Abstract
We consider the Hilbert-type operator defined by
where \(\{B^{\omega }_\zeta \}_{\zeta \in \mathbb {D}}\) are the reproducing kernels of the Bergman space \(A^2_\omega \) induced by a radial weight \(\omega \) in the unit disc \(\mathbb {D}\). We prove that \(H_{\omega }\) is bounded on the Hardy space \(H^p\), \(1<p<\infty \), if and only if
and
where \(\widehat{\omega }(r)=\int _r^1 \omega (s)\,ds\). We also prove that \(H_\omega : H^1\rightarrow H^1\) is bounded if and only if (\(\dag \)) holds and
As for the case \(p=\infty \), \(H_\omega \) is bounded from \(H^\infty \) to \(\mathord \textrm{BMOA}\), or to the Bloch space, if and only if (\(\dag \)) holds. In addition, we prove that there does not exist radial weights \(\omega \) such that \(H_{\omega }: H^p \rightarrow H^p \), \(1\le p<\infty \), is compact and we consider the action of \(H_{\omega }\) on some spaces of analytic functions closely related to Hardy spaces.
Similar content being viewed by others
Avoid common mistakes on your manuscript.
1 Introduction
For \(0<p<\infty \), let \(L^p_{ [0,1)}\) be the Lebesgue space of measurable functions such that
and let \(\mathcal {H}(\mathbb {D})\) denote the space of analytic functions in the unit disc \(\mathbb {D}=\{z\in \mathbb {C}:|z|<1\}\). The Hardy space \(H^p\) consists of \(f\in \mathcal {H}(\mathbb {D})\) for which
where
and
For a nonnegative function \(\omega \in L^1_{[0,1)}\), the extension to \(\mathbb {D}\), defined by \(\omega (z)=\omega (|z|)\) for all \(z\in \mathbb {D}\), is called a radial weight. Let \(A^2_{\omega }\) denote the weighted Bergman space of \(f\in \mathcal {H}(\mathbb {D})\) such that \(\Vert f\Vert _{A^2_\omega }^2=\int _\mathbb {D}|f(z)|^2\omega (z)\,dA(z)<\infty \), where \(dA(z)=\frac{dx\,dy}{\pi }\) is the normalized area measure on \(\mathbb {D}\). Throughout this paper we assume \(\widehat{\omega }(z)=\int _{|z|}^1\omega (s)\,ds>0\) for all \(z\in \mathbb {D}\), for otherwise \(A^2_\omega =\mathcal {H}(\mathbb {D})\).
The Hilbert matrix is the infinite matrix whose entries are \(h_{n,k}=(n+k+1)^{-1},\) \(k,n\in \mathbb {N}\cup \{0\}\). It can be viewed as an operator on spaces of analytic functions, by its action on the Taylor coefficients
called the Hilbert operator. That is, if \(f(z)=\sum _{k=0}^\infty \widehat{f}(k)z^k\in \mathcal {H}(\mathbb {D}) \)
whenever the right hand side makes sense and defines an analytic function in \(\mathbb {D}\).
The Hilbert operator H is bounded on Hardy spaces \(H^p\) if and only if \(1<p<\infty \) [4]. A proof of this result can be obtained using the following integral representation, valid for any \(f\in H^1\),
Going further, the formula (1.2) has been employed to solve a good number of questions in operator theory related to the boundedness, the operator norm and the spectrum of the Hilbert operator on classical spaces of analytic functions [1, 3, 5, 24]. During the last decades several generalizations of the Hilbert operator have attracted a considerable amount of attention [9, 11, 24, 26]. We will focus on the following, introduced in [26]. For a radial weight \(\omega \), we consider the Hilbert-type operator
where \(\{B^\omega _z\}_{z\in \mathbb {D}}\subset A^2_\omega \) are the Bergman reproducing kernels of \(A^2_\omega \). The choice \(\omega =1\) gives the integral representation (1.2) of the classical Hilbert operator, therefore it is natural to think of the features of a radial weight \(\omega \) so that \(H_\omega \) has some of the nice properties of the (classical) Hilbert operator. In this paper, among other results, we describe the radial weights \(\omega \) such that the Hilbert-type operator \(H_\omega \) is bounded on \(H^p\), \(1\le p<\infty \).
In order to state our results some more notation is needed. For \(0<p<\infty \), the Dirichlet-type space \(D^p_{p-1}\) is the space of \(f\in \mathcal {H}(\mathbb {D})\) such that
and the Hardy–Littlewood space HL(p) consists of the \(f(z)=\sum \nolimits _{n=0}^{\infty } \widehat{f}(n) z^n\in \mathcal {H}(\mathbb {D})\) such that
We will also consider the space \(H(\infty ,p)=\{f\in \mathcal {H}(\mathbb {D}): \Vert f\Vert ^p_{H(\infty ,p)}=\int _0^1\,M^p_\infty (r,f)\,dr<\infty \}.\) These spaces satisfy the well-known inclusions
and
See [6, 7, 14] for proofs of (1.3) and (1.4), and [27, p. 127] and [8, Lemma 4] for a proof of (1.5).
The Bergman reproducing kernels, induced by a radial weight \(\omega \), can be written as \(B^\omega _z(\zeta )=\sum \overline{e_n(z)}e_n(\zeta )\) for each orthonormal basis \(\{e_n\}\) of \(A^2_\omega \), and therefore using the basis induced by the normalized monomials,
Here \(\omega _{2n+1}\) are the odd moments of \(\omega \), and in general from now on we write \(\omega _x=\int _0^1r^x\omega (r)\,dr\) for all \(x\ge 0\). A radial weight \(\omega \) belongs to the class \(\widehat{\mathcal {D}}\) if \(\widehat{\omega }(r)\le C\widehat{\omega }(\frac{1+r}{2})\) for some constant \(C=C(\omega )>1\) and all \(0\le r <1\). If there exist \(K=K(\omega )>1\) and \(C=C(\omega )>1\) such that \(\widehat{\omega }(r)\ge C\widehat{\omega }\left( 1-\frac{1-r}{K}\right) \) for all \(0\le r<1\), then \(\omega \in \check{\mathcal {D}}\). Further, we write \(\mathcal {D}=\widehat{\mathcal {D}}\cap \check{\mathcal {D}}\) for short. Recall that \(\omega \in \mathcal {M}\) if there exist constants \(C=C(\omega )>1\) and \(K=K(\omega )>1\) such that \(\omega _{x}\ge C\omega _{Kx}\) for all \(x\ge 1\). It is known that \(\check{\mathcal {D}}\subset \mathcal {M}\) [23, Proof of Theorem 3] but \(\check{\mathcal {D}}\subsetneq \mathcal {M}\) [23, Proposition 14]. However, [23, Theorem 3] ensures that \(\mathcal {D}=\widehat{\mathcal {D}}\cap \check{\mathcal {D}}=\widehat{\mathcal {D}}\cap \mathcal {M}\). These classes of weights arise in meaningful questions concerning radial weights and classical operators, such as the differentiation operator \(f^{(n)}\) or the Bergman projection \( P_\omega (f)(z)=\int _{\mathbb {D}}f(\zeta ) \overline{B^\omega _{z}(\zeta )}\,\omega (\zeta )dA(\zeta )\) [23]. We will also deal with the sublinear Hilbert-type operator
If \(X,Y \subset \mathcal {H}(\mathbb {D})\) are normed vector spaces, and T is a sublinear operator, we denote \(\Vert T\Vert _{X\rightarrow Y}=\sup _{\Vert f\Vert _X\le 1}\Vert T(f)\Vert _{Y}\).
Theorem 1
Let \(\omega \) be a radial weight and \(1<p<\infty \). Let \(X_p,Y_p\in \{H(\infty ,p),H^p, D^p_{p-1}, HL(p)\}\) and \(T\in \{ H_{\omega }, \widetilde{H_{\omega }}\}\). Then the following statements are equivalent:
-
(i)
\(T:\,X_p \rightarrow Y_p\) is bounded;
-
(ii)
\(\omega \in \mathcal {D}\) and \( M_p(\omega )=\sup \nolimits _{N\in \mathbb {N}} \left( \sum \nolimits _{n=0}^N \frac{1}{(n+1)^2 \omega _{2n+1}^p}\right) ^{\frac{1}{p}}\left( \sum \nolimits _{n=N}^{\infty } \omega _{2n+1}^{p'}(n+1)^{p'-2}\right) ^{\frac{1}{p'}}<\infty ;\)
-
(iii)
\(\omega \in \widehat{\mathcal {D}}\) and \( M_p(\omega )=\sup \nolimits _{N\in \mathbb {N}} \left( \sum \nolimits _{n=0}^N \frac{1}{(n+1)^2 \omega _{2n+1}^p}\right) ^{\frac{1}{p}}\left( \sum \nolimits _{n=N}^{\infty } \omega _{2n+1}^{p'}(n+1)^{p'-2}\right) ^{\frac{1}{p'}}<\infty ; \)
-
(iv)
\(\omega \in \widehat{\mathcal {D}}\) and \( M_{p,c}(\omega )= \sup \nolimits _{0<r<1} \left( \int _0^r \frac{1}{\widehat{\omega }(t)^p} dt\right) ^{\frac{1}{p}} \left( \int _r^1 \left( \frac{\widehat{\omega }(t)}{1-t}\right) ^{p'}\,dt\right) ^{\frac{1}{p'}}<\infty \).
The proof of (i)\(\Rightarrow \)(iii) of Theorem 1 has two steps. Firstly, we prove that \(\omega \in \widehat{\mathcal {D}}\), and later on the condition \(M_p(\omega )<\infty \) is obtained by using polynomials of the form \(f_{N, M}(z)=\sum \nolimits _{k=N}^M \omega _{2k}^{\alpha } (k+1)^{\beta }z^k,\, N,M\in \mathbb {N},\; \alpha , \beta \in \mathbb {R}\) as test functions. Then, we see that any radial weight \(\omega \) satisfying the condition \(M_p(\omega )<\infty \), belongs to \(\mathcal {M}\). This proves (ii)\(\Leftrightarrow \)(iii). The proof of (iii)\(\Leftrightarrow \)(iv) is a calculation based on known descriptions of the class \(\widehat{\mathcal {D}}\) [21, Lemma 2.1]. Finally, we prove (iv)\(\Rightarrow \)(i) which is the most involved implication in the proof of Theorem 1. In order to obtain it, we merge techniques coming from complex and harmonic analysis, such as a very convenient description of the class \(\mathcal {D}\), see Lemma 14 below, precise estimates of the integral means of order p of the derivative of the kernels \(K^\omega _u(z)=\frac{1}{z}\int _0^z B^\omega _u(z)\,du \), decomposition norm theorems and classical weighted inequalities for Hardy operators.
Observe that both, the discrete condition \(M_p(\omega )<\infty \) and its continuous version \(M_{p,c}(\omega )<\infty \), are used in the proof of Theorem 1. The first one follows from (i), and the condition \(M_{p,c}(\omega )<\infty \) is employed to prove that \(T:\,X_p \rightarrow Y_p\) is bounded.
As for the case \(p=1\) we obtain the following result.
