Ikeda-Nakayama modules (Q1610967)

Let \(R\) and \(S\) be rings and let \(_SM_R\) be an \((S,R)\)-bimodule. For any \(X\subseteq M\) and any \(T\subseteq S\), denote \(l_S(X)=\{s\in S:sX=0\}\) and \(r_M(T)=\{m\in M:Tm=0\}\). A module \(M_R\) is called \(\pi\)-injective if every submodule is essential in a direct summand and for any two direct summands \(M_1\), \(M_2\) with \(M_1\cap M_2=0\), \(M_1\oplus M_2\) is also a direct summand. Denote by \(\lambda\colon S\to\text{End}(M_R)\) the canonical ring homomorphism given by \(\lambda(s)(x)=sx\) for \(x\in M\) and \(s\in S\). A module \(M_R\) is called an Ikeda-Nakayama module (IN-module) if \(l_S(A\cap B)=l_S(A)+l_S(B)\) for any submodules \(A\) and \(B\) of \(M_R\) where \(S=\text{End}(M_R)\). The authors characterize these modules and establish the following main results: (1) Let \(_SM_R\) be a bimodule such that \(_SM\) is faithful. The following are equivalent: (i) for any submodules \(A\) and \(B\) of \(M_R\) with \(A\cap B=0\), \(S=l_S(A)+l_S(B)\); (ii) if \(A\) and \(B\) are complements of each other in \(M_R\), then \(S=l_S(A)+l_S(B)\); (iii) if \(A\) and \(B\) are complements of each other in \(M_R\), then \(S=l_S(A)\oplus l_S(B)\); (iv) \(M\) is \(\pi\)-injective and, for any \(f^2=f\in\text{End}(M_R)\), \(f=\lambda(s)\) for some \(s\in S\). (2) The following are equivalent for a module \(M_R\) with \(S=\text{End}(M_R)\): (i) \(M\) is \(\pi\)-injective; (ii) for any submodules \(A\) and \(B\) of \(M_R\) with \(A\cap B=0\), \(S=l_S(A)+l_S(B)\).