Boundedness and compactness of Hankel operators on the sphere (Q935042)

From MaRDI portal
Revision as of 00:35, 20 March 2024 by Openalex240319060354 (talk | contribs) (Set OpenAlex properties.)
scientific article
Language Label Description Also known as
English
Boundedness and compactness of Hankel operators on the sphere
scientific article

    Statements

    Boundedness and compactness of Hankel operators on the sphere (English)
    0 references
    0 references
    31 July 2008
    0 references
    Let \(S\) denote the unit sphere in \({\mathbb C}^n\) and \(\sigma\) denote the normalized positive, regular Borel measure on \(S\) invariant under the orthogonal group \(O(2n)\). The Cauchy projection is denoted by \(P\), i.e., \((Pf)(w)=\int_S \frac{f(\xi)}{(1-\langle w,\xi\rangle)^n}\,d\sigma(\xi)\), \(| w| <1\), the multiplication operator by \(M_f\), i.e., \(M_f(g)=fg\), and the Hankel operator by \(H_f\), i.e., \(H_f=(1-P)M_f| _{H^2(S)}\), where \(H^2(S)\) stands for the Hardy space. In this clearly written and interesting paper, the author contributes to the study of the compactness of \(H_f\). Although the compactness result given by \textit{D\,Zheng} [J.~Funct.\ Anal.\ 149, No.\,1, 1--24 (1997; Zbl 0909.47020)] which says that, for \(f\in BMO\), \(H_f\) is compact if and only if \(\lim_{| z| \to 1} \| H_f k_z\| =0\), where \(k_z\) is the normalized Cauchy kernel \(k_z(w)=\frac{(1-| z| ^2)^{n/2}}{(1-\langle w, z\rangle)^n}\), the condition appearing there does not give much information about the symbol \(f\). Due to the connection between \(H_f\) and commutators, from the fundamental result by Coifman, Rochberg and Weiss saying that the commutator \([P,M_f]\) is compact on \(L^2(S)\) if and only if \(\text{VMO}(S)\) and using the fact the \(H_f=H_{f-Pf}\), the author observes that if \(f\in L^\infty(S)\) and \(f-Pf\in \text{VMO}(S)\), then \(H_f\) is compact. In fact, he shows that the converse it true, that is, if \(H_f\) is compact then \(f-Pf\in \text{VMO}(S)\). The main result actually gives a bit more. In fact, denoting the local mean oscillation of \(f\) at the point \(\xi\) by \(\text{LMO}(f)(\xi)= \lim_{\delta\to 0} \sup_{B(\eta,r)\subset B(\xi,\delta)} \frac{1}{\sigma(B(\eta,r))}\int_{B(\eta,r)} | f-f_{B(\eta,r)}| \,d\sigma\), it is proved that if \(f\in \text{BMO}(S)\) and \(\xi\in S\) satisfies that \(\lim_{z\to \xi} \| H_fk_z\| =0\), then the local mean oscillation of \(f-Pf\) at \(\xi\) vanishes, i.e., \(\text{LMO}(f-Pf)(\xi)=0\). He also shows the estimate \(\| f-Pf\|_{\text{BMO}}\leq C\sup_{| z| <1}\| H_fk_z\|\) and the fact that if \(f\in L^2(S)\) and \(\sup_{| z| <1}\| H_fk_z\| <\infty\), then \(f-Pf\in \text{BMO}\). This allows him to conclude that if \(f\in L^2(S)\) and \(\lim_{| z| \to 1}\| H_fk_z\| =0\), then \(f-Pf\in \text{VMO}\) and consequently, \(H_f\) naturally extends to a compact operator mapping \(H^2(S)\) into \(L^2(S) \ominus H^2(S)\).
    0 references
    0 references
    Hankel operator
    0 references
    mean oscillation
    0 references

    Identifiers