Some results on SI-rings (Q1895545)

A ring \(R\) is called a right \(SI\)-ring if every singular right \(R\)-module is injective. Right \(SI\)-rings were introduced and studied by \textit{K. R. Goodearl} [Mem. Am. Math. Soc. 124 (1972; Zbl 0242.16018)] and he proved that a ring \(R\) is a right \(SI\)-ring if and only if \(R\) is right nonsingular and \(R=K\oplus R_1\oplus\dots\oplus R_n\) for some positive integer \(n\), ring \(K\) with \(K/\text{Soc} (K_K)\) semisimple and rings \(R_i\) \((1\leq i\leq n)\) each Morita equivalent to a right \(SI\)-domain. Subsequently \textit{B. L. Osofsky} and the reviewer [J. Algebra 139, No. 2, 342-354 (1991; Zbl 0737.16001)] showed that a ring \(R\) is right \(SI\) if and only if every cyclic singular right \(R\)-module is injective. In this note the authors prove that a ring \(R\) is right \(SI\) if and only if every cyclic semiprimitive singular right \(R\)-module is injective. (A module \(M\) is called semiprimitive if \(\text{Rad } M=0\).) A ring \(R\) is called a right \(CS\)-ring if every right ideal is an essential submodule of a direct summand of \(R_R\). The main theorem states: a ring \(R\) is a ring direct sum of a semiprimary \(SI\)-ring and a right \(CS\) right \(SI\)- ring with zero right socle if and only if every cyclic semiprimitive right \(R\)-module is a direct sum of a projective module and an injective module. In addition a ring \(R\) is the direct sum of a semiprimary \(SI\)- ring and an \(SI\)-ring with zero right (and left) socle if and only if every finitely generated semiprimitive right (or left) \(R\)-module is a direct sum of a projective module and an injective module. Finally a ring \(R\) is the ring direct sum of a semisimple ring and a right \(SI\)-domain if and only if every cyclic semiprimitive right \(R\)-module is projective or injective.