Theorem 2
Let \(\omega \) be a radial weight, \(X_1,Y_1\in \{H(\infty ,1),H^1, D^1_{0}, HL(1)\}\) and \(T\in \{ H_{\omega }, \widetilde{H_{\omega }}\}\). Then the following statements are equivalent:
-
(i)
\(T: X_1\rightarrow Y_1\) is bounded;
-
(ii)
\(\omega \in \widehat{\mathcal {D}}\) and the measure \(\mu _\omega \) defined as \(d\mu _\omega (z)= \omega (z)\left( \int _0^{|z|} \frac{ds}{\widehat{\omega }(s)}\right) \,\chi _{[0,1)}(z)\, dA(z)\) is a 1-Carleson measure for \(X_1\);
-
(iii)
\(\omega \in \widehat{\mathcal {D}}\) and satisfies the condition
$$\begin{aligned} M_{1,c}(\omega )= \sup \limits _{a \in [0,1)} \frac{1}{1-a}\int _a^1 \omega (t)\left( \int _0^t \frac{ds}{\widehat{\omega }(s)}\right) \,dt<\infty \textit{;} \end{aligned}$$ -
(iv)
\(\omega \in \mathcal {D}\) and satisfies the condition \( M_{1,c}(\omega )<\infty \);
-
(v)
\(\omega \in \widehat{\mathcal {D}}\) and satisfies the condition \(M_{1,d}(\omega )= \sup \nolimits _{a \in [0,1)} \frac{\widehat{\omega }(a)}{1-a} \left( \int _0^a \frac{ds}{\widehat{\omega }(s)}\right) <\infty \);
-
(vi)
\(\omega \in \widehat{\mathcal {D}}\) and satisfies the condition
$$\begin{aligned} M_1(\omega )= \sup \limits _{N \in \mathbb {N}} (N+1)\omega _{2N}\sum \limits _{k=0}^{N}\frac{1}{(k+1)^2 \omega _{2k}}<\infty . \end{aligned}$$
We recall that given a Banach space (or a complete metric space) X of analytic functions on \(\mathbb {D}\), a positive Borel measure \(\mu \) on \(\mathbb {D}\) is called a q-Carleson measure for X if the identity operator \(I_d:\, X\rightarrow L^q(\mu )\) is bounded. Carleson provided a geometric description of p-Carleson measures for Hardy spaces \(H^p\), \(0<p<\infty \), [6, Chapter 9]. These measures are called classical Carleson measures. The proof of Theorem 2 uses characterizations of Carleson measures for \(X_1\)-spaces, universal Cesàro basis of polynomials and some of the main ingredients of the proofs of Theorem 1 and [26, Theorem 2].
Concerning the classes of radial weights \(\widehat{\mathcal {D}}\) and \(M_{p,c}=\{ \omega : M_{p,c}(\omega )<\infty \}\), \(1\le p<\infty \), a standard weight, \(\omega (z)=(1-|z|)^{\beta }\), \(\beta >-1\), satisfies the condition \(M_{p,c}(\omega )<\infty \) if and only if \(\beta >\frac{1}{p}-1\), so \(H_\omega : H^p\rightarrow H^p\) is bounded if and only if \(\beta >\frac{1}{p}-1\). Moreover, a calculation shows that the exponential type weight \(\omega (r)=\exp \left( -\frac{1}{1-r}\right) \in M_{p,c}\) for any \(p\in [1,\infty )\), but \(\omega \notin \widehat{\mathcal {D}}\), see [28, Example 3.2] for further details. So, \(\widehat{\mathcal {D}}\) and \(M_{p,c}\) are not included in each other.
The study of the radial weights \(\omega \) such that \(H_\omega : H^p\rightarrow H^p\) is bounded, has been previously considered in [26]. Indeed, Theorem 2 improves [26, Theorem 2], by removing the initial hypothesis \(\omega \in \widehat{\mathcal {D}}\). On the other hand, [26, Theorem 3] describes the weights \(\omega \in \widehat{\mathcal {D}}\) such that \(H_\omega : L^p_{[0,1)}\rightarrow H^p\) is bounded, and consequently gives a sufficient condition for the boundedness of \(H_\omega : H^p\rightarrow H^p\), \(1<p<\infty \). The following improvement of [26, Theorem 3] is a byproduct of Theorem 1.
Corollary 3
Let \(\omega \) be a radial weight and \(1<p<\infty \). Let \(Y_p\in \{H(\infty ,p),H^p, D^p_{p-1}, HL(p)\}\) and \(T\in \{ H_{\omega }, \widetilde{H_{\omega }}\}\). Then the following statements are equivalent:
-
(i)
\(T:L^p_{[0,1)} \rightarrow Y_p\) is bounded;
-
(ii)
\(\omega \in \mathcal {D}\) and satisfies the condition
$$\begin{aligned} m_p(\omega )=\sup \limits _{0<r<1}\left( 1+\int _0^r \frac{1}{\widehat{\omega }(t)^p} dt\right) ^{\frac{1}{p}} \left( \int _r^1 \omega (t)^{p'}\,dt\right) ^{\frac{1}{p'}} <\infty ; \end{aligned}$$ -
(iii)
\(\omega \in \widehat{\mathcal {D}}\) and satisfies the condition \( m_p(\omega ) <\infty \).
In relation to an analogous result to Corollary 3 for \(p=1\), Theorem 26 below shows that the radial weights such that \(T:L^1_{[0,1)} \rightarrow Y_1\) is bounded, where \(Y_1\in \{H(\infty ,1),H^1, D^1_{0}, HL(1)\}\) and \(T\in \{ H_{\omega }, \widetilde{H_{\omega }}\}\), are the weights \(\omega \in \mathcal {D}\) such that \(m_1(\omega )= \mathop {\mathrm {ess\,sup}}\limits _{t \in [0,1)} \omega (t)\left( 1+\int _0^t \frac{ds}{\widehat{\omega }(s)}\right) <\infty .\)
In view of the above findings, we compare the conditions \(M_{p,c}(\omega )<\infty \), \(M_{p,d}(\omega )<\infty \) and \(m_p(\omega )<\infty \) in order to put the boundedness of \(T:\,X_p \rightarrow Y_p\) alongside the boundedness of \(T:L^p_{[0,1)} \rightarrow Y_p\), where \(X_p, Y_p\in \{H(\infty ,p),H^p, D^p_{p-1}, HL(p)\}\) and \(T\in \{ H_{\omega }, \widetilde{H_{\omega }}\}\) for \(1\le p<\infty \). Bearing in mind (1.5), it is clear that the condition \(m_p(\omega )<\infty \) implies that \(M_{p,c}(\omega )<\infty \), for any weight \(\omega \in \widehat{\mathcal {D}}\). Moreover, observe that \(M_{p,c}(\omega )<\infty \) if and only if
and \(\sup \nolimits _{a \in [0,1)} \frac{\widehat{\omega }(a)}{1-a} \left( 1+\int _0^a \frac{ds}{\widehat{\omega }(s)}\right) <\infty \) if and only if \(M_{1,d}(\omega )<\infty \). So, the conditions \(M_{p,c}(\omega )<\infty \) and \(m_p(\omega )<\infty \), are equivalent for any \(1\le p < \infty \) whenever \(\omega \) satisfies the pointwise inequality
and \(\omega \in \widehat{\mathcal {D}}\). The condition (1.7) implies restrictions on the decay and on the regularity of the weight, in fact if \(\omega \) fulfills (1.7) then \(\omega \) cannot decrease rapidly and cannot oscillate strongly. For instance, the exponential type weight \(\omega (r)=\exp \left( -\frac{1}{1-r}\right) \), which is a prototype of rapidly decreasing weight (see [18]), has the property
so it does not satisfy (1.7). On the other hand, any regular or rapidly increasing weight satisfies (1.7). Regular and rapidly increasing weights are large subclasses of \(\widehat{\mathcal {D}}\), see [25, Section 1.2] for the definitions and examples of these classes of radial weights. However, we construct in Corollaries 19 and 28 weights \(\omega \in \mathcal {D}\) with a strong oscillatory behaviour so that \(M_{p,c}(\omega )<\infty \) and \(m_p(\omega )=\infty \), and consequently they do not satisfy (1.7).
With the aim of discussing some results concerning the case \(p=\infty \), we recall that the space \(\mathord \textrm{BMOA}\) consists of those functions in the Hardy space \(H^1\) that have bounded mean oscillation on the boundary of \(\mathbb {D}\) [10], and the Bloch space \(\mathcal {B}\) is the space of all analytic functions on \(\mathbb {D}\) such that
We also consider the space \(HL(\infty )\) of the \(f(z)=\sum _{n=0}^\infty \widehat{f}(n) z^n \in \mathcal {H}(\mathbb {D})\) such that
The following chain of inclusions hold [10]
If \(\omega \) is a radial weight
so \(H_{\omega }\) is not bounded on \(H^\infty \). As for the classical Hilbert matrix H, it is bounded from \(H^\infty \) to BMOA [13, Theorem 1.2]. So, it is natural wondering about the radial weights such that \(H_{\omega }:H^{\infty }\rightarrow \mathord \textrm{BMOA}\) is bounded. The next result answers this question.
Theorem 4
Let \(\omega \) be a radial weight and let \(T\in \{H_{\omega },\widetilde{H_{\omega }}\}\). Then, the following statements are equivalent:
-
(i)
\(T: H^{\infty }\rightarrow HL(\infty )\) is bounded;
-
(ii)
\(T:H^{\infty }\rightarrow \mathord \textrm{BMOA}\) is bounded;
-
(iii)
\(T: H^{\infty }\rightarrow \mathcal {B}\) is bounded;
-
(iv)
\(\omega \in \widehat{\mathcal {D}}\).
The equivalence (iii)\(\Leftrightarrow \)(iv) was proved in [26, Theorem 1], so our contribution in Theorem 4 consists on proving the rest of equivalences.
Bearing in mind Theorems 1, 2 and 4, we deduce that \(T\in \{H_{\omega },\widetilde{H_{\omega }}\}\) is bounded from \(H^\infty \) to \(HL(\infty )\) if \(T: X_p\rightarrow Y_p\) is bounded, where \(X_p,Y_p\in \{H(\infty ,p),H^p, D^p_{p-1}, HL(p)\}\), \(1\le p<\infty \). We prove that this is a general phenomenon for Hilbert-type operators and parameters \(1\le q<p\).
Theorem 5
Let \(\omega \) be radial weight, \(T\in \{H_\omega ,\widetilde{H_\omega }\}\) and \(1\le q<p< \infty \). Further, let \(X_q,Y_q\in \{H^q, D^q_{q-1}, HL(q), H(\infty ,q)\}\) and \(X_p,Y_p\in \{H^p, D^p_{p-1}, HL(p), H(\infty ,p)\}\). If \(T: X_q\rightarrow Y_q\) is bounded, then \(T: X_p\rightarrow Y_p\) is bounded.
We also prove that that there does not exist radial weights \(\omega \) such that \(H_\omega : X_p\rightarrow Y_p\) is compact, where \(X_p,Y_p\in \{H^p, D^p_{p-1}, HL(p), H(\infty ,p)\}\) and \(1\le p< \infty \), neither radial weights such that \(H_{\omega }:H^{\infty } \rightarrow \mathcal {B}\) is compact, see Theorems 22, 31, 34 below.
The letter \(C=C(\cdot )\) will denote an absolute constant whose value depends on the parameters indicated in the parenthesis, and may change from one occurrence to another. We will use the notation \(a\lesssim b\) if there exists a constant \(C=C(\cdot )>0\) such that \(a\le Cb\), and \(a\gtrsim b\) is understood in an analogous manner. In particular, if \(a\lesssim b\) and \(a\gtrsim b\), then we write \(a\asymp b\) and say that a and b are comparable. We remark that if a or b are quantities which depends on a radial weight \(\omega \), the constant C such that \(a\lesssim b\) or \(a\gtrsim b\) may depend on \(\omega \) but it does not depend on a neither on b.
