PlanetMath (more info)
 Math for the people, by the people.
Encyclopedia | Requests | Forums | Docs | Wiki | Random | RSS  
Login
create new user
name:
pass:
forget your password?
Main Menu
Owner confidence rating: High Entry average rating: No information on entry rating
injective $C^*$-algebra homomorphism is isometric (Theorem)

Theorem - Let $ \mathcal{A}$ and $ \mathcal{B}$ be $ C^*$-algebras and $ \Phi : \mathcal{A} \longrightarrow \mathcal{B}$ an injective *-homomorphism. Then $ \Vert\Phi(x)\Vert=\Vert x\Vert$ and $ \sigma(\Phi(x)) = \sigma(x)$ for every $ x \in \mathcal{A}$, where $ \sigma(y)$ denotes the spectrum of the element $ y$.

$ \,$

Proof: It suffices to prove the result for unital $ C^*$-algebras, since the general case follows directly by considering the minimal unitizations of $ \mathcal{A}$ and $ \mathcal{B}$. So we assume that $ \mathcal{A}$ and $ \mathcal{B}$ are unital and we will denote their identity elements by $ e$, being clear from context which one is being used.

Let us first prove the second part of the theorem for normal elements $ x \in \mathcal{A}$. It is clear that $ \sigma(\Phi(x)) \subseteq \sigma(x)$ since if $ x - \lambda e$ invertible for some $ \lambda \in \mathcal{C}$, then so is $ \Phi(x) - \lambda e = \Phi(x - \lambda e)$. Suppose the inclusion is strict, then there is a non-zero function $ f \in C(\sigma(x))$ whose restriction to $ \sigma(\Phi(x))$ is zero (here $ C(\sigma(x))$ denotes the $ C^*$-algebra of continuous functions $ \sigma(x) \longrightarrow \mathbb{C}$). Thus we have, by the continuous functional calculus, that $ f(x) \neq 0$ and also that

$\displaystyle \Phi(f(x))=f(\Phi(x))=0$    

by the continuous functional calculus and the result on this entry. Thus, we conclude that $ \Phi$ is not injective and which is a contradiction. Hence we must have $ \sigma(\Phi(x)) = \sigma(x)$.

Let $ R_{\sigma}(z)$ denote the spectral radius of the element $ z$. From the norm and spectral radius relation in $ C^*$-algebras we know that, for an arbitrary element $ x \in \mathcal{A}$, we have that

$\displaystyle \Vert x\Vert^2=R_{\sigma}(x^*x)$    

Since the element $ x^*x$ is normal, from the preceding paragraph it follows that $ R_{\sigma}(x^*x)=R_{\sigma}(\Phi(x^*x))$, and hence we conclude that
$\displaystyle \Vert x\Vert^2=R_{\sigma}(x^*x)=R_{\sigma}\big(\Phi(x)^*\Phi(x)\big)= \Vert\Phi(x)\Vert^2$    

i.e. $ \Vert\Phi(x)\Vert=\Vert x\Vert$.

Since $ \Phi$ is isometric, $ \Phi(\mathcal{A})$ is closed *-subalgebra of $ \mathcal{B}$, i.e. $ \Phi(\mathcal{A})$ is a $ C^*$-subalgebra of $ \mathcal{B}$, and it is isomorphic to $ \mathcal{A}$. Using the spectral invariance theorem we conclude that $ \sigma(x)=\sigma(\Phi(x))$ for every $ x \in \mathcal{A}$. $ \square$



Anyone with an account can edit this entry. Please help improve it!

"injective $C^*$-algebra homomorphism is isometric" is owned by asteroid.
(view preamble)

View style:

Log in to rate this entry.
(view current ratings)

Cross-references: spectral invariance theorem, isomorphic, closed, isometric, normal, spectral radius, contradiction, continuous functional calculus, continuous functions, restriction, function, strict, inclusion, invertible, normal elements, clear, identity elements, minimal unitizations, unital, proof, spectrum, *-homomorphism, injective
There are 2 references to this entry.

This is version 3 of injective $C^*$-algebra homomorphism is isometric, born on 2008-04-20, modified 2008-04-21.
Object id is 10524, canonical name is InjectiveCAlgebraHomomorphismIsIsometric.
Accessed 209 times total.

Classification:
AMS MSC46L05 (Functional analysis :: Selfadjoint operator algebras :: General theory of $C^*$-algebras)

Pending Errata and Addenda
None.
Discussion
Style: Expand: Order:
forum policy

No messages.

Interact
post | correct | update request | prove | add result | add corollary | add example | add (any)