Let C be a near-group fusion category of type Z/4Z+k. We show k∈{0,3,4}, which then gives a complete characterization of near-group fusion categories determined by Z/4Z.
Introduction
The fusion rules of a near-group fusion category are determined by a finite group G and a nonnegative integer k, with the additional simple object ρ satisfying ρ⊗ρ=⨁g∈Gg⊕kρ. We denote this fusion category by G+k. For every finite abelian group G, near-group fusion categories of type G+0 are exactly Tambara–Yamagami categories [12]. A natural question is therefore to determine which positive values of k admit a near-group fusion category of type G+k. Ostrik proved that, for G={e} and G=Z/2Z, the admissible levels are precisely k∈{0,1} and k∈{0,1,2}, respectively [7, 8]. For G=Z/3Z, Larson showed that the admissible levels are exactly k∈{0,2,3,6}[5]. More recently, Schopieray classified the admissible levels for all nontrivial elementary abelian 2-groups [10 Theorem 1.1]: a near-group fusion category of type G+k exists if and only if k=0, or G=Z/2Z and k∈{1,2}, or G=Z/2Z×Z/2Z and k=4.
Let C be a near-group fusion category of type Z/4Z+k, O(C)={1,g,g2,g3,ρ} and fusion rules are defined as
g4=1,gjρ=ρgj=ρ,ρ2=j=0∑3gj+kρ,k∈Z≥0.(1)
The main results of this paper are as follows:
Theorem 1.1 (Theorem 4.3). Let C be a fusion category with fusion rules (1). Then k∈{0,3,4}.
For k=0, these are the Tambara–Yamagami categories TY(Z/4Z,χ,τ), where χ is a nondegenerate symmetric bicharacter on Z/4Z and τ=±21[12]. For k=3, a categorification is provided by Rep(Z/5Z⋊Z/4Z) and Z/4Z acts faithfully on Z/5Z. For k=4, Izumi’s construction gives unitary near-group categories of type Z/4Z+4, two inequivalent unitary categorifications are recorded in [3 Table 2] and are complex conjugate to one another [4 Example 12.12]. Thus the possible values of k in the pseudo-unitary setting are exactly 0, 3, and 4.
The method follows Larson’s use of central induction, twist traces, and sums of roots of unity [5]. We also use the dimension row formula recorded in [11 Lemma 2.1]. Its application here is an integrality obstruction for a fusion matrix. The numerical inequality in [11 Theorem 5.1] by itself does not exclude k=8; this distinction is explained in § 4.
Central induction and preliminary reductions
Standard identities
For objects U,V in a fusion category C, let [U,V]:=dimHom(U,V). We recall the following standard facts about a spherical fusion category and its Drinfeld center Z(C); see [8 Section 2], [6].
Let F:Z(C)→C be the forgetful functor, and let I be its right adjoint. By [2 Proposition 9.2.2], adjunction and the induction formula give
Let D:=dim(C). Notice that the scalar twists θX of simple objects X∈O(Z(C)) are roots of unity, let νn(V) be the n-th Frobenius-Schur indicator of object V∈C, then [6 Theorem 4.1] states that
Tr(θI(V))=D⋅νn(V)(n≥1),(3)
Let g be an invertible object of order n. Then the vector space Hom(1,g⊗n) is one-dimensional. Since νn(g) is the trace of an endomorphism of finite order on this space, it is a root of unity. Hence ∣νn(g)∣=1. Since dimI(g)=D and dimensions of an objects are positive, Equation (3) forces θX=νn(g) for every simple summand X of I(g).
Let C be a fusion category with Grothendieck ring K0(C). Then the algebra K0(C)⊗ZC is semisimple. Given an irreducible representation χ of K0(C)⊗ZC, there is a positive algebraic integer fχ, called the formal codegree of C, satisfying ∑χfχdimχ=1 by [8 proposition 2.10]. For example, FPdim(C) and dim(C) are formal codegrees of C. Moreover, [8 Theorem 2.13] gives the decomposition
I(1)=χ⨁Aχ,dimAχ=fχD,θAχ=1.(4)
Computation of the bi-adjoint functor and twists
Let C be a near-group fusion category of type Z/4Z+k, and let d=dim(ρ). Then d is the positive root of x2−kx−4, i.e. d=2k+k2+16. Set D=dim(C)=4+d2=8+kd and its Galois conjugate D′=8+k(k−d).
