On the equivariant Tamagawa number conjecture in tame CM-extensions (Q543308)

The setting in this paper is \(L/K\), a finite Galois CM extension of number fields with Galois group \(G\). The so-called Equivariant Tamagawa Number Conjecture (ETNC; sometimes now called more simply Leading Term Conjecture) for the pair \(h^0(\text{Spec}(L))(0),\mathbb Z[G])\) is a far-reaching, \(G\)-invariant generalisation of the analytic class number formula. It is stronger than Chinburg's \(\Omega_3\)-conjecture; in fact it can be considered as a lift of that conjecture from \(K_0(\mathbb Z[G])\) to a much finer relative class group, and it has the decisive advantage over the latter of decomposing canonically into \(p\)-parts. ETNC (more accurately, the special case at hand) is only known to hold for absolutely abelian fields and under some hypotheses for abelian extensions of imaginary quadratic fields. The validity of ETNC has many consequences. For example, \textit{D. Burns} [Invent. Math. 169, No. 3, 451--499 (2007; Zbl 1133.11063)] showed that it implies the Rubin-Stark conjecture, and the reviewer [Compos. Math. 143, No. 6, 1399--1426 (2007; Zbl 1135.11059)] showed that it permits to calculate Fitting ideals of class groups in the minus part. The remarkable article under review does (roughly speaking) two things: It shows that in the minus part, ETNC is equivalent to an explicit statement about the Fitting ideals of certain ray class groups \(A^T_L\), in terms of generalised Stickelberger elements; and second it succeeds in proving this explicit statement for all tame extensions \(L/K\), assuming the vanishing of Iwasawa's \(\mu\)-invariant. (The tameness hypothesis can be somewhat relaxed.) For clarity let us repeat: This gives, under mild technical hypotheses, a proof of the minus part of ETNC. Both main results require new ideas and considerable technique. In establishing the equivalence of ETNC and the Fitting criterion for the ray class group, it is instrumental to have Tate sequences for small sets \(S\) at one's disposal. This is work of \textit{J. Ritter} and \textit{A. Weiss} [Compos. Math. 102, No. 2, 147--178 (1996; Zbl 0948.11041)]. In these, the standard module \(\Delta S\) (for a large set \(S\)) is replaced by something more complicated, which somehow encodes the local fundamental classes at the ramified primes. Apparently this is how the classes of 2-extensions enter into the calculation. The proof of the Fitting criterion, on the other hand, involves ascent, Iwasawa theory, and descent. The latter is fairly delicate; it uses previous work of \textit{A. Wiles} [Ann. Math. (2) 131, No. 3, 555--565 (1990; Zbl 0719.11082)] and the reviewer [Math. Z. 233, No. 3, 515--534 (2000; Zbl 0965.11047)].