Regularity of projection operators attached to worm domains (Q273549): Difference between revisions

In 1996, \textit{M. Christ} [J. Am. Math. Soc. 9, No. 4, 1171--1185 (1996; Zbl 0945.32022)] proved that the Bergman projection does not preserve smoothness of functions up to the boundary on the so called worm domains of Diederich and Fornæss. A crucial ingredient in the proof is the fact, shown by the first author [\textit{D. E. Barrett}, Acta Math. 168, No. 1--2, 1--10 (1992; Zbl 0779.32013)], that on certain unbounded model worm domains, the Bergman projection does not preserve \(L^{2}\)-Sobolev spaces. In the paper under review, the authors address the question of whether there may be other (oblique) projections form \(L^{2}\) onto the Bergman space of these model worms that do preserve Sobolev spaces. Denote by \(D_{\beta}\) one of the unbounded model worm domains (\(\beta\) is as usual the parameter that measures the winding). Because of the rotational symmetry in the second variable, the Bergman space on \(D_{\beta}\) decomposes as an orthogonal direct sum Fourier decomposition \(\bigoplus _{j\in \mathbb{Z}} B_{j}(D_{\beta})\). The authors construct projections onto the subspaces \(B_{j}(D_{\beta})\) which are continuous in Sobolev norms, on the subspace \(W^{s}_{\mathcal{D}}(D_{\beta})\). Namely, for each \(j\in \mathbb{Z}\), there exists a projection \(T_{j}: L^{2}(D_{\beta}) \rightarrow B_{j}(D_{\beta})\) which satisfies \(T_{j}: W^{s}_{\mathcal{D}}(D_{\beta}) \rightarrow W^{s}_{j}(D_{\beta})\). Here \(W^{s}_{\mathcal{D}}(D_{\beta})\) is the closure of \(C^{\infty}_{0}(D_{\beta})\) in the Sobolev-\(s\) norm, and \(W^{s}_{j}(D_{\beta}) = L^{2}_{j}(D_{\beta})\cap W^{s}(D_{\beta})\). The projections \(T_{j}\) are constructed via their kernels. These kernels are modifications of the ordinary Bergman kernel. Namely, the work in [Zbl 0779.32013] allows to explicitly identify how the obstructions to regularity of the Bergman projection arise in the Bergman kernel. The authors find a way to modify the Bergman kernel so that these obstructions disappear (for \(j\) fixed). As the authors point out, two questions arise immediately. First, are there operators so that \(W^{s}_{\mathcal{D}}(D_{\beta})\) can be replaced by \(W^{s}(D_{\beta})\)? Second, is there a projection \textit{onto} the whole Bergman space, \(T: L^{2}(D_{\beta}) \rightarrow B(D_{\beta})\), that is continuous in Sobolev norms (the norm estimates available on the \(T_{j}\) do not allow to sum them up).