On the Behrens radical of matrix rings and polynomial rings. (Q932285)

From MaRDI portal
scientific article
Language Label Description Also known as
English
On the Behrens radical of matrix rings and polynomial rings.
scientific article

    Statements

    On the Behrens radical of matrix rings and polynomial rings. (English)
    0 references
    10 July 2008
    0 references
    For a given associative ring \(R\), the polynomial ring over \(R\) in the set \(X\) of commuting indeterminates is denoted by \(R[X]\), the polynomial ring in the set \(X\) of non-commuting indeterminates over \(R\) is denoted by \(R\langle X\rangle\) and \(M_n(R)\) denotes the ring of all \(n\times n\) matrices over \(R\). A ring \(R\) is Brown-McCoy radical if and only if \(R\) cannot be homomorphically mapped onto a nonzero simple ring with identity. A ring \(R\) is Behrens radical if and only if \(R\) cannot be homomorphically mapped onto a nonzero subdirectly irreducible ring having a nonzero idempotent in its heart. The Behrens radical of a ring \(A\) is denoted by \(\beta(A)\). The authors show that for any set \(X\) of commuting indeterminates over a ring \(R\), \(\beta(R[X])=\beta(R[X]\cap R)[X]\). They also show that for any set \(X\) of non-commuting indeterminates over \(R\) with cardinality \(\text{card}(X)\geqslant 2\), \(\beta(R\langle X\rangle)=\beta(R\langle X\rangle\cap R)\langle X\rangle\) and \(\beta(R\langle X\rangle\cap R)\) is an ideal of \(R\) independent of \(X\). They show that, in contrast to the Brown-McCoy radical, the Behrens radical is quite different for polynomial rings in commuting and non-commuting indeterminates. Namely, they show that there exists a simple ring \(R\) such that \(R[X]\) is Behrens radical for every set \(X\) of commuting indeterminates over \(R\) and \(R\langle Y\rangle\) is Behrens semisimple for every set \(Y\) of non-commuting indeterminates over \(R\) with cardinality \(\text{card}(Y)\geqslant 2\). This is closely connected with the fact that matrix rings over Behrens radical rings need not be Behrens radical. In fact, the authors show more. Namely, if \(D\) is a simple domain with unity and \(I\) a proper nonzero one-sided ideal of \(D\), then \(I\) is a simple Behrens radical ring but \(M_n(I)\) is Behrens semisimple for some \(n\). A radical \(\alpha\) is called an \(N\)-radical, if \(N\) contains all nilpotent rings, \(N\) is closed under left ideals and for any left ideal \(L\) of a ring \(R\), \(L\in\alpha\) implies \(LR^*\in\alpha\), where \(R^*\) denotes the usual extension of a ring \(R\) obtained by adjoining unity. For the class \(\beta\) of all Behrens radical rings, the authors consider two classes: \[ \overline\beta=\{R:M_n(R)\in\beta,\;n=1,2,\dots\} \] and \[ \mathcal C=\{A:A\text{ is a simple ring such that }M_n(A)\text{ has a nonzero idempotent for some }n\}. \] They show that \(R\in\overline\beta\) if and only if \(R\) cannot be homomorphically mapped onto a subdirectly irreducible ring whose heart is in \(\mathcal C\). Moreover, \(\overline\beta\) is the largest \(N\)-radical class contained in \(\beta\) and \(\overline\beta\) is a special radical. The authors conclude the paper with five related questions which stimulate further research in this subject.
    0 references
    0 references
    0 references
    0 references
    0 references
    Behrens radical
    0 references
    Brown-McCoy radical
    0 references
    matrix rings
    0 references
    polynomial rings
    0 references
    Amitsur property
    0 references
    0 references
    0 references
    0 references