Since the fusion ring is commutative, its formal codegrees are the eigenvalues of multiplication by ∑X∈O(C)XX∗=4⋅1+ρ2 on K0(C)⊗ZC[8 Remark 2.11]. With respect to the ordered basis (1,g,g2,g3,ρ), this operator has matrix
R=5111k1511k1151k1115kkkkk8+k2.
Then the formal codegrees of C are precisely the roots of the characteristic polynomial of R, which are D,D′,4,4,4. In particular, Equation (4) means
dimX≤4D,for all simple objects X⊂I(1).(5)
If d is rational, it is an integer, then k=0,3. Assume d is irrational below, then it follows from [8 Theorem A.6(i)] that k=4t.
Proposition 2.1. Let C be a near-group fusion category of type Z/4Z+k, then each I(gj) has exactly five distinct simple summands, for 0≤j≤3.
Proof. For I(1), it is proved by Equation (4). For any h∈{g,g2,g3}, the induction formula gives
F(I(h))=4h⊕(j=0⨁3gj)⊕kρ.(6)
Write I(h)=⨁ZmZZ with mZ∈Z≥0. By definition, we have ∑ZmZ2=5. If I(h) were not multiplicity-free, the only possible decomposition would then be I(h)=X⊕2Y, X≅Y, for some X,Y∈O(Z(C)). Since [I(h),I(1)]=1 and I(1) is multiplicity-free, we have 1=[I(1),X]+2[I(1),Y]. Consequently, X⊂I(1) and Y⊂I(1). In particular, θX=1.
By [6 Theorem 4.1], we know 0=ν1(h)dim(C)=Tr(θI(h))=dim(X)+2θYdim(Y). Hence, dim(X)=∣dim(X)∣=∣2dimY∣=2dim(Y) as C is pseudo-unitary, consequently, θY=−1. Since dimI(h)=D, it follows that D=dim(X)+2dim(Y)=2dim(X). Thus dim(X)=D/2, contradicting Equation (5). This contradiction proves multiplicity-freeness for every h.
Finally, Equation (6) shows that each induction has five distinct simple summands, any two different inductions have exactly one summand in common.
For all X∈C and Z∈Z(C), the adjointness condition gives [I(X),Z]=[X,F(Z)].
Lemma 2.2. There are pairwise non-isomorphic simple objects 1,A,B01,B02,B03∈O(Z(C)) such that
Proof. It follows from Equation (4) and Proposition 2.1 that the decomposition I(1)=1⊕A⊕C1⊕C2⊕C3 is multiplicity-free and
dimA=D′D=4d2=1+td,dimCi=4D=2+td.
Notice that d is irrational, so F(A) contains exactly one invertible summand and t copies of ρ, whereas each F(Ci) contains exactly two invertible summands and t copies of ρ. Meanwhile, Equation (2) implies F(I(1))≅4⋅1⊕(⨁j=03gj)⊕4tρ. Each of g,g2,g3 occurs exactly once in this decomposition. Thus we may label the objects Ci as B01,B02,B03 so that F(B0j)≅1⊕gj⊕tρ(1≤j≤3).
By Proposition 2.1, for any simple object X∈O(Z(C)), every invertible summand of F(X) has multiplicity at most one. Let k=4t with t≥1 as in Section 2. Denote the simple objects of Z(C) by 1,A,Bij(0≤i<j≤3), Pj,Qj(1≤j≤3), Lα(α∈J), where J is a finite index set. Here Bij denotes the unique common simple summand of I(gi) and I(gj); the six objects Bij are shown to be pairwise non-isomorphic in the following proposition.
Proposition 2.3. There are integers p,q,r,s≥0 and γα>0, and η∈{i,−i}, such that p+q=r+s=t. The block above the twist row is the matrix W of the forgetful functor F, with entries WV,Y=[F(Y),V]. Thus each column records the decomposition of F(Y), by adjunction, each row records the decomposition of I(V). In particular, the first four rows determine the decompositions of I(1),I(g),I(g2),I(g3). The last
row records the twists and is not part of W. The final column is repeated for every α∈J, and each θα is a root of unity. Zero entries are omitted.
