5
$\begingroup$

If $A$ is unital C$^*$-algebra, is it true that the multiplier algebra of $A \otimes \mathcal{K} $ is $ A \otimes \mathcal{B}(\mathcal{H})$? Where $\mathcal{K}$ is C$^*$-algebra of compact operators on the Hilbert space $\mathcal{H}$.

$\endgroup$

2 Answers 2

10
$\begingroup$

The fact stated in the answer by vap is proven in the paper "Multipliers of C*-algebras" by Akemann, Pedersen and Tomiyama (see Theorem 3.3, I guess). Moreover, they prove in Theorem 3.8 that multiplier algebras are not very well behaved with respect to minimal tensor products:

Let $A$ and $B$ be $C^*$-algebras and assume that $B$ has a countable approximate unit, but no unit (think of $\mathcal{K}$ here in your case) and that $A$ is infinite dimensional. Then
$$ M(A) \otimes M(B) \subsetneq M(A \otimes B) $$ where the tensor product is the minimal one.

So, in particular, for any infinite dimensional unital $C^*$-algebra $A$, the tensor product $A \otimes \mathcal{B}(\mathcal{H})$ is always a proper subalgebra of $M(A \otimes \mathcal{K})$.

$\endgroup$
8
$\begingroup$

If $A=C_0(X)$ and $B$ is a $C^\ast$-algebra then $M(A\otimes B)$ is the set of strictly continuous functions $\beta X\to M(B)$, where $\beta$ stands for Stone-Čech compactification.

If you take $X$ to be compact and $B=\mathcal{K}$ then we are in your setting.

But $C(X)\otimes\mathcal{B(H)}$ is the set of norm-continuos functions $\beta X=X\to\mathcal{B(H)}$. The strict topology is the $\sigma$-strong-$^\ast$ topology on $\mathcal{B(H)}$, which is different form the norm topology. This should answer your question in the negative.

$\endgroup$
4
  • $\begingroup$ What space $X$ do you have in mind? (It must be infinite, of course.) $\endgroup$
    – Yemon Choi
    May 25, 2016 at 15:06
  • 3
    $\begingroup$ The one-point compactification of the natural numbers should work. We can use this space to index a sequence converging in the strict topology of bounded operators, but not converging in norm. $\endgroup$
    – vap
    May 25, 2016 at 16:15
  • 1
    $\begingroup$ For future reference, as pointed out in the comments of mathoverflow.net/questions/375906 this is not true. $M(A\otimes B)$ may be identified with the space of strictly continuous, bounded functions $X\rightarrow M(B)$, and not from $\beta X$. $\endgroup$ Nov 8, 2020 at 12:13
  • 1
    $\begingroup$ @MatthewDaws Thanks. The error probably stems from the universal property of the Stone-Čech compactification: one would like to upgrade bounded functions on $X$ to continuous functions on the compactification. But this only works if the bounded sets in $M(B)$ are compact in the strict topology, which I think can fail even when $B$ is commutative. $\endgroup$
    – vap
    Nov 9, 2020 at 13:08

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.