Fixed point free locally nilpotent derivations of \(\mathbb A^2\)-fibrations (Q2376708): Difference between revisions

Locally nilpotent derivations are important tools in the study of many problems in affine algebraic geometry and theory of polynomial algebras. Let \(R\subset A\) be commutative domains such that \(R\) contains \(\mathbb Q\). In the paper under review the authors study locally nilpotent \(R\)-derivations \(\xi\) of \(A\). The derivation is fixed point free if the ideal of \(A\) generated by the image of \(\xi\) coincides with \(A\). This holds when \(\xi\) has a slice, i.e., an element \(a\in A\) such that \(\xi(a)=1\) (but the inverse is not always true). The \(R\)-algebra \(A\) is called an \({\mathbb A}^m\)-fibration if \(A\) is finitely generated as an \(R\)-algebra, flat as an \(R\)-module, and the tensor product \(A\otimes_RK({\mathfrak p})\) is isomorphic to the polynomial algebra \(K({\mathfrak p})^{[m]}\) in \(m\) variables over the field of fractions \(K({\mathfrak p})=R_{\mathfrak p}/{\mathfrak p}R_{\mathfrak p}\) for every prime ideal \(\mathfrak p\) of \(R\). The \({\mathbb A}^m\)-fibration \(A\) over \(R\) is trivial, if \(A=R^{[m]}\), it is stably trivial if \(A^{[n]}=R^{[m+n]}\) for some \(n\), and is locally stably trivial if \(A\otimes_RR_{\mathfrak m}\) is stably trivial for any maximal ideal \(\mathfrak m\) of \(R\). The main result of the paper is the following. Let \(R\) be a Noetherian normal domain and \(A\) be a locally stably trivial \({\mathbb A}^2\)-fibration over \(R\). Then the fixed point free locally nilpotent \(R\)-derivation \(\xi\) of \(A\) has a slice if and only if the algebra of constants \(A^{\xi}\) is an \({\mathbb A}^1\)-fibration over \(R\). This result is motivated by a problem of \textit{G. Freudenburg} [J. Algebra 322, No. 9, 3078--3087 (2009; Zbl 1220.13018)] which can be restated as follows: Let \(K\) be a field and let \(R=K^{[n]}\), \(n\geq 2\), and let \(A\) be an \({\mathbb A}^2\)-fibration ofver \(R\). Does every fixed point free locally nilpotent \(R\)-derivation of \(A\) have a slice?