Abstract
In this paper, using some properties about Toeplitz kernels, we present some results about finite-rank properties of the commutator \([A_f,~A_g]\). Firstly, we show that \([A_{B_n},~A_v^*]\) must have a finite rank on the model space \(K_u^2\), where \(B_n\) is a finite Blaschke product and v is an inner function. Next, we present that when \(\text {ker}~T_{\overline{u}B_n}\) is an invariant subspace of \(T_\phi ^*\), then \([A_{B_n},~A_\phi ^*]\) has a finite rank on \(K_u^2\) for \(\phi \in H^\infty \). Finally, we prove that \([A_{B_n},~A_\phi ^*]\) must have a finite rank on \(K_u^2\) when \(u=B_nu_1\) for an inner function \(u_1\).
Similar content being viewed by others
Avoid common mistakes on your manuscript.
1 Introduction
Let \({\mathbb {D}}\) denote the open unit disk in the complex plane \({\mathbb {C}}\) and \({\mathbb {T}}\) denote the unit circle. Denote by \(L^2=L^2({\mathbb {T}},~\text {d}m)\) the Hilbert space of square integrable functions with respect to the Lebesgue measure \(\text {d}m\) on \({\mathbb {T}}\), normalized so that the measure of the entire circle is 1. Let \(L^\infty \) be the space of essentially bounded functions on the unit circle. The Hardy space \(H^2\) denotes the Hilbert space of all holomorphic functions in \({\mathbb {D}}\) having square-summable Taylor coefficients at the origin, and it will be identified with the space of boundary functions, the subspace of \(L^2\) consisting of functions whose Fourier coefficients with negative indices vanish. Let \(H^\infty \) denote the space of all bounded holomorphic functions in \({\mathbb {D}}\) and \(C({\mathbb {T}})\) denote the space of all continuous functions on \({\mathbb {T}}\).
Every function in \(H^2\), other than the constant function 0, can be factorized into the product of an inner function and an outer function. An inner function is a function \(u\in H^\infty \) such that \(|u(e^{i\theta })|=1\) almost everywhere with respect to the Lebesgue measure. Every inner function can be factorized into the product of a Blaschke product and a singular inner function. A Blaschke product is an analytic function \(B\in H^\infty \) of the form
where \(\{z_k\}\) are zeros of B counting multiplicity which satisfy that \(\sum \limits _k(1-|z_k|)< \infty .\) A nonconstant inner function that has no zeros in \({\mathbb {D}}\) is called a singular inner function \(S_\mu \), which has the following form
where \(\mu \) is a finite positive regular Borel measure on \([0,2\pi ]\), singular with respect to the Lebesgue measure and c is a constant of modulus 1. The function \(F\in H^2\) is an outer function if F is a cyclic vector of the unilateral shift S. That is,
For more properties about Hardy spaces, we can refer to [14].
By Beurling’s theorem [4], the invariant subspace of the unilateral shift operator \(Sf=zf\) on \(H^2\) has the form \(uH^2\), where u is an inner function. It is easy to check that \(K_u^2=H^2\ominus uH^2\) is the invariant subspace of the backward shift operator \(S^*\) on \(H^2\), which is called the model space. Let P denote the orthogonal projection from \(L^2\) onto \(H^2\) and \(P_u\) denote the orthogonal projection from \(L^2\) onto \(K_u^2\). For \(f\in L^\infty \), the Toeplitz operator \(T_f\) induced by the symbol f is defined on \(H^2\) by
Obviously, \(T_f^*=T_{\overline{f}}\). Toeplitz operators acting on \(H^2\) have very simple and natural matrix representations via infinite Toeplitz matrices that have constant entries on diagonals parallel to the main one. The Hankel operator \(H_f\) induced by the symbol f is defined on \(H^2\) by
Then \(H^*_f h=P(\overline{f}h)\) for \(h\in (H^2)^\perp \). Compressions of Toeplitz operators on \(K_u^2\) are called truncated Toeplitz operators, for \(\psi \in L^\infty \), which are defined by
The function \(\psi \) is called the symbol of \(A_\psi \). Clearly, \(A_\psi ^*=A_{\overline{\psi }}\).
Truncated Toeplitz operators represent a far reaching generalization of classical Toeplitz matrices. Although a particular case had appeared before in the literature, the general theory has been initiated in the seminal paper [21]. Since then, truncated Toeplitz operators have constituted an active area of research. We mention only a few relevant papers [5, 11, 17, 20] and so on. On the operator theory level, Nagy and others showed that \(A_z\) is a model for a certain class of contraction operators [24]. Every contraction operator T on the Hilbert space H having defect indices (1, 1) and such that \(\lim \limits _{n\rightarrow \infty }T^{*n}=0~ (\text {SOT})\) is unitarily equivalent to \(A_z\) for an inner function u, where \(\text {SOT}\) denotes the strong operator topology. In terms of function theory, Douglas et al. [9] in 1970 showed that functions in the model space and inner functions have the same analytic continuation neighborhood, that is, assuming that u could be analytically extended on \(E\subseteq {\mathbb {T}}\), then all functions in the model space \(K_u^2\) could be analytically extended on E. Therefore, the rational functions approximate the functions in the model space, the boundary value of functions in the model space, the angular derivative of inner functions and the relationship between them are deeply studied, refer to [2] and [3]. Thus, the research on truncated Toeplitz operators is of representative significance.
