On Greenberg's generalized conjecture for CM-fields (Q2408336)

Let \(k\) be a number field and let \(p\) be a prime. Denote the composite of all \(\mathbb Z_p\)-extensions of \(k\) by \(\tilde k\) and for any subfield \(L\) of \(\tilde k\) let \(X_L\) be the Galois group over \(L\) of the maximal unramified abelian pro-\(p\)-extension of \(L\). If \(k\) is totally real and \(L = k_\infty\) is the cyclotomic \(\mathbb Z_p\)-extension of \(k\), \textit{R. Greenberg} [Am. J. Math. 98, 263--284 (1976; Zbl 0334.12013)] conjectured that \(X_{k_\infty}\) is a pseudo-null Iwasawa module. Note that under Leopoldt's conjecture one has \(k_\infty=\tilde k\) in this case. More generally, for an arbitrary number field \textit{R. Greenberg} [Adv. Stud. Pure Math. 30, 335--385 (2001; Zbl 0998.11054)] conjectures that the Iwasawa module \(X_{\tilde k}\) is pseudo-null. The main result of this article is as follows. Let \(k\) be a CM-field and let \(p\) be an odd prime which splits completely in \(k/\mathbb Q\) and such that \(p\) does not divide the class number of \(k\). Let \(k^+\) be the maximal totally real subfield of \(k\). Assume that Leopoldt's conjecture holds for \(k^+\) at \(p\) and that \(X_{k_\infty^+}\) is pseudo-null. Then \(X_{\tilde k}\) is also pseudo-null. This generalizes results of \textit{J. Minardi} [CMS Conf. Proc. 7, 237--242 (1987; Zbl 0625.12004)] and \textit{T. Itoh} [J. Number Theory 131, No. 1, 59--66 (2011;Zbl 1211.11119)] who treated the case \([k:\mathbb Q] = 2\) and \([k:\mathbb Q] = 4\), respectively. However, the proof is not independent of these results, as the authors often assumes that \([k:\mathbb Q] \geq 4\). The proof mainly follows Itoh's approach. One defines a chain of intermediate fields \(K^{(i)} \subseteq K^{(i+1)}\), \(1 \leq i \leq m\) such that \(K^{(i)}/k\) is a sub-\(\mathbb Z_p^i\)-extension and \(K^{(m+1)} = \tilde k\), where \([k:\mathbb Q] = 2m\). Then one shows by induction that each \(X_{K^{(i)}}\) is pseudo-null using class field theory.