You are on page 1of 87

MEMOIRS of the

American Mathematical Society

Number 974

Composition Operators
on Hardy-Orlicz Spaces
Pascal Lefèvre
Daniel Li
Hervé Queffélec
Luis Rodrı́guez-Piazza

September 2010 • Volume 207 • Number 974 (fourth of 5 numbers) • ISSN 0065-9266

American Mathematical Society
Number 974

Composition Operators
on Hardy-Orlicz Spaces
Pascal Lefèvre
Daniel Li
Hervé Queffélec
Luis Rodrı́guez-Piazza

September 2010 • Volume 207 • Number 974 (fourth of 5 numbers) • ISSN 0065-9266
Library of Congress Cataloging-in-Publication Data
Composition operators on Hardy-Orlicz spaces / Pascal Lefèvre ... [et al.].
p. cm. — (Memoirs of the American Mathematical Society, ISSN 0065-9266 ; no. 974)
“Volume 207, number 974 (fourth of 5 numbers).”
Includes bibliographical references.
ISBN 978-0-8218-4637-7 (alk. paper)
1. Composition operators. 2. Hardy spaces. I. Lefèvre, Pascal, 1969-.
QA329.2.C654 2010
515.7246—dc22 2010022778

Memoirs of the American Mathematical Society
This journal is devoted entirely to research in pure and applied mathematics.
Publisher Item Identifier. The Publisher Item Identifier (PII) appears as a footnote on
the Abstract page of each article. This alphanumeric string of characters uniquely identifies each
article and can be used for future cataloguing, searching, and electronic retrieval.
Subscription information. Beginning with the January 2010 issue, Memoirs is accessi-
ble from www.ams.org/journals. The 2010 subscription begins with volume 203 and consists of
six mailings, each containing one or more numbers. Subscription prices are as follows: for paper
delivery, US$709 list, US$567 institutional member; for electronic delivery, US$638 list, US$510 in-
stitutional member. Upon request, subscribers to paper delivery of this journal are also entitled to
receive electronic delivery. If ordering the paper version, subscribers outside the United States and
India must pay a postage surcharge of US$65; subscribers in India must pay a postage surcharge of
US$95. Expedited delivery to destinations in North America US$57; elsewhere US$160. Subscrip-
tion renewals are subject to late fees. See www.ams.org/customers/macs-faq.html#journal for
more information. Each number may be ordered separately; please specify number when ordering
an individual number.
Back number information. For back issues see www.ams.org/bookstore.
Subscriptions and orders should be addressed to the American Mathematical Society, P. O.
Box 845904, Boston, MA 02284-5904 USA. All orders must be accompanied by payment. Other
correspondence should be addressed to 201 Charles Street, Providence, RI 02904-2294 USA.
Copying and reprinting. Individual readers of this publication, and nonprofit libraries
acting for them, are permitted to make fair use of the material, such as to copy a chapter for use
in teaching or research. Permission is granted to quote brief passages from this publication in
reviews, provided the customary acknowledgment of the source is given.
Republication, systematic copying, or multiple reproduction of any material in this publication
is permitted only under license from the American Mathematical Society. Requests for such
permission should be addressed to the Acquisitions Department, American Mathematical Society,
201 Charles Street, Providence, Rhode Island 02904-2294 USA. Requests can also be made by
e-mail to reprint-permission@ams.org.

Memoirs of the American Mathematical Society (ISSN 0065-9266) is published bimonthly (each
volume consisting usually of more than one number) by the American Mathematical Society at
201 Charles Street, Providence, RI 02904-2294 USA. Periodicals postage paid at Providence, RI.
Postmaster: Send address changes to Memoirs, American Mathematical Society, 201 Charles
Street, Providence, RI 02904-2294 USA.

c 2010 by the American Mathematical Society. All rights reserved.
Copyright of individual articles may revert to the public domain 28 years
after publication. Contact the AMS for copyright status of individual articles.
This publication is indexed in Science Citation Index R , SciSearch 
R , Research Alert 
R,

CompuMath Citation Index  R , Current Contents 
R /Physical, Chemical & Earth Sciences.

Printed in the United States of America.

∞ The paper used in this book is acid-free and falls within the guidelines
established to ensure permanence and durability.
Visit the AMS home page at http://www.ams.org/
10 9 8 7 6 5 4 3 2 1 15 14 13 12 11 10
Contents

Chapter 1. Introduction 1
Chapter 2. Notation 5
1. Growth conditions 5
2. Some specific functions 7
Chapter 3. Composition operators on Hardy-Orlicz spaces 9
1. Hardy-Orlicz spaces 9
2. Preliminary results 13
3. Composition operators 15
4. Order bounded composition operators 17
5. Weakly compact composition operators 21
6. p-summing operators 26
Chapter 4. Carleson measures 33
1. Introduction 33
2. Compactness on H Ψ versus compactness on H 2 36
3. General measures 38
4. Characterization of the compactness of composition operators 55
Chapter 5. Bergman spaces 63
1. Bergman-Orlicz spaces 63
2. Compact composition operators on Bergman-Orlicz spaces 66
Acknowledgement 71
References 73

iii
Abstract

We investigate composition operators on Hardy-Orlicz spaces when the Orlicz
function Ψ grows rapidly: compactness, weak compactness, to be p-summing, order
bounded, . . . , and show how these notions behave according to the growth of Ψ. We
introduce an adapted version of Carleson measure. We construct various examples
showing that our results are essentially sharp. In the last part, we study the case
of Bergman-Orlicz spaces.

Received by the editor June 26, 2006.
Article electronically published on March 17, 2010; S 0065-9266(10)00580-6.
2000 Mathematics Subject Classification. Primary: 47B33, 46E30.
Key words and phrases. Bergman-Orlicz space, Carleson measure, composition operator,
Hardy-Orlicz space.

2010
c American Mathematical Society

v
CHAPTER 1

Introduction

Composition operators on the classical Hardy spaces H p have been widely
studied (see [39], and [10], and references therein; see also [18] and [19], and [6],
[9], [17], [31], [36], [40], [41] for some more recent works), but it seems that one
has not paid much attention to the Hardy-Orlicz spaces (in [43] and [44], J.-O.
Strömberg studied Hardy-Orlicz spaces in the case when the Orlicz function Ψ
increases smoothly; see also [29] for composition operators). We shall investigate
what happens when the Orlicz function grows more rapidly than a power function.
Recall that, given an analytic self-map φ : D → D of the unit disk D, the
composition operator associated to φ is the map Cφ : f → f ◦ φ. This map may
operate on various Banach spaces X of analytic functions on D (Hardy spaces,
Bergman spaces, . . . , and their weighted versions (see [45] for instance), Bloch
spaces B and B0 , BM OA and V M OA, Dirichlet spaces (see [1]), or some more
general spaces as Nevanlinna or Smirnov classes: [7], [8], [16], [20]; see also [2],
[3], [13] for composition operators on H p spaces of Dirichlet series, though they
are not induced by an analytic self-map of D). The goal is to link properties of the
composition operator Cφ : X → X (compactness, strong or weak, for example) to
properties of the symbol φ (essentially its behaviour near the frontier of D). For
that study, one can roughly speaking (see [10], Chapter 4, though their notions are
different from ours) distinguish two kind of spaces.
1) The small spaces X; those spaces are in a sense close to the Hardy space
H ∞ : the compactness of Cφ : X → X is very restrictive and it imposes severe
restrictions on φ. For example, if X = H ∞ , a theorem of H.J. Schwartz ([37])
implies that Cφ : H ∞ → H ∞ is compact if and only if φ∞ < 1 (weakly compact
suffices: see [23]), which in turn implies reinforced compactness properties for Cφ .
For example, Cφ : H ∞ → H ∞ is nuclear and 1-summing as soon as it is compact.
2) The large spaces X; those spaces are in a sense close to the Hardy space H 1 :
the compactness of Cφ : X → X can take place fairly often, and in general implies
no self-improvement. For example, for X = H 2 , Cφ : H 2 → H 2 can be compact
without being Hilbert-Schmidt, even if φ is injective ([42], Theorem 6.2). Another
formulation (which lends better to generalizations in the non-Hilbertian case) is
that Cφ can be non-order bounded (see [15], and our Chapter 3) and yet compact.
In this paper, we shall rather be on the small space side, since we shall work
in spaces associated to a very large Orlicz function Ψ (typically: Ψ(x) = Ψ2 (x) =
2
ex − 1), and the previous situation will not take place: our operators will be e.g.
order-bounded as soon as they are (weakly) compact, even if the situation is not
so extreme as for H ∞ . However, for slightly smaller Orlicz functions (for instance
 2 
Ψ(x) = exp log(x + 1) − 1), the situation is closer to the H 2 case: the compo-
sition operators may be compact on H Ψ , but not order-bounded (Theorem 4.22).

1
2 PASCAL LEFÈVRE, DANIEL LI, HERVÉ QUEFFÉLEC, LUIS RODRÍGUEZ-PIAZZA

This paper is divided into five parts. Chapter 1 is this Introduction. In Chapter
2, which is essentially notational, we recall some more or less standard facts on
Orlicz functions Ψ, on associated Orlicz spaces LΨ , and the “little” Orlicz space
M Ψ , and their Banachic properties, associated with various slow growth conditions
(indicated by subscripts: Δ1 , Δ2 , . . . ) or fast growth conditions (indicated by
superscripts: Δ0 , Δ1 , Δ2 , . . . ).
In Chapter 3, we introduce the Hardy-Orlicz space H Ψ , and its (or his) “little
brother” HM Ψ . These spaces have already been studied (see [14], [34]), but rather
for slowly growing functions Ψ (having Δ2 most of the time), and their definition
is not so clearly outlined, so we give a detailed exposition of the equivalence of the
two natural definitions that one has in mind (if one wants to extend the case of
Hardy spaces H p associated to Ψ(x) = xp ), as well as of the automatic boundedness
(through the Littlewood subordination principle and the case of inner self-maps of
the disk) of composition operators on those spaces. Two of the main theorems
are Theorem 3.24 and Theorem 3.27. Roughly speaking, Theorem 3.24 says the
following: if Ψ is very fast growing (having Δ2 more precisely), H Ψ is a small
space, the (weak) compactness of Cφ is very restrictive, and even if the situation
is not so extreme as for H ∞ (φ∞ < 1), φ has to tend to the boundary very
slowly, and Cφ is automatically order-bounded into M Ψ . However, Theorem 3.27
shows the limits of this self-improvement: Cφ may be order-bounded into M Ψ (and
hence compact), but p-summing for no finite p. We also show that, when Ψ has
Δ2 growth, there are always symbols φ inducing compact composition operators on
H 2 (even Hilbert-Schmidt), but not compact on H Ψ .
Chapter 4 is devoted to the use of Carleson measures. The usefulness of those
measures in the study of composition operators is well-known (see [11], [10], [5])
and, to our knowledge, first explicitly used for compactness in [30]. In particular,
we recall the following necessary and sufficient condition for Cφ : H 2 → H 2 to be
compact: if h > 0 and w ∈ ∂D, consider the Carleson window :
W (w, h) = {z ∈ D ; |z| ≥ 1 − h and | arg(z w̄)| ≤ h}.
If φ is an analytic self-map of D with boundary values φ∗ , and μφ = φ∗ (m) denotes
the image under φ∗ of the normalized Lebesgue measure (Haar measure) on T = ∂D,
the measure μφ is always a Carleson measure, i.e.:
 
sup μφ W (w, h) = O (h).
w∈∂D
Now, we can state:
Theorem 1.1 (B. MacCluer [30]). The composition operator Cφ : H 2 → H 2 is
compact if and only if μφ satisfies the “ little-oh” condition, i.e. if and only if:
 
(1.1) sup μφ W (w, h) = o (h) as h → 0.
w∈∂D

There is another famous necessary and sufficient compactness condition, due
to J. Shapiro ([38]): Let us denote by Nφ the Nevanlinna counting function of φ,
i.e.:   1
log |z| if w = φ(0) and w ∈ φ(D)
Nφ (w) = φ(z)=w
0 if w = φ(0) or w ∈/ φ(D).
By Littlewood’s inequality, one always has (see [10], page 33):
Nφ (w) = O (1 − |w|).
1. INTRODUCTION 3

Now, Shapiro’s Theorem reads:

Theorem 1.2 (J. Shapiro [38]). The composition operator Cφ : H 2 → H 2 is
compact if and only if Nφ satisfies the “ little-oh” condition, i.e.:

(1.2) Nφ (w) = o (1 − |w|) as |w| → 1− .

Theorem 1.2 is very elegant, and probably more “popular” than Theorem 1.1.
Yet, it is difficult to apply because the assumption (1.2) is difficult to check. Here,
we shall appeal to Theorem 1.1 to prove (Theorem 4.1) that the compactness of
Cφ : H 2 → H 2 cannot be read on |φ∗ | when φ is not finitely valent; more precisely,
there are two analytic self-maps φ1 and φ2 : D → D such that: |φ∗1 | = |φ∗2 | m-a.e.,
but Cφ1 : H 2 → H 2 is not compact, though Cφ2 : H 2 → H 2 is compact.
We show then that every composition operator which is compact on H Ψ is
necessarily compact on H p for all p < ∞. However, there exist (see above, or
Chapter 3) symbols φ inducing compact composition operators on H 2 but which
are not compact on H Ψ , when Ψ has Δ2 growth. Hence condition (1.1) does not
suffice to characterize the compact composition operators on H Ψ . We have to
replace Carleson measures and condition (1.1) by what we may call “Ψ-Carleson
measures”, and an adaptated “little-oh” condition, which allows us to characterize
compactness for composition operators. It follows that if Ψ ∈ Δ0 , then the weak
compactness of Cφ : H Ψ → H Ψ implies its compactness.
We also show that the above example φ2 induces, for an Orlicz function Ψ which
does not satisfy Δ2 , but which satisfies Δ1 , a composition operator on H Ψ which
is compact, but not order bounded into M Ψ (T) (Theorem 4.22), showing that the
assumption that Ψ ∈ Δ2 in Theorem 3.24 is not only a technical assumption.
In Chapter 5, we introduce the Bergman-Orlicz spaces. Let us remind that, in
the Hilbertian case, the study of compactness of composition operators is simpler
for the Bergman space B 2 than for the Hardy space H 2 . For example, we have the
following:

Theorem 1.3 (see [39]).
i) Cφ : B 2 → B 2 is compact if and only if

1 − |φ(z)|
(1.3) lim = +∞.
|z|→1− 1 − |z|

ii) (1.3) is always necessary for Cφ : H 2 → H 2 to be compact, and it is sufficient
when φ is injective, or only finitely valent.
iii) There are Blaschke products φ satisfying (1.3) for which Cφ : H 2 → H 2 is
(in an obvious manner) non-compact.

