Star order automorphisms on the poset of type 1 operators (Q2207657)

Let \(\mathcal{H}\) be a Hilbert space and \(\mathcal{B(H)}\) the algebra of all bounded linear operators on \(\mathcal{H}\). For \(A,B\in \mathcal{B(H)}\) we say that \(A\) is below \(B\) with respect to the star partial order and write \[ A\leq ^{\ast}B\quad \text{if}\quad A^{\ast}A=A^{\ast}B\text{ and } AA^{\ast}=BA^{\ast}. \] Let \(\mathcal{A}\) be a subset of \(\mathcal{B(H)}\). A map \(\Phi : \mathcal{A}\rightarrow \mathcal{A}\) is said to be a bi-preserver of the star partial order when \[ A\leq ^{\ast}B\quad \text{if and only if}\quad \Phi (A)\leq ^{\ast}\Phi (B) \] for every \(A,B\in \mathcal{A}\). \textit{P. Legiša} characterized in [Linear Multilinear Algebra 54, No. 3, 157--188 (2006; Zbl 1113.15003)] surjective bi-preservers of the star partial order on \(\mathcal{B(H)}\) when \(3\leq \dim \mathcal{H<\infty}\). The infinite-dimensional case was studied by \textit{G. Dolinar} et al. [Oper. Matrices 7, No. 1, 225--239 (2013; Zbl 1295.47038)]; however, the authors found an example of a Hilbert space \(\mathcal{H}\) and an operator \(A\in \mathcal{B(H)}\) such that there is no rank one operator \(C\in \mathcal{B(H)}\) with \(C\leq ^{\ast}A\) and thus they limited their research of star partial order bi-preservers to the subset of all compact operators in \(\mathcal{B(H)}\) where this ``rank-one problem'' does not occur. Several authors continued to study preservers of the star partial order and tried to answer the question what is the infinite-dimensional version of Legiša's result and the paper under review is one of such papers. Let \(x,y\in \mathcal{H}\) be nonzero vectors. We denote by \(x\otimes y\in \mathcal{B(H)}\) a rank one operator defined by \((x\otimes y)z=\left\langle z,y\right\rangle x\), \(z\in \mathcal{H}\). Note that every rank one operator in \(\mathcal{B(H)}\) can be written in this form. Let \(A\in \mathcal{B(H)}\). If \(A=0\) or \(A=\sup \left\{ x\otimes y:x\otimes y\leq ^{\ast}A\right\}\), then we say that \(A\) is of type 1. If there are no rank one operators \(x\otimes y\) such that \(x\otimes y\leq ^{\ast}A\), then we say that \(A\) is of type 2. We denote by \(\mathcal{B(H)}_{I}\) and \(\mathcal{B(H)}_{II}\) the sets of all type 1 and type 2 operators in \(\mathcal{B(H)}\), respectively. It turns out that \(\mathcal{B(H)}_{I}\) is dense in \(\mathcal{B(H)}\) in the norm topology and that for any \(A\in \mathcal{B(H)}\) there exists a unique pair of operators \(A_{1}\in \mathcal{B(H)}_{I}\) and \(A_{2}\in \mathcal{B(H)}_{II}\) such that \(A_{i}\leq ^{\ast}A\) for \(i=1,2\) and \(A=A_{1}+A_{2}\). The authors show that \(\Phi\) is a bijective bi-preserver of the star partial order if and only if there exist bijective bi-preservers of the star partial order \(\Phi_{1}\) and \(\Phi_{2}\) on \(\mathcal{B(H)}_{I}\) and \(\mathcal{B(H)}_{II}\), respectively, such that \(\Phi (A)=\Phi_{1}(A)+\Phi_{2}(A)\) for all \(A=A_{1}+A_{2}\), where \(A_{1}\in \mathcal{B(H)}_{I}\) and \(A_{2}\in \mathcal{B(H)}_{II}\). With the main result of the paper, the authors characterize bijective bi-preservers on \(\mathcal{B(H)}_{I}\) and as a corollary obtain the form of continuous, bijective bi-preservers on \(\mathcal{B(H)}\).