Noether's problem over an algebraically closed field (Q798390)

Let \(F\) be a field, \(G\) a finite group, \(\{x(g)\mid g\in G\}\) a set of indeterminates. Let \(G\) act on the rational field \(K=F(x(g)\mid g\in G\) by \(g(x(h))=x(gh)\), and let \(F(G)\) be the fixed field of \(K\). Is \(F(G)\) rational over \(F)\)? This is Noether's problem. It arose from an attempt to realize \(G\) as the Galois group of a field extension of \(F\) [see \text it{R. G. Swan} in the book ''Emmy Noether in Bryn Mawr'', Proc. Symp., Bryn Mawr/USA 1982, 21--40 (1983; Zbl 0538.12012)]. A result of \text it{E. Fischer} [Gött. Nachr. 1915, 77--80 (1915; JFM 45.0200.01)] shows that \(F(G)\) is rational over \(F\) if \(G\) is abelian of exponent \(n\) and \(F\) contains a primitive \(n\)-th root of unity, solving the abelian Noether problem or such fields. In this paper the author finds a non-abelian group \(G\) of order \(q^9\) for any prime \(q\) so that if \(F\) is algebraically closed of characteristic not \(q\), then \(F(G)\) is not rational over \(F\). The proof makes essential use of results in the author's previous paper [1] [Isr. J. Math. 47, 165--215 (1984; Zbl 0546.14013)]. In fact, the author shows that \(F(G)/F\) is not retract rational, in outline as follows: Let \(K=F(a,b,c,d)\) be purely transcendental, let \(L\) be the generic splitting field for five given central simple \(K\)-algebras of degree \(q\), chosen so that the Brauer group of \(L\) is not zero. Then \(L\) is rational over the quotient field \(M\) of a certain affine ring \(R\) over which there is a Galois extension \(S/R\) which is locally projective and densely representing in the terminology of [1], where \(G\) is a central extension of an elementary abelian \(q\)-group of rank 4 by an elementary abelian \(q\)-group of rank 5. \(M\) is retract rational if and only if \(G\)-Galois extensions of \(F\) lift to local rings \((T,M)\) with \(T/M=F\). But since \(\mathrm{Br}(L)\ne 0\), \(L\) is not retract rational, so \(M\) is not, so \(G\)-Galois extensions do not lift. But then by another result in [1], \(F(G)\) is not retract rational over \(F\), which implies that \(F(G)/F\) is not rational.