We perform here a similar study for the Bergman-Orlicz space B Ψ , and com-
pare the situation with that of the Hardy-Orlicz space H Ψ . We are naturally led
to a reinforcement of (1.3) under the form:

1
Ψ−1
(1 − |φ(a)|)2
(1.4)  −→ 0
−1
1 |a|→1−
Ψ
(1 − |a|)2
4 PASCAL LEFÈVRE, DANIEL LI, HERVÉ QUEFFÉLEC, LUIS RODRÍGUEZ-PIAZZA

(always necessary, and sufficient when Ψ is Δ2 ), which reads, in the case Ψ(x) =
2
Ψ2 (x) = ex − 1:
(1.5) 1 − |φ(z)| ≥ cε (1 − |z|)ε for all ε > 0.
In [39], the construction of a Blaschke product satisfying (1.3) is fairly delicate, and
appeals to Frostman’s Lemma and Julia-Caratheodory’s Theorem on non-angular
derivatives at the boundary. Here, we can no longer use these tools for the rein-
forcement (1.5), so we do make a direct construction, using the Parseval formula
for finite groups. In passing, the construction gives a simpler proof for (1.3). Oth-
erwise, the theorem which we obtain is similar to Shapiro’s one, if one ignores some
technical difficulties due to the non-separability of B Ψ : we have to “transit” by the
smaller Bergman-Morse-Transue space BM Ψ , which is the closure of H ∞ in B Ψ ,
is separable, and has B Ψ as its bidual.
CHAPTER 2

Notation

Let D be the open unit disk of the complex plane, that is the set of complex
numbers with modulus strictly less than 1, and T the unit circle, i.e. the set of
complex numbers with modulus 1.
We shall consider in this paper Orlicz spaces defined on a probability space
(Ω, P), which will be the unit circle T, with its (normalized) Haar measure m
(most often identified with the normalized Lebesgue measure dx/2π on the interval
[0, 2π]), or the open unit disk D, provided with the normalized area measure A .
By an Orlicz function, we shall understand that Ψ : [0, ∞] → [0, ∞] is a non-
decreasing convex function such that Ψ(0) = 0 and Ψ(∞) = ∞. To avoid patholo-
gies, we shall assume that we work with an Orlicz function Ψ having the following
additional properties: Ψ is continuous at 0, strictly convex (hence increasing), and
such that
Ψ(x)
−→ ∞.
x x→∞
This is essentially to exclude the case of Ψ(x) = ax.
If Ψ 
is the left (or instead, if one prefers, the right) derivative of Ψ, one has
x
Ψ(x) = 0 Ψ (t) dt for every x > 0.
The Orlicz space LΨ (Ω) is the space of all (equivalence classes of) measurable
functions f : Ω → C for which there is a constant C > 0 such that
|f (t)|
Ψ dP(t) < +∞
Ω C
and then f Ψ (the Luxemburg norm) is the infimum of all possible constants C
such that this integral is ≤ 1. The Morse-Transue space M Ψ (Ω) is the subspace
generated by L∞ (Ω), or, equivalently, the subspace of all functions f for which the
above integral is finite for all C > 0.
To every Orlicz function is associated the complementary Orlicz function Φ =
Ψ∗ : [0, ∞] → [0, ∞] defined by:
 
Φ(x) = sup xy − Ψ(y) ,
y≥0

The extra assumptions on Ψ ensure that Φ is itself strictly convex.
When Φ satisfies the Δ2 condition (see the definition below), LΨ is (isomorphi-
cally, if LΦ is itself normed by the Luxemburg norm) the dual space of LΦ , which
is, in turn, the dual of M Ψ .

1. Growth conditions
We shall have to use various growth conditions for the Orlicz function Ψ. These
conditions are usually denoted as Δ-conditions. Our interest is in Orlicz functions
which have a somewhat fast growth. Usually, some of these conditions are defined
5
6 PASCAL LEFÈVRE, DANIEL LI, HERVÉ QUEFFÉLEC, LUIS RODRÍGUEZ-PIAZZA

through a moderate growth condition on the complementary function Φ of Ψ, and
the condition Δ for the Orlicz function is translated as a ∇-condition for the com-
plementary function. So we shall distinguish between moderate growth conditions,
that we shall define for the complementary Orlicz function, and fast growth condi-
tions. To emphasize this distinction, we shall denote, sometimes in changing the
usual notation (see [22, 33]), the moderate growth conditions with a subscript, and
the fast growth conditions with a superscript.
Moderate growth conditions
• The Orlicz function Φ satifies the Δ1 -condition (Φ ∈ Δ1 ) if, for some
constant c > 0, one has:
Φ(xy) ≤ c Φ(x)Φ(y)
for x, y large enough.
This is equivalent to say that
Φ(axy) ≤ Φ(x)Φ(y)
for some constant a > 0 and x, y large enough.
This condition is usually denoted by Δ (see [33], page 28).
• Φ satisfies the Δ2 -condition (Φ ∈ Δ2 ) if
Φ(2x) ≤ K Φ(x)
for some constant K > 1 and x large enough.
One has:
Φ ∈ Δ1 ⇒ Φ ∈ Δ2 .
Fast growth conditions
• The Orlicz function Ψ satisfies the Δ0 -condition (Ψ ∈ Δ0 ) if (see [25]),
for some β > 1:
Ψ(βx)
lim = +∞.
x→+∞ Ψ(x)
 
A typical example is Ψ(x) = exp log(x + 2) log log(x + 2) − 2log log 2 ;
 
another is Ψ(x) = exp log(x + 1)3/2 − 1.
• The Orlicz function Ψ satisfies the Δ1 -condition (Ψ ∈ Δ1 ) if there is some
β > 1 such that:
xΨ(x) ≤ Ψ(βx)
for x large enough.
Note that this latter condition is usually written as Δ3 -condition, with a subscript
(see [33], §2.5). This notation fits better with our convention, and the superscript
1 agrees with the fact that this Δ1 -condition is between the Δ0 -condition and the
following Δ2 -condition. Ψ ∈ Δ1 implies that
 
Ψ(x) ≥ exp α (log x)2
for some α > 0 and x large enough (see [33], Proposition 2, page 37). A typical
2
example is Ψ(x) = e(log(x+1)) − 1.
2. SOME SPECIFIC FUNCTIONS 7

• The Orlicz function Ψ : [0, ∞) → [0, ∞) is said to satisfy the Δ2 -condition
(Ψ ∈ Δ2 ) if there exists some α > 1 such that:
Ψ(x)2 ≤ Ψ(αx)
for x large enough.
This implies that
Ψ(x) ≥ exp(xα )
for some α > 0 and x large enough ([33], Proposition 6, page 40). A typical example
2
is Ψ(x) = Ψ2 (x) = ex − 1.
Conditions of regularity
• The Orlicz function Ψ satisfies the ∇2 -condition (Ψ ∈ ∇2 ) if its comple-
mentary function Φ satisfies the Δ2 -condition.
This is equivalent to say that for some constant β > 1 and some x0 > 0,
one has Ψ(βx) ≥ 2β Ψ(x) for x ≥ x0 , and that implies that Ψ(x)
x −→ ∞.
x→∞
In particular, this excludes the case LΨ = L1 .
• The Orlicz function Ψ satisfies the ∇1 -condition (Ψ ∈ ∇1 ) if its comple-
mentary function Φ satisfies the Δ1 -condition.
This is equivalent to say that
Ψ(x)Ψ(y) ≤ Ψ(bxy)
for some constant b > 0 and x, y large enough.
All power functions Ψ(x) = xp satisfy ∇1 , but Ψ(x) = xp log(x + 1) does
not.
One has (see [33], page 43):
+3 Ψ ∈ Δ0
7? ∈ Δ
1
Ψ FFFFF
x
xxxxx FFFF
x
xx FFF
xxxxx &
Ψ ∈ Δ SS
2
:2 Ψ ∈ ∇2
SSS
SSSS lllll
ll
SS %- lll
Ψ ∈ ∇1
But Δ1 does not imply ∇1 . That ∇1 does not even imply Δ0 is clear since any
power function Ψ(x) = xp (p ≥ 1) is in ∇1 .

2. Some specific functions
In this paper, we shall make a repeated use of the following functions:
• If Ψ is an Orlicz function, we set, for every K > 0:
 
(2.1) χK (x) = Ψ KΨ−1 (x) , x > 0.
For example, if Ψ(x) = ex − 1, then Ψ−1 (x) = log(1 + x), and χK (x) =
(1 + x)K − 1.
Note that:
χβ (u)
– Ψ ∈ Δ0 means that −→ +∞, for some β > 1.
u u→∞
– Ψ ∈ Δ means that χβ (u) ≥ uΨ−1 (u) for u large enough, for some
1

β > 1.
– Ψ ∈ Δ2 means that χα (u) ≥ u2 for u large enough, for some α > 1.
8 PASCAL LEFÈVRE, DANIEL LI, HERVÉ QUEFFÉLEC, LUIS RODRÍGUEZ-PIAZZA

– Ψ ∈ ∇1 means that χA (u) ≥ (Ψ(A)/b) u for u large enough and for
every A large enough, for some b > 0.
• For |a| = 1 and 0 ≤ r < 1, ua,r is the function defined on the unit disk D
by:
1 − r 2
(2.2) ua,r (z) = , |z| < 1.
1 − ārz
Note that ua,r ∞ = 1 and ua,r 1 ≤ 1 − r.
CHAPTER 3

Composition operators on Hardy-Orlicz spaces

1. Hardy-Orlicz spaces
It is well-known that the classical H p spaces (1 ≤ p ≤ ∞) can be defined in
two equivalent ways:
1) H p is the space of analytic functions f : D → C for which, setting fr (t) =
f (reit ):
f H p = sup fr p
0≤r<1

is finite (recall that the numbers fr p increase with r). When f ∈ H p , the Fatou-
Riesz Theorem asserts that the boundary limits f ∗ (t) = limr→1− fr (t) exist almost
everywhere and f H p = f ∗ p . One has f ∗ ∈ Lp ([0, 2π]), and its Fourier coeffi-
cients f∗ (n) vanish for n < 0.
2) Conversely, for every function g ∈ Lp ([0, 2π]) whose Fourier coefficients 
vanish nfor n < p0, the analytic extension P [g] : D →∗ C defined by P [g](z) =
ĝ(n)
n≥0 ĝ(n)z is in H and g is the boundary limit (P [g]) of P [g].
Hardy-Orlicz spaces H Ψ are defined in a similar way. However, we did not find
very satisfactory references, and, though the reasonings are essentially the same
as in the classical case, the lack of homogeneity of Ψ and the presence of the two
spaces M Ψ and LΨ gives proofs which are not so obvious and therefore we shall
give some details.
It should be noted that our definition is not exactly the same as the one given
in [33], § 9.1.
We shall begin with the following proposition.
Proposition 3.1. Let f : D → C be an analytic function. For every Orlicz
function Ψ, the following assertions are equivalent:
1) sup0≤r<1 fr Ψ < +∞, where fr (t) = f (reit );
2) there exists f ∗ ∈ LΨ ([0, 2π]) such that f∗ (n) = 0 for n < 0 and for which

f (z) = n≥0 f∗ (n)z n , z ∈ D.
When these conditions are satisfied, one has f ∗ Ψ = sup0≤r<1 fr Ψ .
Let us note that, since
Ψ is convex and increasing, Ψ(a|f |) is subharmonic on
D, and hence the numbers T Ψ(a|fr |) dm increase with r, for every a > 0.
This proposition leads to the following definition.
Definition 3.2. Given an Orlicz function Ψ, the Hardy-Orlicz space H Ψ as-
sociated to Ψ is the space of analytic functions f : D → C such that one of the
equivalent conditions of the above proposition is satisfied. The norm of f is defined
by f H Ψ = f ∗ Ψ . We shall denote by HM Ψ the Hardy-Morse-Transue space,
i.e. the subspace {f ∈ H Ψ ; f ∗ ∈ M Ψ (T)}.
9
10 PASCAL LEFÈVRE, DANIEL LI, HERVÉ QUEFFÉLEC, LUIS RODRÍGUEZ-PIAZZA

In the sequel, we shall make no distinction between f and f ∗ , unless there may
be some ambiguity, and shall write f instead of f ∗ for the boundary limit. Hence
we shall allow ourselves to write f (eit ) instead of f ∗ (t), or even f ∗ (eit ). Moreover,
we shall write f Ψ instead of f H Ψ .
It follows that H Ψ becomes a subspace of LΨ (T) and HM Ψ = H Ψ ∩ M Ψ (T).
These two spaces are closed (hence Banach spaces) since Proposition 3.1 gives:
Corollary 3.3. H Ψ is weak-star closed in LΨ = (M Φ )∗ . When Ψ satisfies
∇2 , it is isometrically isomorphic to the bidual of HM Ψ .
Proof. The weak-star closure of H Ψ is obvious with Proposition 3.1, 2).
Suppose now that Φ satisfies ∇2 , it is plainly seen that (HM Ψ )⊥ is the closed
subspace of LΦ = (M Ψ )∗ generated by all characters en with n < 0, where
en (t) = eint (for convenience, we define the duality between f ∈ LΨ and g ∈ LΦ by
integrating the product f ǧ, where ǧ(t) = g(−t)). As H Ψ ⊆ LΨ = (LΦ )∗ = (M Φ )∗
is the orthogonal of this latter subspace. So, we have (HM Ψ )⊥⊥ = H Ψ . 
Proof of Proposition 3.1. Assume that 1) is satisfied. Since fr 1 ≤ CΨ fr Ψ ,
one has f ∈ H 1 , and hence, by Fatou-Riesz Theorem, f has almost everywhere a
boundary limit f ∗ ∈ L1 (m). If C = sup0≤r<1 fr Ψ , one has:

|fr |
Ψ dm ≤ 1
T C
for every r < 1; hence Fatou’s lemma implies:

|f |
Ψ dm ≤ 1,
T C
i.e. f ∗ ∈ LΨ and f ∗ Ψ ≤ C.
Conversely, assume that 2) is satisfied. In particular f ∗ ∈ L1 (m); hence f ∈ H 1
and f ∗ = limr→1 fr almost everywhere. One has fr = f ∗ ∗ Pr , where Pr is the
Poisson kernel at r. Hence, using Jensen’s inequality for the probability measure
Pr (θ − t) 2π
dt
, we get:
2π ∗ 2π  2π ∗ 
|(f ∗ Pr )(θ)| dθ |f (t)| dt dθ
Ψ ≤ Ψ Pr (θ − t)
0 f ∗ Ψ 2π 0 0 f ∗ Ψ 2π 2π

2π 2π ∗ 
|f (t)| dt dθ
≤ Ψ Pr (θ − t)
0 0 f ∗ Ψ 2π 2π
2π  2π  ∗
dθ |f (t)| dt
= Pr (θ − t) Ψ
0 0 2π f ∗ Ψ 2π
2π ∗
|f (t)| dt
= Ψ ≤ 1,
0 f ∗ Ψ 2π
so that fr Ψ ≤ f ∗ Ψ . Hence we have 1), and f H Ψ ≤ f ∗ Ψ .
The two parts of the proof actually give f H Ψ = f ∗ Ψ . 