D. Sarason once proposed in [22]: what \(f,~g\in H^2\) can make \(T_fT_{\overline{g}}\) to be a bounded operator. From this, many scholars begin to study the boundedness of the product of Toeplitz operators and Hankel operators. By \(T_{fg}-T_fT_g=H_{\overline{f}}^*H_g\), properties of commutators of Toeplitz operators also gradually begin to enter the research line of sight. The map \(\tau :~f\rightarrow T_f\) is a contractive \(*\)-linear mapping from \(L^\infty \) to \(L(H^2)\), where \(L(H^2)\) is the algebra of all bounded linear operators on \(H^2\). But this mapping is not multiplicative. When f is a continuous function, by Hartman’s theorem (see Corollary 4.3.3 in [18]) we get that \(T_{fg}-T_fT_g=H_{\overline{f}}^*H_g\) is compact for any \(g\in L^\infty \). Thus, \(\tau (fg)=\tau (f)\tau (g)+K\) for \(f\in C({\mathbb {T}})\), where K is a compact operator. This makes the multiplicability problem very interesting, and many scholars begin to study the compactness of commutators of Toeplitz operators. In addition, the compactness of commutators of Toeplitz operators also originated from the research on Fredholm theory of Toeplitz operators in 1970s by Douglas, Sarason and others.
The results about compact or finite-rank (semi-)commutators of \(T_f\) or \(H_f\) are quite complete, corresponding results are summarized in [26]. But there are very few results for compact or finite-rank (semi-)commutators of truncated Toeplitz operators. In [11], Garcia concluded that \(A_{fg}-A_fA_g\) is compact for \(f,~g\in C({\mathbb {T}})\). In [7], authors described the kernels and ranks of commutators of truncated Toeplitz operators with symbols induced by finite Blaschke products. By \(L^2=K_u^2\bigoplus (K_u^2)^\bot ,\) for \(f\in L^\infty \), the multiplication operator \(M_f\) is expressed as an operator matrix
where \(D_f\) denotes the dual truncated Toeplitz operator on \(L^2\ominus K_u^2\) defined by
It is easy to see that \(D_f^*=D_{\overline{f}}.\) The truncated Hankel operator \(B_f\) is defined by
Moreover, \(B_f^*h=P_u(\overline{f}h)~\text {for}~h\in L^2\ominus K_u^2.\) By \(M_f M_g=M_{fg},\) we get that
In [19], authors introduced sufficient and necessary conditions for \(B_f\) to be of finite-rank or compact. By (1), we know some sufficient conditions for \(A_f A_g-A_{fg}\) to be of finite-rank or compact. From this, we can study the compact commutator \([A_f,~A_g]\) by truncated Hankel operators \(B_f\). In [26], authors of this article and others gave some results that commutators of truncated Toeplitz operators are compact or of finite-rank operators on model spaces. In our paper, using the structure of Toeplitz kernels, we present some results about finite-rank properties of commutators \([A_f,~A_g]\).
The paper is organized as follows. In Sect. 2, we recall some necessary definitions and properties about model spaces and truncated Toeplitz operators. In Sect. 3, we obtain that \([A_{B_n},~A_v^*]\) must have a finite rank on \(K_u^2\), where \(B_n\) is a finite Blaschke product and v is an inner function. In Sect. 4, using some properties of Toeplitz kernels, when \(\text {ker}~T_{\overline{u}B_n}\) is an invariant subspace of \(T_\phi ^*\), we show that \([A_{B_n},~A_\phi ^*]\) has a finite rank on \(K_u^2\) for \(\phi \in H^\infty \). In particular, we present that \([A_{B_n},~A_\phi ^*]\) has a finite rank on \(K_u^2\) when \(u=B_nu_1\) for an inner function \(u_1\).
2 Preliminaries
In this section we introduce some basic properties of truncated Toeplitz operators. The reproducing kernel of \(H^2\) at \(\lambda \in {\mathbb {D}}\) is the function \(k_\lambda (z)=\frac{1}{1-\overline{\lambda } z}\), and it is easy to check that the reproducing kernel of \(K_u^2\) at \(\lambda \in {\mathbb {D}}\) is the function
It is well known that \(K_u^2\) carries a natural conjugation C, antiunitary, involution operator, defined by \(Cf=\overline{zf}u\) for \(f\in K_u^2\). We have that
which is the conjugation reproducing kernel of \(K_u^2\) at \(\lambda \in {\mathbb {D}}\). That is,
for \(f\in K_u^2\). A bounded linear operator A on \(K_u^2\) is called C-symmetric if
S. R. Garcia and M. Putinar introduced some properties of C-symmetric operators in [12], and they showed that all truncated Toeplitz operators are C-symmetric. More complex symmetric operators can be found in [13].