The rest of the paper is organized as follows. Section 2 is devoted to prove some auxiliary results. We prove Theorem 1 and Corollary 3 in Sect. 3, and Theorem 2 is proved in Sect. 4. Section 5 contains a proof of Theorem 4 and Theorem 5 is proved in Sect. 6 together with some reformulations of the condition \(M_{p,c}(\omega )<\infty \).
2 Preliminary results
In this section, we will prove some convenient preliminary results which will be repeatedly used throughout the rest of the paper. The first auxiliary lemma contains several characterizations of upper doubling radial weights. For a proof, see [21, Lemma 2.1].
Lemma 6
Let \(\omega \) be a radial weight on \(\mathbb {D}\). Then, the following statements are equivalent:
-
(i)
\(\omega \in \widehat{\mathcal {D}}\);
-
(ii)
There exist \(C=C(\omega )\ge 1\) and \(\beta _0=\beta _0(\omega )>0\) such that
$$\begin{aligned} \widehat{\omega }(r)\le C \left( \frac{1-r}{1-t}\right) ^{\beta }\widehat{\omega }(t), \quad 0\le r\le t<1; \end{aligned}$$for all \(\beta \ge \beta _0\).
-
(iii)
$$\begin{aligned} \int _0^1 s^x \omega (s) ds\asymp \widehat{\omega }\left( 1-\frac{1}{x}\right) ,\quad x \in [1,\infty ); \end{aligned}$$
-
(iv)
There exists \(C=C(\omega )>0\) and \(\beta =\beta (\omega )>0\) such that
$$\begin{aligned} \omega _x\le C \left( \frac{y}{x}\right) ^{\beta }\omega _y,\quad 0<x\le y<\infty ; \end{aligned}$$ -
(v)
\( \widehat{\mathcal {D}}(\omega )=\sup \nolimits _{n\in \mathbb {N}}\frac{\omega _n}{\omega _{2n}}<\infty .\)
We will also use the following characterizations of the class \(\check{\mathcal {D}}\), see [23, (2.27)].
Lemma 7
Let \(\omega \) be a radial weight. The following statements are equivalent:
-
(i)
\(\omega \in \check{\mathcal {D}}\);
-
(ii)
There exist \(C=C(\omega )>0\) and \(\alpha _0=\alpha _0(\omega )>0\) such that
$$\begin{aligned} \widehat{\omega }(s)\le C \left( \frac{1-s}{1-t}\right) ^{\alpha }\widehat{\omega }(t), \quad 0\le t\le s<1 \end{aligned}$$for all \(0<\alpha \le \alpha _0\);
-
(iii)
There exist \(K=K(\omega )>1 \) and \(C=C(\omega )>0 \) such that
$$\begin{aligned} \int _r^{1-\frac{1-r}{K}}\omega (s)ds \ge C \widehat{\omega }(r), \quad 0\le r<1. \end{aligned}$$(2.1)
Embedding relations among spaces \(X_p,Y_p\in \{H^p, D^p_{p-1}, HL(p), H(\infty ,p)\}\) are quite useful in the study of operators acting on them. In particular, we recall that
for \(X_p\in \{ H^p,D^p_{p-1}\}\), see [27, p. 127] and [8, Lemma 4].
This inequality is no longer true for \(X_p=HL(p)\) if \(0<p<1\). In fact, take \(f(z)=\sum _{n=0}^\infty 2^{\frac{n}{p}}z^{2^n}\). A calculation shows that \(f\in HL(p)\), if \(0<p<1\). However, using [15, Theorem 1],
Our following result extends the inequality (2.2) to \(X_p=HL(p)\) and \(1\le p<\infty \).
Lemma 8
Let \(1\le p<\infty \). Then, there is \(C_p>0\) such that
where \(X_p\in \{ H^p,D^p_{p-1}, HL(p)\}\).
Proof
By (2.2) it is enough to prove the inequality for \(X_p=HL(p)\). By [15, Theorem 1] and Hölder’s inequality
This finishes the proof.\(\square \)
For \(0<p<\infty \) and \(\omega \) a radial weight, let \(L^p_{\omega , [0,1)}\) be the Lebesgue space of measurable functions such that
Next, we will prove that the sublinear operator \(\widetilde{H_{\omega }}\) does not distinguish the norm of the spaces \(H(\infty ,p), HL(p), D^p_{p-1}, H^p,\) when \(1<p<\infty \) and \(\omega \in \widehat{\mathcal {D}}\).
Lemma 9
Let \(\omega \in \widehat{\mathcal {D}}\), \(1<p<\infty \) and \(X_p, Y_p\in \{H(\infty ,p), HL(p), D^p_{p-1}, H^p\}\). Then,
Proof
Here and on the following, let us denote \(I(n)=\{k\in \mathbb {N}: 2^n\le k<2^{n+1}\}\), \(n\in \mathbb {N}\cup \{0\}\). By Lemma 6
The above equivalences and [15, Theorem 1], yield
This, together with [26, Lemma 8], finishes the proof.\(\square \)
3 Hilbert-type operators acting on \(X_p\)-spaces, \(1<p<\infty \)
3.1 Necessity part of Theorem 1
We begin this section with the construction of appropriate families of test functions to be used in the proof of Theorem 1. To do this, some notation and previous results are needed. Let \(g(z)=\sum \nolimits _{k=0}^{\infty } \widehat{g}(k) z^k \in \mathcal {H}(\mathbb {D})\), and denote \({\Delta }_n g(z) =\sum \nolimits _{k \in I(n)} \widehat{g}(k)z^k\). In the particular case \(g(z)=\frac{1}{1-z}\), we simply write \({\Delta }_n(z)={\Delta }_n(g)(z) =\sum \nolimits _{k \in I(n)}z^k\). We recall that
see [2, Lemma 2.7].
For any \(n_1, \,n_2 \in \mathbb {N}\cup \left\{ 0\right\} \), \(n_1<n_2\), write \( S_{n_1,n_2} g(z)=\sum \nolimits _{k=n_1}^{n_2-1} \widehat{g}(k) z^k\). The next known result can be proved mimicking the proof of [13, Lemma 3.4] (see also [24, Lemma E]), that is, by summing by parts and using the M. Riesz projection theorem.
Lemma 10
Let \(1<p<\infty \) and \(\lambda =\left\{ \lambda _k\right\} _{k=0}^{\infty }\) be a positive and monotone sequence. Let \( g(z)=\sum \nolimits _{k=0}^{\infty }b_k z^k\) and \((\lambda g)(z)=\sum \nolimits _{k=0}^{\infty }\lambda _k b_k z^k\).
-
(a)
If \(\left\{ \lambda _k\right\} _{k=0}^{\infty }\) is nondecreasing, there exists a constant \(C>0\) such that
$$\begin{aligned} C^{-1}\lambda _{n_1}\Vert S_{n_1,n_2}g\Vert _{H^p}\le \Vert S_{n_1,n_2}(\lambda g)\Vert _{H^p}\le C \lambda _{n_2}\Vert S_{n_1,n_2}g\Vert _{H^p}. \end{aligned}$$ -
(b)
If \(\left\{ \lambda _k\right\} _{k=0}^{\infty }\) is nonincreasing, there exists a constant \(C>0\) such that
$$\begin{aligned} C^{-1}\lambda _{n_2}\Vert S_{n_1,n_2}g\Vert _{H^p}\le \Vert S_{n_1,n_2}(\lambda g)\Vert _{H^p}\le C \lambda _{n_1}\Vert S_{n_1,n_2}g\Vert _{H^p}.\end{aligned}$$
Lemma 11
Let \(\omega \in \widehat{\mathcal {D}}\), \(1<p<\infty \), \(\alpha ,\beta \in \mathbb {R}\) and \(M, N \in \mathbb {N}\cup \{0\}\) such that \(0\le N<4N+1\le M\). Let us consider the function
Then,
where the constants involved do not depend on M or N. In particular, if \(\alpha =0\) then (3.2) holds for any radial weight.
Proof
Firstly, let us show that for all \(N, M \in \mathbb {N}\), \(M>N\),
[16, Theorem 2.1(b)] (see also [20, 7.5.8]), Lemma 10, (2.3) and (3.1) implies
A similar calculation shows that
Next, if \(N> 2\), there is \(N^\star , M^\star \in \mathbb {N}\) such that \(2^{N^\star }\le N-1<2^{N^\star +1}\) and \(2^{M^\star }\le M-1<2^{M^\star +1}\), so \(N^\star +1<M^\star \). Then, by [16, Theorem 2.1(b)] and the boundedness of the Riesz projection, (3.3) and (3.4)
On the other hand,
Then, bearing in mind (1.3) and (1.4), we obtain \(\Vert f_{N, M}\Vert _{HL(p)}\asymp \Vert f_{N, M}\Vert _{H^p}\asymp \Vert f_{ N, M}\Vert _{D^p_{p-1}}\) for each \(N> 2\).
If \(N\in \{0,1,2\}\), the previous argument together with minor modifications implies (3.2). This finishes the proof. \(\square \)
Now we are ready to prove the necessity part of Theorem 1.
Proposition 12
Let \(\omega \) be a radial weight and \(1<p<\infty \). If \(X_p, Y_p\in \{H(\infty ,p),H^p, D^p_{p-1}, HL(p)\}\), \(T\in \{H_{\omega }, \widetilde{H_{\omega }}\}\), and \(T: X_p\rightarrow Y_p\) is a bounded operator. Then, \(\omega \in \mathcal {D}\) and
Proof
In order to obtain both conditions, \(\omega \in \mathcal {D}\) and \(M_p(\omega )<\infty \), we are going to work with families of test functions constructed in Lemma 11. Since they have non-negative Maclaurin coefficients, it is enough to prove the result for \(T= H_{\omega }\). Take \(f \in \mathcal {H}(\mathbb {D})\) such that \(\widehat{f}(n) \ge 0\) for all \(n \in \mathbb {N}\).
First Step. We will prove that \(\omega \in \widehat{\mathcal {D}}\). By Lemma 8, it is enough to deal with the case \(Y_p=H(\infty ,p)\).
Observe that \(M_\infty (r,H_\omega (f))=\sum _{n=0}^\infty \frac{1}{2(n+1)\omega _{2n+1}}\left( \sum _{k=0}^\infty \widehat{f}(k)\omega _{n+k}\right) r^n\). Now, consider the test functions \(f_N(z)=\sum \nolimits _{n=0}^N \frac{1}{(n+1)^{1-\frac{1}{p-1}}} z^n, N\in \mathbb {N}\). Given that \(\sum \nolimits _{k=0}^N \frac{1}{(k+1)^{1-\frac{1}{p-1}}}\asymp (N+1)^{\frac{1}{p-1}}\),
So,
Consequently,
Therefore, there is \(C=C(\omega ,p)\) such that \(\omega _{8N}\le C \omega _{12N},\quad N \in \mathbb {N}\). From now on, for each \(x\in \mathbb {R}\), \(\lfloor x\rfloor \) denotes the biggest integer \(\le x\). For any \(x\ge 120\), take \(N\in \mathbb {N}\) such that \(8N\le x<8N+8\), and then
So, \(\omega \in \widehat{\mathcal {D}}\) by Lemma 6.
Second Step. We will prove that \(M_p(\omega ) < \infty \).
Case \(\varvec{Y_p=HL (p)}\). Set an arbitrary \(N \in \mathbb {N}\). Then, bearing in mind that \(\{\omega _k\}_{k=0}^\infty \) is decreasing,
Take \(M, N \in \mathbb {N}\), \(M>4N+1\), and consider the family of test polynomials
where the constants do not depend on M or N.
