Characters and a Verlinde-type formula for symmetric Hopf algebras. (Q2378607)

Let \(H\) be a finite-dimensional symmetric Hopf algebra over an algebraically closed field \(k\), \(t\) a centralizing form in \(H^*\) (i.e., \(t(ab)=t(ba)\), all \(a,b\) in \(H\)) such that \(t\) generates \(H^*\) as cyclic right \(H\)-module (i.e., via left translations). The Higman ideal \(\text{Hig}(H)\) of \(H\) is the image of the trace map sending \(a\) in \(H\) to the sum \(b_iaa_i\) for \(\{a_i\}\) and \(\{b_i\}\) dual bases with respect to the bilinear form \((a,b)=t(ab)\) on \(H\). Let \(I(H)\) be the right \(\text{Hig}(H)\)-module generated by \(t\) (again by left translations). \(I(H)\) is the set of characters of projective \(H\)-modules, and is an ideal of the algebra of cocommutative elements of \(H^*\). Its elements are of the form \(p_a\), \(a\) in \(H\), where \(p_a(b)=\text{Trace}(l(a)r(b))\), \(l,r\) left and right multiplications, respectively. We will list some of the main results of this paper. If \(H\) is unimodular, then \(I(H)\) is stable under the transpose of \(H^*\), and if \(k\) has characteristic zero, then \(I(H)\) is non-zero and its dimension is the rank of the Cartan matrix of \(H\). In general, \(\text{Hig}(H)\) is the cyclic left \(H\)-module of \(H\) (under the left adjoint action) generated by an integral \(i\) of \(H\). \(H\) is semisimple if and only if \(i\) is in \(\text{Hig}(H)\). If \(H\) is a factorizable ribbon Hopf algebra and \(k\) has characteristic zero, then \(\text{Hig}(H)\) is stable under the quantum Fourier transform \(F\) introduced by \textit{V. Lyubashenko} and \textit{S. Majid} [J. Algebra 166, No. 3, 506-528 (1994; Zbl 0810.17006)] and is an ideal of the center of \(H\) containing a primitive idempotent. Finally, a Verlinde-type formula is given for factorizable ribbon Hopf algebras over \(k\) of characteristic zero, which generalizes the semisimple case of \textit{E. Verlinde} [Nucl. Phys., B 300, No. 3, 360-376 (1988)] (also discussed in algebraic terms by the authors [Contemp. Math. 433, 115-133 (2007; Zbl 1162.16025)] using \(F\)). In the case considered in the paper under review, the fusion rules arise from the product of characters of irreducible and of indecomposable projective modules, and the diagonalizing matrix is replaced accordingly. The result coincides with the usual formula in the semisimple case.