1 Introduction and preparation
In this paper, $\mathbb {D}$ denotes the open unit disk and $\mathbb {T}$ be the unit circle. Let $H(\mathbb {D})$ be the space of all analytic functions on $\mathbb {D}$ . For $0<p<\infty $ , the Hardy space $H^p$ consists of analytic functions $f\in \mathbb {D}$ such that
If $p=\infty $ , then $H^{\infty }$ is the space of bounded analytic functions f on $H(\mathbb {D})$ with
For $0<p<\infty $ , the Bergman space $A^p$ consists of all functions f analytic on $\mathbb {D}$ such that
where $dA(z)$ is the normalized Lebesgue area measure on $\mathbb {D}$ . It is clear that $H^p\subset A^p$ . Moreover, $H^p\subset A^{2p}$ and $\|f\|_{A^{2p}}\leq \|f\|_{H^p}$ for $0<p<\infty $ . See [Reference Vukotić44] for example.
The Dirichlet-type space $\mathcal {D}^p$ is the set of all functions $f\in H(\mathbb {D})$ with
The space of all conformal automorphisms of $\mathbb {D}$ forms a group, called the Möbius group, and is denoted by $\mathrm {Aut}\, (\mathbb {D})$ . It is well known that $\varphi $ belongs to $ \mathrm {Aut}\, (\mathbb {D})$ if and only if there exists a real number $\theta $ and a point $a\in \mathbb {D}$ such that
Let X be a Banach space of analytic functions on $\mathbb {D}$ . Then X is said to be Möbius-invariant whenever $f\circ \varphi \in X$ for all $f\in X$ and $\varphi \in \mathrm {Aut}\, (\mathbb {D})$ and $\|f\circ \varphi \|_{X}=\|f\|_{X}$ .
For $1<p<\infty $ , the Besov space $B_p$ consists of analytic functions f on $\mathbb {D}$ such that
The norm of $B_p$ is defined as
When $p=\infty $ , $B_{\infty }=: \mathcal {B}$ is called the classical Bloch space. We define a norm on $\mathcal {B}$ as
When $p=2$ , $B_2=:\mathcal {D}$ is the classical Dirichlet space. When $p=1$ , we get the analytic Besov space $B_1$ , which is the minimal Möbius-invariant space consisting of all functions $f\in H(\mathbb {D})$ with
where the sequences $\{c_k\}_{k\geq 1}\in \ell ^1$ and $\{a_k\}_{k\geq 1}\in \mathbb {D}$ . An equivalent norm of $B_1$ is defined as
Arazy, Fisher, and Peetre [Reference Arazy, Fisher and Peetre5] first studied minimal Möbius-invariant space systematically. More results related to minimal Möbius-invariant space may be seen from [Reference Bao and Wulan6, Reference Blasco7, Reference Colonna and Li9, Reference Mitsis and Papadimitrakis32, Reference Ohno and Pavlović33, Reference Wulan and Xiong45].
Now, we define several operators on $B_1$ . For $g\in H(\mathbb {D})$ , the multiplication operator $M_g$ on $B_1$ is defined as
The differentiation operator is given by $Df=f'$ for each $f\in H(\mathbb {D})$ . Given $g\in H(\mathbb {D})$ , the Volterra-type operator $T_g$ is defined as
When $g(z)=z$ , the operator $T_zf(z)=\int ^{z}_{0}f(w)dw$ becomes the simplest Volterra operator. An integral operator related to $T_g$ , denoted by $I_g$ , is defined as
The Volterra-type operator $T_g$ was originally studied by Pommerenke [Reference Pommerenke38]. Later, a series of articles appeared on the study of Volterra-type integration operators on classical spaces of analytic functions, such as Hardy spaces, Bergman spaces, and Dirichlet-type spaces. For more details, please refer to [Reference Aleman and Cima1, Reference Aleman and Siskakis3, Reference Aleman and Siskakis4, Reference Galanopoulos, Girela and Peláez16, Reference Girela and Peláez18]. In [Reference Čučković and Paudyal13], Čučković and Paudyal describe the lattice of the closed invariant subspaces of Volterra-type operators. Lin, Liu, and Wu [Reference Lin, Liu and Wu22] generalized some of the works of [Reference Čučković and Paudyal13] to the general case when $1\leq p<\infty $ , and obtained the boundedness of the Volterra-type operators $T_g$ and $I_g$ on the derivative Hardy space $S_p(\mathbb {D})$ . And then, they also considered strict singularity of Volterra-type operators on Hardy spaces in [Reference Lin, Liu and Wu23]. Meanwhile, Lin [Reference Lin21] characterized the boundedness and compactness of the Volterra-type operators between Bloch-type spaces and weighted Banach spaces. In [Reference Miihkinen, Pau, Perälä and Wang31], Miihkinen et al. completely characterize the boundedness of the Volterra-type operators acting from the weighted Bergman spaces to the Hardy spaces of the unit ball.
