The IA-congruence kernel of high rank free metabelian groups (Q2281275)

Let \((\Gamma ) \) be a group and \(G\leq \mathrm{Aut}(\Gamma )\). For a finite index characteristic subgroup \(M\) of \(\Gamma \), let \(G(M)=\ker(G\,\longrightarrow\mathrm{Aut}(\Gamma /M))\). The (generalized) congruence subgroup problem (CSP) asks whether for every finite index subgroup, let \(L\), of \(G\) there exists a finite index characteristic subgroup \(M\,=\,M_{L}\) of \(\Gamma \) such that \(G(M)\) belongs in \(L\). Let \(\Gamma \) and \(G\) be as before. Let \(\hat{G}\) denotes the inverse limit of the family \(\{\,G/U\,\}\), where \(U\) ranges over all finite index normal subgroups of \(G\) and \(\bar{G}\) denotes the inverse limit of the family \(\{\,G /G(M)\,\}\), where \(M\) ranges over all finite index characteristic subgroups of \(\Gamma \). Then the CSP is equivalent to the question: Is the congruence map \(\hat{G}\,\longrightarrow\,\bar{G}\) injective? In case where the group \(\Gamma \) is finitely generated there are finite many subgroups of given index \(m\). Therefore the subgroups \(M_{m}\,=\,\cap \{\,\Delta \mid [\Gamma :\Delta ]\,=\,m\,\}\) are of finite index characteristic subgroups of \(\Gamma \). Let \(\hat{\Gamma }\) be the inverse limit of the family \(\{\,\Gamma /M_{m}\,\,m\in \mathbb{N}\,\}\). Then it is easy to see that CSP is equivalent to the question: Is the congruence map \(\hat{G}\,\longrightarrow\,\mathrm{Aut}(\hat{\Gamma })\) injective? More generally the CSP asks what is the kernel \(C(G ,\,\Gamma )\) of this map? In case where the group \(\Gamma\,=\,\mathbb{Z}^{n}\) is the free abelian of rank \(n\) then the (classical) CSP has affirmative answer for \(n\geq 3 \) (namely \(C(\mathbb{Z}^{n})\,=\,C(\mathrm{Aut}(Z^{n}),\,\mathbb{Z}^{n})\,=\,\{\,e\,\}\) ). In case where \(n\,=\,2\) it is proved that \(C(\mathbb{Z}^{2})\,=\,\hat{F}_{\omega }\), where where \(\hat{F}_{\omega }\) is the free nonabelian profinite group on a countable number of generators. [\textit{O. V. Mel'nikov}, Sov. Math., Dokl. 17, 867--870 (1976; Zbl 0369.20027); translation from Dokl. Akad. Nauk SSSR 228, 1034--1036 (1976); \textit{A. Lubotzky}, J. Algebra 77, 411--418 (1982; Zbl 0495.20021)]. In case where the group \(\Gamma \) is not abelian the most results are related to \(\Gamma \,=\,\pi (S_{g},\,n)\), the fundamental group of the closed surface of genus \(g\) with \(n\) punctures. Via these results some partial results have been obtained in the case of finitely generated free groups. Let now \(\Gamma\,=\,\Phi _{n}\,=\,F_{n}/F_{n}^{\prime \prime }\) be the free metabelian group on \(n\) generators in [\textit{D. E. C. Ben-Ezra} and \textit{A. Lubotzky}, J. Algebra 500, 171--192 (2018; Zbl 1386.20027); \textit{D. E. C. Ben-Ezra}, Groups Geom. Dyn. 10, No. 2, 583--599 (2016; Zbl 1353.20013)] it was proved that \(C(\Phi _{2})\,=\,\hat{F_{\omega }}\) and that \(C(\Phi _{3})\,\supseteq\,\hat{F_{\omega }}\). This is the motivation for studying \(C(\Phi _{n})\) for \(n\geq 4\) in this paper. Let \(IA(\Phi _{n})\,=\ker(\mathrm{Aut}(\Phi _{n}\longrightarrow\,GL_{n}(\mathbb{Z} )\), then, using and the fact that \(\widehat{GL_{n}(\mathbb{Z})}\,\longrightarrow\,GL_{n}(\hat{\mathbb{Z}})\) is injective (\(n\geq 3\)) , it is obtained that \(C(\Phi _{n})\) is an image of \(C(IA(\Phi),\,\Phi _{n})\). Therefore the study of \(C(\Phi _{n})\) is related to the study of \(C(IA(\Phi),\,\Phi _{n})\). Theorem 1. For every \(n\geq 4\), the group \(C(IA(\Phi _{n}),\,\Phi _{n})\) contains a subgroup \(C\) which satisfies the following properties: \(\bullet \,\,\, C \) is isomorphic to a product \(C\,=\,\prod _{i\,=\,1}^{n}C_{i}\) of \(n\) copies of \(C_{i}\,\cong\ker (\widehat{ SL_{n-1}(\mathbb{Z}[x^{\pm }])} \, \longrightarrow\, SL_{n-1}(\widehat{\mathbb{Z}[x^{\pm }])} ) \). \(\bullet \,\,\, C\) is a direct factor of \(C(IA(\Phi _{n}),\,\Phi _{n})\) that is, there is a normal subgroup \(N\vartriangleleft C(IA(\Phi _{n}),\,\Phi _{n})\) such that \(C(IA(\Phi _{n}),\,\Phi _{n})\,=\,N\times C\). Using techniques of \textit{M. Kassabov} and \textit{N. Nikolov} [Invent. Math. 165, No. 1, 209--224 (2006; Zbl 1139.19003)], one can show that the subgroups \(C_{i}\) are not finitely generated. So as an immediate corollary, it is obtained the following: Theorem 2. For every \(n\geq 4\), the group \(C(IA(\Phi _{n}),\,\Phi _{n})\) is not finitely generated. This theorem gives the sensation that this group is ``big''. On the other hand it is proved the theorem. Theorem 3. For every \(n\geq 4\), the group \(C(IA(\Phi _{n}),\,\Phi _{n})\) is central in \(\widehat{ IA(\Phi _{n})}\). The fact that \(C(\Phi _{n})\) is an image of \(C(IA(\Phi),\,\Phi _{n})\) allow to deduce: Theorem 4. For every \(n\geq 4\), the group \(C(\Phi _{n})\) is abelian. Never the less the author points out that remain the questions: Question 1. Is \(C(\Phi _{n})\) not finitely generated? trivial? Question 2. Is \(C(IA(\Phi _{n}),\,\Phi _{n})\,=\,\prod _{i\,=\,1}^{n}C_{i}\) or does it contain more elements? Theorem 3 is the Theorem 1.2 and Theorem 4 is the Theorem 1.3 in [\textit{D. E.-C. Ben-Ezra}, ``The congruence subgroup problem for the free metabelian group on \(n\geq 4\) generators'', Preprint, \url{arXiv:1701.02459}] As it is pointed by the author there is an overlap of the present paper and the above paper. We refer on page 386 of the present paper where it is tried an explanation/clarification for the interrelation of these papers.