The systematic study of truncated Toeplitz operators was initiated by Sarason [21], and an intensive study revealed many interesting and different properties about these operators. For example, unbounded symbols may give bounded truncated Toeplitz operators, see Sarason’s example in [21]. For \(f\in L^2\), Sarason in [21] proved that \(A_f=0\) if and only if \(f\in uH^2+\overline{uH^2}\). Therefore, the symbols of truncated Toeplitz operators are not unique.
It is well known that the model space \(K_u^2\) is the kernel space of \(T_{\overline{u}}\). There are many important connections between model spaces and Toeplitz kernels. In [6], authors presented some classical results about the relationship between Toeplitz kernels and model spaces. Using the relationship between Toeplitz kernels and model spaces, authors showed maximal vectors for model spaces by maximal vectors for Toeplitz kernels. Moreover, they also discussed the multiplier between Toeplitz kernels by the multiplier between model spaces.
For Toeplitz kernels, one classical result is the Coburn theorem (see Proposition 7.24 in [8]). It is said that either \(\text {ker}~T_g=\{0\}\) or \(\text {ker}~T_g^*=\{0\}\) for \(g\in L^\infty .\) In 1986, Hayashi [15] showed that the kernel of \(T_g\) can be written as \(\phi K_\eta ^2\), where \(\phi \) is an outer function and \(\eta \) is an inner function with \(\eta (0)=0\), and the function \(\phi \) multiplies \(K_\eta ^2\) isometrically onto \(\text {ker}~T_g\). It is easy to check that the Toeplitz kernel is nearly \(S^*\)-invariant. In 1988, Hitt [16] showed that any nearly \(S^*\)-invariant of subspace M is of form \(hK_\theta ^2\), where \(h\in M\) meeting some conditions and \(\theta \) is an inner function. Sarason [23] gave a new proof of Hitt’s theorem and presented better description of h and \(\theta \). More about the research process of Toeplitz kernels can be found in [6].
The following lemma is well known and we provide a proof for the sake of completeness.
Lemma 1
If u and v are inner functions, then
Moreover, \(\text {ker}~T_{\overline{u}v}\ne \{0\}\) if and only if \(v H^2\cap K_u^2\ne \{0\}.\)
Proof
Denote \(E=\{\varphi \in K_u^2:~v\varphi \in K_u^2\}.\) For \(\phi \in E\subset K_u^2\), since the model space \(K_u^2\) has the conjugation, there exists \(\eta \in K_u^2\) such that \(v\phi =u\overline{z\eta }.\) Then
It implies that
For \(\varphi \in \text {ker}~T_{\overline{u}v},\) we have that \(T_{\overline{u}v}\varphi =P(\overline{u}v\varphi )=0,\) and there exists \(x\in H^2\) such that
Then \(v\varphi =u\overline{zx}\in H^2.\) Since
we get that \(v\varphi \in K_u^2.\) By (3), we have that \(\varphi =u\overline{zxv}\in H^2.\) By (4), we conclude that \(\varphi \in K_u^2~\text {and}~\varphi \in E.\) Thus \(\text {ker}~T_{\overline{u}v} \subseteq E\) and we have proved that \(\text {ker}~T_{\overline{u}v} =E.\)
Suppose that \(v H^2\cap K_u^2\ne \{0\}.\) There exists \(0\ne h\in H^2\) such that \(vh\in K_u^2\). Since \(K_u^2\) has a conjugation, there is \(0\ne g\in K_u^2\) such that \(vh=u\overline{zg}.\) Then
It implies that \(h\in \text {ker}~T_{\overline{u}v},\) and \(\text {ker}~T_{\overline{u}v}\ne \{0\}\). By (2), it is easy to get that \(v H^2\cap K_u^2\ne \{0\}\) when \(\text {ker}~T_{\overline{u}v}\ne \{0\}\). The proof is completed. \(\square \)
3 The Finite-Rank Property of \([A_{B_n},~A_v^*]\) for an Inner Function v
By [11] we know that \(A_fA_g-A_{fg}\) is compact for \(f,~g\in C({\mathbb {T}})\). Thence \([A_{B_n},~A_{v}^*]\) must be compact for any finite Blaschke product \(B_n\) and any inner function v. This makes us want to discuss when \([A_{B_n},~A_{v}^*]\) has a finite rank on \(K_u^2\). In the following we show that \([A_{B_n},~A_{v}^*]\) must have a finite rank for any inner function v.
In the following we will frequently use the following relationship:
for any \(\psi \in H^\infty \) and an inner function u.
We use \(\text {Hol}({\mathbb {D}})\) to denote the set of all holomorphic functions in \({\mathbb {D}}\). For a pair of inner functions v and \(\eta \), we explore multipliers
between model spaces \(K_v^2\) and \(K_\eta ^2\).
