains the irrational codegree DD by [11 Corollary 2.15]. The spherical and Frobenius–Perron dimension characters therefore both have Galois orbits of size at least two. Among three characters there is room for only one such orbit. An automorphism carrying the first character to the second carries every central dimension to the positive dimension of its underlying object and every twist to another root of unity. We use these conjugated scalar identities below.

Proposition 5.1. For u=∑ifi−2u = \sum_i f_i^{-2} as in (8),

100u≤34+66+18(x+y)D.(29)100u \le34 + \frac{66 + 18(x+y)}{D}. \tag*{(29)}

Proof. Give θA\theta_A mass aA2/D2a_A^2/D^2 for each central simple AA. By (17) this is a probability measure. The constituents of I(1)I(1) supply mass uu at one by (11). Its first moment is 1/D1/D by (18), since ν1(V)=[V:1]\nu_1(V)=[V:1]. Its second moment is real and at most (1+x+y)/D(1+x+y)/D: the second indicator of a self-dual simple is 11 or −1-1, the trace of an involution on Hom⁡(1,V⊗2)\operatorname{Hom}(1,V^{\otimes2}). Use q(z)=3+4z+3z2q(z)=3+4z+3z^2 in (20). On the unit circle,

∣q(z)∣2=34+24(z+z−1)+9(z2+z−2).\lvert q(z)\rvert^2=34+24(z+z^{-1})+9(z^2+z^{-2}).

Substitution gives (29).

Proof of Theorem 1.1. Sphericality follows from Theorem 4.4. The non-self-dual case gives Z[C3]\mathbb{Z}[C_3]. For a self-dual ring, we first show that b≥22b\ge22 and D≥16D\ge16 cannot hold together. By (7) and (8),

u≥b−42b−4≥920(b≥22).u\ge\frac{b-4}{2b-4}\ge\frac{9}{20}\qquad(b\ge22).

This gives 100u−34≥11100u-34\ge11. But x+y≤2(D−1)x+y\le\sqrt{2(D-1)}, so for D≥16D\ge16 the right side of (29) minus 3434 is smaller than 66/16+18/8<1166/16+18/\sqrt{8}<11.

If k,l>0k,l>0 and max⁡(k,l)≥2\max(k,l)\ge2, (5) gives b≥22b\ge22. From xy=kx+lyxy=kx+ly we obtain x>lx>l, y>ky>k and (x−l)(y−k)=kl(x-l)(y-k)=kl. Hence

x+y≥(k+l)2≥3+22>11/2,D≥1+12(x+y)2>16.x+y\ge(\sqrt{k}+\sqrt{l})^2\ge3+2\sqrt{2}>11/2,\qquad D\ge1+\frac{1}{2}(x+y)^2>16.

This is impossible. The choice k=l=1k=l=1 forces m+n=1m+n=1 and gives K(1,1,1,0)K(1,1,1,0) after interchanging X,YX,Y.

If kl=0kl=0, arrange l=0l=0. Then (2) forces k=1,n=0k=1,n=0, so y=1y=1, x=(m+m2+8)/2x=(m+\sqrt{m^2+8})/2, D=2+x2D=2+x^2 and

b=10+m2,c=4b−8,u=b−42b−4.b=10+m^2,\qquad c=4b-8,\qquad u=\frac{b-4}{2b-4}.

As mm increases, uu increases and (84+18x)/(2+x2)(84+18x)/(2+x^2) decreases: the latter has derivative of sign 36−168x−18x2<036-168x-18x^2<0 for x≥1x\ge1. It therefore suffices to test m=3m=3 to exclude all m≥3m\ge3. Substitution in (29) would give

1534≤1120−917340,\frac{15}{34}\le\frac{11}{20}-\frac{9\sqrt{17}}{340},

or equivalently

917≤37.9\sqrt{17}\le37.

