Embeddings of Danielewski surfaces (Q1434527)

One way to study affine algebraic varieties is to consider their closed embeddings in various affine spaces. This leads naturally to the problem of classifying these embeddings up to automorphisms of the affine space, two embeddings \(\alpha,\beta:Y\hookrightarrow\mathbb A^n\) of a variety \(Y\) being called {equivalent} if there exists a polynomial automorphism \(\phi\) of \(\mathbb A^n\) such that \(\phi\circ\alpha=\beta\). One of the starting points of this problem was the famous result of \textit{S. S. Abhyankar} and \textit{T. Moh} [J. Reine Angew. Math. 276, 148--166 (1975; Zbl 0332.14004)] and \textit{M. Suzuki} [J. Math. Soc. Japan 26, 241--257 (1974; Zbl 0276.14001)] which says that all embeddings of the complex affine line \(\mathbb A^1_{\mathbb C}\) in the plane are equivalent. The higher-dimensional version of this result, which is known as the Abhyankar-Sathaye conjecture, asks if all embeddings \(\mathbb A^p\hookrightarrow \mathbb A^n\), where \(n\geq 3\) and \(p<n\), are equivalent, and to date this question is completely open. One of the most important result in this direction is a very general embedding theorem due to \textit{S. Kaliman} [Proc. Am. Math. Soc. 113, No.2, 325--334 (1991; Zbl 0743.14011)] which says that all embeddings of a smooth affine variety \(V\) into \(\mathbb A^n\), where \(n\geq 2\dim(V)+2\) are equivalent. In the paper under review, the authors consider embeddings in the affine 3-space \(\mathbb C^3\) of a particular class of affine surfaces called {Danielewski surfaces.} Here {}the term Danielewski surface refers to an affine surface which is isomorphic to a surface in \(\mathbb C^3\) defined by an equation of the form \(x^nz-P\left(y\right)=0\), where \(n\geq 0\) and \(P\left(y\right)\in\mathbb C\left[y\right]\). These surfaces were first considered by Danielewski in a celebrated counter-example to the Zariski cancellation problem [see also \textit{K.-H. Fieseler}, Comment. Math. Helv. 69, No.1, 5--27 (1994; Zbl 0806.14033)]. Over the past 10 years, they became one of the most ubiquitous and intensively studied class of affine surfaces. The study of embeddings of Danielewski surfaces in \(\mathbb C^3\) up to equivalence originated from a paper of \textit{M. Masuda} and \textit{T. Petrie} [in: Algebraic groups and their generalizations, Proc. Symp. Pure Math. 56, Pt. 1, 347--354 (1994; Zbl 0854.22016)], in which the authors discovered that inequivalent embeddings of certain singular Danielewski surfaces arise naturally as obstructions to linearize certain actions of the orthogonal group \(\text{O}\left(2\right)\) on \(\mathbb C^4\). In the present article, the authors establish that every Danielewski surface \(x^nz-P\left(y\right)=0\), where \(n\geq 2\) and \(\deg\left(P\left(y\right)\right)\geq 2\) admits inequivalent embeddings in \(\mathbb C^3\). They also construct inequivalent embeddings of surfaces \(xz-\left(y^{d_n}-1\right)=0\) for infinite sequences of integers \(d_n\). In turns out that the picture is completely different in the holomorphism category. Indeed, the authors establish that if \(n\geq 2\) then the surfaces \(S_1=\left\{x^nz-y^2=0\right\}\) and \(S_{1+x}=\left\{x^nz-\left(1+x\right)y^2=0\right\}\) are algebraically inequivalent embeddings of isomorphic Danielewski surfaces. But they construct an explicit holomorphic automorphism \(\Phi\) of \(\mathbb C^3\) such that \(\Phi(S_1)=S_{1+x}\).