Proposition 3.4. For every f ∈ HM Ψ , one has fr − f ∗ Ψ −→ 0. Therefore
r→1
the polynomials on D are dense in HM Ψ .
Equivalently, on T = ∂D, the analytic trigonometric polynomials are dense in
HM Ψ .
1. HARDY-ORLICZ SPACES 11


Proof. Let f ∈ HM Ψ and ε > 0. Since M Ψ = C(T) , there exists a continuous
function h on T such that f − hΨ ≤ ε. We have, for every r < 1: 
Pr ∗ f − f Ψ ≤ Pr ∗ (f − h)Ψ + Pr ∗ h − hΨ + h − f Ψ ≤ 2ε + Pr ∗ h − hΨ
because Pr ∗ gΨ ≤ gΨ , for every r < 1 and every g ∈ LΨ .
But now, Pr ∗ h −→ h uniformly. The conclusion follows. 
r→1

Remark. We do not have to use a maximal function to prove the existence of
boundary limits because we use their existence for functions in H 1 . However, as in
the classical case, the Marcinkiewicz interpolation Theorem, or, rather, its Orlicz
space version ensures that the maximal non-tangential function is in LΨ . This result
is undoubtedly known, but perhaps never stated in the following form. Recall that
Nα is defined, for every f , say in L1 (T), as
(Nα f )(eiθ ) = sup |(f ∗ Pr )(eit )| = sup |f (z)|,
reit ∈Sθ z∈Sθ

where Sθ is the Stolz domain at eiθ with opening α (see [4], page 177); here f
defines a harmonic function in D.
Proposition 3.5. Assume that the complementary function Φ of the Orlicz
function Ψ satisfies the Δ2 condition (i.e. Ψ ∈ ∇2 ). Then every linear, or sublinear,
operator which is of weak-type (1, 1) and (strong) type (∞, ∞) is bounded from LΨ
into itself. In particular, for every f ∈ LΨ (T), the maximal non-tangential function
Nα f is in LΨ (T) (0 < α < 1).
Proof. If Ψ ∈ ∇2 , then ([33], Theorem 3, 1 (iii), page 23), there exists some
β > 1 such that xΨ (x) ≥ βΨ(x) for x large enough. Integrating between u and
 u β
v, for u < v large enough, we get Ψ(u)
Ψ(v) ≥ v . Hence, for s, t large enough
Ψ−1 (s)
Ψ−1 (s/t)≤ t1/β . This means (see [4], Theorem 8.18) that the upper Boyd index
of L is ≤ 1/β < 1. Hence ([4], Theorem 5.17), Nα is bounded on LΨ (it is well-
Ψ

known that Nα f is dominated by the Hardy-Littlewood maximal function M f ). 

The following, essentially well-known, criterion for compactness of operators
will be very useful.
Proposition 3.6. 1) Every bounded linear operator T : H Ψ → X from H Ψ
into a Banach space X which maps every bounded sequence which is uniformly
convergent on compact subsets of D into a norm convergent sequence is compact.
2) Conversely, if T : H Ψ → X is compact and weak-star to weak continuous, or
if T : H Ψ → Y ∗ is compact and weak-star continuous, then T maps every bounded
sequence which is uniformly convergent on compact subsets of D into a norm con-
vergent sequence.
Though well-known (at least for the classical case of H p spaces), the link with
the weak (actually the weak-star) topology is usually not highlighted. Indeed, the
criterion is an easy consequence of the following proposition. Note that Proposi-
tion 3.6 will apply to the composition operators on H Ψ since they are weak-star
continuous.
Proposition 3.7. On the unit ball of H Ψ , the weak-star topology is the topology
of uniform convergence on every compact subset of D.
12 PASCAL LEFÈVRE, DANIEL LI, HERVÉ QUEFFÉLEC, LUIS RODRÍGUEZ-PIAZZA

Proof. First we notice that the topologies are metrizable. Indeed, this is known
for the topology of uniform convergence on every compact subset of D and, on the
other hand, M Φ is separable, so that the weak-star topology is metrizable on the
unit ball of its dual space LΨ , and a fortiori on that of H Ψ . Now, it is sufficient to
prove that the convergent sequences in both topologies are the same.
Let (fk )k≥1 be in the unit ball of H Ψ and weak-star convergent to f ∈ H Ψ .
Let us fix a compact subset K of D. We may suppose that K is the closed ball of
center 0 and radius r < 1. First, testing the weak-star convergence on characters,
we have fk (n) −→ fˆ(n) for every n ∈ Z. Then:
k→∞

sup |fk (z) − f (z)| = sup |fk (z) − f (z)| ≤ r n |fk (n) − fˆ(n)|.
|z|≤r |z|=r n≥0

The last term obviously tends to zero when k tends to infinity. The result follows.
Conversely, let (fk )k≥1 be in the unit ball of H Ψ and converging to some holo-
morphic function f uniformly on every compact subset of D. We first notice that
f actually lies in the unit ball of H Ψ (by Fatou’s lemma). Fix h ∈ M Φ and ε > 0.
There exists some r < 1 such that Pr ∗ h − hΦ ≤ ε/8, where Pr is the Poisson
kernel with parameter r. Then (see [33], page 58, inequality (4) for the presence of
the coefficient 2):
|h, fk − f | ≤ |Pr ∗ h − h, fk − f | + |Pr ∗ h, fk − f |
≤ 2 Pr ∗ h − hΦ fk − f Ψ + |h, Pr ∗ (fk − f )|
ε
≤ + 2 hΦ [fk ]r − fr Ψ
2
ε
≤ + 2α hΦ [fk ]r − fr ∞
2
ε
= + 2α hΦ sup |fk (z) − f (z)|.
2 |z|=r

where α is the norm of the injection of L∞ into LΨ .
Now, by uniform convergence on the closed ball of center 0 and radius r, there
exists kε ≥ 1 such that for every integer k ≥ kε , one has
hΦ sup |fk (z) − f (z)| ≤ ε/4.
|z|=r

It follows that (fk )k weak-star converges to f . 
However, we shall have to use a similar compactness criterion for Bergman-
Orlicz spaces, and it is worth stating and proving a general criterion. We shall
say that a Banach space X is a Banach space of holomorphic functions on an
open subset Ω of the complex plane if X is continuously embedded (through the
canonical injection) in H (Ω), the space of holomorphic functions on Ω, equipped
with its natural topology of compact convergence. We shall say that a Banach
space X of holomorphic functions on an open set Ω ⊆ C, has the Fatou property
if: for every bounded sequence (fn )n in X which converges uniformly on compact
subsets of Ω to a function f , one has f ∈ X. Then:
Proposition 3.8 (Compactness criterion). Let X, Y be two Banach spaces of
analytic functions on an open set Ω ⊆ C, X having the Fatou property. Let φ be
an analytic self-map of Ω such that Cφ = f ◦ φ ∈ Y whenever f ∈ X.
2. PRELIMINARY RESULTS 13

Then Cφ : X → Y is compact if and only if for every bounded sequence (fn )n in
X which converges to 0 uniformly on compact subsets of Ω, one has Cφ (fn )Y → 0.
Note that Hardy-Orlicz H Ψ and Bergman-Orlicz B Ψ (see Chapter 5) spaces
trivially have the Fatou property, because of Fatou’s Lemma.
Proof. Assume that the above condition is fulfilled. Let (fn )n≥1 be in the unit
ball of X. The assumption on X implies that (fn )n is a normal family in H (Ω).
Montel’s Theorem allows us to extract a subsequence, that we still denote by (fn )n
to save notation, which converges to some f ∈ H (Ω), uniformly on compact subsets
of Ω. Since X has the Fatou property, one has f ∈ X. Now, since (fn − f )n is a
bounded sequence in X which converges to 0 uniformly on compact subsets of Ω,
one has Cφ (fn ) − Cφ (f )Y = Cφ (fn − f )Y → 0. Hence Cφ is compact.
Conversely, assume that Cφ is compact. Let (fn )n be a bounded sequence in
X which converges to 0 uniformly on compact subsets of Ω. By the compactness
of Cφ , we may assume that Cφ (fn ) → g ∈ Y . The space Y being continuously
embedded in H (Ω), (fn ◦ φ)n converges pointwise to g. Since (fn )n converges to
0 uniformly on compact subsets of Ω, the same is true for (fn ◦ φ)n . Hence g = 0.
Therefore, since Cφ is compact, we get Cφ (fn )Y → 0. 

2. Preliminary results
Lemma 3.9. Let (Ω, P) be any probability space. For every function g ∈ L∞ (Ω),
one has:
g∞
gΨ ≤ −1 ·
Ψ (g∞ /g1 )
Proof. We may suppose that g∞ = 1.
Since Ψ(0) = 0, the convexity of Ψ implies Ψ(ax) ≤ aΨ(x) for 0 ≤ a ≤ 1. Hence,
for every C > 0, one has, since |g| ≤ 1:

Ψ(|g|/C) dP ≤ |g|Ψ(1/C) dP = g1 Ψ(1/C).
Ω Ω

But g1 Ψ(1/C) ≤ 1 if and only if C ≥ 1/Ψ−1 (1/g1 ), and that proves the lemma. 

Corollary 3.10. For |a| = 1 and 0 ≤ r < 1, one has:
1
ua,r Ψ ≤ ·
Ψ−1 ( 1−r
1
)
Proof. One has ua,r ∞ = 1, and:
2π  
 1 − r 2
ua,r 1 =   dm(t)
0 1 − āreit

+∞
(1 − r)2 1−r
= (1 − r)2 r 2n = = ·
n=0
1 − r 2 1 +r
Hence ua,r Ψ ≤ 1/Ψ−1 ((1 + r)/(1 − r)), by using Lemma 3.9, giving the result
since (1 + r)/(1 − r) ≥ 1/1 − r. 
Remark. We hence get actually ua,r Ψ ≤ 1/Ψ−1 ((1 + r)/(1 − r)); the term 1 + r
has no important meaning, so we omit it in the statement of Corollary 3.10, but
14 PASCAL LEFÈVRE, DANIEL LI, HERVÉ QUEFFÉLEC, LUIS RODRÍGUEZ-PIAZZA

sometimes, for symmetry of formulae, or in order to be in accordance with the
classical case, we shall use this more precise estimate.

For every f ∈ L1 (T) and every z = r eiθ ∈ D, one has

(P [f ])(z) = f (eit )Pz (t) dm(t),
0

where Pz is the Poisson kernel:
1 − r2 1 − |z|2
Pz (t) = = it ,
1 − 2r cos(θ − t) + r 2 |e − z|2
and f (z) = (P [f ])(z) when f is analytic on D. Since Pz ∈ L∞ (T) ⊆ LΦ (T), it
follows that the evaluation in z ∈ D:
δz (f ) = f (z)
Ψ
is a continuous linear form on H . The following lemma explicits the behaviour of
its norm.
Lemma 3.11. For |z| < 1, the norm of the evaluation functional at z is:
1
δz (HM Ψ )∗ = δz (H Ψ )∗ ≈ Ψ−1 ·
1 − |z|
More precisely:
1 −1 1 + |z| 1 + |z|
Ψ ≤ δz (H Ψ )∗ ≤ 2Ψ−1 ·
4 1 − |z| 1 − |z|
Remark. In particular:
1 −1 1 1
Ψ ≤ δz (H Ψ )∗ ≤ 4Ψ−1 ,
4 1 − |z| 1 − |z|
which often suffices for our purpose.

Proof. The first equality δz (HM Ψ )∗ = δz (H Ψ )∗ comes from the fact that fr ∈
HM Ψ , for every f ∈ H Ψ and r < 1 (thus f (rz) −→ f (z), when z ∈ D and f ∈ H Ψ ).
r→1
On the one hand, we have, when |z| = r, using [33], inequality (4) page 58,
and Lemma 3.9, since Pz 1 = 1 and Pz ∞ = 1−r
1+r
:
1+r 1
δz (H Ψ )∗ ≤ 2 Pz Φ ≤ 2  1+r  ,
1−rΦ −1
1−r

which is less than 2Ψ−1 ((1+r)/(1−r)), by using the inequality (see [33], Proposition
1 (ii), page 14, or [22], pages 12–13):
Ψ−1 (x)Φ−1 (x) ≥ x , x > 0.
On the other hand, one has, using Corollary 3.10, with r = |z| and āz = r:
|ua,r (z)| 1/(1 + r)2 1 1 + r
δz (H Ψ )∗ ≥ ≥  1+r  ≥ Ψ−1 ,
ua,r Ψ 1/Ψ−1 1−r 4 1−r
and that ends the proof. 
3. COMPOSITION OPERATORS 15

3. Composition operators
We establish now some estimations for the norm of composition operators.
Proposition 3.12. 1) Every analytic self-map φ : D → D induces a bounded
composition operator Cφ : H Ψ → H Ψ by setting Cφ (f ) = f ◦ φ. More precisely:
1 + |φ(0)|
Cφ  ≤ ·
1 − |φ(0)|
In particular, Cφ  ≤ 1 if φ(0) = 0.
2) One has:  
1 −1 1 + |φ(0)|
Cφ  ≥ Ψ .
8Ψ−1 (1) 1 − |φ(0)|
3) When Ψ ∈ ∇1 globally: Ψ(x)Ψ(y) ≤ Ψ(bxy) for all x, y ≥ 0, we also have:
 
1 + |φ(0)|
Cφ  ≤ bΨ−1 .
1 − |φ(0)|
4) Moreover, Cφ maps HM Ψ into HM Ψ . Hence, if Ψ ∈ ∇2 , then Cφ : H Ψ →
H is the bi-adjoint of the composition operator Cφ : HM Ψ → HM Ψ .
Ψ

