GENERALIZED HILBERT OPERATOR ACTING ON BERGMAN

SPACES

PENGCHENG TANG˚,1 XUEJUN ZHANG2

ř8
Abstract. Let µ be a positive Borel measure on r0, 1q and f pzq “ n“0 an z n P
arXiv:2208.10747v2 [math.FA] 24 Dec 2024

HpDq. For α ą ´1, the generalized Hilbert operator is defined as follows:

˜ ¸
8 8
α
ÿ Γpn ` 1 ` αq ÿ
Hµ pf qpzq “ µn,k ak z n , z P D,
n“0
Γpn ` 1qΓpα ` 1q k“0
where, for n ě 0, µn denotes the n-th moment of the measure µ, and µn,k “ µn`k .
In this paper, we characterize the measures µ for which Hµα is a bounded op-
erator acting from Bergman space Ap p0 ă p ă 8q into Aq pq ě 1q. We determine
the Hilbert-Schmidt class on A2 for all α ą ´1. In addition, we also study the
boundedness (resp. compactness) of Hµα acting from Ap to the Bloch space.

1. Introduction

Let D “ tz P C : |z| ă 1u denote the open unit disk of the complex plane C
and HpDq denote the space of all analytic functions in D. H 8 pDq denote the set of
bounded analytical functions.
The Bloch space B consists of those functions f P HpDq for which
||f ||B “ |f p0q| ` supp1 ´ |z|2 q|f 1 pzq| ă 8.
zPD

The little Bloch space B0 is the closed subspace of B consists of those functions
f P HpDq such that
lim´ p1 ´ |z|2 q|f 1 pzq| “ 0.
|z|Ñ1

For 0 ă p ă 8, the Bergman space Ap consists of f P HpDq such that

ż
p
||f ||Ap “ |f pzq|p dApzq ă 8,
D

2010 Mathematics Subject Classification. 47B35, 30H30, 30H20.

Key words and phrases. Hilbert operator. Bloch space. Bergman space. Carleson measure.
˚
Corresponding author.
Email:www.tang-tpc.com@foxmail.com(Pengcheng Tang) xuejunttt@263.net(Xuejun Zhang)
The first author was supported by the Scientific Research Fund of Hunan Provincial Education De-
partment (NO. 24C0222) and the second author was supported by the Natural Science Foundation
of Hunan Province (No. 2022JJ30369).
1
where dApzq “ dxdyπ
is the normalized area measure on D. We refer to [9, 33] for the
notation and results about Bergman spaces and Bloch spaces.
Let f pzq “ 8
ř n
n“0 an z P HpDq. For any complex parameters t and s with the
property that neither 1 ` t nor 1 ` t ` s is a negative integer, we define the fractional
differential operator Rt,s and the fractional integral operator Rt,s as follows:
8
t,s
ÿ Γp2 ` tqΓpn ` 2 ` t ` sq
R f pzq “ an z n ,
n“0
Γp2 ` t ` sqΓpn ` 2 ` tq
8
ÿ Γp2 ` t ` sqΓpn ` 2 ` tq
Rt,s f pzq “ an z n .
n“0
Γp2 ` tqΓpn ` 2 ` s ` sq
Let µ be a finite Borel measure on r0, ş 1q and n P N. We use µn to denote the
sequence of order n of µ, that is, µn “ r0,1q tn dµptq. Let Hµ be the Hankel matrix
pµn,k qn,kě0 with entries µn,k “ µn`k . The matrix Hµ induces an operator on HpDq
ÿ8
by its action on the Taylor coefficients : an Ñ µn,k ak , n “ 0, 1, 2, ¨ ¨ ¨ .
k“0
8
ÿ
If f pzq “ an z n P HpDq, the generalized Hilbert operator is defined as follows:
n“0
˜ ¸
8
ÿ 8
ÿ
Hµ pf qpzq “ µn,k ak zn.
n“0 k“0

If µ is the Lebesgue measure on r0, 1q, then Hµ reduce to the classic Hilbert operator
H. It’s known that the generalized Hilbert operator Hµ is closely related to the
following integral operator
ż1
f ptq
Iµ pf qpzq “ dµptq.
0 1 ´ tz

