Extensions and pullbacks in QB-rings. (Q1780991)

The authors introduced the concept of a QB-ring, generalizing the notion of a ring with K-theoretic stable rank 1 [in J. Algebra 230, No. 2, 608--655 (2000; Zbl 0963.16008)]. In particular, they established necessary and sufficient conditions for the middle ring of an extension \(0\to I\to R\to R/I\to 0\) to be a QB-ring. However, those conditions are difficult to verify, aside from the cases where \(I\) or \(R/I\) has stable rank 1. Here, they reformulate their extension theorem to make it more easily applicable in wider circumstances. For example, if \(I\) is a purely infinite, simple, essential ideal of \(R\), then \(R\) is QB if and only if \(R/I\) is QB and the quasi-invertible elements of \(R/I\) are right or left invertible. The authors also apply their extension results to pullbacks and multiplier rings. In particular, they prove that (a) Any finite subdirect product of QB-rings is QB; and (b) If \(R\) is a purely infinite, simple, (von Neumann) regular ring with a countable unit, then its multiplier ring is QB. They then study multiplier rings more closely, using the first and third authors' description of the monoid \(V({\mathcal M}(R))\) (consisting of equivalence classes of idempotent matrices over \({\mathcal M}(R)\)) for certain simple regular rings \(R\) [Commun. Algebra 28, No. 7, 3359--3385 (2000; Zbl 0963.16010)]. This approach yields conditions for \({\mathcal M}(R)\) or \({\mathcal M}(R)/R\) to be a QB-ring. Some results in the paper require certain purely infinite simple rings to be exchange rings. That hypothesis is now redundant, thanks to a recent result of the first author [Proc. Am. Math. Soc. 132, No. 9, 2543--2547 (2004; Zbl 1055.16012)]. (Also submitted to MR.)