On \(k\)-decomposability of positive maps (Q2493837): Difference between revisions

A bounded linear operator \(\phi : A \to B\) between two \(C^*\)-algebras is called \(k\)-positive if \(\phi_k : M_{k} (A) \to M_{k} (B),\) \(\phi_k ([a_{i,j}]) = [\phi(a_{i,j})]\), defines a positive linear operator; where \(M_k (A)\) and \(M_{k} (B)\) denote the \(C^*\)-algebras of \(k\times k\) matrices with entries on from \(A\) and \(B,\) respectively. \(\phi\) is said to be completely positive if it is \(k\)-positive for all \(k\in \mathbb{N}\). Stinespring's theorem affirms that every completely positive map \(\phi : A\to B(H)\) has the form \(\phi (a) = W^* \pi (a) W\), where \(\phi\) is a *-representation of \(A\) on some Hilbert space \(K\) and \(W\) is a bounded linear operator from \(H\) to \(K\). In [Acta Math.\ 110, 233--278 (1963; Zbl 0173.42105)], \textit{E.\,Størmer} introduced the following definition: a map \(\phi : A\to B(H)\) is \textit{decomposable} if there exist a Hilbert space \(K\), a Jordan morphism \(J: A \to B(K)\) and a bounded linear operator \(W: H\to K\) satisfying \(\phi (a) = W^* J (a) W\). In the paper under review, the authors also consider \(k\)-copositive and completely copositive linear operators. A continuous linear operator \(\phi : A \to B(H)\) is \(k\)-copositive (respectively, completely copositive) if the operator \(a\mapsto a^t := \sigma a^* \sigma\) is \(k\)-positive (respectively, completely copositive), where \(\sigma\) is a conjugation on \(H\). A complete description of decomposable operators was made by \textit{E.\,Størmer} in [Proc.\ Am.\ Math.\ Soc.\ 86, 402--404 (1982; Zbl 0526.46054)]. The classification of decomposable operators remains open up today, even in the finite-dimensio\-nal case (\(A=B(\mathbb{C}^{m})\) and \(H=\mathbb{C}^{m}\)). Concerning this problem, we can find find some previous contributions by Størmer, Choi and Woronowicz. In the authors' words, the present paper is a step toward a canonical prescription for the construction of decomposable and non decomposable maps. The paper presents very good descriptions of \(k\)-positive and \(k\)-copositive operators. Section~4 is devoted to a modification of the Tomita--Takesaki theory.