So, testing this family of functions in (3.6), there exists \(C=C(p, \omega )>0\) such that
By letting \(M\rightarrow \infty \), and taking the supremum in \(N\in \mathbb {N}\), (3.5) holds.
Case \(\varvec{Y_p\in \{H(\infty ,p),H^p,D^p_{p-1}\}}\). Let \(f_{ N,M}\) be the functions defined in (3.7), then \(H_{\omega }(f_{N,M})=\widetilde{H_{\omega }}(f_{N,M})\). This together with the fact that \(\omega \in \widehat{\mathcal {D}}\) and Lemma 9, yields
where the constants in the inequalities do not depend on M or N.
Therefore, using Lemmas 8, 9 and 11, there exists \(C=C(p, \omega )>0\) such that
So, arguing as in the case \(Y_p=HL(p)\), we obtain \(M_p(\omega )<\infty .\)
Third Step. We will prove that the condition \(M_p(\omega )<\infty \) implies that \(\omega \in \mathcal {M}\). Indeed, set \(K,M \in \mathbb {N}\), \(K,M >1\) and \(N \in \mathbb {N}\). By (3.5),
So, there is \(C=C(p)>0\) such that
Now, fix \(K>1\) and take \(M\in \mathbb {N}\) large enough such that
Let \(x\ge 1\) and take \(N \in \mathbb {N}\) such that \(2N-2\le x <2N\). Then, by (3.8)
so \(\omega \in \mathcal {M}\). Since \(\omega \in \widehat{\mathcal {D}}\), [23, Theorem 3] yields \(\omega \in \mathcal {D}\). The proof is finished. \(\square \)
3.2 Sufficiency part of Theorem 1
For the purpose of proving Theorem 1 we need some additional preparations. In particular, we aim for reformulating the necessary discrete condition on the moments of the radial weight \(\omega \), \(M_p(\omega )<\infty \), as a continuous inequality in terms of \(\widehat{\omega }(r)\). Observe that a radial weight \(\omega \) satisfies the condition
if and only if \(M_{p,c}(\omega )<\infty .\) This fact will be used repeatedly throughout the paper.
Lemma 13
Let \(1<p<\infty \) and \(\omega \in \widehat{\mathcal {D}}\). Set
Then,
and
Proof
Let \(0<r<1\) and set \(N \in \mathbb {N}\) such that \(1-\frac{1}{N}\le r < 1-\frac{1}{N+1}\). Then, by using Lemma 6,
In addition, by Lemma 6 again,
Therefore, \(K_{p,c}(\omega )\lesssim M_p(\omega )\).
Conversely, in order to obtain the reverse inequality, a similar argument to (3.9) yields
Now, on the one hand, if \(r\le \frac{1}{2}\) then \(\sum \nolimits _{k=0}^N \frac{1}{(k+1)^2 \omega _{2k+1}^p} \lesssim 1\le 1+ \int _0^{r} \frac{1}{\widehat{\omega }(s)^p} ds.\) On the other hand, if \(\frac{1}{2}\le r <1\),
So, Lemma 6 yields
Next,
and consequently, \( M_p(\omega )\lesssim K_{p,c}(\omega )\). Finally, (3.9) and (3.10) imply
This finishes the proof. \(\square \)
We will also need the following description of the class \(\mathcal {D}\).
Lemma 14
Let \(\omega \) be a radial weight. Then the following conditions are equivalent:
-
(i)
\(\omega \in \mathcal {D}\);
-
(ii)
The function defined as \( \widetilde{\omega }(r)=\frac{\widehat{\omega }(r)}{1-r}\), \(0\le r<1\), is a radial weight and satisfies
$$\begin{aligned} \widehat{\omega }(r)\asymp \widehat{\widetilde{\omega }}(r),\quad 0\le r<1. \end{aligned}$$
Proof
(i)\(\Rightarrow \)(ii). By Lemma 7, there is \(\alpha >0\) such that
which, in particular, implies that \(\widetilde{\omega }\) is a radial weight. On the other hand, by Lemma 6, there is \(\beta >0\) such that
Reciprocally, if (ii) holds, there are \(C_1,C_2>0\) such that
So, for any \(K>1\),
Therefore, taking K such that \(\frac{\log K}{C_2}>1\), \(\omega \in \check{\mathcal {D}}\).
Moreover, for any \(K>1\)
If \(\frac{\log K}{C_1}<1\), then
So, \(\omega \in \widehat{\mathcal {D}}\). This finishes the proof. \(\square \)
The previous lemma may be used to prove that a differentiable non-decreasing function \(h:[0,1)\rightarrow [0,\infty )\) belongs to \(L^p_{\omega , [0,1)}\) if and only if it belongs to \(L^p_{\widetilde{\omega }, [0,1)}\). This result is essential for our purposes. In particular, bearing in mind Lemma 14 and two integration by parts,
for any differentiable non-decreasing function \(h:[0,1)\rightarrow [0,\infty )\).
Bearing in mind Lemma 13, our next result ensures that the Hilbert-type operators \(H_\omega \) and \(\widetilde{H_\omega }\) are well defined on \(X_p\in \{H(\infty ,p),H^p, D^p_{p-1}, HL(p)\}\), \(1<p<\infty \) when \(\omega \in \mathcal {D}\) and \(M_p(\omega )<\infty \).
Lemma 15
Let \(\omega \in \mathcal {D}\) and \(1<p<\infty \) such that \(\int _0^1 \left( \frac{\widehat{\omega }(t)}{1-t}\right) ^{p'}\,dt<\infty \).
Then
In particular, \(T(f)\in \mathcal {H}(\mathbb {D})\) for any \(f\in X_p\), where \(X_p\in \{H(\infty ,p),H^p, D^p_{p-1}, HL(p)\}\) and \(T\in \{H_\omega , \widetilde{H_\omega }\}\).
Proof
By (3.11)
Then, by Hölder’s inequality
Joining the above chain of inequalities with Lemma 8, the proof is finished. \(\square \)
Next, for \(p,q>0\) and \(\alpha >-1\), let \(H^1(p,q, \alpha )\) denote the space of \(f \in \mathcal {H}(\mathbb {D})\) such that
It is worth mentioning that \(H^1\left( p,p, p-1\right) =D^p_{p-1}\).
The following inequality will be used in the proof of Theorem 1. It was proved in [16, Corollary 3.1].
Lemma 16
Let \(1<q<p<\infty \). Then,
Now, we are ready to prove the main result of this section.
Proof of Theorem 1
The implication (i)\(\Rightarrow \)(ii) was proved in Proposition 12. The implication (ii)\(\Rightarrow \)(iii) is clear, and (iii)\(\Rightarrow \)(ii) follows from the third step in the proof of Proposition 12. On the other hand, bearing in mind that \(M_{p,c}(\omega )<\infty \) if and only if \(K_{p,c}(\omega )<\infty \), (iii)\(\Leftrightarrow \)(iv) follows from Lemma 13. Then, it is enough to prove (ii)\(\Rightarrow \)(i).
(ii)\(\mathbf {\Rightarrow }\)(i).
First Step. We will prove the inequality
By Lemmas 8 and 16, it is enough to prove
Let \(f \in X_p\). Then, Lemmas 13 and 15 ensure that \(H_\omega (f)\in \mathcal {H}(\mathbb {D})\). By [16, Theorem 2.1]
Due to
and using the proof of [8, Lemma 7], Lemma 10 and (3.1),
Hence, by using Lemma 6,
In addition, by Lemmas 8 and 15
Therefore, by putting together (3.15), (3.16) and (3.17)
The above inequality, together with Lemma 9, yields (3.14).
Second Step. We will prove the inequality
We denote by
By [26, Lemma B]
which together with Minkowski’s inequality yields
Now, by (3.11)
Next, by Lemma 13, \(M_{p,c}(\omega )<\infty \) holds, so [17, Theorem 2] yields
On the other hand, by [17, Theorem 1],
where in the last inequality we have used that \(\sup \nolimits _{0<r<1} \left( \int _r^1(1-t)^{p-1}\right) ^{\frac{1}{p}} \left( \int _0^r(1-t)^{-1-p'}\right) ^{\frac{1}{p'}}<\infty \). So, joining Lemma 15, (3.20), (3.21) and (3.22), we get (3.18).
Third Step. Since \(\omega \in \widehat{\mathcal {D}}\), by Lemma 9
for \(Y_p\in \{H(\infty ,p),H^p, D^p_{p-1}, HL(p)\}\). This, together with (3.13), (3.18) and Lemma 8 implies
This finishes the proof. \(\square \)
3.3 \(\varvec{H_\omega : X_p\rightarrow Y_p}\) versus \(\varvec{H_\omega : L^p_{[0,1)}\rightarrow Y_p,\, 1<p<\infty }\)
An additional byproduct of Theorem 1 is the following improvement of [26, Theorem 3].
Proof of Corollary 3
(i)\(\Rightarrow \)(ii). By Lemma 8, \(T:Y_p \rightarrow Y_p\) is bounded, and so by Theorem 1, \(\omega \in \mathcal {D}\). Next, by the proofs of [26, Theorems 3 and 4] we obtain \(m_p(\omega )<\infty \).
(ii)\(\Rightarrow \)(iii) is clear. Finally, (iii)\(\Rightarrow \)(i) follows from Lemma 8 and [26, Theorem 3].
Putting together Lemma 8, Theorem 1 and Corollary 3, we deduce the following result.
Corollary 17
Let \(\omega \) be a radial weight and \(1<p<\infty \). Let \(X_p,Y_p\in \{H(\infty ,p),H^p, D^p_{p-1}, HL(p)\}\) and let \(T \in \{H_{\omega }, \widetilde{H_{\omega }}\}\). If there exists \(C>0\) such that
Then, the following statements are equivalent:
-
(i)
\(T:L^p_{[0,1)}\rightarrow Y_p\) is bounded;
-
(ii)
\(T: X_p \rightarrow Y_p\) is bounded;
-
(iii)
\(\omega \in \mathcal {D}\) and \(M_{p,c}(\omega )<\infty \);
-
(iv)
\(\omega \in \widehat{\mathcal {D}}\) and \(M_{p,c}(\omega )<\infty \);
-
(v)
\(\omega \in \mathcal {D}\) and \(m_p(\omega )<\infty \);
-
(vi)
\(\omega \in \widehat{\mathcal {D}}\) and \(m_p(\omega )<\infty \).
Proof
The implication (i)\(\Rightarrow \)(ii) follows from Lemma 8, and (ii)\(\Leftrightarrow \)(iii)\(\Leftrightarrow \)(iv) follows from Theorem 1. Next, (iii)\(\Rightarrow \)(v) is a byproduct of the hypothesis (3.23). The equivalence (v)\(\Leftrightarrow \)(vi) and the implication (vi)\(\Rightarrow \)(i) have been proved in Corollary 3. This finishes the proof. \(\square \)
Next, we will prove that there are weights \(\omega \in \mathcal {D}\), such that \(M_{p,c}(\omega )<\infty \) and \(m_p(\omega )=\infty \), so in particular they do not satisfy (3.23). Consequently, the boundedness of the operator \(H_\omega : L^p_{[0,1)}\rightarrow Y_p\) is not equivalent to the boundedness of the the operator \(H_\omega : X_p\rightarrow Y_p\), where \(X_p,Y_p\in \{H(\infty ,p),H^p, D^p_{p-1}, HL(p)\}\). With this aim we prove the next result, which shows that despite its innocent looking condition, the class \(\mathcal {D}\) has in a sense a complex nature.