1
A
B01
B02
B03
B12
B13
P1
Q1
B23
P2
Q2
P3
Q3
Lα
1
1
1
1
1
1
g
1
1
1
1
1
g2
1
1
1
1
1
g3
1
1
1
1
1
ρ
t
t
t
t
t
t
p
q
t
r
s
p
q
γα
θ
1
1
1
1
1
−1
η
−η
−η
−1
1
1
−η
−η
θα
Table 1.
Proof. By Proposition 2.1, I(gj) are multiplicity-free and each have five simple summands, and meet pairwise in exactly one summand. The decomposition of I(1) is obtained in Lemma 2.2, and I(gj) contains a simple summand whose twist is 1. The twist-power constraint from Section 2 gives
θY2=1(Y⊂I(g2)),θY4=1(Y⊂I(g) or I(g3)).
The three pairwise intersections among I(g),I(g2),I(g3) are either three distinct simple objects or a single simple object T common to all three inductions. In the latter case, the invertible summands of F(T) are exactly g,g2,g3. Moreover, I(g) consists of B01, T, and three further simple objects whose restrictions contain g as their only invertible summand. Denote their twists by ξ1,ξ2,ξ3. Since θT2=1 and d∈/Q(i), comparison of coefficients in Tr(θI(g)2)=0 gives 2+3+ξ12+ξ22+ξ32=0, ξℓ2∈{1,−1}, which is impossible. Thus the three intersections are distinct. Denote the common summand of I(gi) and I(gj) by Bij for 1≤i<j≤3. The invertible summands of F(Bij) are exactly gi,gj. Each I(gj) therefore has exactly two remaining summands, denoted Pj,Qj, whose restrictions contain gj as their only invertible summand. This proves the first four rows of the table.
Set ε=θB12∈{1,−1}, η=θB13, ξ=θP1, and ξ′=θQ1. The coefficients of 1 in the second and first twist traces of I(g) give, respectively, 4+2η2+ξ2+(ξ′)2=0, 2+2ε+2η+ξ+ξ′=0. The first equation forces η2=ξ2=(ξ′)2=−1. The real part of the second then gives ε=−1, and its imaginary part gives ξ=ξ′=−η. The same argument for I(g3) yields θB23=−1 and θP3=θQ3=−η. Finally, the coefficient of 1 in the first twist trace of I(g2) is 2−2−2+θP2+θQ2=0, so θP2=θQ2=1.
Let c12,c13,p,q be the multiplicities of ρ in F(B12),F(B13),F(P1),F(Q1), respectively. The ρ coefficient in F(I(g)) and the d-coefficients in the second and first twist traces give
Hence, c12=c13=p+q=t. Duality identifies B23=B12∗ and allows the labelling P3=P1∗,Q3=Q1∗, proving the corresponding entries in the ρ row. Writing r,s for the ρ multiplicities of F(P2),F(Q2), the ρ coefficient of F(I(g2)) gives 3t+r+s=4t. Thus
The restriction of the remaining simple objects Lα (α∈J) do not contain invertible summands, so F(Lα)=γαρ for some γα∈Z>0. Its twist θα is a root of unity. This completes the table.
Notice that ρ is self-dual, so the second Frobenius–Schur indicator ν2(ρ)=±1[6].
Proposition 2.4. Let ν=ν2(ρ)∈{1,−1}, u=pq, v=rs, w=t2+1, N=∑αγα2, Sj=∑αγα2θαj, j=1,2. Then we have
N=8+6t2+4u+2v.(7)
S1=−t2+2v−2tw+η(t2−4u),(8)
S2=−4(t−ν)w−4u+2v.(9)
In particular,
0≤u,v≤4t2,N≤8+215t2.(10)
Proof. By Proposition 2.3, I(ρ)=tA⊕i<jtBi,j⊕p(P1⊕P3)⊕q(Q1⊕Q3)⊕rP2⊕sQ2⊕αγαLα. Equations (2) implies F(I(ρ))=4t(⊕j=03gj)⊕(8+16t2)ρ, which then gives
After simplification, we prove Equations (8) and (9). The Inequalities (10) follow from p+q=r+s=t.
