On normal integral bases in ray class fields over imaginary quadratic fields

From MaRDI portal





The question which motivated this paper is the following: Given an imaginary quadratic field \(K\) and a prime ideal \({\mathfrak p}\) of \(K\), does the ray class field \(K({\mathfrak p})\) have normal integral basis over \(K(1)\), the Hilbert class field? This is an instance of the ``relative tame problem in Galois module structure. Thanks to a counterexample of \textit{E. J. Gómez Ayala} and \textit{R. Schertz} [J. Number Theory 44, 41-46 (1993; Zbl 0777.11046)] using quadratic subextensions, one knows the answer to the problem is no; however this seemed to be due to a peculiarity of the number 2, as witnessed by positive results of \textit{E. J. Gómez Ayala} [Acta Arith. 72, 375-383 (1995; Zbl 0857.11060)] for cases where the degree of the extension is three. Thus a positive answer, under suitable restrictions or modifications, seemed a possibility. In the present paper it is shown, however, that the answer to the initial question is a solid no in general, and remains so even after relaxing somewhat the notion of a normal integral basis. A systematic way of finding counterexamples is presented. This involves some descent theory, manipulation of Stickelberger ideals, and some explicit calculations in class groups, which were done by the program package PARI.











This page was built for publication: On normal integral bases in ray class fields over imaginary quadratic fields

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3130219)