Lemma 18
Let \(1<p<\infty \) and \(\nu \in \mathcal {D}\). Then, there exists \(\omega \in \mathcal {D}\) such that
\(\omega \in L^{p'}_{[0,r_0]}\) for any \(r_0\in (0,1)\) and \(\omega \notin L^{p'}_{[0,1)}\).
Proof
By Lemma 14, \(\widetilde{\nu }\in \mathcal {D}\). So, we can choose \(K>1\) so that \(\widetilde{\nu }\) satisfies (2.1). Next, consider the sequences \( r_n= 1-\frac{1}{K^n},\,t_n=r_n+a_n\), with
Let
Observe that the sequence \(\{h_n\}_{n=0}^\infty \) is well-defined because
Moreover, \(\omega \) is non-negative and
where in the last equivalence we have used Lemma 14.
Next, take \(t\in [0,1)\) and \(N\in \mathbb {N}\cup \{ 0\}\) such that \(r_N\le t<r_{N+1}\). By Lemmas 14 and 6,
so \(\widehat{\omega }(t)\asymp \widehat{\nu }(t)\) and hence \(\omega \in \mathcal {D}\).
It is clear that \(\omega \in L^{p'}_{[0,r_0]}\) for any \(r_0\in (0,1)\), so it only remains to prove that \(\omega \notin L^{p^\prime }_{[0,1)}\). Bearing mind (2.1), we get that
This, together with Lemma 6 and Hölder’s inequality, implies
\(\square \)
Corollary 19
Let \(1<p<\infty \) and \(X_p,Y_p\in \{H(\infty ,p),H^p, D^p_{p-1}, HL(p)\}\). For each radial weight \(\nu \) such that \( Q: L^p_{[0,1)}\rightarrow Y_p\) is bounded, where \(Q\in \{ H_\nu ,\widetilde{H_\nu }\}\), there is a radial weight \(\omega \) such that
\(\omega \in L^{p'}_{[0,r_0]}\) for any \(r_0\in (0,1)\), \(T: X_p\rightarrow Y_p\) is bounded and \(T: L^p_{[0,1)}\rightarrow Y_p\) is not bounded. Here \(T\in \{ H_\omega ,\widetilde{H_\omega }\}\).
Proof
Since \( Q: L^p_{[0,1)}\rightarrow Y_p\) is bounded, by Theorem 1, \(\nu \in \mathcal {D}\) and \(M_{p,c}(\nu )<\infty \). Now, by Lemma 18 there is a radial weight \(\omega \) such that \(\widehat{\omega }(t)\asymp \widehat{\nu }(t)\), \(\omega \in L^{p'}_{[0,r_0]}\) for any \(r_0\in (0,1)\) and \(\omega \not \in L^{p'}_{[0,1)}\). So, \(m_p(\omega )=\infty \) and by Corollary 3, \(T: L^p_{[0,1)}\rightarrow Y_p\) is not bounded. Moreover, \(\omega \in \mathcal {D}\) and \(M_{p,c}(\omega )<\infty \) because \(\nu \) satisfies both properties, so Theorem 1 yields \(T: X_p\rightarrow Y_p\) is bounded. \(\square \)
3.4 Compactness of Hilbert-type operators on \(X_p\)-spaces. Case \(\varvec{1< p<\infty }\)
For X, Y two Banach spaces, a sublinear operator \(L: X\rightarrow Y\) is said to be compact provided L(A) has compact closure for any bounded set \(A\subset X\). Once it has been understood the radial weights \(\omega \) such that \(H_\omega : X_p\rightarrow Y_p\) is bounded, \(X_p,Y_p\in \{ H(\infty ,p),D^p_{p-1}, H^p, HL(p)\}\), \(1<p<\infty \), it is natural to consider the analogous problem, replacing boundedness by compactness. Theorem 22 in this section answers this question, but firstly we need some previous results.
Lemma 20
Let \(1< p< \infty \) and \(\omega \in \mathcal {D}\) such that \(\Vert \widetilde{\omega }\Vert _{L^{p'}_{[0,1)}}<\infty \). Let \(\{f_k\}_{k=0}^\infty \subset X_p \in \{ H(\infty ,p),D^p_{p-1}, H^p, HL(p)\}\) such that \(\sup \nolimits _{k\in \mathbb {N}} \Vert f_k\Vert _{X_p}<\infty \) and \(f_k\rightarrow 0\) uniformly on compact subsets of \(\mathbb {D}\). Then the following statements hold:
-
(i)
\(\int _0^1 |f_k(t)|\omega (t)dt\rightarrow 0\) when \(k\rightarrow \infty \).
-
(ii)
If \(T\in \{H_\omega , \widetilde{H_\omega }\}\), then \(T(f_k)\rightarrow 0\) uniformly on compact subsets of \(\mathbb {D}\).
Proof
(i). Let \(\varepsilon >0\). By hypothesis \(\int _0^1 \widetilde{\omega }(t)^{p'} dt < \infty \), so there exists \(0<\rho _0<1\) such that \(\int _{\rho _0}^1 \widetilde{\omega }(t)^{p'} dt < \varepsilon \). Moreover, there exists \(k_0\) such that for every \(k\ge k_0\) and \(z\in M=\overline{D(0,\rho _0)}\), \(|f_k(z)|<\varepsilon \). Then, by Lemma 14, (3.11), and Hölder inequality
where in the last step we have used Lemma 8.
(ii). Let be \(M\subset \mathbb {D}\) a compact set and \(K_t^\omega (z)=\frac{1}{z}\int _0^z B_t^\omega (u)\,du\). If \(z\in M\)
Since, \(M\subset \overline{D(0,\rho _0)}\), for some \(\rho _0\in (0,1)\), then
so, by (i), \(T(f_k)\rightarrow 0\) uniformly on M. This finishes the proof. \(\square \)
Theorem 21
Let \(\omega \) be a radial weight, \(1<p<\infty \), \(X_p, Y_p \in \{ H(\infty ,p),D^p_{p-1}, H^p, HL(p)\}\) and let \(T\in \{H_\omega , \widetilde{H_\omega }\}\). Then, the following assertions are equivalent:
-
(i)
\(T: X_p \rightarrow Y_p\) is compact;
-
(ii)
For every sequence \(\{f_k\}_{k=0}^\infty \subset X_p \) such that \(\sup \nolimits _{k\in \mathbb {N}} \Vert f_k\Vert _{X_p}<\infty \) and \(f_k\rightarrow 0\) uniformly on compact subsets of \(\mathbb {D}\), \(\lim \nolimits _{k\rightarrow \infty } \Vert T(f_k)\Vert _{Y_p} =0\).
Proof
(i)\(\Rightarrow \)(ii). Let \(\{f_n\}_{n=0}^\infty \subset X_p\) such that \(\sup \nolimits _{n\in \mathbb {N}} \Vert f_n\Vert _{X_p}<\infty \) and \(f_n\rightarrow 0\) uniformly on compact subsets of \(\mathbb {D}\). Assume there exist \(\varepsilon >0\) and a subsequence \(\{n_k\}_k \subset \mathbb {N}\) such that
Since T is compact, there exists a subsequence \(\{n_{k_j}\}_j\subset \mathbb {N}\) and \(g \in Y_p\) such that \(\lim \nolimits _{j\rightarrow \infty } \Vert T(f_{n_{k_j}})-g\Vert _{Y_p}=0\). Moreover, Theorem 1 ensures that \(\omega \in \mathcal {D}\) and \(M_{p,c}(\omega )<\infty \), so \(\Vert \widetilde{\omega }\Vert _{L^{p'}_{[0,1)}}<\infty \). Therefore Lemma 20, implies that \(T(f_{n_{k_j}})\rightarrow 0\) uniformly on compact subsets of \(\mathbb {D}\), so \(\lim \nolimits _{j\rightarrow \infty } \Vert T(f_{n_{k_j}})\Vert _{Y_p}=0\) which yields a contradiction with (3.24).
(ii)\(\Rightarrow \)(i). Let \(\{f_n\}\subset X_p\) such that \(\sup \nolimits _{n\in \mathbb {N}}\Vert f_n\Vert _{X_p}<\infty .\) Then, \(\{f_n\}\) is uniformly bounded on compact subsets of \(\mathbb {D}\). Then, by Montel’s Theorem there exists \(\{f_{n_k}\}_k\) and \(f\in \mathcal {H}(\mathbb {D})\) such that \(f_{n_k}\rightarrow f\) uniformly on compact subsets of \(\mathbb {D}\). Let \(g_{n_k}=f_{n_k}-f\), then \(g_{n_k}\rightarrow 0\) uniformly on compact subsets of \(\mathbb {D}\) and \(\sup \nolimits _{k\in \mathbb {N}}\Vert g_{n_k}\Vert _{X_p}<\infty \). Therefore, by hypothesis \(\lim \nolimits _{k\rightarrow \infty } \Vert T(g_{n_k})\Vert _{Y_p} =0\), that is, T is compact. \(\square \)
Theorem 22
Let \(\omega \) be a radial weight, \(1< p<\infty \), \(X_p,Y_p\in \{ H(\infty ,p),D^p_{p-1}, H^p, HL(p)\}\), and let \(T\in \{H_{\omega }, \widetilde{H_{\omega }}\}\). Then, \(T:X_p\rightarrow Y_p\) is not compact.
Proof
Assume that \(T:X_p\rightarrow Y_p\) is compact. For each \(0<a<1\), set
where \(\widehat{f_a}(n)=(1-a^2)^{1/p}\frac{\Gamma (n+2/p)}{n!\Gamma (2/p)} a^n \ge 0\). So, by Stirling’s formula
Consequently, \(\Vert f_a\Vert _{HL(p)}\asymp 1,\, a\in (0,1)\). Moreover, \(\Vert f_a\Vert _{H(\infty ,p)}\asymp \Vert f_a\Vert _{D^p_{p-1}}\asymp \Vert f_a\Vert _{H^p}=1.\) Furthermore, it is clear that \(f_a\rightarrow 0\), as \(a\rightarrow 1\) uniformly on compact subsets of \(\mathbb {D}\), and \(H_\omega (f_a)=\widetilde{H_\omega }(f_a)\). Since \(T:X_p\rightarrow Y_p\) is compact, \(\omega \in \widehat{\mathcal {D}}\) by Theorem 1. So, Lemma 9 implies that
Therefore, by using (3.25) we have
so using Theorem 21 we deduce that \(T: X_p\rightarrow Y_p\) is not a compact operator. \(\square \)
4 Hilbert type operators acting on \(X_1\)-spaces
The first result of this section gives the equivalence of conditions (iii)–(vi) of Theorem 2.
Lemma 23
Let \(\omega \in \widehat{\mathcal {D}}\). Then, the following conditions are equivalent:
-
(i)
\(K_{1,c}(\omega )=\sup \nolimits _{a \in [0,1)} \frac{1}{1-a}\int _a^1 \omega (t)\left( 1+\int _0^t \frac{ds}{\widehat{\omega }(s)}\right) \,dt<\infty ;\)
-
(ii)
\(K_{1,d}(\omega )= \sup \nolimits _{a \in [0,1)} \frac{\widehat{\omega }(a)}{1-a} \left( 1+\int _0^a \frac{ds}{\widehat{\omega }(s)}\right) <\infty ;\)
-
(iii)
\( M_1(\omega )=\sup \nolimits _{N \in \mathbb {N}} (N+1)\omega _{2N}\sum \nolimits _{k=0}^{N}\frac{1}{(k+1)^2 \omega _{2k}}<\infty \).
