On the necessity of new ramification breaks (Q2378046)

Let \(p\) be a prime number, \(K\) a finite extension of the field of \(p\)--adic numbers \({\mathbb Q}_ p\). Let \(N\) be a finite fully ramified Galois \(p\)--extension of \(K\) with Galois group \(G\) and let \({\mathfrak P}_ N\) be the maximal ideal of the valuation ring \({\mathfrak D}_ N\) of \(N\). Let \(T\) be the maximal unramified subfield of \(K\). In the study of the \({\mathfrak D}_ T [G]\)-module structure of the powers \({\mathfrak P}_ N^ r\), the usual ramification numbers are necessary but not sufficient to find such structure. If \(b\) is a break number then the quotient of two consecutive ramification groups \(G_ b/G_{b+1}\cong C_ p^ s\) is an elementary abelian group and a module over the cyclic group ring \({\mathfrak D}_ T[C_ p]\). There are exactly three indecomposable modules and the multiplicities of these three modules are parametrized by \(b\). In [Can. Math. Bull. 45, No. 2, 168--179 (2002; Zbl 1033.11054)] the authors considered biquadratic extensions and observed that the usual ramification invariants are not sufficient. In [J. Théor. Nombres Bordx. 17, No. 1, 87--107 (2005; Zbl 1162.11394)] they introduced new ramification breaks. In the paper under review it is seen that the insufficiency of the usual ramification ramification filtration is tied to the elementary quotients \(G_ b/ G_ {b+1}\). They introduce the second refined break and study the Galois module structure of ideals in bicyclic extensions under maximal refined ramification, when the second refined break achieves a natural upper bound.