Torsion-free nilpotent groups and E-modules (Q915861): Difference between revisions

The authors define certain subgroups of the matrix group \(\left[\begin{matrix} 1 & b & a \\ 0 & 1 & c \\ 0 & 0 & 1 \end{matrix}\right]\) where a,b,c\(\in R\) and R is an arbitrary ring which is torsion-free as an additive abelian group. Let G be such a group. Then G is nilpotent of class 2, its commutator subgroup G' and its center \({\mathfrak Z}G\) coincide, Inn \(G\triangleleft Stab G\triangleleft Aut G\) where Stab G\(=\{\alpha \in Aut G:\) \(x^{-1}(x\alpha)\in {\mathfrak Z}G\) for all \(x\in G\), and \(x^{- 1}(x\alpha)=1\) for all \(x\in {\mathfrak Z}G\}\), Stab G is abelian, and Aut G\(=Aut R\ltimes Stab G\), a semi-direct product. Hence, by choosing the ring R appropriately one can construct nilpotent groups of class 2 with an automorphism group of the stated type. Using Shelah's ``Black Box'', the authors prove that, given any group H, there is a torsion-free ring R such that Aut \(R\cong H\). This yields nilpotent groups G of class 2 with Aut G\(=H\ltimes Stab G\). But more than this is desired. One wants to achieve further that Stab G\(=Inn G\). This leads to the requirement that R be an ``E-ring'', a concept which has appeared earlier in other contexts. When \(H=1\) a ``good'' E-ring is constructed, and the authors obtain arbitrarily large torsion-free nilpotent groups of class 2 without outer automorphisms. This paper is another demonstration that methods from the theory of abelian groups are capable of solving long-standing questions in other parts of algebra.