Moreover,
and \(\omega \in \check{\mathcal {D}}\) when \(\omega \) satisfies any of the three previous conditions.
Observe that for any radial weight, \(K_{1,c}(\omega )<\infty \) holds if and only if \(M_{1,c}(\omega )<\infty \), and analogously \(K_{1,d}(\omega )<\infty \) if and only if \(M_{1,d}(\omega )<\infty \). This fact will be used repeatedly throughout the paper.
Proof
On the one hand,
so (i)\(\Rightarrow \)(ii) and \(K_{1,d}(\omega )\lesssim K_{1,c}(\omega )\).
On the other hand assume that (ii) holds. Since \(\omega \in \widehat{\mathcal {D}}\), [22, Lemma 3(ii)] (for \(\nu (t)=1\)) yields
that is (i) holds and \(K_{1,c}(\omega )\lesssim K_{1,d}(\omega )\). Finally, by mimicking the proof of Lemma 13,
so (ii)\(\Leftrightarrow \)(iii) and (4.1) holds.
Next, for any \(K>1\) and \(r\in (0,1)\)
that is
so taking \(K>M_{1,d}(\omega )+1\), we get \(\omega \in \check{\mathcal {D}}\). This finishes the proof.\(\square \)
The following result will be used to prove the equivalence (ii)\(\Leftrightarrow \)(iii) of Theorem 2.
Proposition 24
Let \(\mu \) be a finite positive Borel measure on [0, 1) and \(X_1\in \{H(\infty ,1), HL(1)\}\). Then \(\mu \) is a 1-Carleson measure for \(X_1\) if and only if \(\mu \) is a classical Carleson measure. Moreover,
Proof
If \(\mu \) is a 1-Carleson measure for \(X_1\), then by (2.2) and (1.3), \(\mu \) is a 1-Carleson measure for \(H^1\). So, by [6, Theorem 9.3] and its proof, \(\mu \) is a classical Carleson measure and
Conversely, if \(\mu \) is a classical Carleson measure, two integration by parts yield
This inequality, together with Lemma 8, finishes the proof. \(\square \)
We introduce some more notation to prove Theorem 2. For any \(C^\infty \)-function \(\Phi :\mathbb {R}\rightarrow \mathbb {C}\) with compact support, define the polynomials
A particular case of the previous construction is useful for our purposes. Some properties of these polynomials have been gathered in the next lemma, see [12, Section 2] or [19, p. 143–144] for a proof.
Lemma 25
Let \(\Psi :\mathbb {R}\rightarrow \mathbb {R}\) be a \(C^\infty \)-function such that \(\Psi \equiv 1\) on \((-\infty ,1]\), \(\Psi \equiv 0\) on \([2,\infty )\) and \(\Psi \) is decreasing and positive on (1, 2). Set \(\psi (t)=\Psi \left( \frac{t}{2}\right) -\Psi (t)\) for all \(t\in \mathbb {R}\). Let \(V_{0}(z)=1+z\) and
Then,
and for each \(0<p<\infty \) there exists a constant \(C=C(p,\Psi )>0\) such that
In addition
Let us denote \(f_r(z)=f(rz)\), \(z\in \mathbb {D}\), \(r\in (0,1)\). Now we are ready to prove the main theorem of this section.
Proof of Theorem 2
First of all, recall that \(M_{1,c}(\omega )<\infty \) if and only if \(K_{1,c}(\omega )<\infty \) and analogously \(M_{1,d}(\omega )<\infty \) if and only if \(K_{1,d}(\omega )< \infty \), so that the equivalences (iii)\(\Leftrightarrow \)(iv)\(\Leftrightarrow \)(v)\(\Leftrightarrow \)(vi) follow from Lemma 23. The equivalence between (ii) and (iii) is a consequence of [6, Theorem 9.3] when \(X_1=H^1\), [29, Theorem 2.1] when \(X_1=D^1_0\) and Proposition 24 when \(X_1\in \{H(\infty ,1),HL(1)\}\).
(i)\(\Rightarrow \)(iii). In order to obtain both conditions, \(\omega \in \widehat{\mathcal {D}}\) and \(M_{1,c}(\omega )<\infty \), we are going to deal with functions \(f \in \mathcal {H}(\mathbb {D})\) such that \(\widehat{f}(n) \ge 0\) for all \(n \in \mathbb {N}\cup \{0\}\), so it is enough to prove the result for \(T= H_{\omega }\).
First Step. Let us prove \(\omega \in \widehat{\mathcal {D}}\). Bearing in mind Lemma 8 and (1.4)
for any \(f\in \mathcal {H}(\mathbb {D})\) such that \(\widehat{f}(n)\ge 0\), \(n\in \mathbb {N}\cup \{0\}\). Next, for each \(N \in \mathbb {N}\), consider the test functions \(f_{\alpha , N}(z)=\sum \nolimits _{k=0}^N (k+1)^{\alpha }z^k,\, \alpha >0\). Set \(M \in \mathbb {N}\) such that \(2^M <N \le 2^{M+1}\). Then, bearing in mind (4.2),
which together with [16, Lemma 3.1], [19, Lemma 5.4] and (4.4) gives
Testing the functions \(f_{\alpha , N}\) in (4.5), \(\sup \nolimits _{N\in \mathbb {N}}(N+1)\sum \nolimits _{n=0}^{\infty }\frac{\omega _{n+N}}{(n+1)^2 \omega _{2n+1}} <\infty . \) Therefore, there exists \(C=C(\omega )>0\)
So, arguing as in the first step proof of Proposition 12, \(\omega \in \widehat{\mathcal {D}}.\)
Second Step. We will prove \(M_{1,c}(\omega )<\infty \). Let us consider the test functions \(f_a(z)=\frac{1-a^2}{(1-az)^2}\), \(a\in (0,1)\). A calculation shows that \( \Vert f_a\Vert _{D^1_0}\asymp 1\), \(a\in (0,1)\). Then, by Lemma 8 and (1.3),
Consequently, using that \(\omega \in \widehat{\mathcal {D}}\) and mimicking the proof (4.2) of [26, Theorem 2], we get \(M_{1,c}(\omega )<\infty .\)
Now let us prove (iv) \(\Rightarrow \) (i). Firstly, observe that the condition \( M_{1,c}(\omega )<\infty \) implies \( K_{1,c}(\omega )<\infty \) so that \(\widetilde{\omega }(t)=\frac{\widehat{\omega }(t)}{1-t}\) is bounded on [0, 1). So, using Lemma 8 and (3.11),
that is \(H_\omega (f)\in \mathcal {H}(\mathbb {D})\) for any \(f\in X_1\). Secondly, by (1.4) and Lemma 8, it is enough to prove the inequality
to end the proof. Indeed,
Then by (3.19) and [26, Lemma B]
Bearing in mind that \( M_{1,c}(\omega )<\infty \) implies \( K_{1,c}(\omega )<\infty \) and applying Proposition 24, the measure \(\mu _\omega \) defined as \(d\mu _{\omega }(z)= \omega (z)\left( 1+\int _0^{|z|} \frac{ds}{\widehat{\omega }(s)}\right) \,\chi _{[0,1)}(z)\, dA(z)\) is a 1-Carleson measure for \(H(\infty , 1)\), so by Tonelli’s theorem,
This finishes the proof. \(\square \)
4.1 \(\varvec{H_\omega : X_1\rightarrow Y_1}\) versus \(\varvec{H_\omega : L^1_{[0,1)}\rightarrow Y_1}\)
Firstly, we will study the boundedness of \(T: L^1_{[0,1)}\rightarrow Y_1\), \(T\in \{H_\omega ,\widetilde{H_{\omega }}\}\), \(Y_1\in \{H(\infty ,1),H^1, D^1_{0}, HL(1)\}\).
Theorem 26
Let \(\omega \) be a radial weight, let \(Y_1\in \{H(\infty ,1),H^1, D^1_{0}, HL(1)\}\) and let \(T \in \{ H_{\omega }, \widetilde{H_{\omega }}\}\). Then the following statements are equivalent:
-
(i)
\(T: L^1_{[0,1)}\rightarrow Y_1\) is bounded;
-
(ii)
\(\omega \in \mathcal {D}\) and \(m_1(\omega )= \mathop {\mathrm {ess\,sup}}\limits \nolimits _{t \in [0,1)} \omega (t)\left( 1+\int _0^t \frac{ds}{\widehat{\omega }(s)}\right) <\infty .\)
-
(iii)
\(\omega \in \widehat{\mathcal {D}}\) and \(m_1(\omega )= \mathop {\mathrm {ess\,sup}}\limits \nolimits _{t \in [0,1)} \omega (t)\left( 1+\int _0^t \frac{ds}{\widehat{\omega }(s)}\right) <\infty .\)
Proof
(i)\(\Rightarrow \)(ii). If (i) holds, then \(\omega \in \mathcal {D}\) by Lemma 8 and Theorem 2. Next, using Lemma 8 again and making minor modifications in the proof of [26, Theorem 2] we get
for any \(f\ge 0\), \(f\in L^1_{[0,1)}\). So,
which implies that \(m_1(\omega )<\infty \).
(ii)\(\Rightarrow \)(iii) is clear.
(iii)\(\Rightarrow \)(i). If (iii) holds, then \(H_\omega (f)\in \mathcal {H}(\mathbb {D})\) for any \(f\in L^1_{[0,1)}\) and arguing as in (4.6)
This together with (1.3) and Lemma 8 gives that \(H_{\omega }: L^1_{[0,1)}\rightarrow Y_1\) is bounded. This finishes the proof. \(\square \)
Joining Theorems 2, 26 and Lemma 8 we deduce the following.
Corollary 27
Let \(\omega \) be a radial weight, \(X_1,Y_1\in \{H(\infty ,1),H^1, D^1_{0}, HL(1)\}\) and let \(T\in \{H_{\omega }, \widetilde{H_{\omega }}\}\). If \(\omega \) satisfies the condition (3.23), then the following statements are equivalent:
-
(i)
\(T:L^1_{[0,1)}\rightarrow Y_1\) is bounded;
-
(ii)
\(T: X_1 \rightarrow Y_1\) is bounded;
-
(iii)
\(\omega \in \widehat{\mathcal {D}}\) and \(M_{1,c}(\omega )<\infty \);
-
(iv)
\(\omega \in \mathcal {D}\) and \(M_{1,c}(\omega )<\infty \);
-
(v)
\(\omega \in \widehat{\mathcal {D}}\) and \(m_1(\omega )<\infty \);
-
(vi)
\(\omega \in \mathcal {D}\) and \(m_1(\omega )<\infty \).
Proof
(i)\(\Rightarrow \)(ii) follows from Lemma 8, and (ii)\(\Leftrightarrow \)(iii)\(\Leftrightarrow \)(iv) were proved in Theorem 2. Next, since \(M_{1,c}(\omega )<\infty \) implies \(K_{1,c}(\omega )<\infty \), (iii)\(\Rightarrow \)(v) is a byproduct of the hypothesis (3.23). Finally, the equivalences (v)\(\Leftrightarrow \)(vi)\(\Leftrightarrow \)(i) follow from Theorem 26. This finishes the proof. \(\square \)
A similar comparison between the conditions \(M_{1,c}(\omega )<\infty \) and \(m_1(\omega )<\infty \), to that made for the conditions \(M_{p,c}(\omega )<\infty \) and \(m_p(\omega )<\infty \), \(1<p<\infty \), can also be considered. The following result shows that they are not equivalent.
