2
$\begingroup$

Let A and B be two unital $C^∗$-algebras, and $x∈A⊗B$ (the algebraic tensor product $*$-algebra), different from $0$. Is there states $ω_x∈A^∗_+$ and $φ_x∈B^∗_+$ such that, for the product state $ω_x×φ_x$, $(ω_x×φ_x)(x^∗x)>0$? In other words, do the product states separate the points of the tensor product $A⊗B$?

$\endgroup$
2
$\begingroup$

Yes. Lets write $$ 0 \neq x = \sum_{i=1}^n a_i \otimes b_i, $$ so that the $b_i$ are linearly independent. For the moment, fix a state $\varphi$ on $A$. This gives rise to the so-called slice map $$ \eta_\varphi \colon A \otimes B \to B : a \otimes b \mapsto \varphi(a)b. $$ For this $\varphi$, we get $$ \eta_\varphi(x) = \sum_{i=1}^n\varphi(a_i)b_i. $$ Clearly, one can choose $\varphi$ so that $\eta_\varphi(x) \neq 0$, since otherwise, by the linear independence of the $b_i$, $\varphi(a_i) = 0$, for any state $\varphi$ on $A$ and $i \in \{1,2,\cdots,n\}$. That would imply that $a_1=a_2=\cdots=a_n=0$ and therefore $x= 0$. Now, we may take a state $\psi$ on $B$ such that $$ \psi(\eta_{\varphi}(x)) > 0. $$ But $\psi(\eta_{\varphi}(x)) = (\varphi \otimes \psi)(x)$.

$\endgroup$
  • $\begingroup$ The question is essentially answered in: mathoverflow.net/questions/108254/… $\endgroup$ – francesco fidaleo Oct 7 '18 at 22:25
  • 1
    $\begingroup$ That answer uses way more machinery (and it depends on looking a result in a book), while the above is a straightforward elementary argument, that doesn't use tensor norms nor representations. All it requires is the fact that states separate points. $\endgroup$ – Martin Argerami Oct 8 '18 at 6:17
  • $\begingroup$ The above is not answering that I asked. $\endgroup$ – francesco fidaleo Oct 8 '18 at 7:27
  • 2
    $\begingroup$ Yes it does. It shows that product states separate points, which is exactly what you asked. $\endgroup$ – Martin Argerami Oct 8 '18 at 15:51
  • $\begingroup$ better to read the details of my question: $x\neq0$ implies $\omega\times\varphi(x^*x)>0$, this is usually meant for "separate points" in operator algebras. By the way, I asked for that and the answer didn't respond about that question. $\endgroup$ – francesco fidaleo Oct 8 '18 at 20:26

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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