The residual finiteness of completely infinite polycyclic groups (Q1059147)

A group is polycyclic if it has a finite series with cyclic factors and a group is completely infinite if it has a finite series with aperiodic (torsion-free) abelian factors. A completely infinite polycyclic group G has a finite normal series in which every factor is a free abelian group of finite rank which has no infinite factor group by a subgroup which is normal in G. The greatest value n of the ranks of the factors in such a series is called the species of G, and n is an invariant of G. In Ill. J. Math. 8, 536-542 (1964; Zbl 0124.263), \textit{A. Learner} proved that, for a polycyclic group G, there exists a finite set of primes \(\Pi\) such that G is residually \(\Pi\)-finite. When G is also completely infinite with species n, he showed that, for an arbitrary prime p, the set \(\Pi\) could be taken as \(\Delta\) (n,p), where \(\Delta\) (n,p) denotes the set of all the primes which divide \((p^ n-1)(p^{n- 1}-1)...(p-1)p.\) The aim of this paper is to show that Learner has obtained the best possible general result for completely infinite polycyclic groups. The intersection of the sets \(\Delta\) (n,p) for all primes p is the set \(\Psi\) (n) which consists of all the primes less than \(n+2\), therefore \(\Psi\) (n) is the largest set of primes that could be obligatory for G, in the sense that if G is residually \(\Pi\)-finite then \(\Psi\) (n) is a subset of \(\Pi\). It is proved that if n is an odd prime number, then there exists a completely infinite polycyclic group H of species n such that \(\Psi\) (n) is obligatory for H but so that H is not residually \(\Psi\) (n)-finite. Further, there is a completely infinite polycyclic group K of species 3 which is residually \(\Pi\)-finite for \(\Pi =\Delta (3,2)\) or \(\Delta\) (3,3) or \(\Delta\) (3,5), but not for any of their proper subsets.