Quadratic hyponormality and 2-hyponormality for Toeplitz operators (Q2494149)

This paper is concerned with the problem of bridging the gap between subnormality and hyponormality of Toeplitz operators on the Hardy space \(H^2(\mathbb T)\). For a complex Hilbert space \(\mathcal H\), let \({\mathcal L}({\mathcal H})\) denote the algebra of bounded linear operators on \(\mathcal H\). An operator \(T\in{\mathcal L}({\mathcal H}C)\) is said to be normal if \(T^*T = TT^*\), hyponormal if the commutator \([T^*,T] := T^*T - TT^*\geq 0\), and subnormal if \(T\) has a normal extension to some Hilbert space \({\mathcal K}\supseteq {\mathcal H}\). It is known that \(\text{normal}\Rightarrow \text{subnormal}\Rightarrow\text{hyponormal}\). In order to measure the gap between hhyponormality and subnormality, operator theorists have introduced different concepts of hyponormality. An operator \(T\) is called \(k\)-hyponormal if the \(k\times k\) operator matrix \(([T^*,T^i])^k_{i,j=1}\) is positive. By the Bram-Halmos criterion, we know that \(T\) is subnormal if and only if \(T\) is \(k\)-hyponormal for every \(k > 1\). Also, we say \(T\in{\mathcal L}(\mathcal H)\) is weakly \(k\)-hyponormal if \(\{\sum^k_{j=1}\alpha_jT^j :\alpha = (\alpha_1,\dots,\alpha_k)\in \mathbb C^*\}\) consists entirely of hyponormal operators. If \(k = 2\), then \(T\) is called quadratically hyponormal. For \(\varphi\in L^\infty(\mathbb T)\), the Toeplitz operator \(T\) is defined on \(H^2(\mathbb T)\) by \(T_\varphi q=P(\varphi g)\), where \(P : L^2(\mathbb T)\to H^2(\mathbb T)\) is the orthogonal projection. \textit{P.~R.\ Halmos} [Bull.\ Am.\ Math.\ Soc.\ 76, 887--933 (1970; Zbl 0204.15001)] asked if every subnormal Toeplitz operator is either normal or analytic. This problem was answered in the negative by \textit{C.~C.\ Cowen} and \textit{J.~J.\ Long} [J.~Reine Angew.\ Math.\ 351, 216--220 (1984; Zbl 0532.47019)]. They used a very special symbol; for \(0 < \alpha < 1\), let \(\psi\) be a conformed map of the unit disk onto the interior of the ellipse with vertices \(\pm(1+\alpha)i\) and passing through \(\pm(1-\alpha)\). If \(\varphi=(1-\alpha^2)^{-1}(\psi+\alpha\overline\psi)\), then \(T_\varphi\) is subnormal but neither normal nor analytic. In [Integral Equations Oper.\ Theory 44, No.~2, 138--148 (2002; Zbl 1042.47016)], \textit{R.~E.\ Curto} and the authors used the symbols \(\psi+\lambda\overline \psi\) to show that there is a gap between 2-hyponormality and subnormality for Toeplitz operators. In the present paper, the authors consider the relationship between quadratic hyponormality and 2-hyponormality. They prove that for \(0 <\alpha < 1\) and \(\psi\) as above, if \(\varphi=\psi +\lambda\overline \psi\), then \(T_\varphi\) is quadratically hyponormal if and only if \(T_\varphi\) is 2-hyponormal.