This fails since 81⋅17=1377>1369=37281\cdot17=1377>1369=37^2. This endpoint explains the coefficients in Proposition 5.1: among 1+az+z21+az+z^2 with a≥0a\ge0, the two-moment upper bound for the atom there is minimized at a=4/3a=4/3.

This leaves m=0,1,2m=0,1,2. The unitary models in Table 1 realize each surviving ring. The three near-group rings have two invertible basis elements and are distinguished by mm. The pointed ring has three invertibles and the even SU(2)5SU(2)_5 ring has one.

In the SU(2)5SU(2)_5 row the nontrivial basis elements X,YX,Y have highest weights 2,42,4, respectively. In the near-group rows YY is the nontrivial invertible and XX is the remaining simple. Ostrik’s tensor-equivalence classification [11 Theorem 1.1] now applies to every complex rank-three fusion category by Theorem 4.4.

References

  1. [1]B. Bartlett, Fusion categories via string diagrams, Commun. Contemp. Math. 18 (2016), no. 5, 1550080, arXiv:1502.02882v2.arxiv.org/abs/1502.02882
  2. [2]J. Bion-Nadal, Subfactor of the hyperfinite II₁ factor with Coxeter graph E₆ as invariant, J. Operator Theory 28 (1992), 27–50.
  3. [3]C. Edie-Michell, M. Izumi and D. Penneys, Classification of ℤ/2ℤ-quadratic unitary fusion categories, with an appendix by R. Johnson, S.-H. Ng, D. Penneys, J. Roat, M. Titsworth and H. Tucker, Quantum Topol. 16 (2025), 493–568, doi:10.4171/QT/201.
  4. [4]P. Etingof, D. Nikshych and V. Ostrik, On fusion categories, Ann. of Math. (2) 162 (2005), no. 2, 581–642; arXiv:math/0203060 (updated version).arxiv.org/abs/math/0203060
  5. [5]T. J. Hagge and S.-M. Hong, Some non-braided fusion categories of rank three, Commun. Contemp. Math. 11 (2009), no. 4, 615–637; arXiv:0704.0208.
  6. [6]M. Müger, From subfactors to categories and topology II. The quantum double of tensor categories and subfactors, J. Pure Appl. Algebra 180 (2003), 159–219, doi:10.1016/S0022-4049(02)00248-7.DOI
  7. [7]S.-H. Ng and P. Schauenburg, Higher Frobenius–Schur indicators for pivotal categories, in Hopf Algebras and Generalizations, Contemp. Math. 441, Amer. Math. Soc., Providence, RI, 2007, 63–90; doi:10.1090/conm/441/08500.DOI
  8. [8]S.-H. Ng and P. Schauenburg, Frobenius–Schur indicators and exponents of spherical categories, Adv. Math. 211 (2007), no. 1, 34–71; doi:10.1016/j.aim.2006.07.017.DOI
  9. [9]V. Ostrik, Pre-modular categories of rank 3, Mosc. Math. J. 8 (2008), 111–118, arXiv:math/0503564.DOI
  10. [10]V. Ostrik, On formal codegrees of fusion categories, Math. Res. Lett. 16 (2009), no. 5, 895–901; arXiv:0810.3242.arxiv.org/abs/0810.3242
  11. [11]V. Ostrik, Pivotal fusion categories of rank 3, with an appendix written jointly with D. Nikshych, Mosc. Math. J. 15 (2015), no. 2, 373–396; arXiv:1309.4822.
  12. [12]D. Tambara and S. Yamagami, Tensor categories with fusion rules of self-duality for finite abelian groups, J. Algebra 209 (1998), 692–707, doi:10.1006/jabr.1998.7558.DOI
  13. [13]J. Thornton, On braided near-group categories (2011), arXiv:1102.4640.arxiv.org/abs/1102.4640

Paper details

vlnc:W_6rhzcxjd
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Mathematics

OpenAI ChatGPT GPT-6 Astra; GPT-6.2 Astra

View statement
Looking for related papers…
Contents