Bounds for sums of roots of unity
We establish the arithmetic estimates for the identities from § 2. We write φ for Euler’s function and μ for the Möbius function, φ(1)=μ(1)=1, and let ζm=exp(2πi/m). For a prime p and a positive integer m, let vp(m) denote the exponent of p in the prime factorization of m.
Normalized Galois averages
Let c>1 be square-free and set K=Q(c). Its conductor is f=c if c≡1(mod4), and f=4c if c≡2,3(mod4). In particular, K⊆Q(ζm) if and only if f∣m.
Choose a cyclotomic field E containing K and all roots under consideration. For w∈E, define its normalized relative trace by
TK(w)=[E:K]TrE/K(w)=∣G∣1σ∈G∑σ(w),G=Gal(E/K).
Obviously, TK(−) is K-linear and satisfies TK(a)=a for a∈K. Moreover, for every τ∈G, τ(TK(w))=∣G∣1∑σ∈G(τσ)(w)=TK(w) and TK(τ(w))=TK(w), so TK(w)∈EG=K.
The normalization makes TK(w) independent of the choice of cyclotomic field containing K and w. Indeed, if E⊆E′ and w∈E, transitivity of the trace gives [E′:K]TrE′/K(w)=[E′:E][E:K]TrE/K([E′:E]w)=[E:K]TrE/K(w). Thus we may compute TK(w) in any cyclotomic field containing K and w. Since K⊂R, complex conjugation belongs to G. Consequently
Since β is a primitive rth root of unity, the classical evaluation of Ramanujan sums [9 Section 1] gives ∑1≤j≤r(j,r)=1βj=μ(r). For every positive integer r, we have ∣μ(r)∣≤1 and φ(r)≥1. Hence φ(r)μ(r)≤φ(r)∣μ(r)∣≤1, with equality if and only if r=1.
Let ℓ,r be positive integers such that f∣ℓ and (ℓ,r)=1, and put K1=Q(ζℓ) and K2=Q(ζr). Let K1K2 denote the compositum of K1 and K2, namely the smallest subfield of C containing both. By the Chinese remainder theorem, Gal(K1K2/Q)≅Gal(K1/Q)×Gal(K2/Q). Since K⊆K1, an automorphism of K1K2 fixes K if and only if its restriction to K1 fixes K. Its restriction to K2 can be any element of Gal(K2/Q). Hence Gal(K1K2/K)≅Gal(K1/K)×Gal(K2/Q). Writing these two factors as G1 and G2, respectively, for u∈K1 and v∈K2, we obtain the formula
We will use this factorization only for coprime cyclotomic factors.
For a root of unity ξ, write TK(ξ)=xξ+yξc, xξ,yξ∈Q, and let m=ord(ξ). If f∤m, then K⊈Q(ξ) and K∩Q(ξ)=Q. Restriction identifies Gal(KQ(ξ)/K) with Gal(Q(ξ)/Q), so the K-orbit of ξ contains all primitive mth roots. Therefore TK(ξ)=φ(m)μ(m)∈Q.
Next assume that f∣m and that vp(m)>vp(f) for a prime factor p of f. These assumptions imply p2∣m and f∣m/p. Every prime divisor of m divides m/p, so a=1+m/p is relatively prime to m. The automorphism σ(ξ)=ξa of Q(ξ) fixes K, because a≡1(modf). Let ω=ξm/p. Then σ(ξ)=ωξ and σ(ω)=ω, since p∣m/p. It follows by induction that σj(ξk)=ωjkξk. For every positive integer k with p∤k, trace invariance gives
In particular, TK(ξ)=0. Combining rationality with this cancellation proves the necessary conditions: if yξ=0, then f∣m and vp(m)=vp(f) for every prime p∣f.
We use the following theorem without proof.
Theorem 3.1 ([5 Theorem 6.2]). Let c>1 be square-free and let a,b∈Z. If roots of unity ξ1,…,ξn satisfy ∑j=1nξj=a+bc, then n≥∣b∣φ(2c).