The study of the operator Hµ on analytic function spaces was initiated by Widom
[25] in 1966. He proved that the operator Hµ is bounded (resp. compact) on the
Hardy space H 2 if and only if µ is a Carleson measure (resp. vanishing Carleson
measure). In 2000, Diamantopoulos and Siskakis [6] studied the classic Hilbert
operator H on Hardy spaces H p for 1 ă p ă 8. Subsequently, Diamantopoulus [7]
considered the boundedness of H on the Bergman spaces Ap for 2 ă p ă 8. In 2010,
Galanopoulos and Peláez [10] investigated the boundedness of the operator Hµ on
Hardy space H 1 and on Bergman space A2 . In 2014, Chatzifountas, Girela and
Peláez [4] systematically studied the boundedness (resp. compactness) of Hµ from
H p to H q for 0 ă p, q ă 8. Since then, the study of generalized Hilbert operator has
attracted the attention of many scholars. See [2,3,8,11,12,14,16,17,20,24,27–29] for
more information on generalized Hilbert operator on spaces of analytic functions.
The derivative or fractional derivative paly a basic role in the theory of analytic
function spaces. This motivates us to consider the derivative or fractional derivative
2
of the operator Hµ . In this paper, we consider the generalized Hilbert operator as
follows: ˜ ¸
8 8
α
ÿ Γpn ` 1 ` αq ÿ
Hµ pf qpzq “ µn,k ak z n , pα ą ´1q.
n“0
Γpn ` 1qΓpα ` 1q k“0
The operator Hµα can be regarded as the fractional derivative of Hµ , since it is easy
to obtain that Hµα pf q “ R´1,α Hµ pf q. Hence, we may say that Hµα is the fractional
derivative Hilbert operator. If α “ 0, the operator Hµ0 is just the Hµ . If α “ 1, then
the operator Hµα is called Derivative-Hilbert operator. Naturally, the operator Hµα
is closely related to the operator Iµα`1 which is defined by
ż1
f ptq
Iµα`1 pf qpzq “ α`1
dµptq, pα ą ´1q.
0 p1 ´ tzq
If α “ 0, then Iµα`1 is just the integral operator Iµ .
The connection between Hµ and Hµα motivates us to consider the operator Hµα
for all α ą ´1 in a unified manner. The Bergman space is one of the most basic
analytic function spaces. The operator Hµ on Bergman spaces has achieved a few
results, but there are still many unsolved problems. The following question is raised
naturally.
Question: what are the necessary and sufficient conditions for Hµα : Ap Ñ Aq p0 ă
p, q ď 8q to be a bounded (or compact) operator? Here, the space A8 should be
interpreted as the Bloch space B since the Bloch space B can be viewed as the limit
case of Ap as p Ñ `8.
The study of the operator Hµα between Bergman spaces did not proceed well
even for α “ 0. In 2019, Girela and Merchán [13] proved that Hµ is bounded on
Ap p2 ă p ă 8q if and only if µ is a Carleson measure. Ye and Zhou solved the
problem for α “ 1 and 1 ď q ă p “ 8 in [30]. They also solved the problem for
α “ 1, p ď q and 1 ď q ď 8 in [31]. The first author of this paper solved the problem
for all α ą ´1 and 1 ď q ă p “ 8 in [23]. Aguilar-Hernández, Galanopoulos and
Girela [1] considered the operator Hµ on Ap for 1 ď p ă 2. They solved the problem
for α “ 0 and p “ q “ 1. They also obtain a sufficient condition and a necessary
condition for the case 1 ă p ă 2. Recently, the authors were informed that Sun et
al. solved the case α “ 0 and 1 ă p “ q ă 2. A complete solution for all indicators
is far away. In this paper, we will continue to explore this problem, giving some
results for certain cases.
This paper is organized as follows. In section 2, we list some of the lemmas that
will be used. In section 3, we mainly investigate the operator Hµα acting between
the Bergman spaces. We give the necessary and sufficient conditions such that the
operator Hµα acting from Ap p0 ă p ď 1q into Aq pq ě 1q to be bounded. By means
of the Marcinkiewicz interpolation theorem, we give the necessary and sufficient
conditions for the operator Hµα to be bounded on Ap p1 ď p ď 2q. Finally, we
determine the Hilbert-Schmidt class on A2 for all α ą ´1. Section 4 is devote to
3
study the boundeness and compactness of Hµα acting from Bergman space Ap to the
Bloch sapce B. By using the sublinear generalized integral type Hilbert operator and
a decomposition theorem of the Bloch space B, we have completely characterized
the boundedness (resp. compactness) of Hµα : Ap Ñ B.
As mentioned previously, the Carleson-type measures play a basic role in the
studies of the generalized Hilbert operator. Let I Ă BD be an arc, and |I| denote
the length of I. The Carleson square SpIq is defined as
|I|
SpIq “ treiϑ : eiϑ P I, 1 ´ ď r ă 1u.
2π
Let µ be a positive Borel measure on D. For 0 ď β ă 8 and 0 ă t ă 8,
we say that µ is a β-logarithmic t-Carleson measure (resp.a vanishing β-logarthmic
t-Carleson measure) if
˜ ¸
µpSpIqqplog 2π
|I|
qβ
µpSpIqqplog 2π β
|I|
q
sup t
ă 8, resp. lim t
“0 .
IĂBD |I| |I|Ñ0 |I|
If β “ 0 and t “ 1, we say that µ is a Carleson measure. See [32] for more about
logarithmic type Carleson measure.
A positive Borel measure µ on r0, 1q can be seen as a Borel measure on D by
identifying it with the measure µ defined by
µpEq “ µpE X r0, 1qq, for any Borel subset E of D.
In this way, a positive Borel measure µ on r0, 1q is a β-logarithmic t-Carleson
measure if and only if there exists a constant M ą 0 such that
e
µprs, 1qq logβ ď M p1 ´ sqt , 0 ď s ă 1.
1´s
Throughout the paper, the letter C 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 “P À Q” if there exists a constant
C “ Cp¨q such that “P ď CQ”, and “P Á Q” is understood in an analogous manner.
In particular, if “P À Q” and “P Á Q” , then we will write “P — Q”. We shall use
the notation that for any given s ą 1, s1 will denote the conjugate exponent of s,
that is, 1s ` s11 “ 1, or s1 “ s´1
s
.

2. Preliminary Results

Mateljević and Pavlović [19] proved the following result.

Lemma 2.1. Let 1 ă p ă 8 and γ ą ´1. For a function f P HpDq we define:
ż 8
ÿ
p γ
K1 pf q :“ |f pzq| p1 ´ |z|q dApzq, K1 pf q :“ 2´npγ`1q ||∆n f ||pp .
D n“0

Then, K1 pf q — K2 pf q.
4
 The following lemma can be found in [9, page 83 ].
Lemma 2.2. Let f pzq “ 8
ř n
n“0 an z P HpDq, then
8
ÿ 8
ÿ
´1 1
n |an | ă 8 ñ f P A ñ n´2 |an | ă 8.
n“1 n“1

The following lemma comes form Theorem 3.5 in [33].

Lemma 2.3. Let t ą 0 and f P HpDq. If s is a real parameter such that neither
1 ` s nor 1 ` s ` t is a negative integer, then the following conditions are equivalent:
(a) f P B.
(b) The function p1 ´ |z|2 qt Rs,t f pzq is bounded in D.
Moreover,
||f ||B — supp1 ´ |z|2 qt |Rs,t f pzq|.
zPD

The following integral estimates follows from the Proposition 1.4.10 in [22].
Lemma 2.4. Let t be real and δ ą ´1, z P D. Then the integral
p1 ´ |u|2 qδ dApuq
ż
Jpzq “
D |1 ´ zu|2`t`δ
have the following asymptotic properties.
(1) If t ă 0, then Jpzq — 1.
e
(2) If t “ 0, then Jpzq — log .
1 ´ |z|2
(3) If t ą 0, then Jpzq — p1 ´ |z|2 q´t .
Lemma 2.5. Let µ be a positive Borel measure on r0, 1q, β ą 0, γ ą 0. Let τ be
the Borel measure on r0, 1q defined by
dµptq
dτ ptq “ .
p1 ´ tqγ
Then, the following statements are equivalent.
(a) µ is a β ` γ-Carleson measure.
(b) τ is a β-Carleson measure.
Proof. pbq ñ paq. Assume (b). Then there exists a positive constant C ą 0 such
that ż1
dµprq
γ
ď Cp1 ´ tqβ , t P r0, 1q.
t p1 ´ rq
1
Using this and the fact that the function x ÞÑ p1´xqγ is increasing in r0, 1q, we obtain

ż1
µprt, 1qq dµprq
ď ď Cp1 ´ tqβ , t P r0, 1q.
p1 ´ tqγ t p1 ´ rq γ

This shows that µ is a β ` γ-Carleson measure.

