On Tate's refinement for a conjecture of Gross and its generalization (Q2483705)

Let \(k\) be a global field, let \(S,T\) be disjoint finite sets of places of \(k\) and assume that \(S\) contains all archimedean places if they exist. If \(K/k\) is a finite Abelian extension with Galois group \(G\), unramified outside \(S\) and \(\chi\) is a character of \(G\), then define \[ L_{S,T}(\chi,s)=\prod_{v\in T}(1-\chi(Fr_v)N(v)^{1-s})\prod_{v\not\in S}(1-\chi(Fr_v)N(v)^{-s})^{-1}, \] \(Fr_v\) being the Frobenius element. The Stickelberger element \(\theta_G\) is defined as a unique element of \({\mathbb C}[G]\) satisfying \(\chi(\theta_G)=L_{S,T}(\chi,0)\) for all \(\chi\). \textit{B. Gross} [J. Fac. Univ. Tokyo 35, 177--197 (1988; Zbl 0681.12005)] showed \(\theta_G\in {\mathbb Z}[G]\) and defined a kind of regulator, which we shall denote here by \(R(S,T)\), related to the group \(U(S,T)\) of \(S\)-units of \(k\) congruent to unity \(\bmod\;v\) for \(v\in T\). This regulator is an element of \(I_G^n/I_G^{n+1}\), where \(n\) is the \({\mathbb Z}\)-rank of \(U(S,T)\), and \(I_G\) is the augmentation ideal of \({\mathbb Z}[G]\). The conjecture of Gross states that \(\theta_G\) lies in \(I_G^n\) and one has \(\theta_G\equiv\pm hR(S,T)\pmod{I_G^{n+1}}\), \(h\) being the class-number of the ring of \(S\)-integers of \(k\). This conjecture is known to be true in several cases, in particular if either \(K/k\) is quadratic, or \(k\) is the field of rationals, or \(k\) is a number field and \(S\) consists of archimedean places. Recently \textit{J. Tate} [Contem. Math. 358, 189--192 (2004; Zbl 1072.11087)] formulated a refined version of the conjecture for certain \(l\)-extensions. In this paper the author establishes first the truth of this refinement in the case \(l=2\), and then proceeds with a further generalization, stating a conjecture extending Tate's conjecture to arbitrary cyclic extensions. He is then able to establish his conjecture in certain cases, in particular for cyclic extensions of number fields. This result implies immediately the truth of the Gross's conjecture in this case.