Corollary 28
Let \(X_1,Y_1\in \{H(\infty ,1),H^1, D^1_{0}, HL(1)\}\). For each radial weigth \(\nu \) such that \( Q: L^1_{[0,1)}\rightarrow Y_1\) is bounded, where \(Q\in \{ H_\nu ,\widetilde{H_\nu }\}\), there is a radial weight \(\omega \) such that
\(\omega \in L^\infty _{[0,r_0]}\) for any \(r_0\in (0,1)\), \(T: X_1\rightarrow Y_1\) is bounded and \(T: L^1_{[0,1)}\rightarrow Y_1\) is not bounded. Here \(T\in \{ H_\omega ,\widetilde{H_\omega }\}\).
Proof
Since \( Q: L^1_{[0,1)}\rightarrow Y_1\) is bounded, by Theorem 2, \(\nu \in \mathcal {D}\) and \(M_{1,c}(\nu )<\infty \). Now, by Lemma 18 and its proof, there is a radial weight \(\omega \) such that \(\widehat{\omega }(t)\asymp \widehat{\nu }(t)\), \(\omega \in L^\infty _{[0,r_0]}\) for any \(r_0\in (0,1)\) and \(\omega \not \in L^{\infty }_{[0,1)}\). So, \(m_1(\omega )=\infty \) and by Theorem 26, \(T: L^1_{[0,1)}\rightarrow Y_1\) is not bounded. Moreover, \(\omega \in \mathcal {D}\) and \(M_{1,c}(\omega )<\infty \) because \(\nu \) satisfies both properties, consequently \(T: X_1\rightarrow Y_1\) is bounded. \(\square \)
4.2 Compactness of Hilbert-type operators on \(X_1\)-spaces
Lemma 29
Let \(\omega \in \mathcal {D}\) such that \(M_{1,d}(\omega )<\infty \). Let \(\{f_k\}_{k=0}^\infty \subset X_1 \in \{ H(\infty ,1),D^1_{0}, H^1, HL(1)\}\) such that \(\sup \nolimits _{k\in \mathbb {N}} \Vert f_k\Vert _{X_1}<\infty \) and \(f_k\rightarrow 0\) uniformly on compact subsets of \(\mathbb {D}\). Then the following statements hold:
-
(i)
\(\int _0^1 |f_k(t)|\omega (t)dt\rightarrow 0\) when \(k\rightarrow \infty \).
-
(ii)
If \(T\in \{H_\omega , \widetilde{H_\omega }\}\), then \(T(f_k)\rightarrow 0\) uniformly on compact subsets of \(\mathbb {D}\).
Proof
Firstly, let us prove that
Since \(M_{1,d}(\omega )<\infty \), then
So \(\lim _{a\rightarrow 1^-} \int _0^{a} \frac{ds}{\widehat{\omega }(s)}=\infty \), and then using again the condition \(M_{1,d}(\omega )<\infty \), (4.7) holds.
From now on, the proof follow the lines of Lemma 20. Let \(\varepsilon >0\). By (4.7) there exists \(0<\rho _0<1\) such that \( \widetilde{\omega }(t) < \varepsilon \) for any \(t\in [\rho _0,1)\). Moreover, there exists \(k_0\) such that for every \(k\ge k_0\) and \(z\in M=\overline{D(0,\rho _0)}\), \(|f_k(z)|<\varepsilon \). Then, by Lemma 14 and (3.11)
where in the second to last step we have used Lemma 8.
The proof of (ii) is analogous to that of Lemma 20 so we omit its proof. \(\square \)
Using the previous lemma and Theorem 2 we obtain the following by mimicking the proof of Theorem 21.
Theorem 30
Let \(\omega \) be a radial weight and \(X_1, Y_1 \in \{ H(\infty ,1),D^1_{0}, H^1, HL(1)\}\) and let \(T\in \{H_\omega , \widetilde{H_\omega }\}\). Then, the following assertions are equivalent:
-
(i)
\(T: X_1 \rightarrow Y_1\) is compact;
-
(ii)
For every sequence \(\{f_k\}_{k=0}^\infty \subset X_1 \) such that \(\sup \nolimits _{k\in \mathbb {N}} \Vert f_k\Vert _{X_1}<\infty \) and \(f_k\rightarrow 0\) uniformly on compact subsets of \(\mathbb {D}\), \(\lim \nolimits _{k\rightarrow \infty } \Vert T(f_k)\Vert _{Y_1} =0\).
Theorem 31
Let \(\omega \) be a radial weight and \(X_1,Y_1\in \{ H(\infty ,1),D^1_{0}, H^1, HL(1)\}\), and let \(T\in \{H_{\omega }, \widetilde{H_{\omega }}\}\). Then, \(T:X_1\rightarrow Y_1\) is not compact.
Proof
The proof is analogous to that of Theorem 22, so we provide a sketch of the proof. Assume that \(T:X_1\rightarrow Y_1\) is compact. For each \(0<a<1\), set \( f_a(z)=\frac{1-a^2}{(1-az)^2},\, z\in \mathbb {D}. \) A calculation shows that \(\sup _{a\in (0,1)}\Vert f_a\Vert _{X_1}\asymp 1\) and \(f_a\rightarrow 0\), as \(a\rightarrow 1\) uniformly on compact subsets of \(\mathbb {D}\). Moreover, since \(T(f_a)\) has non-negative Taylor coefficients,
where the last inequality follows taking \(p=1\) in (3.26). So, using Theorem 30 we deduce that \(T: X_1\rightarrow Y_1\) is not a compact operator. \(\square \)
5 Hilbert-type operators acting on \(H^{\infty }\)
We will prove a result which includes Theorem 4. With this aim we need some more notation. The space \(Q_p\), \(0\le p<\infty \), consists of those \(f\in H(\mathbb {D})\) such that
where \(\varphi _a(z)= \frac{a-z}{1-\overline{a} z}, \; z, a \in \mathbb {D}.\) If \(p>1\), \(Q_p\) coincides with the Bloch space \(\mathcal {B}\). The space \(Q_1\) coincides with BMOA (see, e. g., [10, Theorem 5.2]). However, if \(0<p<1 \), \(Q_p\) is a proper subspace of BMOA [30]. The space \(Q_0\) reduces to the classical Dirichlet space \(\mathcal {D}\).
We recall that
however if \(0<p<1\), \(H^\infty \not \subset Q_p\), and \(Q_p \not \subset H^\infty \), see [30].
Moreover, \(HL(\infty )\subsetneq Q_p\). This embedding might have been proved in some previous paper, however we include a short direct proof for the sake of completeness.
Lemma 32
Let \(0<p\le \infty \), then \(HL(\infty )\subsetneq Q_p\) and
Proof
Let \(f\in HL(\infty )\), then
So for any \(0<p<\infty \),
so \(\Vert f\Vert _{Q_p}\lesssim \Vert f\Vert _{HL(\infty )}\). The lacunary series \(f(z)=\sum _{k=0}^\infty 2^{-k}\log {(k+2)}z^{2^k}\in \bigcap \nolimits _{0<p} Q_p{\setminus } HL(\infty )\). This finishes the proof.\(\square \)
Now we will prove the main result of this section, which is an extension of Theorem 4.
Theorem 33
Let \(\omega \) be a radial weight and let \(T \in \{H_\omega , \widetilde{H_\omega }\}\). Then, the following statements are equivalent:
-
(i)
\(T: H^{\infty }\rightarrow HL(\infty )\) is bounded;
-
(ii)
\(T: H^{\infty }\rightarrow Q_p\) is bounded for \(0<p<1\);
-
(iii)
\(T: H^{\infty }\rightarrow \mathord \textrm{BMOA}\) is bounded;
-
(iv)
\(T: H^{\infty }\rightarrow \mathcal {B}\) is bounded;
-
(v)
\(\omega \in \widehat{\mathcal {D}}\).
Proof of Theorem 33
By Lemma 6,
so (v)\(\Rightarrow \) (i). The implications (i)\(\Rightarrow \)(ii)\(\Rightarrow \)(iii)\(\Rightarrow \)(iv) follow from (5.1) and Lemma 32.
The implication (iv)\(\Rightarrow \)(v) was proved in [26, Theorem 1], and this finishes the proof. \(\square \)
It is worth mentioning that for \(f(z)=\log \frac{1}{1-z}\in HL(\infty )\) and \(\omega \) a radial weight,
so \(H_\omega (f)\notin \mathcal {B}\). So, the space \(H^\infty \) cannot be replaced by \(HL(\infty )\) and by any \(Q_p\) space, \(0<p<\infty \), in the statement of Theorem 33. That is, the remaining cases for \(p=\infty \), analogous to those of Theorems 1 and 2, which do not appear in Theorem 33, simply do not hold for any radial weight.
Finally, we will prove the analogous result to Theorem 22 for \(p=\infty \).
Theorem 34
Let \(\omega \) be a radial weight and let \(T \in \{H_\omega , \widetilde{H_\omega }\}\). Then \(T:H^{\infty }\rightarrow Y_\infty \) is not a compact operator, where \(Y_\infty \in \{Q_p,\mathcal {B},\mathord \textrm{BMOA}, HL(\infty )\}\), \(0<p<1\).
We need the following result, whose proof can be obtained by mimicking the proof Theorem 21.
Theorem 35
Let \(\omega \) be a radial weight and \(T\in \{H_\omega , \widetilde{H_\omega }\}\). Then, the following assertions are equivalent:
-
(i)
\(T: H^{\infty }\rightarrow \mathcal {B}\) is compact;
-
(ii)
For every sequence \(\{f_k\}_{k=0}^\infty \subset H^\infty \) such that \(\sup \nolimits _{k\in \mathbb {N}} \Vert f_k\Vert _{H^\infty }<\infty \) and \(f_k\rightarrow 0\) uniformly on compact subsets of \(\mathbb {D}\), \(\lim \nolimits _{k\rightarrow \infty } \Vert T(f_k)\Vert _\mathcal {B}=0\).
Proof of Theorem 34
By (5.1) and Lemma 32, it is enough to prove that \(T: H^\infty \rightarrow \mathcal {B}\) is not compact. Let consider for every \(k\in \mathbb {N}\) the function \(f_k(z)=z^k\), \(z\in \mathbb {D}\). It is clear that \(\Vert f_k\Vert _{H^\infty }=1\) for every \(k\in \mathbb {N}\) and \(f_k\rightarrow 0\) uniformly on compact subsets of \(\mathbb {D}\). Since
for any \(k\ge 2\)
so \(\lim \nolimits _{k\rightarrow \infty }\Vert T(f_k)\Vert _\mathcal {B}\ne 0\) and hence, by Theorem 35, \(T: H^\infty \rightarrow \mathcal {B}\) is not compact.\(\square \)
Before ending this section, we briefly compare the action of the Hilbert-type operator \(H_\omega \) and the Bergman projection
induced by a radial weight \(\omega \). As a consequence of Theorem 33 and [23, Theorem 1], the condition \(\omega \in \widehat{\mathcal {D}}\) characterizes the boundedness of the operators \(H_{\omega }:H^{\infty }\rightarrow \mathcal {B}\) and \(P_{\omega }:L^{\infty }\rightarrow \mathcal {B}\). Moreover, \(P_{\omega }:L^{\infty }\rightarrow \mathcal {B}\) is bounded and onto if and only if \(\omega \in \mathcal {D}\) [23, Theorem 3]. So, it is natural to think about the radial weights such that the operator \(H_{\omega }:H^{\infty }\rightarrow \mathcal {B}\) is bounded and onto. A straightforward argument proves there is no radial weights such that \(H_{\omega }:H^{\infty }\rightarrow \mathcal {B}\) satisfies both properties: If \(H_{\omega }:H^{\infty }\rightarrow \mathcal {B}\) is bounded, Theorem 33 yields that \(H_{\omega }:H^{\infty }\rightarrow \mathord \textrm{BMOA}\) is also bounded, so if \(g\in \mathcal {B}\setminus \mathord \textrm{BMOA}\), e.g. \(g(z)=\sum \nolimits _{k=0}^\infty z^{2^k}\), there does not exist \(f\in H^\infty \) such that \(H_{\omega }(f)=g\). Consequently, \(H_{\omega }:H^{\infty }\rightarrow \mathcal {B}\) is not surjective.