Lemma 2
(Corollary 3.3 in [10]) If \(v,~\eta \) are inner functions, then
A finite Blaschke product is a function of the form \(B_n(z)=c\prod \limits _{i=1}^n\frac{z-z_i}{1-\overline{z_i}z}\) for \(z_i\in {\mathbb {D}}\). The degree of a finite Blaschke product \(B_n\) is its number of zeros.
Lemma 3
(Theorem 4.3 in [10]) If B is a finite Blaschke product and v is any inner function with the infinite degree, then \({\mathcal {M}}(B,~v)\cap H^\infty \ne \{0\}.\)
Lemma 4
(Lemma 2.1 in [7]) Let \(a_1,~a_2,~\cdot \cdot \cdot ,~a_n\) be points in \({\mathbb {D}}\) and put
then
where \({\mathcal {P}}_k\) denotes the set of all analytic polynomials with the degree less than or equal to k. In particular, \(\text {dim}~K_{B_n}^2=n.\)
Theorem 1
Let u be a nonconstant inner function and \(K_u^2\) be the infinite dimensional model space. If \(B_n\) is a finite Blaschke product with degree n, then \([A_{B_n},~A_{v}^*]\) has a finite rank on \(K_u^2\) for any inner function v and \(\text {rank}[A_{B_n},~A_{v}^*]\le 2n\). Moreover, if \(\text {ker}~T_{\overline{u}B_n v}\ne \{0\}\), then
Proof
Since \(zB_n\) is a finite Blaschke product, by Lemma 3, we get that
Then by Lemma 2,
By Lemma 1, we know that
Then \(B_n g\in K_u^2\) for \(g\in \text {ker}~T_{\overline{u}B_n}.\) By \(H^2=v H^2\oplus K_{v}^2,\) there exist \(g_1\in H^2\) and \(g_2\in K_{v}^2\) such that \(g=v g_1+g_2.\) Then
Since \(g_2\in K_{v}^2\), we get that \(\overline{v}g_2\in \overline{zH^2}\) and \(P(\overline{v}g_2)=0.\) Then
We claim that \(P(\overline{v}B_n K_{v}^2)\subseteq K_{B_n}^2\). In fact, for any \(f\in K_{v}^2\) and \(h\in H^2\), we have that
Thence
By Lemma 4, we obtain that \(\text {dim}~K_{B_n}^2=n\). Then \(\text {dim}~(\text {ran}~P_uP_{B_n})\le n,\) and
Since
in the following we consider the dimension of \([A_{B_n},~A_{v}^*](K_u^2\ominus \text {ker}~T_{\overline{u}B_n}).\) By
there exist \(h_1\in H^2\) and \(h_2\in K_{B_n}^2\) such that \(h=B_n h_1+h_2\) for any \(h\in H^2.\) Then
By Lemma 4, we know that \(\text {dim}~K_{B_n}^2=n\). It follows that
It is easy to check that \(P(\overline{B_n}u\phi )\in K_u^2\) for any \(\phi \in K_{B_n}^2\). Thus
Consequently,
where the abbreviation “ cl " denotes the closure of a set. Since
we get that
By \(P(\overline{B_n}u\phi )\in K_u^2\) for any \(\phi \in K_{B_n}^2\), we get that \(P(\overline{B_n}uK_{B_n}^2)\subset K_u^2.\) It follows that
Then
Then
By (5), (6) and (9), we get that
Thus \([A_{B_n},~A_{v}^*]\) has a finite rank on \(K_u^2\) and \(\text {rank}[A_{B_n},~A_{v}^*]\le 2n\).
Suppose that \(\text {ker}~T_{\overline{u}B_n v}\ne \{0\}\). By Lemma 1, we get that \(vB_n \phi \in K_u^2\) for any \(\phi \in \text {ker}~T_{\overline{u}B_n v}\). Then there exists \(\psi \in K_u^2\) such that
That is, \(v \phi =u\overline{z\psi B_n}.\) By \(K_u^2=H^2\cap u\overline{zH^2}\), we have that \(v\phi \in K_u^2\). Then
Thus \(v \text {ker}~T_{\overline{u}B_n v}\subseteq \text {ker}~[A_{B_n},~A_{v}^*]\). By the same way, we get that
Since
and \([A_{B_n},~A_{v}^*]\) has a finite rank, we obtain that
Then by (10),
\(\square \)
In the following we give an example illustrating Theorem 1 and use \(\text {span}\{h\}\) to denote the space generated by the function h.
