Locally pro-\(p\) contraction groups are nilpotent (Q2054215)

A \textit{contraction group} is a pair \((G,\tau )\), where \(G\) is a locally compact group and \(\tau\) is an automorphism of \(G\) such that \(\tau^n(x)\rightarrow e\) as \(n\rightarrow\infty\) for all \(x\in G\). A topological group is \textit{locally pro-p} if it has an open subgroup which is a pro-\(p\) group. In this paper the authors answer an open question showing that every locally pro-\(p\) contraction group is nilpotent. Any locally pro-\(p\) contraction group is the direct product of a torsion subgroup and a nilpotent \(p\)-adic Lie group. It is natural to ask whether the torsion factor is nilpotent as well. A torsion locally pro-\(p\) contraction group has a composition series in which each factor is isomorphic to \((\mathbb{F}_p((t)),[t\cdot\,])\), the additive group of the field of formal Laurent series over the finite field \(\mathbb{F}_p\) (and thus is a torsion group of exponent \(p\)), and the map \([t\cdot\,]:f\mapsto tf\) is an automorphism of \(\mathbb{F}_p((t))\) such that \([t\cdot\,]^k f\rightarrow 0\) as \(k\rightarrow\infty\) for every \(f\in\mathbb{F}_p((t))\). In this case the group is called \(p\)-\textit{power contraction group}. \textbf{Theorem A}: Every \(p\)-power contraction group \((G,\tau)\) is nilpotent. The proof of this theorem is by induction on the number of composition factors and follows from the next result. \textbf{Theorem B}: Suppose that \(\phi:\mathbb{F}_p((t))\rightarrow\Aut(\mathbb{F}_p((t))^d)\) is a continuous homomorphism of groups and satisfies \(\phi\circ[t\cdot\,]=\text{ad}([t\cdot\,])\circ\phi\), where ad(\([t\cdot\,](\psi)=[t^{\oplus \,d}\cdot\,]\circ\psi\circ[t^{\oplus \,d}\cdot\,]^{-1}\) \((\psi\in\Aut(\mathbb{F}_p((t))^d))\), the automorphism \([t^{\oplus \,d}\cdot\,]\) of \(\mathbb{F}_p((t))^d\) is given by \([t^{\oplus \,d}\cdot\,](f_1,\cdots,f_d)=(tf_1,\cdots,tf_d)\), and \(\Aut(\mathbb{F}_p((t))^d)\) has the Braconnier topology. Then there is \(\xi\in\mathbb{F}_p((t))^d\setminus\{0\}\) such that \(\phi(f)(\xi)=\xi\) for every \(f\in \mathbb{F}_p((t))\). The proof of the second theorem goes by representing endomorphisms in End(\(\mathbb{F}_p((t))^d)\) as infinite block matrices with the blocks being \(d\times d\) matrices over \(\mathbb{F}_p\). Conditions on block matrices for them to correspond to elements of End(\(\mathbb{F}_p((t))^d\)) are determined, and then additional conditions for them to belong to the image of \(\phi\) are derived.