Profinite groups with pronilpotent centralizers (Q2303691)

A profinite group is said to be a CA-group if the centralisers of all non-identity elements are abelian, and a CN-group if these centralisers are pronilpotent. In crucial intermediate steps to \textit{W. Feit} and \textit{J. G. Thompson}'s [Pac. J. Math. 13, 775--1029 (1963; Zbl 0124.26402)] theorem on the solubility of groups of odd order, \textit{M. Suzuki} [Proc. Am. Math. Soc. 8, 626--695 (1957; Zbl 0079.03104)] proved that finite CA-groups of odd order are soluble, and \textit{W. Feit} et al. [Math. Z. 74, 1--17 (1960; Zbl 0103.01402)] extended this result to CN-groups. Finitely generated residually free CA-groups are limit groups, that entered in a crucial way in the solutions of Tarski's problems. This has spawned a remarkable interest in infinite CA-groups. The author et al. [Isr. J. Math. 230, No. 2, 831--854 (2019; Zbl 1482.20018)] have studied profinite CA-groups. The goal of the paper under review is to deal with profinite CN-groups \(G\). The main result (Theorem 1.1) states the following. Let \(F\) be the maximal normal, pronilpotent subgroup of \(G\). Then \(F\) is open, and for the finite group \(G / F\) one of the following holds: \(G/F\) is cyclic; \(G/F\) is is a direct product of a cyclic group of odd order by a generalised quaternion group; \(G/F\) is a Frobenius group with cyclic kernel of odd order and cyclic complement, and \(F\) is pro-\(p\) for some prime \(p\); \(G/F\) is isomorphic to the group \(\mathrm{SL}(2, 3)\), and \(F\) is nilpotent, of order divisible by \(2\) and by at least another prime; \(G/F\) is almost simple, and \(F\) is a pro-\(2\) group. Examples are given to show that all these possibilities actually occur. All known examples of infinite, profinite CA-groups are prosoluble. Problem 1.2 asks for an example of a finitely generated, infinite profinite CN-group which is not prosoluble.