Embeddings of quotient division algebras of rings of differential operators (Q2357012)

The birational classification of surfaces is an important problem in algebraic geometry. In that sense, we can consider also whether there is a dominant rational map from a smooth, irreducible surface to another smooth one, and how it affects the birational invariants of both varieties. In this paper, the authors address the noncommutative analogue of the latter. Let \(k\) be and algebraically closed field of characteristic zero, and let \(A\) be a graded noetherian domain of Gelfand-Kirillov dimension 3 that is generated in degree one. It is known that \(A\) has a graded quotient division ring, denoted by \(Q_{\text{gr}}(A)\), given by inverting the nonzero homogeneous elements of \(A\). In fact, there is a division ring \(D\) and an automorphism \(\sigma\) of \(D\) such that \(Q_{\text{gr}}(A)=D[t^{\pm};\sigma]\). In this sense, one can affirm that there is a dominant rational morphism from \(\text{Proj}(A)\) to \(\text{Proj}(B)\) if and only if there exists an embedding from \(Q_{\text{gr}}(B)\) to \(Q_{\text{gr}}(A)\). There exists a classification of such division rings achieved by \textit{M. Artin} [Lond. Math. Soc. Lect. Note Ser. 238, 1--19 (1997; Zbl 0888.16025)], whose fourth and last type is a quotient division of \(K[t;\delta]\), where \(K\) is a finitely generated extension of \(k\) and \(\delta\) is a \(k\)-linear derivation of \(K\). In this setting, the main result of the paper states that given two finitely generated field extensions \(F\) and \(K\) of \(k\) of trascendence degree one, and two nonzero derivations \(\mu\) and \(\delta\) of \(F\) and \(K\) respectively, then whenever we have that \(F(t;\mu)\) embeds in \(K(x;\delta)\), the genus of \(F\) is less or equal than the genus of \(K\). The paper is structured in three sections. The first one contains an introduction, the second one consists of the proof of the main result, using a strategy based on certain reduction modulo primes and in the third section the authors give some general comments on the embedding problem for other division rings on the aforementioned Artin's list.