5
 paq ñ pbq. Assume (a). Then there exists a positive constant C ą 0 such that
µptq ď Cp1 ´ tqβ`γ , t P r0, 1q.
Integrating by parts and using the above inequality, we obtain
ż1
dµpxq
τ prt, 1qq “ γ
t p1 ´ xq
ż1
1 1 µprx, 1qq
“ γ
µprt, 1qq ´ lim γ
µprx, 1qq ` γ γ`1
dx
p1 ´ tq xÑ1 p1 ´ xq t p1 ´ xq
ż1
1 µprx, 1qq
“ µprt, 1qq ` γ dx
p1 ´ tqγ t p1 ´ xq
γ`1
ż1
À p1 ´ tq ` p1 ´ xqβ´1 dx À p1 ´ tqβ .
β
t
Thus, τ is a β-Carleson measure. □
The following lemma is useful for dealing with the compactness. It is a conse-
quence of Theorem 3.1 in [18].
Lemma 2.6. Suppose that T is a bounded operator from Ap into Y (Y “ B or Ap ).
Then T is compact operator from Ap into Y if and only if for any bounded sequence
thn u in X which converges to 0 uniformly on every compact subset of D, we have
limnÑ8 ||T phn q||Y “ 0.

3. Generalized Hilbert operator acting between Bergman spaces

Let 0 ă p ă 8, for each f P Ap , it is well known that

||f ||Ap
|f pzq| À 2 for all z P D. (3.1)
p1 ´ |z|q p
The following lemma provide a sufficient condition for the operator Hµα to be well
defined on the Bergman space Ap . For a proof, see [28, Theorem 3.3] or [31, Theorem
2.1].
Lemma 3.1. Suppose 0 ă p ă 8 and α ą ´1. Let µ be a positive Borel measure
on r0, 1q. Then Hµα pf q is a well defined analytic function in D for every f P Ap in
any of the three following cases.
(i) µ is a p2 Carleson measure if 0 ă p ď 1.
´ ¯
2´pp´1q2
(ii) µ is a p
Carleson measure if 1 ď p ď 2.
(iii) µ is a p1 Carleson measure if 2 ă p ă 8.
Furthermore, in such cases we have that
ż1
α f ptq
Hµ pf q “ α`1
dµptq, z P D, f P Ap .
0 p1 ´ tzq
6
 We start considering the case α ą ´1, 0 ă p ď 1 and q “ 1. We have the following
result.

Theorem 3.2. Suppose that 0 ă p ď 1 and α ą ´1, let µ be a positive Borel

measure on r0, 1q which satisfies the conditions of Lemma 3.1. Then the following
statements hold:
(a) If α ą 1, then Hµα : Ap Ñ A1 is bounded if and only if µ is a p2 ` α ´ 1 Carleson
measure.
(b) If ´1 ă α ă 1, then Hµα : Ap Ñ A1 is bounded if and only if µ is a p2 Carleson
measure.

Proof. (a). If α ą 1 and Hµα : Ap Ñ A1 is bounded. Let

2
p1 ´ b2 q p 1
fb pzq “ 4 , b P p , 1q, z P D.
p1 ´ bzq p 2

Then it is easy to check that fb P Ap and sup 1 ăbă1 ||fb ||Ap À 1. Since Hµα pfb q P A1 ,
2
Lemma 2.2 yields
8
Γpn ` 1 ` αq ˇˇ 1 n
ÿ ˇż ˇ
ˇ
n ´2
t f b ptqdµptq ˇ À ||Hµα pfb q||A1 À ||fb ||Ap À 1.
n“1
Γpn ` 1qΓpα ` 1q 0
ˇ ˇ

It is clear that
8
ÿ
k 2 2 Γpk ` p4 qbk
fb pzq “ ak,b z , where ak,b “ p1 ´ b q p ,
k“0
Γpk ` 1qΓp p4 q

and
8
ÿ 1 1
nα´2 bn — α´1
, whenever α ą 1, b P p , 1q.
n“1
p1 ´ bq 2
This together with Stirling’s formula imply that
8
Γpn ` 1 ` αq ˇˇ 1 n
ˇż ˇ
ÿ ˇ
´2
1Á n t f b ptqdµptq ˇ
n“1
Γpn ` 1qΓpα ` 1q ˇ
0
ˇ

2 p2
ÿ8
α´2
ÿ8
Γpk ` p4 qbk ż 1 n`k
Á p1 ´ b q n 4 t dµptq
n“1 k“0
Γpk ` 1qΓp p
q b
˜ ¸
2 ÿ8 ÿ Γpk ` p4 qb2k
8
Á p1 ´ b2 q p µprb, 1qq nα´2 bn
n“1 k“0
Γpk ` 1qΓp p4 q
2 1
— p1 ´ b2 q p µprb, 1qq 4
p1 ´ b2 q p `α´1
µprb, 1qq
“ 2 .
p1 ´ b2 q p `α´1
7
 2
Consequently, µprb, 1qq À p1 ´ b2 q p `α´1 for all 1
2
ă b ă 1. This shows that µ is a
2
p
` α ´ 1-Carleson measure.
2 dµptq
On the other hand, if µ is a p
` α ´ 1 Carleson measure. Then is a
p1 ´ tqα´1
2
p
-Carleson measure by Lemma 2.5. The assumption of the measure µ insure that
Hµα pf q “ Iµα`1 pf q for every f P Ap . The well known Carleson embedding theorem
for Ap shows that
ż1
dµptq
|f ptq| α´1
À ||f ||Ap for all f P Ap , 0 ă p ď 1. (3.2)
0 p1 ´ tq
For each 0 ď r ă 1, f P Ap and g P Bo , we have
ż ż ˇ ˇ
ˇ f psqgprzq ˇ
ˇ ˇ dµpsqdApzq
α`1 ˇ
D r0,1q p1 ´ rszq
ˇ
ż ż
1 |f psq|
ď dµpsq |gprzq|dApzq
p1 ´ rq2 r0,1q p1 ´ sqα´1 D
ż
||f ||Ap ||g||B e
À 2
log dApzq
p1 ´ rq D 1 ´ |z|
||f ||B ||g||B
ď ă 8,
p1 ´ rq2
Let gpzq “ 8
ř n
n“0 bn z . Then by Fubini’s theorem and a simple calculation through
polar coordinate, we have that
ż
Iµα`1 pf qprzqgprzqdApzq
D
8
ż1ÿ
Γpn ` 1 ` αq
“ bn pr2 tqn f ptqdµptq
0 n“0
Γpn ` 2qΓpα ` 1q
ż1
“ R0,α´1 gpr2 tqf ptqdµptq.
0