Example 1
Let \(B_1(z)=\frac{z-a}{1-\overline{a}z}\) for \(a\in {\mathbb {D}}\) and \(v(z)=\exp \frac{z+1}{z-1}\) be a singular inner function. If \(u=B_1v\) and \(K_u^2\) is the corresponding model space, then
and \(\text {rank}[A_{B_1},~A_{v}^*]\le 2\), where \(k_a(z)=\frac{1}{1-\overline{a}z}.\)
Proof
By Lemma 4, we know that
Since \(u=B_1v\), we have that \(K_{B_1}^2\subsetneq K_u^2\). Using the proof of Theorem 1 and (11), we get that
It implies that
In fact,
Hence
By (12) and (13), we conclude that
Then by (11),
\(\square \)
Remark 1
From the proof of Theorem 1, we know that \(P_uP_{B_n}\) has a finite rank because \(B_n\) is the finite Blaschke product. In fact, we have the following claim.
Claim: The projection \(P_uP_\theta \) has a finite rank if and only if u or \(\theta \) is a finite Blaschke product.
Using the following lemma, we give a further proof of the claim.
Lemma 5
(Section 6 in [1]) Hankel operators \(H_f\) have a finite rank if and only if \(f\in \overline{b}H^\infty \), where b is the finite Blaschke product.
Proof
If u or \(\theta \) is a finite Blaschke product, it is obvious that \(P_uP_\theta \) has a finite rank. Suppose that \(P_uP_\theta \) has a finite rank. It is easy to get that
By (14), we have that
has a finite rank. Then \(H_{\overline{\theta }}H_{\overline{u}}^*\) has a finite rank. Lemma 1 in [1] says that the operator \(H_{\overline{\theta }}H_{\overline{u}}^*\) is compact if and only if \(H_{\overline{u}}^*H_{\overline{\theta }}\) is compact. By the same way as in Lemma 1 in [1], we can get that \(H_{\overline{\theta }}H_{\overline{u}}^*\) has a finite rank if and only if \(H_{\overline{u}}^*H_{\overline{\theta }}\) has a finite rank. Using the fact that \(H_{\overline{\theta }}H_{\overline{u}}^*\) has a finite rank if and only if \(H_{\overline{\theta }}\) or \(H_{\overline{u}}\) has a finite rank (see Theorem 4 in [1]), by Lemma 5, we get that \(\overline{\theta }=\overline{b}h\) or \(\overline{u}=\overline{b_1}h_1\), where b and \(b_1\) are finite Blaschke products and \(h,~h_1\in H^\infty \). It implies that \(\theta \) or u is a finite Blaschke product. \(\square \)
For u and v inner functions, the compactness of \(P_uP_v\) reflects the asymptotically orthogonal relationship of \(K_u^2\) and \(K_v^2\). We say that model spaces \(K_u^2\) and \(K_v^2\) are asymptotically orthogonal if \(P_uP_v\) is a compact operator. Moreover, the following statements are equivalent.
-
(a)
Model spaces \(K_u^2\) and \(K_v^2\) are asymptotically orthogonal;
-
(b)
\(T_uT_v^*-T_{u\overline{v}}\) is compact;
-
(c)
\(H^\infty [\overline{u}]\cap H^\infty [\overline{v}]=H^\infty +C\), where \(H^\infty [\overline{u}]\) denotes the Douglas algebra generated by \(\overline{u}\) and \(H^\infty \);
-
(d)
For each support set S, either \(u|_s\) or \(v|_s\) is a constant;
-
(e)
\(\lim \limits _{|z|\rightarrow 1}\max (|u(z)|,~|v(z)|)=1\);
-
(f)
\(H_{\overline{v}}|_{K_u^2}\) is compact;
-
(g)
\(T_{\overline{v}}|_{K_u^2}\) is a compact perturbation of an isometry.
In fact, by \(T_uT_v^*-T_{u\overline{v}}=-H_{\overline{u}}^*H_{\overline{v}}\), using the same meaning as in the proof that \(P_uP_v\) has a finite rank, we get that \((a)\Leftrightarrow (b)\). The equivalence of (b) and (c) comes from Theorem 1 in [1] and [25]. By Lemma 3 in [1], we get that \((c)\Leftrightarrow (d)\Leftrightarrow (e)\). The proof of \((b)\Leftrightarrow (f)\Leftrightarrow (g)\) can be found in Theorem 2 in [1].
4 The Finite-Rank Property of \([A_{B_n},~A_\phi ^*]\) for \(\phi \in H^\infty \)
It is well known that \([A_{B_n},~A_\phi ^*]\) is compact when \(B_n\) is a finite Blaschke product and \(\phi \in H^\infty \). In this section assuming that \(\text {ker}~T_{\overline{u}B_n}\) is an invariant subspace of \(T_\phi ^*\), we present that \([A_{B_n},~A_\phi ^*]\) has a finite rank on \(K_u^2\).
Theorem 2
Let u be a nonconstant inner function and \(K_{u}^2\) be the infinite dimensional model space. If \(T_\phi ^*(\text {ker}~T_{\overline{u}B_n})\subseteq \text {ker}~T_{\overline{u}B_n}\), then \([A_{B_n},~A_\phi ^*]\) has a finite rank on \(K_u^2\) and \(\text {rank}[A_{B_n},~A_\phi ^*]\le 2n\), where \(\phi \in H^\infty \) and \(B_n\) is a finite Blaschke product with degree n.