The following lemma is a direct consequence of [1 Lemma 3.11].
Lemma 3.2. Let c>1 be even and square-free. Suppose that roots of unity ξ1,…,ξn satisfy
Rej=1∑nξj=x+yc,j=1∑nξj2=z+qc,x,y,z,q∈Z.
Then n≥φ(2c)(∣y∣+∣q∣).
A joint estimate over Q(5)
Throughout this subsection, put K=Q(5) and T=TK. Let ζ=ζ5. The extension Q(ζ)/K has degree 2, with complex conjugation as its nontrivial automorphism. Its two orbits of primitive fifth roots are therefore {ζ,ζ−1} and {ζ2,ζ−2}. Squaring interchanges them, so for every primitive fifth root β there is a sign ε∈{1,−1} such that T(β)=4−1+ε5, T(β2)=4−1−ε5. The two signs are linked by the choice of β.
Lemma 3.3. Let ξ1,…,ξn be roots of unity such that
Rej=1∑nξj=x+y5,j=1∑nξj2=z+q5,x,y,z,q∈Z.
Then n≥−x−y−4q.
Proof. For a root of unity ξ, write T(ξ):=xξ+yξ5, T(ξ2):=zξ+qξ5, ψ(ξ):=−xξ−yξ−4qξ. We first prove ψ(ξ)≤1 by considering s=ord(ξ).
If 5∤s, then neither s nor ord(ξ2) is divisible by the conductor 5. Both averages are rational, so yξ=qξ=0. Moreover, xξ is the real part of an average of roots of unity, and hence xξ≥−1. Thus ψ(ξ)=−xξ≤1. If 25∣s, the cancellation argument in Subsection 3.1 and Equation (11), with p=5 and k=1,2, gives T(ξ)=T(ξ2)=0. Hence ψ(ξ)=0 in this case.
In the remaining case 5∣s but 25∤s. Removing the power of 2 and the single factor 5 gives s=2a5r, a=v2(s)≥0, (r,10)=1. Suppose first that a≥3. The integers b1=1+s/2 and b2=1+s/4 are odd and are 1 modulo every odd prime divisor of s. Thus they are relatively prime to s. They are also 1 modulo 5, so the automorphisms σi(ξ)=ξbi fix K. Since ξs/2=−1, we have σ1(ξ)=−ξ, σ2(ξ2)=−ξ2. Since T(σ(w))=T(w) for every automorphism σ fixing K, we obtain
It remains to treat a∈{0,1,2}. Let ℓ=2a5. Since (ℓ,r)=1, choose integers b,e satisfying br+eℓ=1 and set δ=ξbr, γ=ξeℓ. Then ξ=δγ, and the congruences br≡1(modℓ) and eℓ≡1(modr) show that δ and γ are primitive roots of orders ℓ and r, respectively. For r=1, this means γ=1. Because r is odd, squaring permutes its primitive roots. Thus γ and γ2 have the same rational average λ=μ(r)/φ(r). The coprime factorization from Subsection 3.1 gives
T(ξ)=λT(δ),T(ξ2)=λT(δ2),ψ(ξ)=λψ(δ).
The last identity follows by comparing the rational and 5 coefficients in the first two identities. Furthermore, −21≤λ≤1. Indeed, ∣μ(r)∣≤1. If μ(r)=−1, then the odd integer r is greater than 1. Hence 1 and r−1 are distinct and both coprime to r, so φ(r)≥2.
We now compute ψ(δ) for the three possible orders. For ℓ=5, take δ=β in the fifth-root formulas above. For ℓ=10, write δ=−β with β primitive of order 5; then T(δ)=−T(β) and T(δ2)=T(β2). For ℓ=20, put η=δ5 and β=δ−4. These roots have orders 4 and 5, satisfy δ=ηβ, and have η2=−1. The rational average of η is zero, so coprime factorization yields
T(δ)=0,T(δ2)=−T(β2)=41+ε5.
