Birational endomorphisms of the affine plane (Q1181047)

Let \(\mathbb{A}^ 2\) be the affine plane over an algebraically closed field \(k\). The birational endomorphisms of \(\mathbb{A}^ 2\) are those maps \(\mathbb{A}^ 2\to\mathbb{A}^ 2\) given by endomorphisms \(\varphi\) of the polynomial algebra \(k[X,Y]\) such that \(k(\varphi(X)\), \(\varphi(Y))=k(X,Y)\). The simplest non-automorphic example is \(\varphi_ 0\) defined by \(\varphi_ 0(X)=X\) and \(\varphi_ 0(Y)=XY\). The birational endomorphisms of \(\mathbb{A}^ 2\) given by \(\varphi=u\circ\varphi_ 0\circ v\), where \(u\) and \(v\) are any automorphisms of \(k[X,Y]\), are called simple affine contractions in \(\mathbb{A}^ 2\). The question whether every birational endomorphism of \(\mathbb{A}^ 2\) is a composite of simple affine contractions arose in the early seventies, in Abhyankar's seminar. That question was answered negatively by \textit{K. P. Russell}. He soon exhibited a whole zoo of irreducible endomorphisms, some of them having infinitely near fundamental points. A reasonably complete classification of all birational endomorphisms of \(\mathbb{A}^ 2\) is interesting and difficult. The aim of this paper is to make some contributions to that problem. The methods by which Russell constructed those endomorphisms and proved their irreducibility consist in a detailed analysis of the configuration of missing curves, that is the one dimensional components of the closure (in \(Y)\) of \(Y\backslash f(X)\). Our approach is essentially an elaboration of Russell's methods, and includes the use of some graph- theoretic techniques.