Kernel of vector-valued Toeplitz operators (Q989961)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Kernel of vector-valued Toeplitz operators
scientific article

    Statements

    Kernel of vector-valued Toeplitz operators (English)
    0 references
    0 references
    23 August 2010
    0 references
    Let \(H^2\) be the Hardy space of analytic functions on the unit disc \(\mathbb{D}\) and let \(S\) denote the shift operator on \(H^2\). A closed subspace \(\mathcal{F}\) of \(H^2\) is called ``nearly \(S^*\)-invariant'' if \(S^*\) maps \(\mathcal{F}_0:=\{ f\in \mathcal{F} : f(0)=0\}\) into \(\mathcal{F}\). Note that, for any \(\varphi\in L^{\infty}\), the kernel of the Toeplitz operator \(T_{\varphi}=p_+M_{\varphi}\) defines a nearly \(S^*\)-invariant subspace of \(H^2\). Here, \(p_+:L^2\rightarrow H^2\) denotes the orthogonal projection and we write \(M_{\varphi}\) for the multiplication by \(\varphi\). In [Pac. J. Math. 134, No.~1, 101--120 (1988; Zbl 0662.30035)], \textit{D. Hitt} classified such type of spaces \(\mathcal{F}\) and proved a factorization of the elements \(f\in \mathcal{F}\) in the form \(f=gf_0\) with a certain function \(g\in \mathcal{F}\) of unit norm and a unique element \(f_0\in K_u:=H^2\ominus uH^2\), where \(u\) denotes an inner function vanishing at the origin. Conversely, given a unit norm function \(g\) in \(H^2\) and an inner function \(u\) with \(u(0)=0\), one may ask when the multiplication by \(g\) acts isometrically from \(K_u\) to \(\mathcal{F}=\{gf_0 : f_0\in K_u\}\). In [Operator theory and functional analysis, Proc. Conf., Mesa/Ariz. 1987, Oper. Theory, Adv. Appl. 35, 481--493 (1988; Zbl 0687.47003)], \textit{D. Sarason} gave two equivalent statements of this problem. Also, the question arises under which conditions a nearly \(S^*\)-invariant subspace \(\mathcal{F}=gK_u\) arises as the kernel of a Toeplitz operator. In [Integral Equations Oper. Theory 9, 588--591 (1986; Zbl 0636.47023)], \textit{E. Hayashi} gave a complete solution. In the present paper, the above results by Hitt, Sarason and Hayashi are generalized to the setting of vector valued functions. In particular, new difficulties arise due to the non-commutativity of the matrix multiplications. If \(H^2(\mathbb{C}^m)\) denotes the \(\mathbb{C}^m\)-valued Hardy space, then the non-trivial \(S^*\)-invariant subspaces \(\mathcal{F}\subset H^2(\mathbb{C}^m)\) are characterized in Theorem 2.1. In Theorem 3.5, the matricial version of Sarason's theorem is given. In Section 4, the author gives various interesting and concrete examples of \(S^*\)-nearly invariant spaces \(\mathcal{F}\) in the multi-dimensional setting and he discusses the question whether \(\mathcal{F}\) arises as the kernel of a Toeplitz operator. Finally, the generalization of Hayashi's result can be found in Section 5 of the paper.
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    de~Branges-Rovnyak spaces
    0 references
    nearly \(S^*\)-invariant space
    0 references
    vector valued functions
    0 references
    Toeplitz operators
    0 references
    0 references
    0 references
    0 references