for all 0 ď r ă 1, f P Ap , g P B0 .
For each 0 ď r ă 1, f P Ap and g P B0 , using Lemma 2.3 and (3.2) we have
ˇż ˇ ˇż 1 ˇ
ˇ ˇ ˇ 0,α´1 2
ˇ
ˇ Iµα`1 pf qprzqgprzqdApzqˇ “ ˇ R gpr tqf ptqdµptqˇˇ
ˇ ˇ ˇ
D 0
ż1 ż1
dµptq dµptq
À ||g||B |f ptq| À ||g||B |f ptq|
0 p1 ´ r2 tqα´1 0 p1 ´ tqα´1
À ||f ||Ap ||g||B .
Thus ˇż ˇ
ˇ ˇ
lim´ ˇ Iµα`1 pf qprzqgprzqdApzqˇˇ À ||f ||Ap ||g||B .
ˇ (3.3)
rÑ1 D
8
Since pB0 q‹ » A1 under the pairing
ż
xF, Gy “ lim´ F przqGprzqdApzq, F P B0 , G P A1 .
rÑ1 D

This together with (3.3) imply that Hµα is a bounded operator from Ap into A1 .
(b). If ´1 ă α ă 1, then it is obvious that
8
ÿ 1
nα´2 bn — 1 whenever b P p , 1q.
n“1
2
Arguing as in the proof of (a), we will obtain the necessity.
On the other hand, suppose that µ is a p2 Carleson measure, then Hµα pf q is well
defined analytic function for each f P Ap by Lemma 3.1. Moreover, the Carleson
embedding for Ap implies that
ż1
|f ptq|dµptq À ||f ||Ap for all f P Ap , 0 ă p ď 1.
0

Using Fubini’s theorem and Lemma 2.4, we have

ż ˇż 1 ˇ
α
ˇ f ptq ˇ
||Hµ pf q||A1 “ ˇ ˇ
α`1
dµptqˇˇ dApzq
D 0 p1 ´ tzq
ż1 ż
dApzq
ď |f ptq| α`1
dµptq
0 D |1 ´ tz|
ż1
À |f ptq|dµptq À ||f ||Ap .
0

This implies that Hµα p 1

: A Ñ A is bounded. □
Remark 3.3. For α “ 0 and p “ 1, Hµ is bounded on A1 if and only if µ is a
2-Carleson measure. This also was proved in [1, Theorem 3]. For α “ 1, then
Hµ1 : Ap Ñ A1 is bounded if and only if µ is a 1-logarithmic p2 -Carleson measure.
This result was proved by Ye and Zhou [31, Theorem 3.2].
As a consequence of Theorem 3.2, we may obtain the following results.
Corollary 3.4. Let 0 ă p ď 1 and let µ be a positive Borel measure on r0, 1q.
If Hµα : Ap Ñ A1 is bounded for some α ą ´1, then for any ´1 ă α1 ă α,
1
Hµα : Ap Ñ A1 is bounded.
Corollary 3.5. Suppose that 0 ă p ă 8 and α ą ´1. Let µ be a positive Borel
measure on r0, 1q.ż Then the following statements hold.
1
dµptq
(a) If α ą 1 and 2
`α´1
ă 8, then Hµα : Ap Ñ A1 is bounded.
0 p1 ´ tq p
ż1
dµptq α p 1
(b) If ´1 ă α ă 1 and 2 ă 8, then Hµ : A Ñ A is bounded.
0 p1 ´ tq p
9
 ş1 dµptq
Proof. (a) If α ą 1 and 0 2 `α´1 ă 8, for every f P Ap and g P B0 , by Lemma
p1´tq p
2.3 and (3.1) we have that
ˇż 1 ˇ ż1
ˇ
ˇ R 0,α´1 2
ˇ dµptq
gpr tqf ptqdµptqˇˇ À ||f ||Ap ||g||B 2
0 p1 ´ r 2 tqα´1 p1 ´ tq p
ˇ
0
ż1
dµptq
À ||f ||Ap ||g||B 2
`α´1
0 p1 ´ tq p

À ||f ||Ap ||g||B .

The rest of the proof is straightforward.
ş1
(b) If ´1 ă α ă 1 and 0 dµptq2 ă 8, then (3.1) means that
p1´tq p
ż1 ż1
dµptq
||Hµα pf q||A1 À |f ptq|dµptq À ||f ||Ap 2 À ||f ||Ap .
0 0 p1 ´ tq p
This finishes the proof. □
ż1
dµptq 2
Remark 3.6. If α ą 1 and 2
`α´1
` α ´ 1-Carleson
ă 8 , then µ is a p
p1 ´ tq 0 p
2
`α´2
measure. But the reverse does not hold. For example, dµptqż 1“ p1 ´ tq
p dt. It is
dµptq
easy to check that µ is a p2 ` α ´ 1-Carleson measure but 2
`α´1
“ 8.
0 p1 ´ tq p

Let us turn to the case 0 ă p ď 1 and 1 ă q ă 8. We have the following result.