Note that when Ψ(x) = xp for 1 ≤ p < ∞, then Ψ ∈ ∇1 globally, with b = 1.
Proof. 1) Assume first that φ(0) = 0. Let f ∈ H Ψ , with f Ψ = 1. Since Ψ is
convex and increasing, the function u = Ψ ◦ |f | is subharmonic on D, thanks to
Jensen’s inequality. The condition φ(0) = 0 allows to use Littlewood’s subordina-
tion principle ([11], Theorem 1.7); for r < 1, one has:
2π 2π
  dt   dt
Ψ |(f ◦ φ)(reit )| ≤ Ψ |f (reit )| ≤ 1.
0 2π 0 2π
Hence f ◦ φ ∈ H Ψ and f ◦ φΨ ≤ 1.
Assume now that φ is an inner function, and let a = φ(0). It is known that
(see [32], Theorem 1) that
φ∗ (m) = Pa .m ,
where φ∗ (m) is the image under φ∗ (the boundary limit of φ) of the normalized
Lebesgue measure m, and Pa .m is the measure of density Pa , the Poisson kernel
at a. Therefore, for every f ∈ H Ψ with f Ψ = 1, one has for 0 ≤ r < 1, setting
1+|a|
Ka = Pa ∞ = 1−|a| :

|(f ◦ φ)(reit )| dt |(f ◦ φ)∗ |
Ψ ≤ Ψ dm
0 Ka 2π T Ka

|f ◦ φ∗ |
= Ψ dm (recall that |φ∗ | = 1)
T Ka
∗ ∗
|f | |f |
(3.1) = Ψ dφ∗ (m) = Ψ Pa dm
Ka Ka
T T
1
≤ Ψ(|f ∗ |) Pa dm , since Ka > 1
T K a

1
≤ Ψ(|f ∗ |) Pa ∞ dm
K
T a

= Ψ(|f ∗ |) dm ≤ 1.
T
16 PASCAL LEFÈVRE, DANIEL LI, HERVÉ QUEFFÉLEC, LUIS RODRÍGUEZ-PIAZZA

Hence (f ◦ φ)r Ψ ≤ Ka , and therefore f ◦ φΨ ≤ Ka .
Then, for an arbitrary φ, let a = φ(0) again, and let φa be the automorphism
z → 1−āz
z−a ,
whose inverse is φ−a . Since φ = φ−a ◦(φa ◦φ), one has Cφ = Cφa ◦φ ◦Cφ−a .
But φ−a is inner and, on the other hand, (φa ◦ φ)(0) = 0; hence parts a) and b) of
the proof give:
1 + |a|
Cφ  ≤ Cφ−a  ≤ Ka = ,
1 − |a|
which gives the first part of the proof.
2) By Lemma 3.11, we have for every f ∈ H Ψ with f Ψ ≤ 1:
|(f ◦ φ)(0)| ≤ δ0 (H Ψ )∗ f ◦ φΨ ≤ 2Ψ−1 (1) Cφ .
In other words:  
|f φ(0) | ≤ 2Ψ−1 (1)Cφ 
for every such f ∈ H Ψ . Hence:
δφ(0) (H Ψ )∗ ≤ 2Ψ−1 (1) Cφ  ,
giving  
1 −1 1 + |φ(0)|
Cφ  ≥ Ψ ,
8Ψ−1 (1) 1 − |φ(0)|
by using Lemma 3.11 again, but the minoration.
3) When Ψ ∈ ∇1 globally, we go back to the proof of 1). We have  only to

modify inequalities (3.1). Setting Ka = Ψ−1 (Ka ), and writing Pa = Ψ Ψ−1 (Pa ) ,
we get, for every f ∈ H Ψ with f Ψ = 1:
2π ∗
|(f ◦ φ)(reit )| dt |f |
Ψ 
≤ Ψ Pa dm
bKa 2π bK 
0
T ∗a
|f |  
= Ψ 
Ψ Ψ−1 (Pa ) dm
bK
T ∗a
|f | −1
≤ Ψ Ψ (P a ) dm
Ka
T
≤ Ψ(|f ∗ |) dm ≤ 1,
T

since Ψ−1 (Pa ) ≤ Ψ−1 (Pa ∞ ) = Ka , giving f ◦ φΨ ≤ bKa .
4) Suppose now that f ∈ HM Ψ . As before, when φ(0) = 0, Littlewood’s
subordination principle gives, for every C > 0:
2π 2π
  dt   dt
Ψ C|(f ◦ φ)(eit )| = sup Ψ C|(f ◦ φ)(reit )|
0 2π r<1 0 2π
2π 2π
  dt   dt
≤ sup Ψ C|f (reit )| = Ψ C|f (eit )| < +∞ ;
r<1 0 2π 0 2π
hence f ◦ φ ∈ HM Ψ . When φ is inner, the same computations as in 1) above, using
that φ∗ (m) = Pa .m, where a = φ(0), give, for every C > 0:

  dt
Ψ C|(f ◦ φ)(reit )| ≤ Ψ(C|f ∗ |)Pa ∞ dm
2π T
0

= Ka Ψ(C|f ∗ |) dm < +∞,
T
4. ORDER BOUNDED COMPOSITION OPERATORS 17

and f ◦ φ ∈ HM Ψ again. The general case follows as in 1) above, since f ∈ HM Ψ
implies f ◦φ−a ∈ HM Ψ , because φ−a is inner, and then f ◦φ = (f ◦φ−a )◦(φa ◦φ) ∈
HM Ψ since (φa ◦ φ)(0) = 0. 

4. Order bounded composition operators
Recall that an operator T : X → Z from a Banach space X into a Banach
subspace Z of a Banach lattice Y is order bounded if there is some positive y ∈ Y
such that |T x| ≤ y for every x in the unit ball of X.
Before studying order bounded composition operators, we shall recall the fol-
lowing, certainly well-known, result, which says that order boundedness can be seen
as stronger than compactness.
Proposition 3.13. Let T : L2 (μ) → L2 (μ) be a continuous linear operator.
Then T is order bounded if and only if it is a Hilbert-Schmidt operator.
The proof is straightforward: if B is the unit ball of L2 (μ), and (ei )i is an
 
2 1/2
orthonormal basis, one has supf ∈B |T f | = i |T ei | . Hence supf ∈B |T f | ∈

   
L2 (μ) if and only if i |T ei | 2
dμ = i T ei 2
< +∞, i.e. if and only if T is
Hilbert-Schmidt.
J. H. Shapiro and P. D. Taylor proved in [42] that there exist composition
operators on H 2 which are compact but not Hilbert-Schmidt. We are going to see
that, when the Orlicz function Ψ grows fast enough, the compactness of composition
operators on H Ψ is equivalent to their order boundedness.
Proposition 3.14. The composition operator Cφ : H Ψ → HΨ is order bounded
(resp. order bounded into M Ψ (T)) if and only if Ψ−1 1−|φ|1
∈ LΨ (T) (resp.
 
Ψ−1 1−|φ|
1
∈ M Ψ (T)). Equivalently, if and only if
1
(OB1) χA ∈ L1 (T) for some A > 0,
1 − |φ|
respectively:

1
(OB2) χA ∈ L1 (T) for every A > 0,
1 − |φ|
 
In other words (recall that χA (x) = Ψ AΨ−1 (x) ), if and only if:
 1
−1
Ψ AΨ dm < +∞ for some (resp. every) A > 0.
T 1 − |φ|

Remark For a sequence (gn )n of elements of LΨ (Ω), one has gn Ψ −→ 0 if and
n→+∞
only if
|g |
n
(3.2) Ψ dP −→ 0 for every ε > 0.
Ω ε n→+∞

In fact, if (3.2) holds, then for n ≥ nε , the above integrals are ≤ 1, and hence 
gn Ψ ≤ ε. Conversely, assume that gn Ψ −→ 0, and let ε > 0 be given.
n→+∞
Let 0 < δ ≤ 1. Since gn /(εδ)Ψ −→ 0, one has gn /(εδ)Ψ ≤ 1, and hence
n→+∞
18 PASCAL LEFÈVRE, DANIEL LI, HERVÉ QUEFFÉLEC, LUIS RODRÍGUEZ-PIAZZA

 
Ω
Ψ |gεδn | dP ≤ 1, for n large enough. Then, using the convexity of Ψ:
|g | |g | |g |
n n n
Ψ dP = Ψ δ dP ≤ δ Ψ dP ≤ δ ,
Ω ε Ω εδ Ω εδ
for n large enough.
Remark If Cφ : H Ψ → H Ψ is order bounded into M Ψ (T), then the composition
operator Cφ : H Ψ → H Ψ is compact.
Indeed, let (gn )n≥0 be a sequence in the unit ball of H Ψ . Thank s to Propo-
sition 3.7, there exists a subsequence, which we still denote by (gn )n≥0 , uniformly
converging to some g (lying in the unit ball of H Ψ ) on every compact subset of D.
Since (OB2) implies that |φ∗ | < 1 almost everywhere, we have for every > 0,
|g ◦ φ∗ |
n
Ψ −→ 0 a.e.
ε n→+∞

On the other hand, the order boundedness hypothesis on Cφ implies that, on T:
|g ◦ φ∗ | G
n
Ψ ≤Ψ a.e.,
ε ε
where G ∈ M Ψ (T). Using Lebesgue’s dominated convergence Theorem and the
previous remark, the conclusion follows.
Proof of Proposition 3.14. As HM Ψ is separable, there exists a sequence
(fn )n≥1 in the unit ball of HM Ψ such that
 
supfn ◦ φ(reiθ ) = δφ(reiθ ) (HM Ψ )∗ .
n

Now, suppose that Cφ is order bounded into LΨ (T) (resp. into M Ψ (T)). Then
there exists some g in LΨ (T) (resp. in M Ψ (T)) such that g ≥ |Cφ (f )| a.e., for every
f in the unit ball of H Ψ . Using Lemma 3.11, we have a.e.
1  
Ψ−1 ≤ 4δφ(reiθ ) (H Ψ )∗ = 4 supfn ◦ φ(reiθ ) ≤ 4g.
1 − |φ(re )|

n
The result hence follows letting r ↑ 1.
The converse is obvious. 

Theorem 3.15. If the composition operator Cφ : H Ψ → H Ψ is order bounded,
then:
1
(OB3) m(1 − |φ| < λ) = O , as λ → 0, for some A > 0.
χA (1/λ)
and if it is order bounded into M Ψ (T), then:
1
(OB4) m(1 − |φ| < λ) = O , as λ → 0, for every A > 0.
χA (1/λ)
Under the hypothesis Ψ ∈ Δ1 , the converse holds.

Proof. The necessary condition follows from Proposition 3.14 and Markov’s in-
equality:
1 1 1
m >t ≤ χA dm.
1 − |φ| χA (t) T 1 − |φ|
For the converse, we shall prove a stronger result, and for that, we define the
weak-LΨ space as follows:
4. ORDER BOUNDED COMPOSITION OPERATORS 19

Definition 3.16. The weak-LΨ space LΨ,∞ (Ω) is the space of measurable func-
tions f : Ω → C such that, for some constant c > 0, one has, for every t > 0:
1
P(|f | > t) ≤ ·
Ψ(ct)
For subsequent references, we shall state separately the following elementary
result.
Lemma 3.17. For every f ∈ LΨ (Ω), one has, for every t > 0:
t
f Ψ ≥  ·
Ψ−1 1
P(|f |>t)

Proof. By Markov’s inequality, one has, for t > 0:
t |f |
Ψ P(|f | > t) ≤ Ψ dP ≤ 1;
f Ψ Ω f Ψ
and that gives the lemma. 

Since Lemma 3.17 can be read:
1
P(|f | > t) ≤ ,
Ψ(t/f Ψ )
we get that LΨ (Ω) ⊆ LΨ,∞ (Ω).
The converse of Theorem 3.15 is now an immediate consequence of the following
proposition.
Proposition 3.18.
1) a) If Ψ ∈ Δ1 , then LΨ (Ω) = LΨ,∞ (Ω).
b) If LΨ (Ω) = LΨ,∞ (Ω), then Ψ ∈ Δ0 .
2) If LΨ (T) = LΨ,∞ (T), then condition (OB3) implies that Cφ : H Ψ → H Ψ is
order bounded, and condition (OB4) implies that it is order bounded into M Ψ (T).
Lemma 3.19. The following assertions are equivalent
i) LΨ (Ω) = LΨ,∞ (Ω).
∞  +∞
Ψ (u) 1
ii) du ≡ dx < +∞, for some B > 1.
1 Ψ(Bu) Ψ(1) χ B (x)
Proof of the Lemma. Assume first that 1/χB is integrable on (Ψ(1), ∞). For
every f ∈ LΨ,∞ (Ω), there is a c > 0 such that P(|f | > t) ≤ 1/Ψ(ct). Then
|f | +∞
Ψ c dP = P(|f | > Bt/c) Ψ (t) dt
Ω B 0
+∞
≤ Ψ(1) + P(|f | > Bt/c) Ψ (t) dt
1
+∞  +∞
Ψ (t) du
≤ Ψ(1) + dt = Ψ(1) + < +∞,
1 Ψ(Bt) Ψ(1) χ B (u)

so that f ∈ LΨ (Ω).
Conversely, assume that LΨ (Ω) = LΨ,∞ (Ω). Since 1/Ψ is decreasing, there is
a measurable function f : Ω → C such that P(|f | > t) = 1/Ψ(t), when t ≥ Ψ−1 (1).
20 PASCAL LEFÈVRE, DANIEL LI, HERVÉ QUEFFÉLEC, LUIS RODRÍGUEZ-PIAZZA

Such a function is in LΨ,∞ (Ω); hence it is in LΨ (Ω), by our hypothesis. Therefore,
there is a B > 1 such that

Ψ(|f |/B) dP < +∞.
Ω
But

+∞

+∞
Ψ (t)
Ψ(|f |/B) dP = P(|f | > Bt) Ψ (t) dt ≥ dt
Ω 0 1

−1 (1) Ψ(Bt)

