A A C*-algebra A and a closed bi-ideal J of A such that [ A/ J] and J are liminal, but A is not liminal.

******************************

Let H be an infinite dimensional Hilbert space and IH be the identity operator on H. Then A = K(H)+C IH isn't liminal (otherwise, since identity representation K(H)+C IH® B(H) is nonzero irreducible ((K(H)+C IH)¢ = K(H)¢ = CIH (see CW17)) we should have IH Î K(H), a contradiction). But K(H) is liminal, since each nonzero irreducible representation of K(H) is unitarily equivalent to identity representation K(H)® B(H) (see [Mur, page 146]). Also [ A/ J] @ CIH which is finite dimensional and so is liminal (Every finite dimensional C*-algebra B is liminal, since if (H1,Y) is a nonzero irreducible representation of B, then for x ¹ 0 in H1, Y(B)x is finite dimensional (for Y(B) is finite dimensional). If (ul)l be any approximate unit for B, then (Y(ul))l strongly converges to IH so x Î [Y(B)x] = Y(B)x. Hence Y(B)x is a nonzero (closed) subspace of H1 invariant for Y(B), so by irreducibility, Y(B)x = H1. Therefore H1 is finite dimensional. Thus Y(B) Í B(H1) = K(H1)).

Ref.
[Mur] G.J. Murphy, C*-algebras and operator theory, Academic Press, 1990.


File translated from TEX by TTH, version 2.70.
On 22 Feb 2001, 00:19.