Theorem 3.7. Suppose that 0 ă p ď 1 ă q ă 8 and α ą ´1. Let µ be a
positive Borel measure on r0, 1q which is a p2 -Carleson measure. Then the following
statements hold:
(a) If α ` 1 ´ 2q ą 0, then Hµα : Ap Ñ Aq is bounded if and only if µ is a p2 ` q21 ` α ´ 1
Carleson measure.
(b) If α ` 1 ´ 2q ă 0, then Hµα : Ap Ñ Aq is bounded.
(c) If α`1´ 2q “ 0 and µ is a 1q -logarithmic p2 -Carleson measure, then Hµα : Ap Ñ Aq
is bounded.
Proof. (a). If Iµα`1 : Ap Ñ Aq is bounded, then arguing as the proof of Theorem
3.2 we may obtain that µ is a p2 ` q21 ` α ´ 1 Carleson measure. Thus, we only need
to prove the sufficiency. By using the integral form of Minkowski’s inequality and
Lemma 2.4, we have
"ż ˆż 1 ˙q * 1q
|f ptq|
||Hµα pf q||Aq ď α`1
dµptq dApzq
D 0 |1 ´ tz|
ż1 ˆż ˙ 1q
dApzq
ď |f ptq| qpα`1q
dµptq (3.4)
0 D |1 ´ tz|
10
 ż1
dµptq
À |f ptq| 2 .
0 p1 ´ tqα`1´ q
2 2 dµptq
If µ is a p
` q1
` α ´ 1 Carleson measure, then Lemma 2.5 shows that α`1´ 2
is
p1´tq q

a p2 -Carleson measure. Hence we have that

ż1
dµptq
|f ptq| 2 À ||f ||Ap , 0 ă p ď 1.
0 p1 ´ tqα`1´ q

This means that Hµα : Ap Ñ Aq is bounded.

(b). If α ` 1 ´ 2q ă 0, by Lemma 2.4 and (3.4) we have
ż1
||Iµα`1 pf q||Aq À |f ptq|dµptq À ||f ||Ap .
0

This implies Hµα : Ap Ñ Aq is bounded.

(c). If α ` 1 ´ 2q “ 0, Lemma 2.4 and and (3.4) show that

ż1 ˆ ˙ 1q
e
||Hµα pf q||Aq À |f ptq| log dµptq.
0 1´t

Since µ is a 1q -logarithmic p2 -Carleson measure, it follows from Lemma 2.5 in [11]

` e
˘ 1q
that log 1´t dµptq is a p2 -Carleson measure. Hence,

ż1 ˆ ˙ 1q
e
|f ptq| log dµptq À ||f ||Ap .
0 1´t

This finishes the proof. □

Theorem 3.8. Suppose that 1 ď p ď 2 and α ą 1. Let µ be a positive Borel measure

2
on r0, 1q which is a 2´pp´1q
p
Carleson measure. Then Hµα is a bounded operator on
Ap if and only if µ is an α ` 1 Carleson measure.
2
Proof. If µ is a 2´pp´1q p
-Carleson measure, then Lemma 3.1 shows that Hµα is well
defined on Ap and Hµα pf q “ Iµα`1 pf q for all f P Ap p1 ď p ď 2q.
If p “ 1, it follows from Theorem 3.2 that Hµα is a bounded operator on A1 if and
only if µ is an α ` 1-Carleson measure.
If p “ 2 and µ is an α ` 1 Carleson measure. Let f pzq “ 8
ř n 2
n“0 an z P A , then
|an |2
||f ||2A2 “ 8
ř
n“0 n`1 . Since µ is an α ` 1 Carleson measure, we have

1
|µn,k | “ |µn`k | À . (3.5)
pn ` k ` 1qα`1
11
Using (3.5) and the Hilbert’s inequality, we have
ˇ ˇ2
8
ÿ ˇÿ 8 ˇ
||Hµα pf q||2A2 pn ` 1q2α´1 ˇ µn,k ak ˇ
ˇ ˇ
—
n“0
ˇ k“0
ˇ
˜ ¸2
8 8
ÿ ÿ |a k |
À pn ` 1q2α´1
n“0 k“0
pn ` k ` 1qα`1
˜ ¸2
8 8
ÿ ÿ |ak |
ď 3
n“0 k“0 pn ` k ` 1q
2

˜ ¸2
8 8
ÿ ÿ |ak |
ď 1
n“0 k“0 pn ` k ` 1qpk ` 1q
2

8
ÿ |ak |2
À .
k“0
k ` 1

The last step above used the Hilbert’s inequality. Thus Hµα is a bounded operator
on A2 . The complex interpolation theorem (see Theorem 2.34 in [33]) shows that
2
Ap “ rA1 , A2 sθ , if 1 ă p ă 2 and θ “ 2 ´ .
p

This implies that Hµα is a bounded operator on Ap p1 ď p ď 2q. The proof of the
necessity is similar to that of Theorem 3.2, we omit the details here. □

In [10], the authors proved that Hµ is a Hilbert-Schmidt operator on A2 if and

only if
ż1
µprt, 1qq e
2
log dµptq ă 8.
0 p1 ´ tq 1´t
We recall that an operator T on a separable Hilbert space X is a Hilbert-Schmidt
operator if for an orthonormal basis tek : k “ 0, 1, 2, ¨ ¨ ¨ u of X the sum 8
ř 2
k“0 ||T pek q||X
is finite. The finiteness of this sum does not depend on the basis chosen. Here, we
consider the analogous problem on A2 for α ‰ 0.

Theorem 3.9. Let µ be a positive Borel measure on r0, 1q which is a 21 Carleson

measure. Then the following statements hold.
(a) If α ą 0, then Hµα is a Hilbert-Schmidt operator on A2 if and only if
ż1
µprt, 1qq
dµptq ă 8. (3.6)
0 p1 ´ tq2`2α

(b) If ´1 ă α ă 0, then Hµα is a Hilbert-Schmidt operator on A2 .

12
Proof. (a). Note that Hµα pf q is well defined on A2 by the assumption of the measure
1
µ. Take the orthonormal basis tek ukě0 “ tpk ` 1q 2 z k ukě0 , we have that
8
ÿ 8
ÿ 8
ÿ
||Hµα pek q||2A2 — pk ` 1q pn ` 1q2α´1 µ2n,k
k“0 k“0 n“0
ÿ8 ÿ8 ż1ż1
2α´1
“ pk ` 1q pn ` 1q ptsqn`k dµpsqdµptq
k“0 n“0 0 0
ż1ż1
1
— dµpsqdµptq
0 0 p1 ´ tsq2`2α
ż1ż1
1
“2 dµpsqdµptq
0 t p1 ´ tsq2`2α
ż1
µprt, 1qq
— dµptq.
0 p1 ´ tq2`2α

So the operator Hµα is Hilbert-Schmidt on A2 if and only if (3.6) holds.