+∞
Ψ (t) +∞
du
≥ dt = ,
Ψ−1 (1) Ψ(Bt) 1 χB (u)
and hence 1/χB is integrable on (1, ∞). 
Proof of Proposition 3.18. 1)a) We first remark that for every Orlicz function
Ψ, one has Ψ(2x) ≥ xΨ (x) for every x > 0, because, since Ψ is positive and
increasing:
2x 2x
Ψ(2x) = Ψ (t) dt ≥ Ψ (t) dt ≥ xΨ (x).
0 x
Assume now that Ψ ∈ Δ1 : xΨ(x) ≤ Ψ(βx) for some β > 0 and for x ≥ x0 > 0. By
Lemma 3.19, it suffices to show that 1/χ2β is integrable on (Ψ(x0 ), +∞). But
+∞ +∞ +∞ +∞
dx Ψ (u) Ψ(2u)/u du
= du ≤ du = < +∞.
Ψ(x0 ) χ2β (x) x0 Ψ(2βu) x0 2uΨ(2u) x0 2u2
1)b) Suppose now that LΨ (Ω) = LΨ,∞ (Ω). By the preceding lemma, there
exists some B > 1 such that
2x 
Ψ (u)
lim du = 0.
x→+∞ x Ψ(Bu)
By convexity, Ψ(2x) ≥ 2Ψ(x) and Ψ is nonnegative, so that
2x  2x
Ψ (u) 1 Ψ(2x) − Ψ(x) Ψ(x)
du ≥ Ψ (u) du ≥ ≥ ·
x Ψ(Bu) Ψ(2Bx) x Ψ(2Bx) Ψ(2Bx)
Therefore Ψ satisfies Δ0 .
2) Assume that LΨ (T) = LΨ,∞ (T) and that condition (OB3) (resp. (OB4))
holds. By Lemma 3.19, there is a B > 0 such that 1/χB is integrable on (1, +∞).
We get, using condition (OB3) (resp. (OB4)) and setting C = A/B:
1 +∞
1
χC dm = m > t χC (t) dt
T 1 − |φ| 0 1 − |φ|
+∞
= m(1 − |φ| < 1/t) χC (t) dt
0
+∞ 
χC (t)
≤ χC (1) + K dt.
1 χA (t)
 
But, if we set u = χC (t), i.e. u = Ψ CΨ−1 (t) , then Ψ−1 (u) = CΨ−1 (t), and hence
   
χA (t) = Ψ AΨ−1 (t) = Ψ BΨ−1 (u) = χB (u).
Therefore:

1 +∞
du
χC dm ≤ χC (1) + K du < +∞.
T 1 − |φ| χC (1) χB (u)
5. WEAKLY COMPACT COMPOSITION OPERATORS 21

It follows from Proposition 3.14 that Cφ is order bounded (resp. into M Ψ (T)).

Remark. The condition Ψ ∈ Δ1 is not equivalent to LΨ (Ω) = LΨ,∞ (Ω). For
example, we can take:
 3/2 
Ψ(x) = exp log(x + 1) − 1.
Then, as x tends to infinity,
3 
Ψ(Kx) ∼ Ψ(x) exp (log K)(log x)1/2 ,
2
1/2
and hence Ψ ∈ Δ1 . On the other hand, Ψ (x) ∼ 32 (log xx) Ψ(x); hence:
+∞ +∞ 
du Ψ (t)
= dt
Ψ(1) χK (u) 1 Ψ(Kt)
+∞ +∞
3u2 3u2
∼ 3 du = du < +∞
0 exp( 2 (log K)u) 0 K 3u/2
for K > 1. Therefore LΨ (Ω) = LΨ,∞ (Ω) by Lemma 3.19.

5. Weakly compact composition operators
We saw in Lemma 3.10 that
1
ua,r Ψ ≤ ·
Ψ−1 ( 1−r
1
)
The next result shows that the weak compactness of Cφ transforms the “big-oh”
into a “little-oh”, when Ψ grows fast enough.
Theorem 3.20. Assume that Ψ ∈ Δ0 . Then the weak compactness of the
composition operator Cφ : H Ψ → H Ψ implies that:
 
1
(W) sup Cφ (ua,r )Ψ = o  1  , as r → 1.
a∈T Ψ−1 1−r

Proof. Actually, we only need to use that the restriction of Cφ : H Ψ → H Ψ to
HM Ψ is weakly compact. We proved in [25], Theorem 4, that, under the hypothesis
Ψ ∈ Δ0 , the operator Cφ : HM Ψ → HM Ψ is weakly compact if and only if for every
ε > 0, we can find Kε > 0 such that, for every f ∈ HM Ψ , one has:
Cφ (f )Ψ ≤ Kε f 1 + ε f Ψ .
Using Corollary 3.10, we get, for every ε > 0, since ua,r 1 ≤ 1 − r:
1
Cφ (ua,r )Ψ ≤ Kε (1 − r) + ε ·
Ψ−1 ( 1−r
1
)
Ψ(x)
But −→ +∞.
x x→+∞ Hence
1
1−r =o as r → 1,
Ψ−1 ( 1−r
1
)
and that proves the theorem. 
22 PASCAL LEFÈVRE, DANIEL LI, HERVÉ QUEFFÉLEC, LUIS RODRÍGUEZ-PIAZZA

We shall see in Chapter 4 that the converse holds when Ψ satisfies Δ0 and more-
over that Cφ is then compact. That will use some other techniques. Nevertheless,
we can prove, from now on, the following result.
Theorem 3.21. If Ψ ∈ Δ2 , then condition:
 
1
(W) sup Cφ (ua,r )Ψ = o  1  , as r → 1,
a∈T Ψ−1 1−r
in Theorem 3.20 implies condition:
1
(OB4) m(1 − |φ| < λ) = O , as λ → 0, for every A > 0,
χA (1/λ)
in Theorem 3.15.
Remark that when Ψ ∈ Δ0 (and in particular when Ψ ∈ Δ2 ), one has, for any
B > βA (where β is given by the definition of Δ0 ): Ψ(Bx)
Ψ(Ax) → +∞ as x → +∞;
 
hence 1/χA (x) = o 1/χB (x) and the “big-oh” condition in (OB4) may be replaced
by a “little-oh” condition.
Before proving this theorem, let us note that:
Proposition 3.22. Condition
1
(W) sup Cφ (ua,r )Ψ = o   as r → 1
a∈T Ψ−1 1−r
1

implies that m(|φ| = 1) = 0.
Proof. Otherwise, one has m(|φ| = 1) = δ > 0. Splitting the unit circle T into N
parts, we get some a ∈ T such that:
m(|φ − a| ≤ π/N ) ≥ δ/N.
But, the inequality |φ − a| ≤ π/N implies, with r = 1 − 1/N (since |φ| ≤ 1):
|1 − ārφ| ≤ |1 − āφ| + (1 − r)|āφ| = |a − φ| + (1 − r)|φ| < 5/N = 5(1 − r).
Hence:
   
m |Cφ (ua,r )| > 1/25) = m |1 − ārφ| < 5(1 − r)
 
≥ m |φ − a| ≤ π(1 − r) ≥ δ(1 − r) ,
and therefore, by Lemma 3.17:
1/25 δ/25
Cφ (ua,r )Ψ ≥  ≥  ·
Ψ−1 1/δ(1 − r) Ψ−1 1/(1 − r)
Since r can be taken arbitrarily close to 1, that proves the proposition. 

Proof of Theorem 3.21. Assume that condition (W) is satisfied, and fix A > 1.
Let ε > 0 to be adjusted later. We can find rε < 1 such that rε ≤ r < 1 implies:
ε
Cφ (ua,r )Ψ ≤ −1  1  , ∀a ∈ T.
Ψ 1−r
Now, Lemma 3.17 also reads:
1
m(|f | > t) ≤  t ,
Ψ f Ψ
5. WEAKLY COMPACT COMPOSITION OPERATORS 23

so that, if one sets B = 1/9ε:
    1
m |1 − ārφ| < 3(1 − r) = m |Cφ (ua,r )| > 1/9 ≤   1  ·
Ψ BΨ−1 1−r
We claim that this implies a good upper bound on m(|φ| > r), even if we lose
a factor 1/(1 − r), due to the effect of a rotation on φ. For that, we shall use the
following lemma.
Lemma 3.23. Let φ : D → D be an analytic function. Then, for every r with
0 < r < 1, there exists a ∈ T such that:
  1−r
m |1 − ār φ| < 3(1 − r) ≥ m(|φ| > r).
8
Admitting this for a while, we are going to finish the proof.
Fix an r such that rε ≤ r < 1, and take an a ∈ T as in Lemma 3.23. We get,
from the preceding, in setting λ = 1 − r:
m(1 − |φ| < λ) = m(|φ| > r)
8  
≤ m |1 − ārφ| < 3(1 − r)
1−r
8 1
≤   1  ,
1 − r Ψ BΨ−1 1−r
i.e., setting x = Ψ−1 (1/1 − r) = Ψ−1 (1/λ):
Ψ(x)
m(1 − |φ| < λ) ≤ 8 ·
Ψ(Bx)
 2
But Ψ satifies the Δ2 -condition: Ψ(y) ≤ Ψ(αy) for some α > 1 and y large
enough. Then, adjusting now ε > 0 as ε = 1/9αA, in order that B = αA, we get,
for x large enough, since A > 1:
Ψ(x)Ψ(Ax) ≤ [Ψ(Ax)]2 ≤ Ψ(Bx).
Therefore, for r close enough to 1:
8 8
m(1 − |φ| < λ) ≤ = ·
Ψ(Ax) χA (1/λ)
We hence get condition (OB4), and that proves Theorem 3.21. 

Proof of Lemma 3.23. Let λ = 1 − r, and let δ > 0 be a number which we shall
specify later. Consider the set:
Cδ = {z ∈ D ; |z| ≥ 1 − λ and |arg z| ≤ δ}
(for δ = λ, Cδ is a closed Carleson window). It is geometrically clear that (1 − λ)Cδ
is contained in the closed disk of center 1 and whose edge contains (1 − λ)2 eiδ ;
hence, for every z ∈ Cδ , one has:
|1 − (1 − λ)z|2 ≤ |1 − (1 − λ)2 eiδ |2 = 2(1 − λ)2 (1 − cos δ) + λ2 (2 − λ)2
≤ (1 − λ)2 δ 2 + λ2 (2 − λ)2 ≤ 9λ2
if δ ≤ λ.
By rotation, one has, for every a ∈ T:
|z| ≥ 1 − λ and |arg (āz)| ≤ δ ⇒ |1 − (1 − λ)āz| ≤ 3λ.
24 PASCAL LEFÈVRE, DANIEL LI, HERVÉ QUEFFÉLEC, LUIS RODRÍGUEZ-PIAZZA

Let now N ≥ 2 be the integer such that:
π π
≤λ< ,
N N −1
and take δ = π/N .
One has, by the previous inequalities, setting ak = e2ikδ :
 
{z ∈ D ; |z| ≥ 1 − λ} = āk Cδ ⊆ {z ∈ D ; |1 − (1 − λ)āk z| ≤ 3λ}.
1≤k≤N 1≤k≤N

it
Hence, with z = φ(e ) (remark that, by Proposition 3.22, we have only to consider
the values of eit for which |φ(eit )| < 1; however, in this lemma, we may replace D
by D), and get:
m(|φ| ≥ 1 − λ) ≤ N sup m(|1 − (1 − λ)āk φ| ≤ 3λ).
1≤k≤N

Therefore, we can find some k such that:
1 λ
m(|1 − (1 − λ)āk φ| ≤ 3λ) ≥
m(|φ| ≥ 1 − λ) ≥ m(|φ| ≥ 1 − λ),
N 8
since λ ≤ 2π/N ≤ 8/N . That proves Lemma 3.23. 

Since the Δ2 -condition implies the Δ1 -condition, which, in its turn, implies the
0
Δ -condition, we get, from Theorem 3.15, Theorem 3.20 and Theorem 3.21 that
the weak compactness of Cφ implies its order boundedness into M Ψ (T), and thanks
to the Remark after Proposition 3.14, its compactness. We get
Theorem 3.24. If Ψ satisfies the Δ2 -condition, then the following assertions
for composition operator Cφ : H Ψ → H Ψ are equivalent:
1) Cφ is order bounded into M Ψ (T);
2) Cφ is compact;
3) Cφ is weakly
 compact;
4) Ψ−1 1−|φ|
1
∈ M Ψ (T) (i.e.: χB (1/1 − |φ|) ∈ L1 (T) for every B > 0);

5) m(1 − |φ| < λ) = O χA (1/λ)
1
as λ → 0, for every A > 0;
1
6) sup Cφ (ua,r )Ψ = o  1  as r → 1 (W )
a∈T Ψ−1 1−r
Moreover these conditions
imply
7) φn p = o n1q , for every q ≥ 1 and every p ≥ 1.
8) Cφ is a nuclear operator on every H p , where p ≥ 1.
Remark. We shall see in the next Chapter (Theorem 4.22) that the assumption
Ψ ∈ Δ2 cannot be removed in general: Theorem 3.24 is not true for the Orlicz
 2 
function Ψ(x) = exp (log(x + 1) − 1 (which nevertheless satisfies Δ1 ).
Proof. Only the two last assertions are not proved yet.
First, we are going to show that 4) ⇒ 7).
Since Ψ ∈ Δ2 , there exists α > 1 such that Ψ(αx) ≥ [Ψ(x)]2 for x large enough.
Hence there exists a > 0 such that Ψ(αx) ≥ a[Ψ(x)]2 for every x > 0, and, for every
integer k ≥ 1, one has Ψ(αk x) ≥ a2 −1 [Ψ(x)]2 for every x > 0.
k k

Let q ≥ 1 and take k ≥ 1 such that 2 k−1
≤ q < 2k .
5. WEAKLY COMPACT COMPOSITION OPERATORS 25

We shall use the following inequality:
1 n n/2q
≥ t for 0 ≤ t < 1, n ≥ 1,
1−t 2q
which follows from the fact that the map t → td (1 − t) is less than 1
d+1 on [0, 1].
Take B = αk in 4) and set
 1 
C = Cq = Ψ αk Ψ−1 dm.
T 1 − |φ|
Since
 1  n 2 k

Ψ αk Ψ−1 ≥ a2 −1
k k
|φ|2 n/2q ,
1 − |φ| 2q
k−1
we get, since |φ| ≤ 1 and q ≥ 2k−1 imply that |φ|2 n/q ≥ |φn |:
n 2k
C ≥ a2 −1
k
φn 1 ,
2q
so:
C(2q)2 1
k 1
φn 1 ≤ 2k −1 2k = o q ,
a n n
since q < 2k .
Now, by interpolation, we actually have φn p = o n1q , for every q, p ≥ 1,
since φ∞ ≤ 1, and that proves 7).
To see that Cφ is nuclear on each H p , it suffices to notice that, for every f ∈ H p :

