I am reading a book about C*-algebra. But i am confused with some of its content. It says

Assume $A$ is a non-unital C*-algebra and $\tilde{A}$ is its unitization (the elements of the form $a+\lambda$). If $\phi: A\rightarrow \mathbb{C}$, is a nonzero homomorphism($\mathbb{C}$ denotes complex field), then $\tilde{\phi}(a+\lambda)=\phi(a)+\lambda~(a\in A$ and $\lambda\in \mathbb{C})$. Let $P(A)$ and $P(\tilde{A})$ denote the nonzero multiplicative linear functional on A and $\tilde{A}$, respectively. Thus, $P(A)$ is a subset of $P(\tilde{A})$. In fact, $P(A)=P(\tilde{A})\cup\{\pi\}$, where $\pi$ is determined by $\tilde{A}/A=\mathbb{C}$, i.e., $\ker\pi=A$. Since $P(\tilde{A})$ is compact Hausdorff space, we conclude that $P(A)$ is a locally compact Hausdorff space.

My question are

  1. Does the "=" in the equation $P(A)=P(\tilde{A})\cup\{\pi\}$ is under the meaning of isomorphism? Is it topological isomorphism? Why?

  2. Why does the $P(\tilde{A})$ is compact Hausdorff space implies that $P(A)$ is a locally compact Hausdorff space?


First of all, you have to consider $P(A)$ as the set of all non-zero multiplicative linear functionals on $A$ with the weak$^*$-topology inherited from $A^*$, the dual of $A$. When $A$ is unital $P(A)$ is compact. So $P(\tilde{A})$ is always compact.

(1) It is $P(\tilde{A})=P(A)\cup \pi$. In fact, this equality reveals a more important fact which is $P(\tilde{A})$ is the one point compactification of $P(A)$.

(2) Yes. The reason is easy, because every compact subset of $P(\tilde{A})$ which does not contain $\{\pi\}$ (which is exactly the point at infinity) is compact in $P(A)$ too.

  • $\begingroup$ Could you explain more specifically ? I am a beginner of C*-algebra. $\endgroup$ – Insomnia Oct 10 '13 at 17:31
  • $\begingroup$ Actually, I have explained this topic in my lecture notes on $C^*$-algebra, which is available at arxiv:1211.3404. More specifically, see Section 3.1. Of course, I denote $P(A)$ by $\Omega(A)$. $\endgroup$ – user56706 Oct 10 '13 at 17:35
  • 1
    $\begingroup$ Thanks you and also thanks for your book~ :) $\endgroup$ – Insomnia Oct 10 '13 at 17:54
  • $\begingroup$ (2) is not quite as obvious as you say; not every subspace of a compact space is locally compact. (Consider $\mathbb{Q} \cap [0,1]$ as a subspace of $[0,1]$.) $\endgroup$ – Nate Eldredge Oct 10 '13 at 18:18
  • $\begingroup$ @NateEldredge: Thanks Nate, you are right. We have to use the fact that $P(\tilde{A})$ is the one point compactification of $P(A)$. Then every compact neighborhood of a point $x\in P(A)$ in $P(\tilde{A})$ which does not meet $\{\pi\}$ is actually compact in $P(A)$ too. $\endgroup$ – user56706 Oct 10 '13 at 18:25

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.