C<sup>*</sup>-Convexity and the Numerical Range (Q4498147)

From MaRDI portal

!

This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:

scientific article; zbMATH DE number 1491115
Language Label Description Also known as
default for all languages
No label defined
    English
    C<sup>*</sup>-Convexity and the Numerical Range
    scientific article; zbMATH DE number 1491115

      Statements

      C<sup>*</sup>-Convexity and the Numerical Range (English)
      0 references
      0 references
      14 August 2000
      0 references
      \(C^*\)-extreme point
      0 references
      numerical range
      0 references
      prime \(C^*\)-algebra
      0 references
      \(C^*\)-convex hull
      0 references
      Krein-Milman theorem
      0 references
      weak\(^*\) compact \(C^*\)-convex subsets of injective factors
      0 references
      The paper starts with a characterization of the numerical range of elements in certain unital \(C^*\)-algebras: If \(a\) is an element of a prime \(C^*\)-algebra \({\mathcal A}\), then \(W_{\mathcal A}(a)\), the numerical range of \(a\) in \({\mathcal A}\), consists of those complex numbers \(\lambda\) such that for any \(\varepsilon> 0\), there exists a positive unit element \(h\) in \({\mathcal A}\) with \(\|h(a- \lambda)h\|< \varepsilon\). The same approximation condition holds if \({\mathcal A}\) is an arbitrary \(C^*\)-algebra and \(\lambda\) is an extreme point of \(W_{\mathcal A}(a)\). Results of this type correspond to the known ones for \({\mathcal A}\) the \(C^*\)-algebra of all bounded linear operators on a Hilbert space or the Calkin algebra (\(\lambda\) is in \(W_{\mathcal A}(a)\) if and only if there is a nonzero projection \(p\) such that \(p(a- \lambda)p= 0\)), although in general, we can only have this approximate version and \(h\) cannot be taken to be a projection. This characterization of the numerical range is then used to determine, for a von Neumann algebra \({\mathcal R}\) (resp., a unital \(C^*\)-algebra \({\mathcal A}\)), the normal elements in the weak\(^*\) closure (resp. norm closure) of the \(C^*\)-convex hull of an element \(a\) of \({\mathcal R}\) (resp., \({\mathcal A}\)). (The \(C^*\)-convex hull of \(a\) in \({\mathcal A}\) is the set of all elements of the form \(\sum^n_{j=1} v^*_j av_j\), where \(n\) is some natural number, and \(v_j\)'s are in \({\mathcal A}\) satisfying \(\sum^n_{j=1} v^*_j v_j= 1\).) For instance, it is proved that if \(a\) and \(b\) are elements in a von Neumann algebra \({\mathcal R}\) with \(b\) normal, then \(b\) is in the weak\(^*\) closure of the \(C^*\)-convex hull of \(a\) if and only if \(W_{{\mathcal R}p}(bp)\subseteq W_{{\mathcal R}p}(ap)\) for every central projection \(p\). From this, the author deduces the existence of \(C^*\)-extreme points for any weak\(^*\) compact \(C^*\)-convex subset of a von Neumann algebra, generalizing the corresponding result for the matrix algebra \(M_n(\mathbb{C})\). A variant of the Krein-Milman theorem for weak\(^*\) compact \(C^*\)-convex subsets of injective factors is also proved.
      0 references

      Identifiers

      0 references
      0 references
      0 references
      0 references
      0 references
      0 references