 ∞

Cφ (f ) = fˆ(n)φn = e∗n (f )φn
n=0 n=0

where e∗n (f ) = fˆ(n) is a norm one functional on H p . Moreover


e∗n (H p )∗ .φn p < ∞
n=0
and that ends the proof. 
xp
If one specializes this corollary to the case where Ψ(x) = Ψp (x) = e −1
(p > 0), which verifies the Δ2 -condition, we get, using Stirling’s formula:
Corollary 3.25. The following assertions are equivalent:
1) Cφ : H Ψp → H Ψp is order bounded into M Ψp (T);
2) Cφ : H Ψp → H Ψp is compact;
3) Cφ : H Ψp → H Ψp is weakly compact;
1
4) 1−|φ| ∈ Lr (T), ∀r ≥ 1;
5) ∀q ≥ 1 ∃Cq > 0: m(1 − |φ| < λ) ≤ Cq λ q
;
 −1/p

6) sup Cφ (ua,r )Ψp = o log(1 − r)  as r → 1
a∈T
7) ∀q ≥ 1 φn 2 =√o (n−q );
8) φn Ψp = o (1/ p log n).
Remark. Observe that conditions 4), 5) and 7) do not depend on p. Hence this is
equivalent to the same properties of Cφ acting on H Ψs for another s > 0.
As a consequence of Theorem 3.24, we obtain the following:
26 PASCAL LEFÈVRE, DANIEL LI, HERVÉ QUEFFÉLEC, LUIS RODRÍGUEZ-PIAZZA

Corollary 3.26. Assume that Ψ ∈ Δ2 . Then there exist compact composition
operators Cφ : H p → H p for 1 ≤ p < ∞ (even Hilbert-Schmidt on H 2 ) which are
not compact as operators Cφ : H Ψ → H Ψ .
Remark. We shall see in Theorem 4.3 that compactness on H Ψ implies com-
pactness on H p for p < ∞. Note that this shows that, though H Ψ is an inter-
polation space between H 1 and H ∞ (see [4], Theorem V.10.8), the compactness
of Cφ : H 1 → H 1 with the continuity of Cφ : H ∞ → H ∞ does not suffice to have
compactness for Cφ : H Ψ → H Ψ .
Remark. Corollary 3.26 will be improved in [27], Theorem 4.1, where we show
that whenever Ψ ∈ / Δ2 , there exists a composition operator which is compact on
H 2 , but not on H Ψ .
 
Proof. As the Δ2 condition implies the Δ0 condition, we have x = o χβ (x) as
x → ∞, for some β > 1. It follows that we can find a positive function a : T → R+
such that a ≥ 2, a ∈ L1 but χβ (a) ∈ / L1 . Set h = 1 − a1 . One has 1/2 ≤ h ≤ 1 and
in particular log h ∈ L (T). Then the outer function φ : D → C defined for z ∈ D
1

by:

u+z
φ(z) = exp log h(u) dm(u)
T u−z
is analytic on D and its boundary limit verifies |φ| = h ≤ 1 on T. By [42],

Theorem
6.2, Cφ : H 2 → H 2 is Hilbert-Schmidt, and hence compact, since T 1−|φ| 1
dm =

T
a dm < +∞.
It is then
 1 compact
from H p
to H p
for every p < ∞ ([42], Theorem
6.1). However, T χβ 1−|φ| dm = T χβ (a) dm = +∞, and hence, by our Theorem
3.24, Cφ is not compact on H Ψ . 

6. p-summing operators
Recall that an operator T : X → Y between two Banach spaces is said to be
p-summing (1 ≤ p < +∞) if there is a constant C > 0 such that, for every choice
of x1 , . . . , xn ∈ X, one has:

n 
n
T xk p ≤ C p sup |x∗ (xk )|p .
x∗ ∈X ∗
k=1 x∗ ≤1 k=1

In other terms, T maps weakly unconditionaly p-summable sequences into
 norm p-
summable sequences. When X ⊆ Y = LΨ , this implies that whenever n≥1 |gn | ∈

LΨ , then n≥1 T gn pΨ < +∞.
For 1 ≤ p < +∞, J. H. Shapiro and P. D. Taylor proved in [42], Theorem 6.2,
that the condition:

dm
(3.3) < +∞
T 1 − |φ|

implies that the composition operator Cφ : H p → H p is p-summing (and condition
(3.3) is necessary for 1 ≤ p ≤ 2; in particular, for p = 2, it is equivalent to say that
Cφ is Hilbert-Schmidt). Actually, they proved that (3.3) is equivalent to the fact
that Cφ is order bounded on H p , and (acknowledging to A. Shields, L. Wallen, and
J. Williams) every order bounded operator into an Lp -space is p-summing. The
6. p-SUMMING OPERATORS 27

counterpart of (3.3) in our setting, are conditions (OB1) and (OB2)
1
χA dm < +∞
T 1 − |φ|
in Proposition 3.14. We are going to see that, if Ψ grows fast enough, order bound-
edeness does not imply that Cφ is p-summing. Note that, for composition operators
on H ∞ , being p-summing is equivalent to being compact ([23], Theorem 2.6), but
H ∞ corresponds to the very degenerate Orlicz function Ψ(x) = 0 for 0 ≤ x ≤ 1
and Ψ(x) = +∞ for x > 1, which does not match in the proof below.
Theorem 3.27. If Ψ ∈ Δ2 , then there exists a composition operator Cφ : H Ψ →
H which is order bounded into M Ψ (T), and hence compact, but which is p-
Ψ

summing for no p < ∞.
Note that every p-summing operator is Dunford-Pettis (it maps the weakly
convergent sequences into norm convergent sequences); therefore, when it starts
from a reflexive space, it is compact. However, when Ψ ∈ Δ2 , being Dunford-Pettis
implies compactness for composition operators on H Ψ , though H Ψ is not reflexive,
thanks to the next proposition and Theorem 3.24. Later (see Theorem 4.21), we
shall see that, under condition Δ0 , every Dunford-Pettis composition operator is
compact.
Proposition 3.28. When Ψ ∈ ∇2 , every Dunford-Pettis composition operator
satisfies condition (W ).
 
Proof. Let ga,r = Ψ−1 1/(1 − r) ua,r . If condition (W) were not satisfied, we
could find a sequence (an )n≥1 in T and a sequence of numbers (rn )n≥1 tending  to
1 such that Cφ (gan ,rn )Ψ ≥ δ > 0 for all n ≥ 1. But (1 − r)2 Ψ−1 1/(1 − r) −→ 0.
  r→1
Therefore gan ,rn (z) = (1 − rn )2 Ψ−1 1/(1 − rn ) /(1 − ān rn z) tends to 0 uniformly
on compact sets of D. Hence, by Proposition 3.7, (gan ,rn )n≥1 tends weakly to 0
(because gan ,rn ∈ HM Ψ and, on HM Ψ, the weak-star Ψ
 topology of H is the weak
topology). Since Cφ is Dunford-Pettis, Cφ (gan ,rn ) n≥1 tends in norm to 0, and we
get a contradiction, proving the proposition. 
Proof of Theorem 3.27. We shall begin with some preliminaries. First, since
 2
Ψ ∈ Δ2 , there exists α > 1 such that Ψ(x) ≤ Ψ(αx) for x large enough. Hence:
Ψ(x) x
≥Ψ −→ +∞.
Ψ(x/α) α x→+∞
Therefore, there exists, for every n ≥ 1, some xn > 0 such that:
Ψ(x)
≥ 2n , ∀x ≥ xn .
Ψ(x/α)
Then
x 1 x 1
n
Ψ ≤ n Ψ(x) + Ψ ≤ n Ψ(x) + Ψ(xn ) , ∀x > 0.
α 2 α 2
For convenience, we shall assume, as we may, that Ψ(xn ) ≥ 1.
Remark also that, setting a = Ψ−1 (1), one has, for every f ∈ L∞ :

|f |
Ψ a dm ≤ 1,
T f ∞
28 PASCAL LEFÈVRE, DANIEL LI, HERVÉ QUEFFÉLEC, LUIS RODRÍGUEZ-PIAZZA

so that:
1
f Ψ ≤f ∞ .
a
We are now going to start the construction.
For n ≥ 1, let Mn = log(n + 1). Choose positive numbers βn which tend to 0
fast enough to have:

βk ≤ βn , ∀n ≥ 1,
k>n
and
Ψ−1 (8/βn )
tn = −→ +∞,
Mn n→+∞

with (tn )n≥1 non-decreasing and Ψ(tn ) > 1.
Set:
1
rn = 1 − ·
Ψ(tn )
One has rn −→ 1 and:
n→+∞
1 8
χM n = ·
1 − rn βn
Actually, for the end of the proof, we shall have to choose the βn ’s decreasing
so fast that:  n
t1 + · · · + tn−1 2 tn 1
Ψ + Ψ(xn ) ≤ n·
α Ψ(tn ) 2
This is possible, by induction, since t/Ψ(t) −→ 0. Note that, since Ψ(xn ) ≥ 1,
t→+∞
 2n tn
one has, in particular: +∞
n=1 Ψ(tn ) < +∞.

 measurable subsets of T with measure m(Bn ) = cβn (where
Let Bn be disjoint
c ≥ 1 is such that n≥1 βn = 1/c), and whose union is T. Define h : T → C by:

h= rn IBn .
n≥1 

has log h ∈ L (T), since h does not vanish, rn ≥ 1/2 for n large enough, and
1
One
n m(Bn ) = 1 < +∞. We can define the outer function:

u+z
φ(z) = exp log h(u) dm(u) , |z| < 1.
T u−z

φ is analytic on D and its boundary limit verifies |φ| = h ≤ 1 on T. Hence φ defines
a composition operator on H Ψ .
For any A > 0, one has, when n is large enough to ensure Mn ≥ A, and when
rn ≤ r < rn+1 :
  8c
m(|φ| > r) = m(|φ| = rk ) = c βk ≤ c βn =  
k>n k>n
χ M n
1/(1 − r n )
8c 8c
≤  ≤  ·
χA 1/(1 − rn ) χA 1/(1 − r)
Since rn −→ 1, it follows from Theorem 3.15 that Cφ : H Ψ → H Ψ is order-
n→+∞
bounded into M Ψ (T) (and hence is compact).
6. p-SUMMING OPERATORS 29 

We are Ψ now going
 to construct a sequence of functions gn ∈ H Ψ such that
p
n |gn | ∈ L , but n Cφ (gn )Ψ = +∞ for all 1 ≤ p < ∞. That will prove that
Cφ is p-summing for no p < ∞.
Since
m(|φ| ≥ rn ) ≥ m(|φ| = rn ) = c βn ≥ βn ,
we can apply Lemma 3.17 and Lemma 3.23 (which remain valid with non-strict
inequalities instead of strict ones), and we are able to find, for every n ≥ 1, some
an ∈ T such that:
1/9
Cφ (uan ,rn )Ψ ≥ ·
8
Ψ−1
(1 − rn )βn
But:
8
= Ψ(tn )Ψ(Mn tn ).
(1 − rn )βn
Since now Ψ satisfies Δ2 : [Ψ(x)]2 ≤ Ψ(αx), for x large enough and one has, for n
large enough, since Mn ≥ 1 (for n ≥ 2):
 2
Ψ(tn )Ψ(Mn tn ) ≤ Ψ(Mn tn ) ≤ Ψ(αMn tn ).
Therefore:
1/9
Cφ (uan ,rn )Ψ ≥ ·
αMn tn
Taking now:

1
gn = Ψ−1 uan ,rn = tn uan ,rn ,
1 − rn
one has gn Ψ ≤ 1 (by Corollary 3.10), and
1/9 1/9
Cφ (gn )Ψ ≥ = ·
αMn α log(n + 1)
Therefore

+∞
Cφ (gn )pΨ = +∞
n=1
for every p ≥ 1. 
It remains to show that g = n |gn | ∈ LΨ .
We shall follow the lines of proof of Theorem II.1 in [24].
By Markov’s inequality, one has:
2n tn
m(|gn | > 2−n ) ≤ 2n tn uan ,rn 1 ≤ 2n tn (1 − rn ) = ·
Ψ(tn )
Set: 
An = {|gn | > 2−n } ; Ãn = An \ Aj ,
j>n
and
ğn = gn I{|gn |>2−n } .
Since:

+∞
1
+∞
1 1
+∞
1
gn − ğn Ψ ≤ gn − ğn ∞ ≤ n
= < +∞,
n=1
a n=1 a n=1 2 a

it suffices to show that ğ = n |ğn | ∈ L .
Ψ
30 PASCAL LEFÈVRE, DANIEL LI, HERVÉ QUEFFÉLEC, LUIS RODRÍGUEZ-PIAZZA

    
But ğ vanishes out of n≥1 Ãn ∪ lim supn An , and m lim supn An = 0, since

+∞ 
+∞
2n tn
m(An ) ≤ < +∞ .
n=1 n=1
Ψ(tn )
Therefore:
|ğ| +∞
 |ğ|
Ψ dm = Ψ dm.
T 2α n=1 Ãn

Since ğj = 0 on Ãn for j > n, we get:
+∞
 |ğ | + · · · + |ğ |
|ğ| 1 n
Ψ dm = Ψ dm.
T 2α n=1 Ãn

Now, by the convexity of Ψ:
|ğ | + · · · + |ğ | 1  |ğ | + · · · + |ğ |ğ |
1 n 1 n−1 | n
Ψ ≤ Ψ +Ψ ·
2α 2 α α
But:
|ğ | 1
n
Ψ ≤ n Ψ(|ğn |) + Ψ(xn ) ,
α 2
and
|ğ | + · · · + |ğ t + · · · + t
1 n−1 | 1 n−1
Ψ ≤Ψ ;

α
α
therefore, using that T Ψ(|ğn |) dm ≤ T Ψ(|gn |) dm ≤ 1:
+∞ 
|ğ|  1 t1 + · · · + tn−1
Ψ dm ≤ Ψ m(Ãn )
T 2α n=1
2 α

1
+ n Ψ(|ğn |) dm + Ψ(xn )m(Ãn )
2 T
+∞ 
 1 t1 + · · · + tn−1  2n t
n 1
≤ Ψ + Ψ(xn ) +
n=1
2 α Ψ(tn ) 2n
+∞ 
 1 1 1
≤ + = 1,
n=1
2 2n 2n

which proves that ğ ∈ LΨ , and ğΨ ≤ 2α.
The proof is fully achieved. 
Remark. In the above  proof, we chose Mn = log(n + 1). This choice was only
used to conclude that n Cφ (gn )pΨ = +∞ for every p < ∞. Therefore, the above
proof shows that, given any increasing function Υ : (0, ∞) → (0, ∞) tending to
∞, we can find, with a suitable choice of a slowly increasing
 sequence (Mn )n≥1 ,
n n | ∈ L , although
|g
Ψ Ψ
a 
symbol φ and a sequence (g )
n n≥1 in H such that
n Υ(C (g
φ n Ψ) ) = +∞.

