Spectral theory of normal C*-algebras
In functional analysis, every C*-algebra is isomorphic to a subalgebra of the C*-algebra of bounded linear operators on some Hilbert space H. This article describes the spectral theory of closed normal subalgebras of .
Resolution of identity
Throughout we fix a Hilbert space H.
Definition: A projection-valued measure on a measurable space , where is a σ-algebra of subsets of , is a mapping such that for all , is a self-adjoint projection on H (i.e. is a bounded linear operator that satisfies and ) such that
(where is the identity operator of H) and for every x and y in H, the function defined by is a complex measure on (that is, a complex-valued countably additive function).
Definition:[1] A resolution of identity on a measurable space is a function such that for every :
- ;
- ;
- for every , is a self-adjoint projection on H;
- for every x and y in H, the map defined by is a complex measure on ;
- ;
- if then ;
If is the -algebra of all Borels sets on a Hausdorff locally compact (or compact) space, then we also add the following requirement:
- for every x and y in H, the map is a regular Borel measure (this is automatically satisfied on compact metric spaces).
Conditions 2, 3, and 4 imply that is a projection-valued measure.
Properties
Throughout let be a resolution of identity.
- For all x in H, is a positive measure on with total variation and that satisfies for all .[1]
For every :
- (since both are equal to ).[1]
- If then the ranges of the maps and are orthogonal to each other and .[1]
- is finitely additive.[1]
- If are pairwise disjoint elements of whose union is and if for all i then .[1]
- However, is countably additive only in trivial situations as we now describe: suppose that are pairwise disjoint elements of whose union is and that the partial sums converge to in (with its norm topology) as ; then since the norm of any projection is either 0 or , the partial sums cannot form a Cauchy sequence unless all but finitely many of the are 0.[1]
- For any fixed x in H, the map defined by is a countably additive H-valued measure on .
- Here countably additive means that whenever are pairwise disjoint elements of whose union is , then the partial sums converge to in H (or succinctly, ).[1]
L∞(π) - space of essentially bounded function
The be a resolution of identity on .
Essentially bounded functions
Suppose we are given a complex-valued -measurable functions . There exists a unique largest open subsets of (ordered under subset inclusion) such that .[2] To see why, let be a basis for 's topology consisting of open disks and suppose that is the subsequence (possibly finite) consisting of those sets such that ; then . Note that, in particular, if D is an open subset of such that then so that (although there are other ways in which may equal 0). Indeed, .
The essential range of f is defined to be the complement of . It is the smallest closed subset of that contains for almost all (i.e. for all except for those in some set such that ).[2] Note that the essential range is a closed subset of so that if it is also a bounded subset of then it is compact.
We say that f is essentially bounded if its essential range is bounded, in which case we define its essential supremum, denoted by , to be the supremum of all as ranges over the essential range of f.[2]
Space of essentially bounded functions
Let be the vector space of all bounded complex-valued -measurable functions , which becomes a Banach algebra when normed by . Note that is a seminorm on , but not necessarily a norm. The kernel of this seminorm, , is a vector subspace of that is a closed two-sided ideal of the Banach algebra .[2] Hence the quotient of by is also a Banach algebra, denoted by where the norm of any element is equal to (since if then ) and this norm makes into a Banach algebra. The spectrum of in is the essential range of f.[2] We will follow the usual practice of writing f rather than to represent elements of .
Theorem:[2] Let be a resolution of identity on . There exists a closed normal subalgebra A of and an isometric *-isomorphism satisfying the following properties:
- for all x and y in H and , which justifies the notation ;
- for all and ;
- an operator commutes with every element of if and only if it commutes with every element of .
- if f is a simple function equal to , where is a partition of X and the are complex numbers, then (here is the characteristic function);
- if f is the limit (in the norm of ) of a sequence of simple functions in then converges to in and ;
- for every .
Spectral theorem
Recall that the maximal ideal space of a Banach algebra A is the set of all complex homomorphisms , which we'll denote by . For every T in A, the Gelfand transform of T is the map defined by . is given the weakest topology making every continuous. With this topology, is a compact Hausdorff space and every T in A, G(T) belongs to , which is the space of continuous complex-valued functions on . Recall that the range of is the spectrum and that the spectral radius is equal to , which is .[3]
Theorem:[4] Suppose A is a closed normal subalgebra of that contains the identity operator and let be the maximal ideal space of A. Let be the Borel subsets of . For every T in A, let denote the Gelfand transform of T so that G is an injective map . There exists a unique resolution of identity that satisfies:
- for all and all
(we use the notation to summarize this situation). Let be the inverse of the Gelfand transform where note that can be canonically identified as a subspace of . Let B be the closure (in the norm topology of ) of the linear span of . The following hold:
- B is a closed subalgebra of containing A;
- There exists a (linear multiplicative) isometric *-isomorphism extending such that for all ;
- Recall that the notation means that for all ;
- Note in particular that for all ;
- Explicitly, satisfies and for every (so if f is real valued then is self-adjoint);
- If is open and nonempty (which implies that ) then ;
- A bounded linear operator commutes with every element of A if and only if it commutes with every element of ;
The above result can be specialized to a single normal bounded operator.
References
- Rudin 1991, pp. 316-318.
- Rudin 1991, pp. 318-321.
- Rudin 1991, p. 280.
- Rudin 1991, pp. 321-325.
- Robertson, A. P. (1973). Topological vector spaces. Cambridge England: University Press. ISBN 0-521-29882-2. OCLC 589250.CS1 maint: ref=harv (link)
- Rudin, Walter (1991). Functional analysis. McGraw-Hill Science/Engineering/Math. ISBN 978-0-07-054236-5.
- Schaefer, Helmut H. (1999). Topological Vector Spaces. GTM. 3. New York, NY: Springer New York Imprint Springer. ISBN 978-1-4612-7155-0. OCLC 840278135.CS1 maint: ref=harv (link)