Blocks inequivalent to their Frobenius twists. (Q2459984)

Let \(k\) be an algebraically closed field of characteristic \(p>0\), and let \(q=p^a\) be a power of \(p\). For a \(k\)-algebra \(A\), one can define a new \(k\)-algebra \(A^{(q)}\) which, as a ring, is the same as \(A\), but the structural map \(k\to A^{(q)}\) is given by \(\lambda\mapsto\lambda^{1/q}1_A\). Then \(A^{(q)}\) is called a Frobenius twist of \(A\). The authors construct a solvable group \(G\) and a (non-principal) block \(B\) of \(kG\) such that \(B\) is not Morita equivalent to \(B^{(p)}\) (but to \(B^{(p^2)}\)). Moreover, \(B\) has elementary Abelian normal defect groups and only one isomorphism class of simple modules, and its basic algebra is a quantum complete intersection. Such a block \(B\) is not even derived equivalent to \(B^{(p)}\). At present, no example of a block \(B\) seems to be known which is not Morita equivalent to \(B^{(p^2)}\). The results in this paper are related to \textit{R. Kessar}'s reformulation of Donovan's conjecture [Arch. Math. 82, No. 5, 391-394 (2004; Zbl 1058.20011)].