Proof
By Lemmas 2 and 3, we get that \(\text {ker}~T_{\overline{u}B_n}\ne \{0\}.\) For \(f\in \text {ker}~T_{\overline{u}B_n}\), by Lemma 1, we have that \(B_nf\in K_u^2\). Then
Since \(T_\phi ^*(\text {ker}~T_{\overline{u}B_n})\subseteq \text {ker}~T_{\overline{u}B_n}\), by Lemma 1, we get that \(B_n T_{\phi }^*f\in K_u^2\). Then
It is easy to check that \(\text {ker}~H_{\overline{B_n}}=B_n H^2\). Then
By Lemma 4, we obtain that \(\text {dim}~(K_{B_n}^2)=n\). It implies that
Since
in the following we consider the dimension of \([A_{B_n},~A_{\phi }^*](K_u^2\ominus \text {ker}~T_{\overline{u}B_n}).\) By
there exist \(h_1\in H^2\) and \(h_2\in K_{B_n}^2\) such that \(h=B_nh_1+h_2\) for any \(h\in H^2.\) Then
We claim that
In fact, for any \(\psi \in H^2\), we have that
Thus
This implies that
Since \((\text {ker}~T_{\overline{u}B_n})^\perp =cl~(\text {ran}~T_{\overline{B_n}u}),\) we get that
We claim that
For \(g\in (uH^2\oplus P(\overline{B_n}uK_{B_n}^2))\cap K_u^2\), there are \(g_1\in H^2\) and \(g_2\in P(\overline{B_n}uK_{B_n}^2)\) such that
we have that
By (17), we obtain that \(\langle ug_1,g_2\rangle =0\). Then \(\Vert g_1\Vert ^2=0\) and \(g_1=0\). This implies that \(g=g_2\) and (19) holds. By (17), (18) and (19), we obtain that
By Lemma 4, we get that \(\text {dim}~K_{B_n}^2=n\) and \(\text {dim}~(K_u^2\ominus \text {ker}~T_{\overline{u}B_n})\le n<\infty .\) Then
By (15), (16) and (20), we get that
Thus \([A_{B_n},~A_{\phi }^*]\) has a finite rank on \(K_u^2\) and \(\text {rank}[A_{B_n},~A_\phi ^*]\le 2n\).
\(\square \)
Remark 2
For the condition of Theorem 2, we give the following explanation. For any \(f\in \text {ker}~T_{\overline{u}B_n}\), we have that \(T_\phi ^*(\text {ker}~T_{\overline{u}B_n})\subseteq \text {ker}~T_{\overline{u}B_n}\) if and only if \( B_nT_{\phi }^*f\in K_u^2\) if and only if \(P_u(\phi {\widetilde{f}})\in B_n\text {ker}~T_{\overline{u}B_n}\), where \({\widetilde{f}}=C(f)\).
Proof
By \(\text {ker}~T_{\overline{u}B_n}=\{f\in K_u^2:~B_nf\in K_u^2\},\) it is easy to get that \(\text {ker}~T_{\overline{u}B_n}\) is invariant under \(T_\phi ^*\) if and only if \(B_nT_{\phi }^*f\in K_u^2\).
Suppose that \(B_nT_{\phi }^*f\in K_u^2\) for \(f\in \text {ker}~T_{\overline{u}B_n}\). There exists \(g\in K_u^2\) such that
By \(K_u^2=H^2\cap u\overline{zH^2}\), we get that
There is \(g_1\in H^2\) such that
Since \(f\in K_u^2\), there exists \(f_1\in K_u^2\) such that \(f=u\overline{zf_1}\). Then \(\overline{\phi }u\overline{zf_1}=u\overline{zgB_n}+\overline{zg_1}.\) This implies that
By (21), we have that
Then \(P_u(\phi f_1)=gB_n.\) By \(f=u\overline{zf_1},\) we get that \(f_1=u\overline{zf}=C(f)={\widetilde{f}}.\) Thus, the function \(P_u(\phi {\widetilde{f}})\) belongs to \(B_n\text {ker}~T_{\overline{u}B_n}.\)
Suppose that \(P_u(\phi {\widetilde{f}})\in B_n\text {ker}~T_{\overline{u}B_n}\) for \(f\in \text {ker}~T_{\overline{u}B_n}\). There exist functions \(g_1\in \text {ker}~T_{\overline{u}B_n}\) and \(g_2\in H^2\) such that
Then
By \(C(f)=u\overline{zf}\), we have that
That is,
This implies that
Since \(g_1\in \text {ker}~T_{\overline{u}B_n}\), we get that \(B_ng_1\in K_u^2\) and \(u\overline{zB_ng_1}\in K_u^2\). Then
By \(K_u^2=H^2\cap u\overline{zH^2}\), we have that \(B_nT_\phi ^*f=u\overline{zg_1}\in K_u^2.\) The proof is completed. \(\square \)
Corollary 1
Let u be a nonconstant inner function and \(K_{u}^2\) be the infinite dimensional model space. If \(u=B_nu_1\) for a finite Blaschke \(B_n\) and an inner function \(u_1\), then \([A_{B_n},~A_\phi ^*]\) has a finite rank on \(K_u^2\) for \(\phi \in H^\infty \).