In this paper, we mainly study the operator-theoretic properties in minimal Möbius-invariant space $B_1$ . The structure of this article is as follows. In Section 2, we discuss the boundedness of the Volterra operator on $B_1$ . In Section 3, it is shown whether the integral operator belongs to Deddens algebras. In Section 4, we will be concerned with the essential norms of integral operators on $B_1$ . Section 5 is devoted to the study of the spectrum of integral operators on $B_1$ .
Throughout this paper, we use the following convention. For two nonnegative functions F and G defined on some function space $X,$ we write $F \lesssim G$ if $F(f) \leq C\cdot G(f)$ for all $f \in X$ and for some positive constant C which is independent of F and $G.$ Denote by $F\approx G$ whenever $F\lesssim G \lesssim F$ .
2 Volterra-type operators on $B_1$
First, we need the following Hardy inequality.
Lemma 1 [Reference Duren15]
If $f\in H^1$ and $f(z)= \sum ^{\infty }_{n=0}a_nz^n$ , then
The following lemma is a classic exercise in mathematical analysis, but it might be worth to give a brief details of the proof for completeness.
Lemma 2 Suppose that $f(x)$ is a continuously increasing function on $[a,b]$ . Then,
Proof As $f(x)$ is monotonically increasing on $[a,b]$ , the “integral mean value theorem” shows that there exists a $\xi \in [a,b]$ such that
This completes the proof of the lemma.
Lemma 3 If $f\in \mathcal {D}^1$ , then $\|f\|_{H^1}\leq \|f\|_{\mathcal {D}^1}.$
Proof Let $f\in \mathcal {D}^1.$ Then, we can see that
Thus,
On the other hand,
By Hardy’s convexity theorem (see [Reference Mashreghi28, Reference Xiao and Zhu46]), we find that
is a nondecreasing function of r. It follows from Lemma 2 that
This shows that $\|f\|_{H^1}\leq \|f\|_{\mathcal {D}^1}$ .
Remark 1 In Lemma 3, set $f(z)=z$ . Then, we see that $1=\|z\|_{H^1}\leq \|z\|_{\mathcal {D}^1}=1$ , showing that the norm estimate is sharp. This improves the previous conclusion, namely, $\|f\|_{H^1}\leq 2\|f\|_{\mathcal {D}^1}$ from the works of Girela and Merchán [Reference Girela and Merchán17].
Lemma 4 If $f\in B_1$ , then $\|f\|_{\infty }\leq \pi \|f\|_{B_1}$ and $B_1\subset H^{\infty }$ .
Proof Assume that $f\in B_1$ and write $f(z)=\sum ^{\infty }_{k=0}a_kz^k$ . By Lemmas 1 and 3, we have
for all $z\in \mathbb {D}$ . Hence, we obtain that $\|f\|_{\infty }\leq \pi \|f\|_{B_1}$ .
Remark 2 In [Reference Lin, Liu and Wu22, Theorem 1], they obtained that $\|f\|_{\infty }\leq \pi \|f\|_{S^1}$ for each $f\in S^1$ , where the space $S^1$ is defined as $S^1=\{f\in H^1:f'\in H^1\}$ . The norm on $S^1$ is given by
Moreover, we obtain the following norm estimate:
In the following, we discuss the boundedness of $T_g$ and $I_g$ on minimal Möbius-invariant space.