6 Comparisons and reformulations of the \(M_{p,c}\)-conditions
In order to prove Theorem 5 we will study the relationship between some of the conditions which describe the boundedness of the Hilbert-type operators \(H_\omega \) and \(\widetilde{H_\omega }\) from \(X_p\) to \(Y_p\), and \(X_q\) to \(Y_q\).
Theorem 5
Firstly, assume \(1<q<p<\infty \). Since \(T:X_q\rightarrow Y_q\) is bounded, Theorem 1 yields \(\omega \in \widehat{\mathcal {D}}\) and \(M_{q,c}(\omega )<\infty \), and as a consequence, \(K_{q,c}(\omega )<\infty \). By using Lemma 6,
On the other hand, Hölder’s inequality with exponents \(x=\frac{q'}{p'}>1\) and \(x'=\frac{x}{x-1}\), implies
Moreover,
So, the identity \(\frac{1}{p'}-\frac{1}{q'}=\frac{1}{q}-\frac{1}{p}\), together with (6.1), (6.2) and (6.3) yield
Consequently \(K_{p,c}(\omega )<\infty \), and by Theorem 1, \(T:X_p\rightarrow Y_p\) is bounded.
Assume that \(q=1\), that is, \(T:X_1\rightarrow Y_1\) is bounded. By Theorem 2, \(\omega \in \widehat{\mathcal {D}}\) and \(M_{1,d}(\omega )<\infty \), so \(K_{1,d}(\omega )<\infty \). Then,
where in the last inequality we have used (6.1) with \(q=1\). Moreover,
So, \(M_{p,c}(\omega )<\infty \), and by Theorem 1, \(T:X_p\rightarrow Y_p\) is bounded. This finishes the proof.
Finally, we present two more conditions which characterize the radial weights \(\omega \) such that \(T: X_p\rightarrow Y_p\), \(1<p<\infty \), is bounded, where \(X_p,Y_p\in \{H(\infty ,p),H^p, D^p_{p-1}, HL(p)\}\) and \(T\in \{ H_{\omega }, \widetilde{H_{\omega }}\}\).
Proposition 36
Let \(\omega \) be a radial weight and \(1<p<\infty \). Then, the following conditions are equivalent:
-
(i)
\(\omega \in \widehat{\mathcal {D}}\) and \(K_{p,d}(\omega )=\sup \nolimits _{0< r<1}\frac{\widehat{\omega }(r)}{(1-r)^{\frac{1}{p}}}\left( 1+\int _0^r \frac{1}{\widehat{\omega }(s)^p}\,ds \right) ^{\frac{1}{p}}<\infty \);
-
(ii)
\(\omega \in \widehat{\mathcal {D}}\) and \(K_{p,e}(\omega )=\sup \nolimits _{0< r<1}\frac{(1-r)^{\frac{1}{p}}}{\widehat{\omega }(r)} \left( \int _r^1 \left( \frac{\widehat{\omega }(s)}{1-s}\right) ^{p'}\,ds \right) ^{\frac{1}{p'}}<\infty \);
-
(iii)
\(\omega \in \widehat{\mathcal {D}}\) and \( K_{p,c}(\omega )= \sup \nolimits _{0<r<1} \left( 1+\int _0^r \frac{1}{\widehat{\omega }(t)^p} dt\right) ^{\frac{1}{p}} \left( \int _r^1 \left( \frac{\widehat{\omega }(t)}{1-t}\right) ^{p'}\,dt\right) ^{\frac{1}{p'}}<\infty .\)
Proof
Assume that (i) holds. A calculation shows that \(F(r)=(1-r)^{\kappa }\left( 1+\int _0^r \frac{ds}{\widehat{\omega }(s)^p}\, \right) \), with \(\kappa =\frac{1}{K^p_{p,d}(\omega )}\), is non-decreasing in [0, 1). So,
where in the last inequality we have used (6.1). That is (ii) holds.
Now, assume that (ii) holds. Since \(\omega \in \widehat{\mathcal {D}}\)
Moreover, \(H(r)=(1-r)^{-\eta }\int _r^1 \left( \frac{\widehat{\omega }(s)}{1-s}\right) ^{p'}\,ds \), with \(\eta =\frac{1}{K^{p'}_{p,e}(\omega )}\), is non-increasing in [0, 1). So,
therefore (i) holds.
Next, if (i) holds, then (ii) holds and so it is clear that (iii) holds. Finally, (iii) together with (6.1) implies (ii). This finishes the proof. \(\square \)
Data Availability
This manuscript has no associate data.
References
Aleman, A., Montes-Rodríguez, A., Sarafoleanu, A.: The eigenfunctions of the Hilbert matrix. Constr. Approx. 36(3), 353–374 (2012)
Contreras, M.D., Peláez, J.A., Pommerenke, C., Rättyä, J.: Integral operators mapping into the space of bounded analytic functions. J. Funct. Anal. 271(10), 2899–2943 (2016)
Diamantopoulos, E.: Hilbert matrix on Bergman spaces. Ill. J. Math. 48(3), 1067–1078 (2004)
Diamantopoulos, E., Siskakis, A.: Composition operators and the Hilbert matrix. Stud. Math. 140(2), 191–198 (2000)
Dostanić, M., Jevtić, M., Vukotić, D.: Norm of the Hilbert matrix on Bergman and Hardy spaces and a theorem of Nehari type. J. Funct. Anal. 254(11), 2800–2815 (2008)
Duren, P.: Theory of \(H^p\) Spaces. Pure and Applied Mathematics, vol. 38. Academic Press, New York (1970)
Flett, T.M.: The dual of an inequality of Hardy and Littlewood and some related inequalities. J. Math. Anal. Appl. 38, 746–765 (1972)
Galanopoulos, P., Girela, D., Peláez, J.A., Siskakis, A.G.: Generalized Hilbert operators. Ann. Acad. Sci. Fenn. Math. 39(1), 231–258 (2014)
Galanopoulos, P., Peláez, J.A.: A Hankel matrix acting on Hardy and Bergman spaces. Stud. Math. 200(3), 201–220 (2010)
Girela, D.: Analytic Functions of Bounded Mean Oscillation. Complex Function Spaces (Mekrijärvi, 1999), 61–170, Univ. Joensuu Dept. Math. Rep. Ser., 4, Univ. Joensuu, Joensuu (2001)
Girela, D., Merchán, N.: A Hankel matrix acting on spaces of analytic functions. Integral Equ. Oper. Theory 89(4), 581–594 (2017)
Jevtić, M., Pavlović, M.: On multipliers from \(H^p\) to \(l^q\), \(0<q<p<1\). Arch. Math. (Basel) 56(2), 174–180 (1991)
Lanucha, B., Nowak, M., Pavlović, M.: Hilbert matrix operator on spaces of analytic functions. Ann. Acad. Sci. Fenn. Math. 37(1), 161–174 (2012)
Littlewood, J.E., Paley, R.E.A.C.: Theorems on Fourier series and power series (II). Proc. Lond. Math. Soc. 2, 52–89 (1937)
Mateljević, M., Pavlović, M.: \(L^p\)-behaviour of power series with positive coefficients and Hardy spaces. Proc. Am. Math. Soc. 87(2), 309–316 (1983)
Mateljević, M., Pavlović, M.: \(L^p\) behaviour of the integral means of analytic functions. Stud. Math. 77(3), 219–237 (1984)
Muckenhoupt, B.: Hardy’s inequality with weights. Stud. Math. 44, 31–38 (1972)
Pau, J., Peláez, J.A.: Embedding theorems and integration operators on Bergman spaces with rapidly decreasing weights. J. Funct. Anal. 259(10), 2727–2756 (2010)
Pavlović, M.: Function Classes on the Unit Disc. An Introduction De Gruyter Studies in Mathematics, vol. 52. De Gruyter, Berlin (2014). (ISBN: 978-3-11-028123-1)
Pavlović, M.: Introduction to function spaces on the disk. Posebna Izdanja [Special Editions], 20. Matematićki Institut SANU, Belgrade (2004) (ISBN: 86-80593-37-0)
Peláez, J.A.: Small weighted Bergman spaces. Proceedings of the Summer School in Complex and Harmonic Analysis, and Related Topics, 29–98, Publ. Univ. East. Finl. Rep. Stud. For. Nat. Sci., 22, Univ. East. Finl., Fac. Sci. For., Joensuu (2016)
Peláez, J.A., Perälä, A., Rättyä, J.: Hankel operators induced by radial Bekollé–Bonami weights on Bergman spaces. Math. Z. 296(1–2), 211–238 (2020)
Peláez, J.A., Rättyä, J.: Bergman projection induced by a radial weight. Adv. Math. 391, Paper No. 107950, 70 pp (2021)
Peláez, J.A., Rättyä, J.: Generalized Hilbert operators on weighted Bergman spaces. Adv. Math. 240, 227–267 (2013)
Peláez, J.A., Rättyä, J.: Weighted Bergman spaces induced by rapidly increasing weights. Mem. Am. Math. Soc. 227, 1066 (2014). (ISBN: 978-0-8218-8802-5)
Peláez, J.A., de la Rosa, E.: Hilbert-type operator induced by radial weight. J. Math. Anal. Appl. 495(1), Paper No. 124689, 22 pp (2021)
Pommerenke, C.: Univalent Functions. With a Chapter on Quadratic Differentials by Gerd Jensen. Studia Mathematica/Mathematische Lehrbücher, Band XXV. Vandenhoeck Ruprecht, Göttingen (1975)
Siskakis, A.G.: Weighted integrals of analytic functions. Acta Sci. Math. (Szeged) 66(3–4), 651–664 (2000)
Vinogradov, S.A.: Multiplication and division in the space of analytic functions with an area-integrable derivative, and in some related spaces. Zap. Nauchn. Sem. S.-Peterburg. Otdel. Mat. Inst. Steklov. 222 (1995), Issled. po Lineĭn. Oper. i Teor. Funktsiĭ. 23, 45–77, 308; translation in J. Math. Sci. 87 (1997), no. 5, 3806–3827
Xiao, J.: Holomorphic \(Q\) Classes. Lecture Notes in Mathematics, vol. 1767. Springer, Berlin (2001)
Funding
Funding for open access publishing: Universidad Málaga/CBUA.
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
This research was supported in part by Ministerio de Ciencia e Innovación, project PID2022-136619NB-100; La Junta de Andalucía, project FQM210 (UMA18-FEDERJA-002).
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Merchán, N., Peláez, J.Á. & de la Rosa, E. Hilbert-type operator induced by radial weight on Hardy spaces. Rev. Real Acad. Cienc. Exactas Fis. Nat. Ser. A-Mat. 118, 2 (2024). https://doi.org/10.1007/s13398-023-01500-z
Received:
Accepted:
Published:
DOI: https://doi.org/10.1007/s13398-023-01500-z