Surjectivity of the Gelfand Transform map

311 Views Asked by At

This is a very basic question on understanding the Gelfand Transform...

Let $A$ be a unital commutative Banach Algebra.

Let $0\neq e\neq 1$ be an idempotent in $A$.

Let $\Omega(A)$ be the set of non-zero homomorphisms from $A$ to $\Bbb C$.

Show the map $\hat{e}:\Omega(A)\rightarrow \{0,1\}$ defined by $\phi\mapsto \phi(e)$ is surjective.

I see that we have $\phi(e ) = \phi(e^2) = \phi(e)^2= \cdots = \phi(e^n)= \phi(e)^n$.

Is this enough to show surjectivity? I'm a bit confused by the definitions.

Thank you for your help.

1

There are 1 best solutions below

0
On BEST ANSWER

The characters of $A$ (the nonzero homomorphisms $A\to\mathbb{C}$) are in a bijective correspondence with the maximal ideals of $A$. More precisely, the maximal ideals are exactly the kernels of the characters of $A$.

Now, since $e$ is an idempotent we have $e^2=e$, and so $e(e-1)=0$. If $e$ was invertible then we would get $e=1$, a contradiction. Thus $e$ is not invertible, and so it belongs to some maximal ideal $M\subseteq A$. (follows from Zorn's lemma) But this means $e$ belongs to the kernel of some character $\phi$, i.e there is $\phi\in\Omega(A)$ such that $\phi(e)=0$.

Also, since $e\ne 0$, the equality $e(e-1)=0$ implies that $e-1$ also cannot be invertible. So $e-1$ also belongs some maximal ideal, i.e to the kernel of some character $\varphi$. But then we get:

$0=\varphi(e-1)=\varphi(e)-\varphi(1)=\varphi(e)-1$

And so $\varphi(e)=1$ for this $\varphi$. So indeed there is a character $\varphi\in\Omega(A)$ such that $\varphi(e)=1$.

Alright, so your map $\Omega(A)\to\{0,1\}$ is indeed surjective.