Semicrossed Products of the Disk Algebra: Contractive Representations and Maximal Ideals

Semicrossed Products of the Disk Algebra: Contractive Representations and Maximal Ideals

pacific journal of mathematics Vol. 185, No. 1, 1998 SEMICROSSED PRODUCTS OF THE DISK ALGEBRA: CONTRACTIVE REPRESENTATIONS AND MAXIMAL IDEALS Dale R. Buske and Justin R. Peters Given the disk algebra A(D) and an automorphism α, there + is associated a non-self-adjoint operator algebra Z ×α A(D) called the semicrossed product of A(D) with α. We consider those algebras where the automorphism arises via composi- tion with parabolic, hyperbolic, and elliptic conformal maps ϕ of D onto itself. To characterize the contractive representa- + tions of Z ×α A(D), a noncommutative dilation result is ob- tained. The result states that given a pair of contractions S, T on some Hilbert space H which satisfy TS = Sϕ(T), there exist unitaries U, V on some Hilbert space K⊃Hwhich dilate S and T respectively and satisfy VU =Uϕ(V). It is then shown that there is a one-to-one correspondence between the contractive + (and completely contractive) representations of Z ×α A(D) on a Hilbert space H and pairs of contractions S and T on H satis- fying TS = Sϕ(T). The characters, maximal ideals, and strong + radical of Z ×α A(D) are then computed. In the last section, we compare the strong radical to the Jacobson radical. I. Introduction. A semicrossed product of the disk algebra is an operator algebra associated to the pair (A(D),α), where A(D) is the disk algebra and α an automor- phism of A(D). Any such α has the form α(f)=f◦ϕ(f∈A(D)) for a linear fractional transformation ϕ. It is well-known there is a one-to-one correspondence between contractions (i.e., bounded linear operators T on some Hilbert space with kT k≤1) and contractive representations of A(D). Here, analogously, there is a one-to-one correspondence between pairs S, T of contractions satisfying the relation TS = Sϕ(T) and contractive repre- + sentations of the semicrossed product, denoted Z ×α A(D), associated to (A(D),α). This is meaningful since linear fractional transformations map contractions to contractions (cf. [Sz-NF]). The question of whether con- + tractive representations of Z ×α A(D) can be dilated to representations of ∗ the C -crossed product Z ×α C(T) is equivalent to the following: given a pair of contractions S, T on some Hilbert space H satisfying TS = Sϕ(T)do 97 98 DALE R. BUSKE AND JUSTIN R. PETERS there exist unitaries U, V on some Hilbert space K⊃Hwhich are dilations of S, T respectively, and satisfy VU = Uϕ(V)? This question, which is of interest in its own right, has an affirmative solution [Theorem II.4]. Fur- thermore, it marks the starting point in our study of semicrossed products + of the disk algebra, by giving faithful representations of Z ×α A(D). Section V deals with the characterization of the maximal ideal space. The character space, or equivalently the space of maximal ideals of codimension one, was easily obtained from the representation theory [Corollary III.11]. It turns out there are no maximal ideals of codimension greater than one – unless the automorphism (i.e., the linear fractional transformation) is elliptic of finite period, say K, in which case the maximal ideals have codimension either 1 or K2. Finally, these results, together with [HPW] are used to com- pare the strong radical with the Jacobson radical: the two radicals coincide except when α is elliptic and nonperiodic (that is, an irrational rotation). (Theorem VI.1.) II. Dilating Noncommuting Contractions. It is well-known that each contraction T on a Hilbert space H can be dilated to a unitary U on a Hilbert space K⊆H[Sz-NF]. That is, n n for n ≥ 1,T h = PHU h ∀ h ∈H. This result then yields the fact that the contractive representations of the disk algebra A(D)onHare in a one-to-one correspondence with contractions on H [DP], [Sz-NF]. Moreover, it shows that each contractive representation is completely con- tractive [DP]. Andˆo then generalized this result by showing that every commuting pair of contractions S and T on H have a unitary dilation on some K⊇H[A]. That is, there exist unitaries U and V on K such that m n m n ∀ m ≥ 1,n≥1,S Th=PHU Vh. Hence the contractive represen- tations of the bidisk algebra A(D2)onHare in one-to-one correspondence with commuting pairs of contractions on H. Furthermore, each representa- tion of A(D2) is completely contractive [DP]. Recently, Sebesty´en showed that every anti-commuting pair of contractions have such a dilation [S]. In this section we show that when ϕ is a conformal automorphism of D and S and T are contractions on H satisfying TS = Sϕ(T) then a unitary dilation exists. This result is then used to characterize the contractive representa- tions of a semicrossed product. Proofs in this section closely resemble those in [S]. Lemma II.1 is directly lifted from [S]. Lemma II.1. Let K and K0 be Hilbert spaces, H⊂Kand H0 ⊂K0 be subspaces and X : H→K0 and X0 : H0 →Kbe given bounded linear transformations. Then, there exists an operator Y : K→K0 extending X so that Y ∗ extends X0 if and only if hXh,h0i = hh, X0h0i∀h∈H,h0 ∈H0. SEMICROSSED PRODUCTS OF THE DISK ALGEBRA 99 Moreover, kY k≤max{kXk, kX0k}. Suppose now that S and T are contractions on H which satisfy TS = Sϕ(T) for some linear fractional transformation ϕ of D. Note that ϕ(T )isa well-defined contraction by the functional calculus found in [Sz-NF]. Let U be the minimal isometric dilation of S acting on a Hilbert space K containing W W ∗ ∗ ∞ n k H. Then, U extends S , where K = n=0 Kn and Kn = k=0 U (H). Lemma II.2. Let S and T be contractions on H such that TS = Sϕ(T). If U is the minimal isometric dilation of S acting on K, then there exists Tϕ ∗ ∗ an operator on K such that Tϕ extends T , kTϕk≤1, and TϕU = Uϕ(Tϕ). Proof. The proof is similar to [S]. At the n-th step of induction, Lemma II.1 is applied to the maps Uϕ(T )U∗| : H ≡ U(K ) →K and n−1 Hn−1 n−1 n−1 n ∗ Tn−1 : Kn−1 →Kn and kTnk≤1 since kϕ(Tn−1)k≤1. Supposing that ϕ = µϕ where |µ| = 1 and ϕ (z)= z−a, the conditions of Lemma II.1 a a 1−az ∗ ∗ ∗ are satisfied since hx, [µϕa(Tn)] yi = hx, µϕa(Tn )yi = hx, µϕa(Tn−1)yi = ∗ hx, [µϕa(Tn−1)] yi for x ∈Hn−1 and y ∈Kn−1 [Sz-NF, I.4]. Lemma II.3. Let S and T be contractions on H such that TS = Sϕ(T).If Uis the minimal unitary dilation of S acting on a Hilbert space K, then there exists Tϕ on K which is a dilation of T with kTϕk≤1and TϕU = Uϕ(Tϕ). Proof. Let U+ be the minimal isometric dilation of S acting on K+; H⊆K+ ⊆ K. By Lemma II.2, there exists a dilation T of T to K with kT k≤1 and ϕ+ + ϕ+ −1 Tϕ+U+ = U+ϕ(Tϕ+). By considering a sequence of polynomials pn → ϕ , −1 it follows from Tϕ+U+ = U+ϕ(Tϕ+) that ϕ (Tϕ+)U+ = U+Tϕ+. Taking ∗ ∗ ∗ −1 ∗ ∗ adjoints yields Tϕ+U+ = U+ϕ (Tϕ+) . Since U is the unique minimal ∗ isometric dilation of U+ it follows by Lemma II.2 and [Sz-NF, I.4] that ∗ ∗ there exists an operator Tϕ on K such that Tϕ extends Tϕ+, kTϕk≤1, and ∗ ∗ ∗ −1 ∗ TϕU = U ϕ (Tϕ) . Reasoning as above, TϕU = Uϕ(Tϕ). Theorem II.4. Let S and T be contractions on H such that TS = Sϕ(T). Then there exists a pair of unitaries U and V such that VU =Uϕ(V) and m n m n S T = PHU V |H for every m, n ∈ N. Proof. As in [S], let U0 be the minimal unitary dilation of S and T0 a con- tractive dilation of T with T0U0 = U0ϕ(T0). Then let V be the minimal unitary dilation of T0 and proceed to extend U0 to a unitary such that VU = Uϕ(V). The proof follows [S] after it is shown that U defined on W∞ n n −1 n K = V (K0)byU(V k0)=ϕ (V)U0k0 is isometric. However, since −∞ −1 V is the minimal unitary dilation of T0, it follows that ϕ (V ) is the mini- −1 m n mal unitary dilation of ϕ (T0)[Sz-NF, I.4.3] and so hU(V h0),U(V k0)i= 100 DALE R. BUSKE AND JUSTIN R. PETERS −1 m −1 n −1 m −1 n hϕ (V ) U0h0,ϕ (V) U0k0i = hϕ (T0) U0h0,ϕ (T0) U0k0i = m n m n hU0T0 h0,U0T0 k0i=hV h0,V k0i. III. The Semicrossed Product. Example III.1. Let D be an operator algebra and α an automorphism of D. Let (ρ, K) be a contractive representation of D and let H = H 2(K), the space L∞ of square summable elements in 0 K. Define a contractive representation 2 π of D on H by π(f )(ξ0,ξ1,ξ2,...)=(ρ(f)ξ0,ρ(α(f))ξ1,ρ(α (f))ξ2,...). If U+ is the unilateral shift on H, then U+π(α(f)) = π(f)U+∀ f ∈D.

View Full Text

Details

  • File Type
    pdf
  • Upload Time
    -
  • Content Languages
    English
  • Upload User
    Anonymous/Not logged-in
  • File Pages
    17 Page
  • File Size
    -

Download

Channel Download Status
Express Download Enable

Copyright

We respect the copyrights and intellectual property rights of all users. All uploaded documents are either original works of the uploader or authorized works of the rightful owners.

  • Not to be reproduced or distributed without explicit permission.
  • Not used for commercial purposes outside of approved use cases.
  • Not used to infringe on the rights of the original creators.
  • If you believe any content infringes your copyright, please contact us immediately.

Support

For help with questions, suggestions, or problems, please contact us