Substitution into the definition of ψ gives the following table:
ℓ
xδ
yδ
qδ
ψ(δ)
5
−41
4ε
−4ε
41+3ε
10
41
−4ε
−4ε
4−1+5ε
20
0
0
4ε
−ε
Table 2.
( ε∈{1,−1}).
In each row the sign is determined by its associated fifth root β. These values give −3/2≤ψ(δ)≤1. If λ≥0, then ψ(ξ)=λψ(δ)≤λ≤1. If λ<0, multiplication reverses the lower bound on ψ(δ), and hence ψ(ξ)=λψ(δ)≤−23λ≤43<1. This proves ψ(ξ)≤1 for every root of unity.
Finally, use one common cyclotomic field for all the given roots. The identity T(w)=T(Rew) and the hypotheses give
Since 1,5 are linearly independent over Q, comparison of coefficients gives ∑jxξj=x, ∑jyξj=y, ∑jqξj=q. Therefore, −x−y−4q=∑j=1n(−xξj−yξj−4qξj)=∑j=1nψ(ξj)≤n.
Lemma 3.4. If there are roots of unity ξj (1≤j≤16) such that ∑j=116ξj=−4−45, then their multiset consists of eight copies of ζ52 and eight copies of ζ53. Consequently, ∑j=116ξj2=−4+45.
Proof. Let T(ξ)=xξ+yξ5. We first identify all roots for which yξ=−1/4. If 5∤ord(ξ), then yξ=0; if 25∣ord(ξ), then T(ξ)=0. Otherwise ord(ξ)=5r with (r,5)=1. Choose integers u,v such that ur+5v=1 and let α=ξur, β=ξ5v. Then ξ=αβ, with α and β primitive of orders 5 and r. The coprime factorization in Subsection 3.1 and the fifth-root calculation in Subsection 3.2 give
T(ξ)=T(α)φ(r)μ(r)=φ(r)μ(r)(4−1+ε5),ε∈{1,−1}.
Here r need not be odd. In particular, yξ=4φ(r)εμ(r)≥−41. Equality requires ∣μ(r)∣=φ(r)=1, which is possible only when r=1 or r=2. For r=1, equality requires ε=−1, giving the roots ζ52,ζ53 and rational coefficient xξ=−1/4. For r=2, we have μ(2)=−1, so equality requires ε=1, giving the roots −ζ5,−ζ54 and rational coefficient xξ=1/4.
Averaging the assumed identity and comparing coefficients yields ∑j=116yξj=−4, ∑j=116xξj=−4. The first equality is a sum of sixteen terms each at least −1/4, so every term equals −1/4. Let a and b count the roots in {ζ52,ζ53} and {−ζ5,−ζ54}, respectively, including multiplicities. Then the number of roots and the rational coefficients give a+b=16, −4a+4b=−4. Thus a=16, b=0, and every original root is either ζ52 or ζ53. Let their multiplicities be a2,a3. The original sum is real, and conjugation exchanges these two roots. Subtracting the conjugate sum gives (a2−a3)(ζ52−ζ53)=0. Since the roots are distinct, a2=a3=8. Finally, ∑j=116ξj2=8(ζ54+ζ56)=8(ζ54+ζ5−1)=−4+45.
Proof of the main theorem
Proposition 4.1. Let C be a near-group fusion category of type Z/4Z+k, k=4t. Then t≤2.
Proof. Write t2+1=m2c, where m∈Z≥1 and c is square-free. For every odd prime divisor p of c, we have t2≡−1(modp) and p∤t. Fermat’s little theorem gives 1≡tp−1=(−1)(p−1)/2(modp). Thus (p−1)/2 is even, so p≡1(mod4). If c is odd, it is a product of such primes, and hence c≡1(mod4). By Proposition 2.4, we have
ReS1=−t2+2v−2tmc,S2=−4u+2v−4(t−v)mc.(12)
Case 1:c is odd and c=5. Every prime divisor of c is congruent to 1 modulo 4. Since c>1 is square-free and c=5, it has a prime divisor p0=5, and hence p0≥13. Moreover, (p−1)/p>1 for every prime divisor p of c. Therefore, cφ(2c)=∏p∣cpp−1≥p0p0−1≥1312. Note ν≤1, applying Theorem 3.1 to S2, we get
N≥4(t−1)mφ(2c)≥1348(t−1)t2+1>13t(t−1).