Theorem 1 The operator $T_g$ is bounded on $B_1$ if and only if $g\in B_1$ . Moreover,
Proof Let $f\in B_{1}$ . By Hölder’s inequality and Lemma 3, we have
Hence, we get
showing that $T_g$ is a bounded operator on $B_1$ .
Conversely, assume that $T_g$ is a bounded operator on $B_1$ and let $f=1$ . Then, we obtain
which gives that $g\in B_1$ . Thus,
For $0<p<\infty $ , $-2<q<\infty $ , and $0\leq s<\infty $ , we define the general family of function spaces $F(p,q,s)$ as the set of all analytic functions f on $\mathbb {D}$ such that
where $g(z,a)=\log \frac {1}{|\sigma _a(z)|}$ . These spaces were introduced by Zhao in [Reference Zhao47]. In 2003, Rättyä provided the following nth derivation characterization of functions in spaces $F(p,q,s)$ .
Lemma 5 [Reference Rättyä39, Theorem 3.2]
Let f be an analytic function on $\mathbb {D}$ , and let $0<p<\infty , -2<q<\infty $ , and $0\leq s<\infty $ . Let $n\in \mathbb {N}$ and $q+s>-1$ ; or $n=0$ and $q+s-p>-1$ . Then, $f\in F(p,q,s)$ if and only if
For $p>0$ , the space $Z_p$ consists of all analytic functions f on $\mathbb {D}$ such that
It is clear that $Z_1=F(1,-1,1)$ . For more results about $Z_{p}$ space, see [Reference Liu, Lou and Zhu25, Reference Zhu48].
Theorem 2 The operator $I_g$ is bounded on $B_1$ if and only if $g\in Z_{1}\cap H^{\infty }$ .
Proof Assume that $g \in Z_{1}\cap H^{\infty }$ . Using Lemma $6$ of [Reference Liu, Lou and Zhu25], we obtain that
Thus, we have
for all $f\in B_1$ . This implies that $I_g$ is a bounded operator on $B_1$ .
Conversely, suppose that $I_g$ is a bounded operator on $B_1$ . For each $a\in \mathbb {D}$ , let $f(z)=\sigma _{a}(z)$ . Then, $\|f\|_{B_1}\lesssim 1$ and
which shows that $\|I_g\|\gtrsim \|g\|_{\infty }$ . Hence, $g\in H^{\infty }$ . Moreover,
from which it follows that
This implies that $g\in Z_1$ . Therefore, $g\in Z_1\cap H^{\infty }$ . This completes the proof of the theorem.
Remark 3 Note that $B_1\subset Z_1\cap H^{\infty }.$ By Lemma 4, we know that $f\in H^{\infty }$ whenever $f\in B_1$ . Using the second derivation characterization of functions on $Z_1$ , we have
Then $f\in Z_1$ , and thus $B_1\subset Z_1\cap H^{\infty }$ .
Now, we define the space $B^0_1$ as
The following theorem gives the connection between the Volterra operator $T_z$ on $\mathcal {D}^1$ and the multiplication operator $M_z$ on $B^0_1$ .
Theorem 3 The Volterra-type operator $T_{z}:\mathcal {D}^1\rightarrow B^0_1$ is bounded and invertible with $T^{-1}_z= D .$
Proof First, we show that $T_z(\mathcal {D}^1)=B^0_1.$ For $f\in \mathcal {D}^1$ , we consider
Clearly, $F'=f\in \mathcal {D}^1$ so that $F\in B^0_1$ and $T_{z}(\mathcal {D}^1)\subseteq B^0_1.$
Conversely, for each $F\in B^0_1$ , we have $F' \in \mathcal {D}^1$ . Then,
Then, $B^0_1 \subseteq T_{z}(\mathcal {D}^1)$ . This implies that $T_{z}(\mathcal {D}^1)=B^0_1$ .
