Global construction of associated orders in complex multiplication (Q2568208): Difference between revisions

The setting of this paper is as follows: \({\mathfrak p}\) is a degree one prime in an imaginary quadratic field \(K\), \(E\) is an elliptic curve with complex multiplication by \(O_K\) defined over some extension \(L/K\), and one studies ``principal homogeneous \(G_m\)-spaces'' obtained from a point \(P\) on \(E\) with order \({\mathfrak p}^{m+r}\). Here \(G_m\) is the set of \({\mathfrak p}^m\)-torsion points on \(E\), and \(E\) is supposed to have good reduction at \({\mathfrak p}\). To this setup one associates a certain \(O_L\)-algebra \(\widetilde{{\mathfrak O}}_P\), and a subalgebra \({\mathfrak A}\) of \({\mathbb Q}^c[G_m]\), which acts on \(\tilde{\mathfrak O}_P\) and is the so-called associated order. The author proves global results concerning algebra generators and Galois generators of \(\widetilde {\mathfrak O}_P\). (By a Galois generator we mean a generator of \(\widetilde {\mathfrak O}_P\) as a module over the associated order.) The latter result can also be interpreted in terms of integral Galois module structure of extensions of ray class fields \(K_{\mathfrak q} {\mathfrak p}^{r+m} / K_{\mathfrak q}{\mathfrak p}^{r}\), under the restriction that \(r\geq m\). In this way, the author obtains improvements upon earlier results of \textit{A. Srivastav} and \textit{M. J. Taylor} [Invent. Math. 99, No. 1, 165--184 (1990; Zbl 0705.14031)] and himself [J. Number Theory 77, No. 1, 97--121 (1999; Zbl 0953.11034)]. The basic algebraic results concerning \(\tilde{\mathfrak O}_P\) and \({\mathfrak A}\) were already shown by \textit{M.~Taylor} [Ill. J. Math. 32, No. 3, 428--452 (1988; Zbl 0631.14033)] for \(E\) having good reduction everywhere. The main ingredients of the proofs in the paper under review are: division values of elliptic functions; resolvents; and calculation of discriminants. For the construction of a global Galois generator of \(\widetilde{\mathfrak O}_P\) over the order \({\mathfrak A}\), the starting point is a normalised \(\sigma\)-function, whereas a normalised Weierstrass \(\wp\)-function is used in order to obtain an algebra generator. The paper is very clearly written and contains rather general results.