An elementary calculation shows 13t(t−1)>8+15t2/2 for t≥3, which contradicts Inequality (10).
Case 2:c is even and c=2. Write c=2h with h>1. Then the smallest possible odd prime divisor of h is 5, so cφ(2c)=2∏p∣hpp−1≥108>25. Apply Lemma 3.2 to (12), we obtain
N≥(6t−4ν)mφ(2c)≥(6t−4)mφ(2c).(13)
Consequently, N>25(6t−4)t=15t2−10t. For t≥3, this is larger than 8+15t2/2, impossible.
Case 3:c=2. Inequality (13) implies N≥2(6t−4)t2+1>2(6t−4)t. For every t≥7, 2(6t−4)t>8+15t2/2. Indeed, the difference is (62−15/2)t2−42t−8; at t=7 it is 2662−751/2>0, and its derivative is positive for t≥7. Thus t≤6. None of t=3,4,5,6 has t2+1 equal to twice a square.
Case 4:c=5. Lemma 3.3 gives N≥t2−2ν+(18t−16ν)m. Using ν≤1, the formula for N, and the bounds on u,v, we obtain
(18t−16)m≤8+5t2+4u+4v≤8+7t2=35m2+1.(14)
If m≥2, then t2=5m2−1>4m2. Since t,m are integers, this implies t≥2m+1. But then
(18t−16)m≥(18(2m+1)−16)m=36m2+2m>35m2+1,
contradicting Inequality (14). Thus m=1 and t=2.
In summary, we conclude that t≤2.
We show that k=8 can be excluded in the following proposition.
Proposition 4.2. The case k=8 is impossible.
Proof. Suppose k=8, and retain the notations of § 2. Set d=4+25, a=1+2d, b=1+d, D=8b. Then d2=4a, b2=5a, D2=320a, and a+a−1=18. Also u=pq, v=rs∈{0,1}, since p+q=r+s=2. Let s denote the normalized S-matrix of the center Z(C), with s1,X=(dimX)/D.
Case 1:v=1. If u=0, there is a central invertible object Z with F(Z)=g, and hence F(Z2)=g2. Since Z2 is simple, the restriction table forces it to be either P2 or Q2. However, r=s=1 gives F(P2)=F(Q2)=g2⊕φ, a contradiction. Thus u=v=1.
For the nontrivial automorphism σ of Q(5), the dimension-row formula [11 Lemma 2.1] gives sA,X=σ(dimX/D). Apart from 1,A, the dimensions of simple objects of Z(C) are b, 2b, or γαd. Since σ(b/D)=b/D and σ(d/D)=−d/D, it follows that sA,A=D1, sA,X2=s1,X2(X∈/{1,A}). Let P be a simple object of dimension b. Then sP,1=sP,A=1/8. The Verlinde formula and orthogonality of the s1,− and sP,− give NA,AP=∑Xs1,XsA,X2−s1,X2sP,X=8Da2−1+a−1−a=8D16(a+1)=21, where the last equalities use a+a−1=18 and D=4(a+1), it is a contradiction.
Case 2:v=0. Relabel the simple objects so that r=0, and let h=P2. Then F(h)=g2, θh=1, and h2=1, since the unit is the only central simple object restricting to 1. For each simple object X, define εX by cX,h∘ch,X=εXidh⊗X, where c denotes the braiding. The hexagon and balancing equation give εX∈{1,−1}, θh⊗X=εXθX, sh,X=εXDdimX. Since h=1, orthogonality of sh,− and s1,− implies ∑εX=−1dim(X)2=2D2=160a. Tensoring by h maps each invertible summand gj of F(X) to gj+2, with exponents taken modulo 4. The table of Proposition 2.3 shows that it preserves the twists of objects whose restrictions contain exactly one invertible summand. It also fixes B02 and B13, since each is uniquely determined by the invertible summands of its restriction. All these objects have positive sign. Tensoring by h exchanges B01 with B23 and B03 with B12, and the two objects in each pair have opposite twists. Thus, among the previously determined objects, precisely these four have negative sign, and each has dimension 2b.