(b). If ´1 ă α ă 0, then 8
ř 2α´1
n“0 pn`1q — 1. The rest of the proof is obvious. □

Corollary 3.10. Let µ be a positive Borel measure on r0, 1q. If Hµα is a Hilbert-
1
Schmidt operator on A2 for some α ą ´1, then for any ´1 ă α1 ă α, Hµα is a
Hilbert-Schmidt operator on A2 .

Remark 3.11. It is not difficult to obtain the results of compactness in Theorem

3.2, Theorem 3.7 and Theorem 3.8. They are vanishing Carleson type measures.
Note that the boundedness (resp. compactness) of Hµ acting from H p p0 ă p ď 1q to
H 1 has been studied in [4, 10]. The proofs of Theorem 3.2 and Theorem 3.7 can be
applied to study the operator Hµα acting from H p p0 ă p ď 1q into H q pq ě 1q, or Hµα
acting from H p p0 ă p ď 1q into Aq pq ě 1q.

4. Generalized Hilbert operator acting from Ap to B

The Bloch space B can be viewed as the limit case of Aq as q Ñ `8. In order to
obtain the boundedness of Hµα : Ap Ñ B for all α ą ´1 and 0 ă p ă 8, let us do
some preparations.
In [15], a sequence tVn u was constructed in the following way: Let ψ be a C 8 -
function on R such that (1) ψpsq “ 1 for s ď 1, (2) ψpsq “ 0 for s ě 2, (3) ψ is
decreasing and positive on the interval p1, 2q.
Let φpsq “ ψp 2s q ´ ψpsq, and let v0 “ 1 ` z, for n ě 1,

8 2n`1
ÿ´1
ÿ k k k
Vn pzq “ φp qz “ φp qz k .
k“0
2n´1 k“2n´1
2n´1
13
The polynomials Vn have the properties:
8
ÿ
(1) f pzq “ Vn ˚ f pzq , for f P HpDq;
n“0
(2) ||Vn ˚ f ||p À ||f ||p , for f P H p , p ą 0;
1
(3) ||Vn ||p — 2np1´ p q , for all p ą 0, where ˚ denotes the Hadamard product and || ¨ ||p
denotes the norm of Hardy space H p .
The following lemma can be found in [21, Theorem 3.1 ].
Lemma 4.1. Let f P HpDq, then f P B if and only if
sup ||Vn ˚ f ||8 ă 8.
ně0

Moreover,
||f ||B — sup ||Vn ˚ f ||8 .
ně0

The sublinear generalized integral type Hilbert operator Irµα`1 is defined by

ż1
|f ptq|
Irµα`1 pf qpzq “ α`1
dµptq, pα ą ´1q.
0 p1 ´ tzq

Theorem 4.2. Let 0 ă p ă 8 and α ą ´1. Suppose µ is a positive Borel measure

on r0, 1q and satisfies the conditions of Lemma 3.1. Then the following statements
are equivalent:
(a) Hµα : Ap Ñ B is bounded;
(b) Irµα`1 : Ap Ñ B is bounded;
(c) µ is a p2 ` α ` 1 Carleson measure.

Proof. paq ñ pcq. A similar discussion to the proof of Theorem 3.2 implies that
Hµα : Ap Ñ B is bounded if and only if
ˇż 1 ˇ
ˇ 0,α´1
ˇ
ˇ R gptqf ptqdµptqˇ À ||f ||Ap ||g||A1 (4.1)
ˇ ˇ
0

for all f P Ap , g P A1 .
Let β ą 0, take the test functions
p1 ´ a2 qβ p1 ´ a2 qβ
fa pzq “ β` p2
and ga pzq “ , a P p0, 1q.
p1 ´ azq p1 ´ azqβ`2
Then fa P Ap , ga P A1 and
sup ||fa ||Ap — 1, sup ||ga ||A1 — 1.
0ăaă1 0ăaă1

Note that
8
ÿ Γpn ` 1 ` αqΓpn ` β ` 2q
R0,α´1 ga ptq “ p1 ´ a2 qβ patqn .
n“0
Γpn ` 2qΓpα ` 1qΓpn ` 1qΓpβ ` 2q
14
By Stirling’s formula we have that
p1 ´ a2 qβ
R0,α´1 ga ptq — . (4.2)
p1 ´ atqβ`α`1
Using (4.1) and (4.2), we have
ˇż 1 ˇ
ˇ 0,α´1
ˇ
1 Á sup ||f ||Ap sup ||g||A1 Á ˇˇ R ga ptqfa ptqdµptqˇˇ
0ăaă1 0ăaă1 0
ż1 β
p1 ´ aq
Á R0,α´1 ga ptq 2 dµptq
a p1 ´ atqβ` p
µpra, 1qq
Á 2 .
p1 ´ a2 q p `α`1
2
This implies that µ is a p
` α ` 1-Carleson measure.
dµptq
pcq ñ pbq. Assume (c), then 2 is an α ` 1-Carleson measure by Lemma 2.5.
p1´tq p
Therefore, for each n P N, we have
ż1
dµptq 1
tn 2 “ Op α`1 q. (4.3)
0 p1 ´ tq p n

For every 0 ı f P Ap , our assumptions guarantee that Irµα`1 pf q converges absolutely

for every z P D and
ż1 8
|f ptq| ÿ
Iµα`1 pf qpzq “
r
α`1
dµptq “ cn z n ,
0 p1 ´ tzq n“0

where ż1
Γpn ` 1 ` αq
cn “ tn |f ptq|dµptq.
Γpn ` 1qΓpα ` 1q 0
Obviously, tcn u8
n“1 is a nonnegative sequence. Theorem 3.3.1 in [26] or [23, Corollary
3.2] shows that
n
1 ÿ
||Iµα`1 pf q||B — |c0 | ` sup
r kck . (4.4)
ně1 n
k“1

Using Stirling’s formula, (3.1) and (4.3), we deduce that

n n
1 ÿ α`1 1 k dµptq
ż
1 ÿ
|c0 | ` sup kck À ||f ||Ap ` ||f ||Ap sup k t 2
ně1 n ně1 n 0 p1 ´ tq p
k“1 k“1
n
1 ÿ α`1 1
À ||f ||Ap ` ||f ||Ap sup k
ně1 n k“1 k α`1
À ||f ||Ap .

Hence, Irµα`1 : Ap Ñ B is bounded.