Proof
Since \(u=B_nu_1\), we get that
By Theorem 2, we obtain that \([A_{B_n},~A_\phi ^*]\) has a finite rank on \(K_u^2\). \(\square \)
Example 2
Let \(B_1(z)=\frac{z-a}{1-\overline{a}z}\) for \(a\in {\mathbb {D}}\) and \(v(z)=\exp \frac{z+1}{z-1}\) be a singular inner function. If \(u=B_1v\) and \(K_u^2\) is the corresponding model space, then the following statements hold.
-
(a)
If \(\phi =1+B_1\) is an outer function, then
$$\begin{aligned}{}[A_{B_1},~A_{\phi }^*]K_u^2\subseteq \text {span}\left\{ k_a\right\} \oplus \text {span}\left\{ P_u\left( B_1P(\overline{B_1}vk_a) \right) \right\} , \end{aligned}$$and \(\text {rank}[A_{B_1},~A_{\phi }^*]\le 2\), where \(k_a(z)=\frac{1}{1-\overline{a}z}.\)
-
(b)
If \(\phi =1+v\) is an outer function, then
$$\begin{aligned} {[}A_{B_1},~A_{\phi }^*{]}K_u^2\subseteq \text {span}\left\{ k_a\right\} \oplus \text {span}\left\{ P_u\left( B_1k_a \right) \right\} , \end{aligned}$$and \(\text {rank}[A_{B_1},~A_{\phi }^*]\le 2\), where \(k_a(z)=\frac{1}{1-\overline{a}z}.\)
Proof
By Lemma 4, we know that
Since \(u=B_1v\), we have that \(K_{B_1}^2\subsetneq K_u^2\). Using the proof of Theorem 2, we get that
and
-
(a)
If \(\phi =1+B_1\), then
$$\begin{aligned} \begin{aligned} {[}A_{B_1},~A_{\phi }^*{]}(vk_a)&=(A_{B_1}A_{\overline{1+B_1}}-A_{\overline{1+B_1}}A_{B_1})(vk_a)\\&=A_{B_1}P(\overline{1+B_1}vk_a)-A_{\overline{1+B_1}}P_u(B_1vk_a)\\&=A_{B_1}(vk_a)+A_{B_1}P(\overline{B_1}vk_a)\\&=P_u\left( B_1P(\overline{B_1}vk_a)\right) . \end{aligned} \end{aligned}$$By (23) and (24), we conclude that
$$\begin{aligned} \begin{aligned} {[}A_{B_1},~A_{\phi }^*{]}K_u^2&= [A_{B_1},~A_{\phi }^*](\text {ker}~T_{\overline{u}B_1} \oplus (K_u^2\ominus \text {ker}~T_{\overline{u}B_1}))\\&\subseteq K_{B_1}^2 \oplus \text {span}\left\{ P_u\left( B_1P(\overline{B_1}vk_a) \right) \right\} . \end{aligned} \end{aligned}$$Then by (22),
$$\begin{aligned} \text {dim}~([A_{B_1},~A_{\phi }^*]K_u^2)\le 2. \end{aligned}$$ -
(b)
If \(\phi =1+v\), then
$$\begin{aligned} \begin{aligned} {[}A_{B_1},~A_{\phi }^*{]}(vk_a)&=(A_{B_1}A_{\overline{1+v}}-A_{\overline{1+v}}A_{B_1})(vk_a)\\&=A_{B_1}P(\overline{1+v}vk_a)-A_{\overline{1+v}}P_u(B_1vk_a)\\&=A_{B_1}(vk_a)+A_{B_1}P(\overline{v}vk_a)\\&=P_u\left( B_1k_a\right) . \end{aligned} \end{aligned}$$By (23) and (24), we conclude that
$$\begin{aligned} \begin{aligned} {[}A_{B_1},~A_{\phi }^*{]}K_u^2&= [A_{B_1},~A_{\phi }^*](\text {ker}~T_{\overline{u}B_1} \oplus (K_u^2\ominus \text {ker}~T_{\overline{u}B_1}))\\&\subseteq K_{B_1}^2 \oplus \text {span}\left\{ P_u\left( B_1k_a\right) \right\} . \end{aligned} \end{aligned}$$Then by (22),
$$\begin{aligned} \text {dim}~([A_{B_1},~A_{\phi }^*]K_u^2)\le 2. \end{aligned}$$
\(\square \)
References
Axler, S., Chang, S.-Y.A., Sarason, D.: Integral equations operator theory. Prod. Toeplitz Op. 1(3), 285–309 (1978)
Ahern, P.R., Clark, D.N.: Radial limits and invariant subspaces. Am. J. Math. 92, 332–342 (1970)
Ahern, P.R., Clark, D.N.: On functions orthogonal to invariant subspaces. Acta Math. 124, 191–204 (1970)
Beurling, A.: On two problems concerning linear transformations in Hilbert space. Acta Math. 81, 239–255 (1948)
Bessonov, R.: Truncated Toeplitz operators of finite rank. Proc. Am. Math. Soc. 142(4), 1301–1313 (2014)
Cristina, C.M., Partington, J.R.: Toeplitz kernels and model spaces. The diversity and beauty of applied operator theory, 139-153, Operator Theory: Advances and Applications, 268, (2018)
