Duality of the Bergman spaces on some weakly pseudoconvex domains (Q5928661): Difference between revisions

The article is a summarized version of the author's dissertation under the direction of R. M. Range. Let \(D\) be a smoothly bounded pseudoconvex domain of finite type in \(\mathbb C^2\), and for \(1 \leq p < \infty\), let \(\mathcal O L^p (D) := L^p (D) \cap \mathcal O(D)\) be the Bergman space of holomorphic functions on \(D\) such that the \(p\)th power of their modulus is integrable with respect to Euclidean volume. Let furthermore \(\mathcal B (D)\) stand for the Bloch space of functions holomorphic on \(D\) such that the norm of the complex normal derivative multiplied by the distance to the boundary is uniformly bounded on \(D\). The main results of the article are that, (A) for \(1<p<\infty\), the dual of \(\mathcal O L^p (D)\) is \(\mathcal O L^q (D)\), where \((1/p)+(1/q)=1\); and (B) that the dual of \(\mathcal O L^1 (D)\) can be identified with \(\mathcal B (D)\). This was known for other classes of domains, in particular strictly pseudoconvex domains in any dimension. The proof of (A) relies (as can be expected) on the fact that the Bergman projection (i.e. the orthogonal projection from \(L^2 (D)\) to \(\mathcal O L^2 (D)\)) is bounded on \(L^p(D)\) (as an operator densely defined on \(L^p(D) \cap L^2(D)\)) for \(1<p<\infty\). This in turn follows from the main Lemma of the paper, estimates about the integral of the Bergman kernel multiplied by powers of the distance to the boundary, and the integral of iterated normal derivatives of the Bergman kernel. Those are derived from precise estimates due to \textit{J. D. McNeal} [Duke Math. J. 58, No. 2, 499-512 (1989; Zbl 0675.32020)] of the partial derivatives of the Bergman kernel in domains of finite type. A corollary of (A) is the density of \(\mathcal O L^r (D)\) in \(\mathcal O L^p (D)\) for \(1<p<2\), \(p<r<q\). A corollary of the boundedness of the Bergman projection and of the density results is the extension to \(\mathcal O L^p (D)\), \(1<p<\infty\), of the integral representation formula. The proof of (B) relies on a theorem of \textit{B. Coupet} [Proc. Am. Math. Soc. 102, No. 3, 493-501 (1988; Zbl 0659.46027)], on the estimates of the main Lemma, and on the fact that holomorphic functions smooth up to the boundary of \(D\) are dense in \(\mathcal O L^1 (D)\) (this is proved with the help of an appropriate solution of the \(\bar \partial\) equation due to \textit{A. Bonami} and \textit{N. Sibony} [J. Geom. Anal. 1, No. 4, 307-327 (1991; Zbl 0743.32015)]). Again, the integral representation formula can be extended to \(\mathcal O L^1 (D)\) (even though the Bergman projection is not bounded on \(\mathcal O L^1 (D)\)).