Remark. Proposition 3.14 obviously implies that, if the composition operator
Cφ : H Ψ → H Ψ is order bounded into M Ψ (T), then Cφ : H 2 → H 2 is order bounded
(i.e. Hilbert-Schmidt). One could expect that under such an hypothesis with a fast
growth condition for Ψ, Cφ : H 2 → H 2 is actually in a “better” Schatten class,
6. p-SUMMING OPERATORS 31

nuclear for instance. This is the case when Ψ ∈ Δ2 : see 8) in Theorem 3.24. Is it
the case too under a weaker growth condition for Ψ?
In the same spirit, if the composition operator Cφ : H Ψ → H Ψ is compact and
Ψ verifies a very fast growth condition (such as Δ2 for instance and Cφ is even
order bounded into M Ψ (T) in that situation), we could expect that Cφ : H Ψ → H Ψ
is actually nuclear as this is the case when H Ψ is replaced by H ∞ . But, this is false
as shown by Theorem 3.27: the composition operator cannot be nuclear since it is
not an absolutely summing operator.
CHAPTER 4

Carleson measures

1. Introduction
B. MacCluer ([30]; see also [10], Theorem 3.12) has characterized compact
composition operators on Hardy spaces H p (p < ∞) in term of Carleson measures.
In this chapter, we shall give an improvement of this result for Hardy-Orlicz spaces
H Ψ , but in terms of “Ψ-Carleson measures”. Indeed, Carleson measures do not
characterize the compactness of composition operators when Ψ grows too quickly,
as it follows from Corollary 3.26.
Before that, we shall recall some definitions (see for example [10], pages 37–38,
or [11], page 157).
Let ξ ∈ T and h ∈ (0, 1). Define

(4.1) S(ξ, h) = {z ∈ D ; |ξ − z| < h}.

The Carleson window W (ξ, h) is the following subset of D:

(4.2) W (ξ, h) = {z ∈ D ; 1 − h < |z| ≤ 1 and ¯ < h}.
|arg(z ξ)|

It is easy to show that we have for every ξ ∈ T and h ∈ (0, 1):

S(ξ, h/2) ⊆ W (ξ, h) and W (ξ, h/2) ⊆ S(ξ, h) ,

so that, in the sequel, we may work equivalently with either S(ξ, h) or W (ξ, h).
Recall that a positive Borel measure μ on D (or D) is called a Carleson measure if
there exists some constant K > 0 such that:
 
μ S(ξ, h) ≤ Kh , ∀ξ ∈ T , ∀h ∈ (0, 1).

Carleson’s Theorem (see [10], Theorem 2.33, or [11], Theorem 9.3) asserts that, for
0 < p < ∞, the Hardy space H p is continuously embedded into Lp (μ) if and only
if μ is a Carleson measure.
Given an analytic self-map φ : D → D, we define the pullback measure μφ on
the closed unit disk D (which we shall denote simply μ when this is unambiguous)
as the image of the Haar measure m of T = ∂D under the map φ∗ (the boundary
limit of φ):
 
(4.3) μφ (E) = m φ∗ −1 (E) ,

for every Borel subset E of D.
The automatic continuity of composition operators Cφ on the Hardy space H p ,
combined with Carleson’s Theorem means that μφ is always a Carleson measure.
33
34 PASCAL LEFÈVRE, DANIEL LI, HERVÉ QUEFFÉLEC, LUIS RODRÍGUEZ-PIAZZA

B. MacCluer ([30], [10], Theorem 3.12) showed that:
The composition operator Cφ is compact on H 2
  if and only if
(MC) μφ S(ξ, h) = o (h) as h → 0, uniformly for ξ ∈ T.
  if and only if
μφ W (ξ, h) = o (h) as h → 0, uniformly for ξ ∈ T.
While the Shapiro’s compactness criterion, via the Nevanlinna counting func-
tion ([39]), deals with the behavior of φ inside the open unit disk, the character-
ization (M C) deals with its boundary values φ∗ . It is natural to wonder whether
the modulus of φ∗ on T = ∂D suffices to characterize the compactness of Cφ . This
leads to the following question: if two functions φ1 and φ2 have the same modulus
on T, are the compactness of the two associated composition operators equivalent?
We have seen in Theorem 3.24 that the answer is positive on H Ψ when Ψ ∈ Δ2 .
However, on H 2 it turns out to be negative. We give the following counterexample.
Theorem 4.1. There exist two analytic functions φ1 and φ2 from D into itself
such that |φ∗1 | = |φ∗2 | on T but for which the composition operator Cφ2 : H 2 → H 2
is compact, though Cφ1 : H 2 → H 2 is not compact.
Remark. Let Ψ be an Orlicz function which satisfies Δ2 . We shall see in Theorem
4.3 that every composition operator Cφ : H 2 → H 2 is compact as soon as Cφ : H Ψ →
H Ψ is compact. Hence, in the above theorem, Cφ1 : H Ψ → H Ψ is not compact.
It follows hence from Theorem 3.24, since φ∗1 and φ∗2 have the same modulus,
that Cφ2 : H Ψ → H Ψ is not compact (and, even, not weakly compact), though
Cφ2 : H 2 → H 2 is compact. We have already seen such a phenomenon in Corollary
3.26. However, the results ofthe next section
 will allow us to conclude (Theorem
4.22) that, when Ψ(x) = exp log(x+1)2 −1, which does not satisfy condition Δ2 ,
but satisfies conditions Δ1 and ∇1 , the composition operator Cφ2 : H Ψ → H Ψ is
compact, but not order bounded into M Ψ (T). That will show that our assumption
that Ψ ∈ Δ2 in Theorem 3.24 is not only a technical one. Nevertheless, it will be
shown in [27] that Theorem 3.24 holds for a larger class of Orlicz functions.
We shall improve Theorem 4.1 in [26], Theorem 4.2, where we show that more-
over, for every p > 2, the composition operator Cφ2 can be taken in the Schatten
class Sp .
Proof. We start simply with φ1 (z) = 1+z 2 . It is well known that Cφ1 is not
compact on H 2 (this was first observed in H. J. Schwartz’s thesis [37]: see [42],
page 471). Now, let:
1 + z
M (z) = exp −
1−z
and
φ2 (z) = φ1 (z) M (z).
For simplicity, we shall write φ = φ2 , and we are going to show that Cφ is a compact
operator on H 2 , using the criterion (MC).
Let ξ = eiα ∈ T, with |α| ≤ π. We are going to prove that:
 
μφ W (ξ, h) = O (h3/2 ).
For θ ∈ (−π, π), one has |φ(eiθ )| = |φ1 (eiθ )| = cos(θ/2), and so the condition
1 − h < |φ(eiθ )| ≤ 1 is equivalent to 1 − h < cos(θ/2) < 1, which implies, since
1. INTRODUCTION 35

cos t = 1 − 2 sin2 (t/2) ≤ 1 − 2t2 /π 2 ≤ 1 − t2 /5 for 0 ≤ t ≤ π/2 (because sin t ≥ π2 t),

that 1 − h < 1 − (θ/2) 
2
/5, i.e.  θ 2 ≤ 20h and so |θ| ≤ 6 h. On the other
hand, M (eiθ ) = exp − i cot(θ/2) ; hence arg φ(eiθ ) = θ/2 − cot(θ/2), modulo 2π.
Therefore, for h small enough:
  √
μφ W (ξ, h) ≤ m({|θ| ≤ 6 h ; | − α + θ/2 − cot(θ/2)| ≤ h, mod 2π})
 √
≤2 m({|t| ≤ 3 h ; | − α + t − cot t + 2πn| ≤ h}).
n∈Z
 √
We have to majorize both m({0 < t ≤ 3 h ; | − α + t − cot t + 2πn| ≤ h}) and 
√ n∈Z
m({0 < t ≤ 3 h ; |α + t − cot t + 2πn| ≤ h}).
n∈Z
The function
F (t) = t − cot(t)
is increasing , and we define an , bn ∈ (0, π) by:
F (an ) = α − 2π(nh + n) − hand F (bn ) = α − 2π(nh + n) + h,

where the integer nh is given large enough to ensure that a0 ≤ 3 h. Of course,
a√ √ . Observe that 2h = F (b0 ) − F (a0 ) ≥ b0 − a0 , and then b0 ≤
n < bn < an−1
3 h + 2h ≤ 4 h for h small enough. One has:
 √ ∞
m({0 < t ≤ 3 h ; | − α + t − cot t + 2πn| ≤ h}) ≤ (bn − an ).
n∈Z n=0
1 1
Since F  (t) = 1 + 2 ≥ t2 , one has, on the one hand:
sin t
bn
bn − an
(4.4) 2h = F (bn ) − F (an ) = F  (t) dt ≥ ;
an an bn
hence:
(4.5) bn − an ≤ 2han bn ≤ 2hb2n , for all n ≥ 0.
On the other hand, let us first point out that, for 0 ≤ t ≤ 1,
1 π 2 /4 4
F  (t) = 1 + 2 ≤ 1 + t2 ≤ 2·
sin t t
Hence, for h small enough:
bn
bn − bn+1
2π = F (bn ) − F (bn+1 ) = F  (t) dt ≤ 4 ,
bn+1 bn+1 bn
and we get:
2
b2n+1 ≤ bn+1 bn ≤ (bn − bn+1 ).
π
Hence, using the fact that (4.4) gives b0 − a0 ≤ 2ha0 b0 ≤ 24 h2 , we get, from (4.5):
∞ ∞
4h 
(bn − an ) ≤ (b0 − a0 ) + (bn − bn+1 )
n=0
π n=0
4h 4h √ √ 16 3/2
≤ 24 h2 + b0 ≤ 24 h2 + 4 h ≤ 24 h + h ≤ 6 h3/2 ,
π π π
for h small enough.
36 PASCAL LEFÈVRE, DANIEL LI, HERVÉ QUEFFÉLEC, LUIS RODRÍGUEZ-PIAZZA

In the same way, we have
∞ √
m({0 < t ≤ 3 h ; |α + t − cot t + 2πn| ≤ h}) ≤ 6 h3/2 .
n=0
 
We can hence conclude that μφ W (ξ, h) ≤ Ch3/2 , where C is a numerical
constant. 
Remark. Cφ actually maps continuously H 2 into H 3 , and compactly H 2 into H p ,
for any p < 3 (see [15] or Theorem 4.10 and 4.11).
However, in some cases, the behaviour of |φ∗ | on the boundary ∂D suffices.
Proposition 4.2. Let φ1 and φ2 be two analytic self-maps of D such that
|φ∗1 | ≤ |φ∗2 | on ∂D. Assume that they are both one-to-one on D, and that there
exists a ∈ D such that φ2 (a) = 0. Then the compactness of Cφ2 : H 2 → H 2 implies
that of Cφ1 : H 2 → H 2 .
Proof. By composing φ1 and φ2 with the automorphism of D which maps 0 into
a, we may assume that a = 0. We can hence write φ2 (z) = zφ(z), with φ : D → C
analytic in D. φ does not vanish in D because of the injectivity of φ2 (this is obvious
for z = 0, and for z = 0, follows from the fact that the injectivity of φ2 implies
φ2 (0) = 0).
Then there is some δ > 0 such that |φ(z)| ≥ δ for every z ∈ D. In fact,
by continuity, there is some α > 0 and some 0 < r < 1 such that |φ(z)| ≥ α
and |φ2 (z)| ≤ α for |z| ≤ r. But being analytic and non constant, φ2 is an open
map, so there is some ρ > 0 such that ρD ⊆ φ2 (rD). Injectivity of φ2 shows that
φ2 (D \ rD) ∩ ρD = ∅, that is to say that |φ2 (z)| ≥ ρ for |z| > r. A fortiori |φ(z)| ≥ ρ
for |z| > r. The claim is proved with δ = min(α, ρ).
 φ∗   φ∗   1 (z) 
Then φ1 ∈ H ∞ , as well as φφ1 . Since  φ1∗  =  φ1∗  ≤ 1, one has  φφ(z)  ≤ 1 for
2
every z ∈ D. Hence:
1 − |φ1 (z)| 1 − |φ(z)| 1 − |φ2 (z)|
≥ = − |φ(z)|.
1 − |z| 1 − |z| 1 − |z|
Now ([39], Theorem 3.5), the compactness of Cφ2 : H 2 → H 2 implies that
1 − |φ2 (z)|
lim− = +∞;
z→1 1 − |z|
we get:
1 − |φ1 (z)|
lim = +∞,
1 − |z|
z→1−
which implies the compactness of Cφ1 : H 2 → H 2 , thanks to the injectivity of φ1
([39], Theorem 3.2). 