Second, we show that $T_z:\mathcal {D}^1 \rightarrow B^0_1$ is a bounded isomorphism, and its inverse $T^{-1}_{z}= D.$ Recall from the above discussion that $(T_z(F'(z)))=F(z)$ for $F\in B^0_1$ . Then, for each $f\in \mathcal {D}^1$ , we have that $(D(T_zf))(z)=f(z)$ . This implies that $T_z$ is a bijective operator from $\mathcal {D}^1$ onto $B^0_1$ , since $T_z$ is linear and $T_z$ is an isomorphism from $\mathcal {D}^1$ onto $B^0_1$ .
Finally, we need to prove that $T_z$ is a bounded operator on $\mathcal {D}^1$ . For $f\in \mathcal {D}^1$ , we have
Therefore, $T_z$ is a bounded isomorphism from $\mathcal {D}^1$ onto $B^0_1$ .
Let us introduce an addition operator P defined as
Theorem 4 Let $T_{z}:\mathcal {D}^1\rightarrow B^0_1$ and $M_z:B^0_1\rightarrow \mathcal {D}^1$ be the Volterra-type operator and the multiplication operator, respectively. Then, P is an operator on $\mathcal {D}^1$ with $P=T^{-1}_{z}M_zT_z.$
Proof Let $f\in \mathcal {D}^1$ and $F=T_zf$ . Then, we get
which shows that $P=T^{-1}_{z}M_zT_z$ .
Theorem 5 If $f,g\in B_1$ , then $\|fg\|_{B_1}\leq (2\pi +2)\|f\|_{B_1}\|g\|_{B_1}$ .
Proof For $f,g\in B_1$ , we get
and the proof is complete.
Remark 4 In [Reference Arazy, Fisher and Peetre5, Theorem 10], Arazy et al. obtained $\|fg\|\leq 7\|f\|\|g\|$ for $f,g\in B_1$ , in which they defined the norm of $f\in B_1$ as
Inspired by their work, we derived Theorem 5 and we are not sure whether the constant $2\pi +2$ in Theorem 5 is optimal or not.
3 Deddens algebras
Let $\mathcal {L}(X)$ denote the algebra of all bounded linear operators on a complex Banach space X. A nontrivial invariant subspace of an operator $A\in \mathcal {L}(X)$ is, by definition, a closed subspace M of X such that $M\neq \{0\}$ , $M\neq X$ , and $Ax\in M$ for every $x\in M$ ; or, briefly, $A(M)\subset M$ .
Let $A\in \mathcal {L}(X)$ . The operator T is said to belong to the Deddens algebra $\mathcal {D}_A$ if there exists $M=M(T)>0$ such that
for each $n\in \mathbb {N}$ and $f\in X$ .
The study of the Deddens algebra was originally introduced by Deddens [Reference Deddens14], where he assumed that A is an invertible operator and $\sup _{n\in \mathbb {N}}\|A^nTA^{-n}\|<\infty $ . Later, it received the attention of many scholars (see [Reference Deddens14, Reference Karaev and Mustafayev19, Reference Lacruz20, Reference Petrovic34–Reference Petrovic and Sievewright37, Reference Sievewright42, Reference Todorov43]). Recently, Petrovic and Sievewright [Reference Petrovic and Sievewright37] studied the Deddens algebra associated with compact composition operators $C_{\varphi }$ on Hardy spaces, where A is not necessarily invertible, and they have demonstrated that the operators $M_g$ and $T_z$ belong to the Deddens algebra $\mathcal {D}_{C_{\varphi }}$ . It is worth to point out that compact operators on $H^2$ are not invertible.
Let us begin to present the boundedness of composition operators and multiplication operators on $B_1$ .
Lemma 6 [Reference Wulan and Xiong45]
Let $\varphi $ be an analytic self-map of $\mathbb {D}$ . Then, the composition operator $C_{\varphi }$ is bounded on $B_1$ if and only if
In particular, we give another description of sufficiency condition for the boundedness of the composition operator in the following theorem.
