On the values of Artin \(L\)-series at \(s=1\) and annihilation of class groups (Q2840499)

Let \(L/K\) be a finite abelian Galois extension of number fields with Galois group \(G\). Using values of Artin \(L\)-series at \(s=0\) one can construct certain Stickelberger elements in the group algebra \(\mathbb Q[G]\). A classical conjecture of Brumer asserts that these Stickelberger elements can be used to construct annihilators of the class group of \(L\). It is known that this conjecture follows from an appropriate special case of the equivariant Tamagawa number conjecture (ETNC) as formulated by \textit{D. Burns} and \textit{M. Flach} [Doc. Math., J. DMV 6, 501--570 (2001; Zbl 1052.11077)].NEWLINENEWLINEUsing leading terms of Artin \(L\)-series at \(s=1\) and a certain \(p\)-adic regulator map, \textit{D. Solomon} [J. Number Theory 128, No. 1, 105--143 (2008; Zbl 1195.11149)] defined a \(\mathbb Z_p[G]\)-submodule \(\mathfrak S_{L/K}\) of \(\mathbb Q_p[G]\). Solomon conjectures that \(\mathfrak S_{L/K}\) actually belongs to \(\mathbb Z_p[G]\) and annihilates the \(p\)-part of the class group.NEWLINENEWLINENEWLINEThe aim of this article is twofold. First, the authors generalize Solomon's conjecture to arbitrary, not necessarily abelian Galois extensions. Second, it is shown that this conjecture also follows from an appropriate special case of the ETNC.NEWLINENEWLINENEWLINEIn this more general context it is no longer expected that \(\mathfrak S_{L/K}\) is always integral. However, the authors show that the ETNC implies that \(\mathfrak S_{L/K}\) belongs to the noncommutative Fitting invariant (introduced by the reviewer [J. Algebra 323, No. 10, 2756--2778 (2010; Zbl 1222.11132)]) of the \(p\)-part of the class group. In particular, the product of \(\mathfrak S_{L/K}\) and a certain `denominator ideal' is integral and annihilates the \(p\)-part of the class group. As for abelian \(G\) the denominator ideal equals the whole of \(\mathbb Z_p[G]\), this thereby generalizes Solomon's conjecture.NEWLINENEWLINEReviewer's remark: The reviewer's approach to noncommutative Fitting invariants allows to define Fitting invariants over \(\mathbb Z_p[G]\), but not over \(\mathbb Z[G]\). Thus each occurrence of \(\operatorname{Fitt}_{\mathbb Z[G]}(M)\otimes \mathbb Z_p\) for some finitely generated \(\mathbb Z[G]\)-module \(M\) should be replaced by \(\operatorname{Fitt}_{\mathbb Z_p[G]}(M\otimes \mathbb Z_p)\).