Hankel operators on the Bergman spaces of Reinhardt domains and foliations of analytic disks (Q777208): Difference between revisions

From the point of view of operator theory, compactness of Hankel operators is interesting, for example, because it implies that the assignment \(\phi\mapsto T_{\phi}\) from the symbol \(\phi\) (say \(\phi\in C(\overline{\Omega})\)) to its associated Toeplitz operator \(T_{\phi}\) is an algebra homomorphism of \(C(\overline{\Omega})\) into the Calkin algebra on the Bergman space of the domain \(\Omega\subset \mathbb{C}^{n}\). This means, among other things, that if the symbol \(\phi\) is nonvanishing, then \(T_{\phi}\) is invertible in the Calkin algebra, that is, it is Fredholm. From the point of view of several complex variables, the interest stems from the connection with the \(\overline{\partial}\)-Neumann problem: the Hankel operator \(H_{\phi}\) on the Bergman space can be expressed as \(H_{\phi}f = \overline{\partial}^{*}N\overline{\partial}(\phi f)\). Here, \(\Omega\) is a bounded pseudoconvex domain in \(\mathbb{C}^{n}\) and \(N\) is the \(\overline{\partial}\)-Neumann operator on \((0,1)\)-forms. It is well known that, if \(N\) is compact, then so is \(H_{\phi}\) for all \(\phi\in C(\overline{\Omega})\). It is natural to ask what happens if the \(\overline{\partial}\)-Neumann operator is not compact. The paper under review makes a contribution to this circle of ideas. The main result is the following. Theorem 1. Let \(\Omega\subset\mathbb{C}^{2}\) be a bounded pseudoconvex complete Reinhardt domain with a \(C^{\infty}\)-smooth boundary. Suppose that \(\Omega\) is locally convex near \(\{(z_{1},z_{2}): z_{1}=0\}\cup\{(z_{1},z_{2}):z_{2}=0\}\). Also, suppose that there exists \(\{\Gamma_{k}\}_{k=1,\dots, l}\subset b\Omega\) so that for each \(k\in\{1,\dots, l\}\), \(\Gamma_{k}\) is a \(C^{\infty}\)-smooth family of analytic discs generated by rotations. Furthermore, \(\Gamma_{\Omega} = \cup_{k \in \{1,\dots, l\}}\Gamma_{k}\). Assume that \(\phi\in C(\overline{\Omega})\). Then \(\phi\circ f\) is holomorphic for any holomorphic \(f:\mathbb{D}\to b\Omega\) if and only if \(H_{\phi}\) is compact on \(A^{2}(\Omega)\). The ``if'' part appeared in [Integral Equations Oper. Theory 90, No. 6, Paper No. 71, 14 p. (2018; Zbl 1459.47013)]; here, the author proves the ``only if'' direction. Roughly speaking, there are four ingredients in the proof. First, on convex domains, one has good compactness estimates for the \(\overline{\partial}\)-Neumann operator away from analytic discs in the boundary. Second, Reinhardt domains are locally convexifiable away from the coordinate axes; the additional assumptions guarantee that they also are near points on the coordinate axes. Third, compactness of Hankel operators localizes suitably; essentially, one is now working on a convex domain. Fourth, and finally, the analyticity of the symbol along the analytic discs in the boundary provides the required estimates near these discs.