Let $X$ be a locally compact Hausdorff topological space, denote by $M_n$ the $C^*$-algebra of complex $n\times n$ matrices, by $C_0(X,M_n)$ the $C^*$-algebra of continuous functions on $X$ with values in $M_n$ vanishing at infinity, and by $C_b(X,M_n)$ the $C^*$-algebra of bounded continuous functions on $X$ with values in $M_n$. Does there exist a description of all sub $C^*$-algebras of $C_0(X,M_n)$ or $C_b(X,M_n)$? I am most interested in the (easier?) case where $X$ is compact and sub $C^*$-algebras of $C(X,M_n)$ containing the unit of $C(X,M_n)$.

For the case $n=1$, i.e., for $C_0(X)$, you can find the answer here: What is the commutative analogue of a C*-subalgebra?.

  • 7
    $\begingroup$ I don't think there is a simple description: you can have $C^*$-subalgebras with non-Hausdorff spectra (simplest example: continuous functions $[0,1]\rightarrow M_n$ which are diagonal at $0$); and if you assume your $C^*$-subalgebra to be homogeneous (i.e. all its irreducible rep's are of the same dimension), then you have the $K$-theory of $C(X)$ showing up (if $p$ is a self-adjoint idempotent in $M_n(C(X))$, look at the subalgebra $pM_n(C(X))p$). $\endgroup$ Mar 6, 2013 at 20:10

1 Answer 1


Every $C^*$-subalgebra $A$ of $C_0(X,M_n)$ has irreps of dimension $\leq n.$ (Just because every irrep of $A$ can be continued to an irrep of $C_0(X,M_n)$). Such $C^*$-algebras are called $n$-subhomogeneous.

There is a complete (rather complicated) description of such $C^*$-algebras in algebraic topological terms:

Vasilʹev, N. B. "$C^∗$-algebras with finite-dimensional irreducible representations." Uspehi Mat. Nauk 21 1966 no. 1 (127), 135–154.

Much easier is to describe the $C^*$-algebras with all irreps of the same dimension $n$. They are exactly the algebras of $C_0$-sections of vector bundles over l.-c. hausdorff spaces with fiber $M_n$ and group $PU_n,$ see

Tomiyama, Jun; Takesaki, Masamichi Applications of fibre bundles to the certain class of $C^∗$-algebras. Tôhoku Math. J. (2) 13 1961 498–522.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

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