$C([0,1])$ is a $C^*$ algebra of complex functionals. $A$ is a $C^*$ algebra. Hence $C([0,1],A)$, the continuous functions from $[0,1]$ to $A$ is also a $C^*$ algebra.

We construct its tensor product in the category of $C^*$ algebras. This is unique as $C([0,1])$ is nuclear.

There is a canonical $*$-homomorphism, $$C[0,1] \otimes A \rightarrow C([0,1],A)$$ $$(f \otimes a) (x) = f(x)a $$ How does one show this is in fact an isomorphism?


The map is $\pi:f\otimes a \longmapsto f a$. It's obvious that it is linear and multiplicative, and preserves adjoints, so it is a $*$-homomorphism. It is injective: if $\pi(\sum_j f_j\otimes a_j)=0$, we may choose the $a_j$ so that they are linearlity independent. Then $$ 0=\pi(\sum_j f_j\otimes a_j)=\sum_j f_j a_j. $$ Evaluating at any $t\in[0,1]$, we have $0=\sum_j f_j(t)a_j$, and the linear independence gives $f_j(t)=0$ for all $j$. As we can do this for any $t$, $f_j=0$ for all $j$, and so $\sum_jf_j\otimes a_j=0$. Thus $\pi$ is injective (in reality one also needs to check that the map is well-defined, see this answer for details).

It remains to show that $\pi$ is onto. Since it is an isometry, it is enough to show that its range is dense. There you use compactness of $[0,1]$ to show that the functions of the form $\sum_j f_j a_j$ are dense. Concretely, let $f\in C([0,1],A)$, Fix $\varepsilon>0$, then there exists a partition $0=t_0<t_1<\ldots<t_m=1$ such that $\|f(t_j)-f(t)\|<\varepsilon$ for $t\in[t_j,t_{j+1}]$. Let $a_j=f(t_j)$ for each $j$; let $g_j=1_{[t_j,t_{j+1}]}$, and let $f_j\in C[0,1]$ such that $\|f_j-g_j\|<\varepsilon/(m\|f\|)$. Then, with $k$ such that $t\in[t_k,t_{k+1}]$, \begin{align} \|f(t)-\sum_j f_j(t)a_j\| &\leq \|f(t)-\sum_j g_j(t)a_j\|+\|\sum_j(g_j(t)-f_j(t))a_j\|\\ &\leq \|f(t)-\sum_j g_j(t)a_j\|+\varepsilon\\ &=\|f(t)-f(t_k)\|+\varepsilon\\ \ \\ &<2\varepsilon. \end{align}

  • $\begingroup$ I added more details. $\endgroup$ – Martin Argerami Apr 13 at 20:09
  • $\begingroup$ Thanks a lot Martin? I wonder if you can comment on my recent $K$ theory question? $\endgroup$ – CL. Apr 13 at 21:45
  • $\begingroup$ No problem. I don't usually chime in the K-theory questions because my knowledge is close to nil. $\endgroup$ – Martin Argerami Apr 13 at 22:35
  • $\begingroup$ Sorry, I didn't mean to type the "?" It was supposed to be "!" !! Your answers are great. $\endgroup$ – CL. Apr 13 at 22:45
  • $\begingroup$ I hadn't noticed :D $\endgroup$ – Martin Argerami Apr 13 at 22:48

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.