The preceding dimension count and the identity N=32+4u of Equation (7) give
N−=d2160a−4(2b)2=20,N+=12+4u≤16.(15)
Tensoring by h pairs the negative-sign Lα with equal multiplicities γα and opposite twists. Indeed, g2ρ=ρ, and a fixed point would have θLα=−θLα. Hence S1,−=0, and the formula for S1 gives
S1,+=S1=−4−45+4η(1−u).
Applying Theorem 3.1 to S1,++S1,+=−8−85 gives 2N+≥8φ(10)=32. Thus N+=16 and u=1. Lemma 3.4 now applies to S1,+=−4−45 and gives S2,+=−4+45. Using the formula for S2 in Proposition 2.4 with u=1 and v=0, and subtracting the preceding expression for S2,+, we obtain S2,−=S2−S2,+=−4(3−ν)5. This is a sum of N−=20 roots of unity, counted with multiplicity. Theorem 3.1 therefore implies 20≥4(3−ν)φ(10)=16(3−ν)≥32, a contradiction. □
Proposition 4.1 and Proposition 4.2 together imply the following theorem.
Theorem 4.3. There is a near-group fusion category of type Z/4Z+k if and only if k∈{0,3,4}.
Acknowledgements
Z. Yu is supported by NSFC (no.12571041) and Qinglan Project of Yangzhou University. The authors acknowledge the use of AI as an auxiliary tool during the preparation of this manuscript, including for mathematical discussion and editorial assistance. All mathematical content was independently verified by the authors, who take full responsibility for the final manuscript.
References
Yue Meng Email: [email protected] School of Mathematics, Yangzhou University, Yangzhou 225002, China
Zhiqiang Yu Email: [email protected] School of Mathematics, Yangzhou University, Yangzhou 225002, China
References
[1]J. Dong, S. Palcoux and K. Wang, Grothendieck rings of near-Tambara–Yamagami fusion categories, arXiv:2610.35574, 2026.arxiv.org/abs/2610.35574
[2]P. Etingof, S. Gelaki, D. Nikshych and V. Ostrik, Tensor categories, Mathematical Surveys and Monographs, vol. 205, American Mathematical Society, Providence, RI, 2015.
[3]D. Evans and T. Gannon, Near-group fusion categories and their doubles, Adv. Math. 255 (2014), 586–640.arxiv.org/abs/1208.1500
[4]M. Izumi, A Cuntz algebra approach to the classification of near-group categories, in Proceedings of the 2014 Maui and 2015 Qinhuangdao Conferences in Honour of Vaughan F. R. Jones’ 60th Birthday, Proc. Centre Math. Appl. Austral. Nat. Univ., vol. 46, Australian National University, Canberra, 2017, pp. 222–343.
[5]H. Larson, Pseudo-unitary non-self-dual fusion categories of rank 4, J. Algebra 415 (2014), 184–213.arxiv.org/abs/1401.1879
[6]S.-H. Ng and P. Schauenburg, Frobenius–Schur indicators and exponents of spherical categories, Adv. Math. 211 (2007), no. 1, 34–71.DOI
[7]V. Ostrik, Fusion categories of rank 2, Math. Res. Lett. 10 (2003), no. 2-3, 177–183.DOI
[8]V. Ostrik, Pivotal fusion categories of rank 3, with an appendix jointly with D. Nikshych, Mosc. Math. J. 15 (2015), no. 2, 373–396.
[9]N. Robles and A. Roy, Moments of averages of generalized Ramanujan sums, Monatsh. Math. 182 (2017), 433–461.DOI
[10]A. Schopieray, Categorification of integral group rings extended by one dimension, J. Lond. Math. Soc. (2) 108 (2023), no. 4, 1617–1641.arxiv.org/abs/2208.07319
[11]A. Schopieray, The dimension block of a spherical fusion category, preprint (2026), arXiv:2609.35287v1.arxiv.org/abs/2609.35287
[12]D. Tambara and S. Yamagami, Tensor categories with fusion rules of self-duality for finite abelian groups, J. Algebra 209 (1998), no. 2, 692–707.DOI