Theorem 6 Let $\varphi $ be an analytic self-map of $\mathbb {D}$ such that $\varphi (0)=0$ . Then, the composition operator $C_{\varphi }$ is bounded on $B_1$ whenever $\varphi '\in Z_1\cap H^{\infty }$ .
Proof Suppose that $\varphi '\in Z_1\cap H^{\infty }$ . Then, we get
which implies that $C_{\varphi }$ is bounded on $B_1$ if $\varphi '\in Z_1\cap H^{\infty }$ .
Theorem 7 Suppose that $M_g$ is a multiplication operator on $B_1$ . Then, $M_g$ is bounded if and only if $g\in B_1$ .
Proof For any $f\in B_1$ , if $g\in B_1$ , we have $M_g$ is bounded on $B_1,$ by Theorem 5.
Conversely, let $M_g$ be a bounded operator on $B_1.$ Then, with $f=1$ , we get $\|M_g\|\geq \|M_g1\|_{B_1}=\|g\|_{B_1}$ , which implies that $g\in B_1$ .
In the following theorem, we will consider the algebra $\mathcal {D}_{C_{\varphi }}$ , in which the operator $C_{\varphi }$ is a bounded composition operator. For $n\in \mathbb {N}$ , it clear that $C^{n}_{\varphi }f=f\circ \varphi \circ \cdots \circ \varphi $ . For simplicity of the notation, we write $\varphi _{n}$ instead of $\varphi \circ \cdots \circ \varphi $ .
Theorem 8 Let $g\in B_1$ , and let $\varphi $ be an analytic self-map of $\mathbb {D}$ with $\varphi (0)=0$ such that $C_{\varphi }$ is bounded on $B_1$ . Then, the operators $M_g$ , $T_g$ , and $I_g$ belong to the Deddens algebra $\mathcal {D}_{C_{\varphi }}$ .
Proof For each $n\in \mathbb {N}$ , we see that
Since $\varphi _{n}(\mathbb {D})\subset \mathbb {D}$ , it follows that
and therefore
where $f\in B_{1}(\mathbb {D})$ . This implies that $M_g\in \mathcal {D}_{C_{\varphi }}$ .
Next, we have
Since $\varphi $ is an analytic self-map of $\mathbb {D}$ satisfying $\varphi (0)=0$ , we have $\varphi _{n}(0)=0$ , and therefore
where $f\in B_1$ . It shows that $C^n_{\varphi }T_g=T_{g\circ \varphi _n}C^{n}_{\varphi }$ . By Theorem 1, we have
which gives $T_g\in \mathcal {D}_{C_{\varphi }}$ .
Finally, we have
and
where $f\in B_1$ . Therefore, $C^n_{\varphi }I_g=I_{g\circ \varphi _{n}}C^n_{\varphi }$ . By Theorem 2 and Remark 3, we find that
for all $f\in B_1$ . We thus deduce that $I_g\in \mathcal {D}_{C_{\varphi }}$ .
4 Essential norms of Volterra-type operators on $B_{1}$
Suppose that X is a Banach space and T is a bounded linear operator on X. The essential norm of T is defined to be
Obviously, the essential norm of T is 0 if and only if T is compact. For more results, we invite the reader to refer to [Reference Liu, Liu, Xia and Yu26, Reference Shargorodsky41]. In this section, we characterize the essential norm of linear operator on $B_1$ , which generalizes the conclusion of Liu et al. [Reference Liu, Lou and Xiong24]
Theorem 9 Every bounded operator $T_g$ on $B_1$ is compact.
Proof By definition, we know $\|T_{g}\|_{e}\geq 0$ .
Next, we show that $\|T_{g}\|_{e}\leq 0$ . To do this, we define the following operators:
where $g_r(z)=g(rz)$ and $r\in (0,1)$ . It is easy to see that $T_{g_r}$ is a compact operator on $B_1$ for $g\in B_1$ . In fact, if $T_{g}$ is a bounded operator on $B_1$ , then $g\in B_1$ . Suppose that $\{ f_{n}\}_{n=1}^{\infty } \subset B_1$ with $\|f_{n}\|_{B_1}\leq 1$ , and $f_{n}\rightarrow 0$ uniformly on compact subsets of $\mathbb {D}$ . Then,
where $\delta _1<1,$ and $\mathbb {D}_{\delta _1}= \{z: |z|<\delta _1\}$ are compact subsets of $\mathbb {D}$ .