15
 pbq ñ paq. The assumption of the measure µ guarantee that Hµα pf q “ Iµα`1 pf q
for every f P Ap . If Irµα`1 : Ap Ñ B is bounded, for each f P Ap , by Lemma 4.1 we
get
sup ||Vn ˚ Irµα`1 pf q||8 — ||Irµα`1 pf q||B À ||f ||Ap ||Irµα`1 ||.
ně1

Since the coefficients of Irµα`1 pf q are non-negative, it is easy to check that

M8 pr, Vn ˚ Iµα`1 pf qq ď M8 pr, Vn ˚ Irµα`1 pf qq for all 0 ă r ă 1.
Therefore,
||Vn ˚ Iµα`1 pf q||8 “ sup M8 pr, Vn ˚ Iµα`1 pf qq ď ||Vn ˚ Irµα`1 pf q||8 .
0ără1

Consequently,
||Hµα pf q||B “ ||Iµα`1 pf q||B — sup ||Vn ˚ Iµα`1 pf q||8 À ||f ||Ap .
ně1

This implies that Hµα p

: A Ñ B is bounded. □

Remark 4.3. The assumption of the measure µ also insures that Irµα`1 is well
defined on Ap for all 0 ă p ă 8.
Theorem 4.4. Let 0 ă p ă 8 and α ą ´1. Suppose µ is a positive Borel measure
on r0, 1q and satisfies the conditions of Lemma 3.1. Then the following statements
are equivalent:
(a) Hµα : Ap Ñ B is compact;
(b) Irµα`1 : Ap Ñ B is compact;
(c) µ is a vanishing p2 ` α ` 1 Carleson measure.
Proof. It is obvious that pbq ñ paq.
paq ñ pcq. Let tan u Ă r0, 1q be any sequence with an Ñ 1. Take
p1 ´ a2n qβ
fn pzq “ 2 , where β ą 0.
p1 ´ an zqβ` p
Then tfn u Ă Ap , supně1 ||fn ||Ap — 1 and tfn u converges to 0 uniformly on every
compact subset of D. Using Lemma 2.6 and the fact that Hµα : Ap Ñ B is compact,
we have that
lim ||Hµα pfn q||B “ 0.
nÑ8
This and (4.1) imply that
ˇż 1 ˇ
ˇ 0,α´1
ˇ
lim ˇ R
ˇ gptqfn ptqdµptqˇˇ “ 0 for all g P A1 .
nÑ8 0

Now, let
p1 ´ a2n qβ
gn pzq “ .
p1 ´ an zqβ`2
16
It is clear that tgn u Ă A1 . Thus, using (4.2) we have
ˇż 1 ˇ
ˇ ˇ
ˇ R0,α´1 gn ptqfn ptqdµptqˇ
ˇ ˇ
0
ż1
dµptq
— p1 ´ a2n q2β 2
0 p1 ´ an tq2β`α`1` p
ż1
dµptq
Á p1 ´ a2n q2β 2
an p1 ´ an tq2β`α`1` p
µpran , 1qq
Á 2 .
p1 ´ an qα`1` p
This together with tan u is an arbitrary sequence on r0, 1q we have that
µprt, 1qq
lim´ 2 “ 0.
tÑ1 p1 ´ tqα`1` p
2
Hence, µ is a vanishing p
` α ` 1 Carleson measure.
dµptq
pcq ñ pbq. Assume pcq, then 2 is a vanishing α ` 1-Carleson measure by
p1´tq p
Lemma 2.5. Therefore, for any ε ą 0, there exists a integer N such that
ż1
α`1 dµptq
n tn 2 ă ε whenever n ą N. (4.5)
0 p1 ´ tq p
p
Let tfk u8
k“1 be a bounded sequence in A which converges to 0 uniformly on every
compact subset of D. For each k P N, we have
ż1 8
|fk ptq| ÿ
Irµα`1 pfk qpzq “ α`1
dµptq “ cn,k z n ,
0 p1 ´ tzq n“0

where
ż1
Γpn ` 1 ` αq
cn,k “ tn |fk ptq|dµptq.
Γpn ` 1qΓpα ` 1q 0

It is obvious that tcn,k u8

n“1 is a nonnegative sequence for each k P N. It is sufficient
to prove that
˜ ¸
n
1ÿ
lim c0,k ` sup jcj,k “0
kÑ8 ně1 n j“1

ş1 dµptq
by (4.4) and Lemma 2.6. Since 0 2 ă 8, there exists a 0 ă t0 ă 1 such that
p1´tq p
ż1
dµptq
2 ă ε. (4.6)
t0 p1 ´ tq p
17
Bearing in mind that tfk u8 k“1 converges to 0 uniformly on every compact subset of
D, so that there exists a integer K such that
sup |fk ptq| ă ε whenever k ą K. (4.7)
tPr0,t0 s

Using (4.1), (4.6) and (4.7) we have that

ż1
lim c0,k “ lim |fk ptq|dµptq “ 0. (4.8)
kÑ8 kÑ8 0

It is obvious that
n n n
1ÿ 1ÿ 1ÿ
sup jcj,k ď sup jcj,k ` sup jcj,k .
ně1 n j“1 1ďnďN n j“1 něN `1 n j“1

By Stirling’s formula and (4.8) we have

n n
1 ÿ α`1 1
ż ż1
1ÿ
sup jcj,k À sup j |fk ptq|dµptq À |fk ptq|dµptq Ñ 0, pk Ñ 8q.
1ďnďN n j“1 1ďnďN n j“1 0 0
(4.9)
On the other hand, by Stirling’s formula, (3.1) and (4.5) we have
n N
1 ÿ α`1 1 j
ż
1ÿ
sup jcj,k À sup j t |fk ptq|dµptq
něN `1 n j“1 něN `1 n j“1 0
n
1 ÿ α`1 1 j
ż
` sup j t |fk ptq|dµptq
něN `1 n j“N `1 0

N
1 ÿ α`1 1
ż
À sup j |fk ptq|dµptq
něN `1 n j“1 0
n
1 ÿ α`1 1 j dµptq
ż
` sup ||fk ||Ap sup j t 2
kě1 něN `1 n j“N `1 0 p1 ´ tq p
ż1 n
1 ÿ α`1 1
À |fk ptq|dµptq ` ε sup ||fk ||Ap sup j
0 kě1 něN `1 n j“N `1 j α`1
ż1
À |fk ptq|dµptq ` ε sup ||fk ||Ap .
0 kě1

This implies that

n
1ÿ
lim sup jcj,k “ 0. (4.10)
kÑ8 něN `1 n
j“1

Then (4.8)-(4.10) imply that

˜ ¸
n
1ÿ
lim c0,k ` sup jcj,k “ 0.
kÑ8 ně1 n j“1
18
Therefore, Irµα`1 : Ap Ñ B is compact. □
Declarations. The authors declare that there are no conflicts of interest regarding
the publication of this paper.
Availability of data and material. Data sharing not applicable to this article
as no datasets were generated or analysed during the current study: the article
describes entirely theoretical research.

