Numerical Ranges of Operators of Class C₀ (Q3147571)

Given an operator \(T\) on a complex Hilbert space \(H\), denote by \(W(T)=\{ (Tx, x): \|x\|=1\}\) the numerical range of \(T\), and by \(w(T)=\sup\{ |\lambda|: \lambda\in W(T)\}\) the numerical radius of \(T\). Assume that \(T\) is a nilpotent contraction, say \(T^n=0\) for some \(n\in{\mathbb N}\), and denote by \(J_n\) the Jordan cell of size \(n\) with eigenvalue zero. \textit{U. Haagerup} and \textit{P. de la Harpe} proved in [Proc. Am. Math. Soc. 115, 371-379 (1992; Zbl 0781.47014)] that \(w(T)\leq w(J_n)\), and that \(J_n\) must be an orthogonal direct summand of \(T\) if there is \(\lambda\in W(T)\) with \(|\lambda|=w(J_n)\). The author examines this theorem in the context of dilatation theory, and generalizes this result to operators of class \(C_0\), i.e., to completely nonunitary operators \(T\) such that \(u(T)=0\) for some nonzero function \(u\) in the Hardy space \(H^{\infty}\) on the unit disk. He shows that the numerical range of an operator \(T\) of class \(C_0\) is contained in the numerical range of the top block in its Jordan model. The operator contains this block as an orthogonal direct summand provided that the top block is finite dimensional, and one of the extreme points of its numerical range belongs to the numerical range of \(T\).