Note that
Using the theorem of absolute continuity of Lebesgue measure, we conclude that
On the other hand, using a basic result of complex analysis (see page 151 of [Reference Conway12]), we know that if $f_{n}\rightarrow 0,$ then $f_{n}' \rightarrow 0$ uniformly on compact subsets of $\mathbb {D}$ . Consequently,
converge to $0$ when $n\rightarrow \infty $ . This implies that $T_{g,r}$ is compact.
Meanwhile, we have
for $r\in (0,1)$ . Similarly, with the above computation, we have
so that $\|T_{g}\|_{e}\leq 0$ . Hence, we deduce that $\|T_{g}\|_{e}=0$ , which completes the proof.
Theorem 10 If $I_{g}$ is bounded operator on $B_1$ , then $\|g\|_{\infty }\leq \|I_{g}\|_{e} \lesssim ( \|g\|_{\infty }+\|g\|_{Z_{1}}).$
Proof From the proof of Theorem 2, we have
where C is a positive constant.
We next show that $\|I_g\|_{e}\geq \|g\|_{\infty }$ . Choose $a_n\in \mathbb {D}$ such that $|a_n|\rightarrow 1$ as $n\rightarrow \infty $ . Let $f_n(z)=\sigma _{a_n}(z)-a_n$ . It is obvious that $\|f_n\|_{B_1}=1$ . Since $\{f_n\}$ converges to zero uniformly on compact subsets of $\mathbb {D}$ , for every compact operator K on $B_1$ , we obtain $\|Kf_n\|_{B_1}\rightarrow 0$ as $n\rightarrow \infty $ . Therefore,
Similarly to the proof of Theorem 2, we get
As the choice of the sequence $\{a_n\}\subset \mathbb {D}$ is arbitrary, we have $\|I_g\|_{e}\geq \|g\|_{\infty }$ , which completes the proof.
Theorem 11 Every bounded operator $M_g$ on $B_1$ is compact.
Proof The proof is similar to Theorem 9, so we omit its details.
5 Spectrum of Volterra-type operators on $B_{1}(\mathbb {D})$
The spectrum of integral operators on different spaces has attracted the attention of many scholars. The spectra of integral operators on weighted Bergman spaces are characterized by Aleman and Constantin [Reference Aleman and Constantin2]. Later, Constantin [Reference Constantin10] obtained the spectrum of Volterra-type operators on Fock spaces. Mengestie [Reference Mengestie29] studied the spectrum of Volterra-type operators on Fock–Sobolev spaces. Mengestie [Reference Mengestie30] also obtained the spectrum of $T_g$ in terms of a closed disk of radius twice the coefficient of the highest degree term in a polynomial expansion of g. For more results, see [Reference Bonet8, Reference Malman27]. Recently, Lin et al. described the spectra of the multiplication operator and the Volterra-type operator $I_g$ in [Reference Lin, Liu and Wu23], respectively. Inspired by the above results, it is natural to study the spectra of the multiplication operator and the Volterra-type operators on $B_{1}$ .
Theorem 12 Suppose that $M_g$ is bounded on $B_1$ . Then, we have $\sigma (M_g)=\overline {g(\mathbb {D})}.$
Proof Suppose that $\lambda \notin \sigma (M_g)$ . Then, $M_g-\lambda I$ is invertible. As $1 \in B_1,$ there exists an $f\in B_1$ such that $(g(z)-\lambda )f(z)=1$ for all $z\in \mathbb {D}$ , which implies that $\lambda \notin g(\mathbb {D})$ . Thus, $g(\mathbb {D})\subset \sigma (M_g)$ .
