Radicals of crossed products of enveloping algebras (Q1098922)

Let \(L\) be a Lie algebra over a commutative ring \(K\) such that \(L\) is a free \(K\)-module and let \(U(L)\) denote its enveloping algebra. Let \(R\) be a \(K\)-algebra such that \(L\) acts on \(R\) as \(K\)-derivations and let \(A\) denote the crossed product \(R*U(L)\) of \(R\) by \(U(L)\). The aim of this paper is to describe the Jacobson radical \(J(A)\) when \(L\neq 0\). In particular it is proved that if \(R\) is either right Noetherian, a PI-algebra or a ring with no nonzero nilpotent elements then \(J(A)=N*U(L)\), where \(N\) is the largest \(L\)-invariant nil ideal of \(R\), and moreover \(J(A)\) is a nil ideal in this case. The conditions on \(R\) arise because the authors require \(R\) to have the property that there exists a family of prime ideals \(P\) such that every non-zero ideal of \(R/P\) contains a regular element. Furthermore, if \(K\) has characteristic zero then \(N*U(L)\subseteq J(A)\subseteq M*U(L)\), where \(N\) is the prime radical of \(R\) and \(M\) the largest \(L\)-stable ideal contained in \(J(R)\); if in addition some \(0\neq x\in L\) acts as an inner derivation on \(R\) then \(M\) can be replaced by the nil radical of \(R\). These results generalize theorems of \textit{S. A. Amitsur} on polynomial rings [Can. J. Math. 8, 355-361 (1956; Zbl 0072.024)], \textit{R. S. Irving} on enveloping algebras [Am. J. Math. 106, 113-135 (1984; Zbl 0547.17007)], \textit{M. Ferrero}, \textit{K. Kishimoto} and \textit{K. Motose} on Ore extensions [J. Lond. Math. Soc., II. Ser. 28, 8-16 (1983; Zbl 0518.16003)] and of \textit{D. A. Jordan} also on Ore extensions [J. Lond. Math. Soc., II. Ser. 10, 281-291 (1975; Zbl 0313.16011)]. In addition, by considering the countable-dimension exterior algebra \(E\) over an arbitrary field \(K\) and a suitable derivation \(\delta\) of \(E\) the authors show that \(E\) is \(\delta\)-prime but not semiprime, thus answering a question of \textit{K. R. Goodearl} and \textit{R. B. Warfield} jun. [Math. Z. 180, 503-523 (1982; Zbl 0495.16002)]. Moreover, \(S=E[y;\delta]\) is a prime ring whose Jacobson radical \(J(S)\) is nil; thus \(J(S)\) is a prime nil ring generated by two elements. Moreover \(E\) is a Jacobson ring whereas \(S\) is not, irrespective of the characteristic of \(K\).