On the Hopf-Schur group of a field. (Q934049)

Let \(L\) be an arbitrary field, \(\text{Br}(L)\) the Brauer group of \(L\), and \(C\) a (finite-dimensional) central simple \(L\)-algebra. We say that \(C\) is a Hopf-Schur \(L\)-algebra, if it is a homomorphic image of a finite-dimensional Hopf algebra over a subfield \(k\) of \(L\). The Hopf-Schur group \(\text{HS}(L)\) is defined to be the subgroup of \(\text{Br}(L)\) generated by (and in fact consisting of) similarity classes of Hopf-Schur \(L\)-algebras. The paper under review shows that if \(C\) is similar to a tensor product of cyclic \(L\)-algebras (conjecturally, this condition always holds), then \(C\) is a homomorphic image of a finite-dimensional Hopf \(L\)-algebra \(X\). Moreover, if the characteristic \(\text{char}(L)\) does not divide the dimension \([C:L]\), then \(X\) may be taken to be semisimple. For the proof of these results, the authors show that the twisted group \(L\)-algebra \(L^\alpha G\) is a homomorphic image of a finite-dimensional Hopf \(L\)-algebra \(A\), whenever \(G\) is a finite group and \(\alpha\) lies in the \(2\)-cocycle group \(Z^2(G,L^*)\); the algebra \(A\) turns out to be semisimple if and only if \(\text{char}(L)\) does not divide the order of \(G\). For the same purpose, they also prove that every finite Abelian extension \(M\) of \(L\) is a homomorphic image of a commutative semisimple Hopf \(L\)-algebra \(H\) of dimension \(2[M:L]\), which is an \(M\)-form of the Hopf \(L\)-algebra \(\text{Fun}(\mathbb{Z}/2\mathbb{Z}\times G(M/L),L)\), where \(G(M/L)\) is the Galois group of \(M/L\) and the product is semidirect with \(\mathbb{Z}/2\mathbb{Z}\) acting on \(G(M/L)\) by inversion. The proof of the main results exhibits the fact that \(\text{HS}(L)\) includes the projective Schur group \(\text{PS}(L)\). As pointed out by the authors, it is known that \(\text{HS}(F(X))\neq\text{PS}(F(X))\) and \(\text{HS}(F)=\text{PS}(F)\), provided that \(F\) is a global field and \(X\) is a transcendental element over \(F\).