Rationally convex domains and singular Lagrangian surfaces in \(\mathbb {C}^2\) (Q2634621): Difference between revisions

A compact set \(X\subset\mathbb C^N\) is called rationally convex if for every point \(p\not\in X\) there exists a complex algebraic hypersurface \(H\subset\mathbb C^N\) such that \(p\in H\) and \(X\cap H=\emptyset\). One reason for interest in rationally convex sets, from a complex analytic point of view, is the property that any holomorphic function on a neighborhood of a rationally convex set \(X\) can be approximated uniformly on \(X\) by rational functions. In this paper, the authors are interested in the case when \(X\) is the closure of a strictly pseudoconvex domain, and for rationally convex \(X\) they try to answer the question of which smooth manifolds can \(X\) be. In [Invent. Math. 199, No. 1, 215--238 (2015; Zbl 1310.32014)], \textit{K. Cieliebak} and \textit{Y. Eliashberg} gave a complete answer to this question in high dimensions by showing that, if \(W\subset\mathbb C^N\) is a smoothly bounded domain, with \(N > 2\), then \(W\) is smoothly isotopic to a strictly pseudoconvex domain with rationally convex closure if and only if it admits a Morse function \(\varphi:\overline{W}\to\mathbb R\) without critical points of index greater than \(N\), and such that \(\partial W\) is the maximal regular level set of \(\varphi\). For \(N=2\), the authors show that there exist strictly pseudoconvex domains in \(\mathbb C^2\) with rationally convex closures diffeomorphic to the following disk bundles: \(D(\chi,0)\) for \(\chi\not=2\) and \(\widetilde D(\chi,e)\) for \((\chi,e)\not=(1,-2)\) or \((0,0)\) and \(e\in\{2\chi-4,2\chi,2\chi+4,\dots,-2\chi-4+4\lfloor\chi/4+1\rfloor\}\), where \(D(\chi,e)\) is the \(D^2\)-bundle over an orientable surface of Euler characteristic \(\chi\), with Euler number \(e\), and \(\widetilde D(\chi,e)\) is the \(D^2\)-bundle over a non-orientable surface. They show that these are the only possibilities. Also, the authors recall some classical obstructions, prove some more general obstructions related to symplectic and contact topology, explain the close connection to Lagrangian surfaces with isolated singularities and develop techniques for constructing such surfaces. Finally, they give a complete characterization of Lagrangian surfaces with open Whitney umbrellas.