More discrete copies of \(\mathbb Z\) in \(\beta \mathbb N\) (Q989097)

In this paper the authors establish that if \(q\) is any minimal idempotent in \(\beta\mathbb N\), then for all exept possibly one \(p\in\text{cl}\{2^n:n\in\mathbb N\}\setminus\mathbb N\), \(q+p+q\) generates an infinite discrete group. Responding to a question of Wis Comfort, the authors extend this result in two directions. They show on the one hand that for a minimal idempotent \(q\), there is at most one prime \(r\) for which there exists \(p\in\text{cl}\{r^n:n\in\mathbb N\}\setminus\mathbb N\) such that the group generated by \(q + p + q\) is not both infinite and discrete. On the other hand, the authors show that for any \(p\in \beta\mathbb N\), if \(p\in \text{cl}(n\mathbb N)\) for infinitely many \(n\in\mathbb N\), then there is some minimal idempotent \(q\) such that the group generated by \(q + p + q\) is infinite and discrete. They also show that if \(G\) is a countable discrete group and if \(p\) is a right cancellable element of \(G^*\), then there is an idempotent \(q\in G^*\) such that \(q\cdot p\cdot q\) generates a discrete copy of \(\mathbb Z\) in \(G^*\). It is unknown whether there exists any minimal idempotent \(q\) and any \(p\) with \(p\in\text{cl}(n\mathbb N)\) for infinitely many \(n\in \mathbb N\) such that the group generated by \(q + p + q\) is not discrete. The authors finish the paper by showing that if such a ``bad'' \(q\) exists, then there are many of them.