On the lattice of matric-extensible radicals. II. (Q858098)

The authors continue their study of radicals of associative rings (which are not necessarily unital rings) [for part I cf. ibid. 101, No. 1-2, 163-172 (2003; Zbl 1053.16016)]. A class \(\sigma\) of rings is called matric-extensible if it is closed with respect to the rings \(R_n\) of all \(n\times n\)-matrices with coefficients in rings \(R\in\sigma\) for all non-negative integers \(n\). In the following we present some interesting results proved in the present paper: for a simple idempotent ring \(S\) the smallest matric-extensible supernilpotent radical containing \(S\) is an atom of the lattice of all matric-extensible supernilpotent radicals (Proposition 3.3), and the smallest supernilpotent radical containing \(S\) is an atom of the lattice of all matric-extensible supernilpotent radicals if and only if \(S\cong S_n\) for all \(n\) and \(S\cong A\) whenever \(S_n\cong A_n\) (Proposition 3.4); if \(A\) is a \(*\)-ring, the smallest special radical containing \(A\) is matric-extensible if and only if the smallest special class containing \(A\) is matric-extensible (Theorem 5.4).