References

[1] T. Aguilar-Hernández, P. Galanopoulos, D. Girela, Hilbert-type operators acting on Bergman

spaces, Comput. Methods Funct. Theory (2024). https://doi.org/10.1007/s40315-024-00560-5
[2] A. Aleman, A. Montes-Rodrı́guez, A. Sarafoleanu, The eigenfunctions of the Hilbert matrix,
Constr. Approx. 36 (3) (2012) 353–374.
[3] G. Bao, K. Guo, F. Sun, Z. Wang, Hankel matrices acting on the Dirichlet space, J. Fourier
Anal. Appl. 30 (2024) no. 53.
[4] C. Chatzifountas, D. Girela, J. Peláez, A generalized Hilbert matrix acting on Hardy spaces,
J. Math. Anal. Appl. 413 (2014) 154–168.
[5] C. Cowen, B. MacCluer, Composition operators on spaces of analytic functions, CRC Press,
Boca Raton, 1995.
[6] E. Diamantopoulos, A. Siskakis, Composition operators and the Hilbert matrix, Studia Math.
140 (2000) 191–198.
[7] E. Diamantopoulos, Hilbert matrix on Bergman spaces, Illinois J. Math. 48 (3) (2004) 1067–
1078.
[8] M. Dostanić, M. Jevtić, D. Vukotić, Norm of the Hilbert matrix on Bergman and Hardy spaces
and a theorem of Nehari type, J. Funct. Anal. 254 (2008) 2800–2815.
[9] P. Duren, A. Schuster, Bergman spaces, Mathematical Surveys and Monographs. American
Mathematical Society, Providence, R.I., 2004.
[10] P. Galanopoulos, J. Peláez, A Hankel matrix acting on Hardy and Bergman spaces, Studia
Math. 200 (3) (2010) 201–220.
[11] D. Girela, N. Merchán, A generalized Hilbert operator acting on conformally invariant spaces,
Banach J. Math. Anal. Appl. 12 (2018) 374–398.
[12] D. Girela, N. Merchán, A Hankel matrix acting on spaces of analytic functions, Integral Equ.
Oper. Theory. 89 (4) (2017) 581–594.
[13] D. Girela, N. Merchán, Hankel matrices acting on the Hardy space H 1 and on Dirichlet spaces,
Rev. Mat. Comput. 32 (2019) 799–822.
[14] M. Jevtić, B. Karapetrović, Generalized Hilbert matrices acting on spaces that are close to the
Hardy space H 1 and to the space BMOA, Complex Anal. Oper. Theory 13 (2019) 2357–2370.
[15] M. Jevtić, M. Pavlović, On multipliers from H p to lq p0 ă q ă p ă 1q, Arch. Math. 56 (1991)
174–180.
[16] S. Li, Generalized Hilbert operator on the Dirichlet-type space, Appl. Math. Comput. 214
(2009) 304–309.
[17] S. Li, S. Stević, Generalized Hilbert operator and Fejér-Riesz type inequalities on the polydisc,
Acta Math. Sci. 29(B) (2009) 191–200.
[18] M. Lindström, D. Norrbo, S. Stević, On compactness of operators from Banach spaces of
holomorphic functions to Banach spaces, J. Math. Inequal. 18(3) (2024) 1153–1158.
19
[19] M. Mateljević, M. Pavlović, Lp behaviour of the integral means of analytic functions, Studia
Math. 77 (1984) 219–237.
[20] N. Merchán, Mean Lipschitz spaces and a generalized Hilbert operator, Collect. Math. 70
(2019) 59–69.
[21] M. Pavlović, Logarithmic Bloch space and its predual, Publ. Inst. Math. (Beograd) (N.S.)
100 (114) (2016) 1–16.
[22] W. Rudin, Function theory in the unit ball of C n , Springer, New York, 1980.
[23] P. Tang, Hilbert type operators acting from the Bloch space into Bergman spaces, Proc. Edinb.
Math. Soc. (2024) 1–21. https://doi:10.1017/S0013091524000786.
[24] P. Tang, X. Zhang, Generalized integral type Hilbert operator acting between weighted Bloch
spaces, Math. Meth. Appl. Sci. 46 (2023) 18458–18472.
[25] H. Widom, Hankel matrices, Trans. Amer. Math. Soc. 121 (1966) 1–35.
[26] J. Xiao, Holomorphic Q Classes, Springer, Lecture Notes in Math, 2001.
[27] S. Ye, G. Feng, A Derivativ-Hilbert operator acting on Hardy spaces, Acta Math. Sci. 43B(6)
(2023) 2398–2412.
[28] S. Ye, G. Feng, Generalized Hilbert operators acting on weighted Bergman spaces and Dirichlet
spaces, Banach J. Math. Anal. 17 (2023) no. 38.
[29] S, Ye, Y. Xu, A Derivative-Hilbert operator acting from logarithmic Bloch spaces to Bergman
spaces, Acta Math. Sci. 44B(5) (2024) 1916–1930.
[30] S. Ye, Z. Zhou, A Derivative-Hilbert operator acting on the Bloch space, Complex Anal. Oper.
Theory. 15 (2021) no. 88.
[31] S. Ye, Z. Zhou, A Derivative-Hilbert operator acting on Bergman spaces, J. Math. Anal. Appl.
506 (2022) no. 125553.
[32] R. Zhao, On logarithmic Carleson measures, Acta Sci. Math. (Szeged) 69 (3–4) (2003) 605–
618.
[33] K. Zhu, Spaces of holomorphic functions in the unit ball, Springer-Verlag (GTM 226), New
York, 2005.

1
College of Mathematics and Statistics, Hunan University of Science and Tech-
nology, Xiangtan, Hunan 411201, China
2
College of Mathematics and Statistics, Hunan Normal University, Changsha,
Hunan 410006, China

20