For the other way inclusion, we let $\lambda \notin \overline {g(\mathbb {D})}$ . Then, we can choose a $t>0$ such that $|g(z)-\lambda |>t$ for all $z\in \mathbb {D}$ . This shows that $h=(g-\lambda )^{-1}$ is a bounded analytic function on $\mathbb {D}$ . For all $g\in B_1$ , we get
Hence, $h\in B_1$ . Then, $M_h$ is bounded on $B_1,$ by Theorem 7. Since $M_h=M_{(g-\lambda )^{-1}}=M_{g-\lambda }^{-1}$ , we see that $M_{g-\lambda }$ is invertible, and thus $\lambda \notin \sigma (M_g)$ . Therefore, $\sigma (M_g)\subset \overline {g(\mathbb {D})}$ . Since the spectrum set is closed, we conclude that $\sigma (M_g)=\overline {g(\mathbb {D})}$ .
Lemma 7 [Reference Rudin40]
Let T be a bounded linear operator on a Banach space X, and let T be compact. If $\dim X=\infty $ , then $\sigma (T)=\{0\} {\cup} $ {eigen values of T}.
Theorem 13 Suppose that $T_g$ is a bounded operator on $B_1$ . Then, $\sigma (T_g)=\{0\}.$
Proof Let $T_g$ be a bounded operator on $B_1.$ Then, $T_g$ is compact, by Theorem 9. By Lemma 7, we obtain $0\in \sigma (T)$ .
Next, we prove that $T_g$ has no nonzero eigenvalue. Assume that $T_g$ has an eigenvalue $\lambda \neq 0$ with eigenvector $f.$ Then,
Differentiating equation (1) with respect to z, we get
All nonzero solutions of this equation are of the form $f(z)=ce^{\frac {g(z)}{\lambda }}$ for some $c\neq 0$ . Setting $z=0$ in (1) shows that $0=\lambda f(0)$ , which contradicts the last relation about $f.$ Therefore, there is no nonzero eigenvalue for $T_g$ . From this, we deduce that $\sigma (T_g)=\{0\}$ .
Theorem 14 If $I_g$ is a bounded operator on $B_1$ , then
Proof For any constant function a, we have
which gives $0\in \sigma (I_g)$ .
Suppose that $\lambda \in \mathbb {C}\backslash \{0\}$ . Note that the equation
has a unique analytic solution
(see [Reference Constantin and Persson11] for more details). Hence, the resolvent set $\rho (I_g)$ of the bounded operator $I_g$ consists precisely of all points $\lambda \in \mathbb {C}$ for which $R_{\lambda ,g}$ is a bounded operator on $B_1$ .
If $\lambda \in \mathbb {C}\backslash (\{0\}\cup \overline {g(D)})$ , then it is clear that $1-\frac {1}{\lambda }g(z)$ is bounded away from $0$ , which implies that $\frac {1}{1-\frac {1}{\lambda }g(z)} \in H^{\infty }$ . If $I_g$ is a bounded operator on $B_1$ , then $g\in H^{\infty }\bigcap Z_{1}$ by Theorem 2. Moreover, it is easy to show that $ \frac {1}{1-\frac {1}{\lambda }g(z)} \in Z_{1}.$ This implies that the operator $R_{\lambda ,g}$ is a bounded operator on $B_1$ . It follows that $\mathbb {C}\backslash (\{0\}\cup \overline {g(\mathbb {D})})\subset \rho (I_g)$ . Thus, $\sigma (I_g)\subset (\{0\}\cup \overline {g(\mathbb {D})})$ .
On the other hand, if $\lambda \in g(\mathbb {D})$ and $\lambda \neq 0$ , then $\frac {1}{1-\frac {1}{\lambda }g(\xi )}$ is not bounded, which shows that the operator $R_{\lambda ,g}$ is not bounded on $B_1$ . Therefore, we obtain that $g(\mathbb {D})\backslash \{0\}\subset \sigma (I_g)$ . This together with the fact that $0\in \sigma (I_g)$ shows that
Since the spectrum $\sigma (I_g)$ is closed, we deduce that $\sigma (I_g)=\overline {g(\mathbb {D})}\cup \{0\}$ .
Acknowledgement
We would like to thank the referee for his/her valuable comments.