Chen, Y., Lee, Y.J., Zhao, Y.: Ranks of commutators for a class of truncated Toeplitz operators. Banach J. Math. Anal. 15(1), 16 (2021)
Douglas, R.G.: Banach algebra techniques in the theory of Toeplitz operators. Regional Conference Series in Mathematics, no. 15, American Mathematical Society, (1973)
Douglas, R.G., Shapiro, H.S., Lowell, Allen: Cyclic vectors and invariant subspaces for the backward shift operator. Université de Grenoble. Ann. de ıInst. Fourier 20(1), 37–76 (1970)
Fricain, E., Hartmann, A., Ross, W.T.: Multipliers between model spaces. Stud. Math. 240(2), 177–191 (2018)
Garcia, S.R., Ross, W.T., Wogen, W.R.: \(C^*\)-algebras generated by truncated Toeplitz operators. Concrete operators, spectral theory, operators in harmonic analysis and approximation, 181-192. Operator Theory Advances and Applications, 236, (2014)
Garcia, S.R., Putinar, M.: Complex symmetric operators and applications. Trans. Am. Math. Soc. 358(3), 1285–1315 (2006)
Garcia, S.R., Putinar, M.: Complex symmetric operators and applications II. Trans. Am. Math. Soc. 359(8), 3913–3931 (2007)
Hoffman, K.: Banach spaces of analytic functions. Englewood Cliffs, Prentice-Hall, Prentice-Hall Series in Modern Analysis, (1962)
Hayashi, E.: The kernel of a Toeplitz operator. Integr. Equs. Oper. Theory 9(4), 588–591 (1986)
Hitt, D.: Invariant subspaces of \(H^2\) of an annulus. Pac. J. Math. 134(1), 101–120 (1988)
Jurasik, J., Łanucha, B.: Asymmetric truncated Toeplitz operators equal to the zero operator Annales. Universitatis Mariae Curie-Sk\({\cal{l}}\)odowska, 7(2), 51–62 (2016)
Martinez-Avenda\({\tilde{n}}\)o, R.A., Rosenthal, P.M.: An introduction to operators on the Hardy-Hilbert space. Graduate Texts in Mathematics, 237. Springer, New York (2007)
Ma, P., Yan, F., Zheng, D.: Zero, finite rank, and compact big truncated Hankel operators on model spaces. Proc. Am. Math. Soc. 146(12), 5235–5242 (2018)
Ma Pan; Zheng Dechao: Compact truncated Toeplitz operators. J. Funct. Anal. 270(11), 4256–4279 (2016)
Sarason, D.: Algebraic properties of truncated Toeplitz operators. Oper. Matrices 1(4), 491–526 (2007)
Sarason, D.: Products of Toeplitz operators. In ‘Linear and Complex Analysis Problem Book 3’, Part I, Vol. 1573, 318–319, Springer-Verlag, (1994)
Sarason, D.: Nearly invariant subspaces of the backward shift. Contributions to operator theory and its applications, 481–493, Operator Theory: Advances and Applications, 35, (1988)
Szőkefalvi-Nagy, B., Foias, C., Bercovici, H., Kérchy, L.: Harmonic analysis of operators on Hilbert space, 2nd edn. Universitext, Springer, New York, Revised and enlarged edition (2010)
Volberg, Alexander: Two remarks concerning the theorem of S. Axler, S.-Y. A. Chang and D. Sarason. J. Oper. Theory 7(2), 209–218 (1982)
Yang, X., Lu, Y., Yang, Y.: Compact commutators of truncated Toeplitz operators on the model space. Ann. Funct. Anal. 13(3), 49 (2022)
Acknowledgements
This research is supported by National Natural Science Foundation of China No. 12031002, 11671065.
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
Not applicable.
Additional information
Communicated by Mohammad Sal Moslehian.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Springer Nature or its licensor holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Yang, X., Lu, Y. Toeplitz Kernels and Finite-Rank Commutators of Truncated Toeplitz Operators. Bull. Malays. Math. Sci. Soc. 45, 2175–2193 (2022). https://doi.org/10.1007/s40840-022-01374-1
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s40840-022-01374-1