Decomposition theorems for \(\mathcal Q_p\) spaces (Q1426914): Difference between revisions

The Möbius invariant spaces \(Q_ p\) and \(Q_{p,0}\) of analytic functions are considered. For example, an analytic function \(f \) in the unit disk \(D \) is said to belong to \(Q_ p \) if \[ \sup_{w\in D} \int_ D | f^\prime(z)|^2 g(z,w)^ p \,dA(z) < \infty \,, \] while \(Q_{p,0}\) is a ``little-oh'' version of this space. Obviously, \(Q_ 0\) yields the Dirichlet space; it is also well known that \(Q_ 1=BMOA\). (The paper might contain a slight inaccuracy here, as the authors seem to suggest that this was proved by \textit{A. Nicolau } and \textit{J. Xiao} in, J. Funct. Anal. 150. No. 2, 383--425 (1997; Zbl 0886.30036); although the paper cited certainly contains several quite relevant results in the field, the two facts mentioned above were certainly known earlier and probably were the starting point in defining the \(Q_ p\) spaces.) It is not difficult to see that \(Q_ p\) coincides with the Bloch space whenever \(p>1 \) while it only contains constant functions when \(-1<p<0 \). Thus, the interesting case is \(0<p<1 \), when an increasing scale of spaces is obtained. The study of these spaces started with the papers of R. Aulaskari, P. Lappan, J. Xiao, and R. Zhao; they have also been investigated by M. Essén, A. Nicolau, D. A. Stegenga, K. M. Dyakonov, and D. Girela, to mention a few authors. Besides the citations given in the paper, good reference are also the survey article by \textit{M. Essén} and \textit{J. Xiao} in [ Rep. Ser., Dept Math. Univ. Joensuu. 4, 41--60 (2001; Zbl 0981.30025)] and \textit{J. Xiao}'s lectures [Lecture Notes in Mathematics 1767, Berlin: Springer (2001; Zbl 0983.30001)]. The main results of the present paper are two atomic decomposition theorems: one for \(Q_ p\) and the other for \(Q_{p,0}\), in both cases when \(0<p\leq 1\). They assert that, given an appropriate lattice of points in the disk with the sufficiently large separation constant, any function in the space can be written as an infinite linear combination of certain generalized kernels corresponding to the points of the lattice. The coefficients of this combination are controlled by the norm of the function. Conversely, any infinite linear combination of given type is in the space provided that an associate discrete measure is of Carleson type. Besides Carleson measures, a key point in the proofs is the boundedness of certain integral operators. There is some resemblance here to the classical proof (due to V. P. Zaharjuta, V. I Judovich and F. Forelli, W. Rudin of the boundedness of the Bergman projection via Schur's theorem. In summary, the main results constitute an analogue of the known theorems for other function spaces in the disk (Hardy, Bergman, etc.). For example, a theorem of this type for Bergman spaces was proved by \textit{R. Coifman} and \textit{R. Rochberg} [Astérisque 77, 1--66 (1980; Zbl 0472.46040)].