2. Compactness on H Ψ versus compactness on H 2
The equivalence (MC) holds actually for every H p space (with p < ∞) instead
of H 2 . We are going to see in this chapter that for Hardy-Orlicz spaces H Ψ ,
one needs a new notion of Carleson measures, which one may call Ψ-Carleson
measures. Before that, we are going to see that condition (MC) allows to get that
the compactness of composition operators on H Ψ always implies that on H p for
p < ∞. Recall that, when Ψ ∈ Δ2 , we have seen in Corollary 3.26 that the converse
is not true.
2. COMPACTNESS ON H Ψ VERSUS COMPACTNESS ON H 2 37

Theorem 4.3. Let φ : D → D be an analytic function. If one of the following
conditions:
i) Cφ is a compact operator on H Ψ
ii) Ψ ∈ Δ0 and Cφ is a weakly compact operator on H Ψ
is satisfied, the composition operator Cφ is compact on H 2 .
Note that we have proved in Theorem 3.24 that the weak compactness of
Cφ : H Ψ → H Ψ implies its compactness only when Ψ satisfies the Δ2 condition.
Nevertheless, we shall show in Theorem 4.21 that when Ψ ∈ Δ0 , the weak compact-
ness of Cφ is equivalent to its compactness. This is obviously false (in particular
when LΨ is reflexive) without any assumption on Ψ.
Proof. We are going to use the characterization (MC) for compact composition
operators on H 2 .
Suppose that the condition on μφ is not fulfilled. Then there exist β ∈ (0, 1),
ξn ∈ T, and hn ∈ (0, 1), with hn −→ 0, such that:
n→+∞
 
μφ S(ξn , hn ) ≥ βhn .
We are now going to use the functions:
h2n
vn (z) = ,
(1 − an z)2
where:
an = (1 − hn )ξn .
Of course, vn is actually nothing but uξn ,1−hn . We have by Corollary 3.10:
1
vn Ψ ≤ −1 ·
Ψ (1/hn )
Define gn = Ψ−1 (1/hn )vn , which is in the
 unit
Ψ
 ball of HM . We have assumed at
the beginning of the paper that x = o Ψ(x) as x → ∞; hence Ψ−1 (x) = o (x) as
x → ∞, and so h2n Ψ−1 (1/hn ) → 0. Therefore (gn )n converges uniformly to zero
on compact subsets of D and gn 1 → 0, because gn 1 ≤ hn Ψ−1 (1/hn ). Then, in
both cases, we should have Cφ (gn )Ψ → 0. Indeed, in case i), this follows from
Proposition 3.6, and in case ii), this follows from [25], Theorem 4.
We are going to show that this is not true. Indeed:

4 4
Ψ |gn ◦ φ| dm ≥ Ψ Ψ−1 (1/hn )|vn (z)| dμφ
T β β
D 4
≥ Ψ Ψ−1 (1/hn )|vn (z)| dμφ .
S(ξn ,hn ) β
But when z ∈ S(ξn , hn ), one has |vn (z)| ≥ 1/4, because
|1 − an z| ≤ |1 − an ξn | + |an (ξn − z)| = hn + (1 − hn )hn ≤ 2hn .
We obtain that by convexity (since β < 1):

4 1  −1
Ψ |gn ◦ φ| dm ≥ Ψ 4Ψ (1/hn )|vn (z)|) dm
T β T β
1  
≥ μφ S(ξn , hn ) ≥ 1.
βhn
This implies that Cφ (gn )Ψ ≥ β/4 and proves the theorem. 
38 PASCAL LEFÈVRE, DANIEL LI, HERVÉ QUEFFÉLEC, LUIS RODRÍGUEZ-PIAZZA

3. General measures
We used several times the criterion (MC) for compactness on H 2 via Carleson
measures. The fact that this provides such a useful tool leads to wonder if the
boundedness and the compactness on Hardy-Orlicz spaces can be expressed in such
a pleasant manner. Theorem 4.10 and Theorem 4.11 are the Orlicz version of
Carleson’s Theorem for H p spaces ([10], Theorem 2.35).
The key for our general characterization is the use of the following functions
(see (4.2) for the definition of the Carleson’s window W (ξ, h)).
Definition 4.4. For any positive finite Borel measure μ on the unit disk D (or
on D), we set, for h ∈ (0, 1]:
 
(4.6) ρμ (h) = sup μ W (ξ, h) ,
ξ∈T
ρμ (t)
(4.7) Kμ (h) = sup
0<t≤h t
 
Hence μ W (ξ, t) ≤ Kμ (h)t for t ≤ h.
The measure μ is a Carleson measure if and only if Kμ (h) is bounded by a
constant K, for h ∈ (0, 1) and this happens as soon as Kμ (h0 ) is finite for some
h0 ∈ (0, 1).
In Theorem 4.11, we shall use the following condition.
Definition 4.5. We say that Ψ satisfies the ∇0 condition if for some x0 > 0,
C ≥ 1 and every x0 ≤ x ≤ y, one has:
Ψ(2x) Ψ(2Cy)
(4.8) ≤ ·
Ψ(x) Ψ(y)
This is a condition on the regularity of Ψ. It is satisfied if
Ψ(2x) Ψ(2y)
≤C ·
Ψ(x) Ψ(y)
Before stating our main theorems, let us give some results about this condition.
Proposition 4.6. The following assertions are equivalent
i) Ψ satisfies the ∇0 condition.
ii) There exists some x0 > 0 satisfying: for every β > 1, there exists Cβ ≥ 1
such that
Ψ(βx) Ψ(βCβ y)
≤ , for every x0 ≤ x ≤ y.
Ψ(x) Ψ(y)
iii) There exist x0 > 0, β > 1 and Cβ ≥ 1 such that
Ψ(βx) Ψ(βCβ y)
≤ , for every x0 ≤ x ≤ y.
Ψ(x) Ψ(y)
Proof. We only have to prove i) ⇒ ii), since iii) ⇒ i) is similar and ii) ⇒ iii) is
trivial.
If β ∈ (1, 2], it is easy, taking Cβ = 2C/β. Now, if β ∈ (2b , 2b+1 ] for some
integer b ≥ 1, we write for every x0 ≤ x ≤ y:
Ψ(βx) Ψ(2b+1 x) Ψ(2b+1 x) Ψ(2x)
≤ = b
··· ·
Ψ(x) Ψ(x) Ψ(2 x) Ψ(x)
3. GENERAL MEASURES 39

But we have for every integer j ≥ 1: 2j−1 x ≤ (2C)j−1 x ≤ (2C)j−1 y, so:
Ψ(2j x) Ψ((2C)j y)
j−1
≤ ,
Ψ(2 x) Ψ((2C)j−1 y)
and we obtain:
Ψ(βx) Ψ((2C)b+1 y) Ψ(βCβ y)
≤ ≤ ,
Ψ(x) Ψ(y) Ψ(y)
where Cβ = 2C b+1 . 
Examples. It isimmediately α seen that the following functions satisfy ∇0 : Ψ(x) =
α
xp , Ψ(x) = exp log(x + 1) − 1, Ψ(x) = ex − 1, α ≥ 1.
Note that when Ψ ∈ ∇0 , with constant C = 1, i.e. Ψ(βx)/Ψ(x) is increasing
for x large enough, then we have the dichotomy: either Ψ ∈ Δ2 , or Ψ ∈ Δ0 .
We shall say that ∇0 is uniformly satisfied if there exist C ≥ 1 and x0 > 0 such
that, for every β > 1:
Ψ(βx) Ψ(Cβy)
(4.9) ≤ for x0 ≤ x ≤ y.
Ψ(x) Ψ(y)

One has:
Proposition 4.7.
1) Condition Δ2 implies condition ∇0 uniformly.
2) If Ψ ∈ ∇0 uniformly, then Ψ ∈ ∇1 .
3) The function κ(x) = log Ψ(ex ) is convex on (x0 , +∞) if and only if, in ( 4.9),
∇0 is satisfied with constant C = 1.
We shall say that Ψ is κ-convex when κ is convex at infinity. Note that Ψ is
κ-convex whenever Ψ is log-convex. In the above examples Ψ is κ-convex; it also
the case of Ψ(x) = x2 / log x, x ≥ e; but, on the other hand, if Ψ(x) = x2 log x for
x ≥ e, then Ψ is not κ-convex. Nevertheless, for β 2 ≤ x ≤ y, one has:
Ψ(βx) log β 3β 2 3 Ψ(βy)
= β2 1 + ≤ ≤ ,
Ψ(x) log x 2 2 Ψ(y)
and hence Ψ ∈ ∇0 .
We do not know whether Ψ ∈ ∇0 uniformly implies that Ψ is equivalent to an
Orlicz function for which the associated function κ is convex.
 2
Proof. 1) Since Ψ ∈ Δ2 , one has Ψ(u) ≤ Ψ(αu) for some α > 1 and x ≥ x0 .
We may assume that Ψ(x0 ) ≥ 1. Then, for y ≥ x ≥ x0 and every β > 1:
 2
Ψ(βx)Ψ(y) ≤ Ψ(βy) ≤ Ψ(αβy) ≤ Ψ(αβy)Ψ(x),
which is (4.9).
2) Suppose that Ψ satisfies condition ∇0 uniformly. We may assume that
Ψ(x0 ) ≥ 1. Let x0 ≤ u ≤ v; we can write u = βx0 for some β ≥ 1. Then
condition (4.9) gives:
 
Ψ(u)Ψ(v) = Ψ(βx0 )Ψ(v) ≤ Ψ(x0 )Ψ(Cβv) ≤ Ψ Ψ(x0 )Cβv = Ψ(buv),
with b = CΨ(x0 )/x0 .
3) Assume that κ is convex on (x0 , +∞). For every β > 1, let κβ (t) =
κ(t log β) = log(Ψ(β t )), which is convex on (x0 / log(β), +∞). Taking y ≥ x ≥ ex0 ,
40 PASCAL LEFÈVRE, DANIEL LI, HERVÉ QUEFFÉLEC, LUIS RODRÍGUEZ-PIAZZA


write x = β θ and y = β θ with θ ≤ θ  . Convexity of κ gives, since θ  ≥ θ ≥
x0 / log(β):
κβ (θ + 1) − κβ (θ) ≤ κβ (θ  + 1) − κβ (θ  ),
which means that:
Ψ(βx) Ψ(βy)
≤ ·
Ψ(x) Ψ(y)
Assume that (4.9) is fulfilled for every β > 1, with C = 1. Then, taking y = βx,
one has:
 2
Ψ(βx) ≤ Ψ(x)Ψ(β 2 x).
Let u < v be large enough. Taking x = u2 and β = v/u, we get:
 2
Ψ(uv) ≤ Ψ(u2 )Ψ(v 2 ),
which means that κ is convex. 
Remark. The growth and regularity conditions for Ψ can be expressed in the
following form:
• Ψ ∈ Δ0 iff κ(x + β  ) − κ(x) −→ +∞, for some β  > 0.
x→∞
• Ψ ∈ Δ1 iff for some β  > 0, one has x + κ(x) ≤ κ(x + β  ), for x large
enough.
• Ψ ∈ Δ2 iff for some α > 0, one has 2κ(x) ≤ κ(x + α ), for x large enough.
• Ψ ∈ ∇1 iff κB (x)+κB (y) ≤ κB (x+y) for x, y large enough, with B = e−b .
• Ψ ∈ ∇0 iff for some c ≥ 1 and A > 1, one has κA (θ + 1) − κA (θ) ≤
κA (θ  + c ) − κA (θ  ) for θ ≤ θ  large enough.
Before proving the main results of this chapter, let us collect some basic facts
on the compactness of the embedding of H Ψ1 into LΨ2 (μ). First:
Lemma 4.8. Let Ψ1 , Ψ2 be two Orlicz functions and μ a finite Borel measure
on D. Assume that the identity maps H Ψ1 into LΨ2 (μ) compactly. Then, μ(T) = 0.
Proof. The sequence (z n )n≥1 is weakly null in M Ψ1 by Riemann-Lebesgue’s
Lemma. Its image by a compact operator is then norm null. This implies that
for every ε ∈ (0, 1), we have, for n large enough,

 
Ψ2 |z|n dμ ≤ ε.
D

Fatou’s Lemma yields Ψ2 (1)μ(T) ≤ ε. 

Now, we summarize what is true in full generality about compactness for canon-
ical embeddings.
Proposition 4.9. Let Ψ1 , Ψ2 be two Orlicz functions and μ a finite Borel
measure on D. The following assertions are equivalent
i) The identity maps H Ψ1 into LΨ2 (μ) compactly.
ii) Every sequence in the unit ball of H Ψ1 , which is convergent to 0 uniformly
on every compact subset of D, is norm-null in LΨ2 (μ).
iii) The identity maps H Ψ1 into LΨ2 (μ) continuously and
lim− Ir  = 0, where Ir (f ) = f ID\rD .
r→1
3. GENERAL MEASURES 41

Proof. i) ⇒ ii): let (fn )n≥1 be a sequence in the unit ball of H Ψ1 , which is
uniformly convergent to 0 on every compact subset of D. In particular, fn (z)
converges to 0 for every z ∈ D. This means that (fn )n≥1 converges to 0 μ-almost
everywhere, since μ(T) = 0 by the preceding lemma. If the conclusion did not hold,
we could assume, up to an extraction, that limfn Ψ2 > 0. Thus by compactness
of the embedding, up to a new extraction, (fn )n≥1 is norm-convergent to some
g ∈ LΨ2 (μ). Necessarily g = 0. A subsequence of (fn )n≥1 would be convergent to
g μ-almost everywhere. This gives a contradiction.
ii) ⇒ iii): if not, there exist a sequence (fn )n≥1 in the unit ball of H Ψ1 and
δ > 0 with fn ID\(1− 1 )D Ψ2 > δ, for every n ≥ 1. Let us introduce the sequence
n
gn (z) = z n fn (z) for z ∈ D. The sequence (gn )n≥1 lies in the unit ball of H Ψ1 and
is convergent to 0 uniformly on every compact subset of D. But
1 n 1 n
gn Ψ2 ≥ z n fn ID\(1− 1 )D Ψ2 ≥ 1 − fn ID\(1− 1 )D Ψ2 ≥ 1 − δ.
n n n n
This contradicts (ii).
iii) ⇒ ii) is very easy.
ii) ⇒ i) follows from Proposition 3.6. 

We can now state some deeper characterizations.
Theorem 4.10. Let μ be a finite Borel measure on the closed unit disk D and
let Ψ1 and Ψ2 be two Orlicz functions. Then:
1) If the identity maps H Ψ1 into LΨ2 (μ) continuously, there exists some A > 0
such that:
1
(R) ρμ (h) ≤   for every h ∈ (0, 1].
Ψ2 AΨ−11 (1/h)

2) If there exists some A > 0 such that:
1/h
(K) Kμ (h) ≤   for every h ∈ (0, 1],
Ψ2 AΨ−1
1 (1/h)

then the identity maps H Ψ1 into LΨ2 (μ) continuously.

Theorem 4.11. Let μ be a finite Borel measure on the closed unit disk D and
let Ψ1 and Ψ2 be two Orlicz functions. Then:
1) If the identity maps H Ψ1 into LΨ2 (μ) compactly, then
1
(R0 ) ρμ (h) = o  −1
 as h → 0, for every A > 0.
Ψ2 AΨ1 (1/h)
2) If μ(T) = 0 and
1/h
(K0 ) Kμ (h) = o  −1
 as h → 0 , for every A > 0,
Ψ2 AΨ1 (1/h)

then the identity maps H Ψ1 into LΨ2 (μ) compactly.
3) When Ψ1 = Ψ2 = Ψ satisfies condition ∇0 , then the above conditions are
equivalent: the identity maps H Ψ1 into LΨ2 (μ) compactly if and only if condition
(R0 ) is satisfied and if and only if condition (K0 ) is satisfied.
42 PASCAL LEFÈVRE, DANIEL LI, HERVÉ QUEFFÉLEC, LUIS RODRÍGUEZ-PIAZZA

Remarks.  
1. a) If ρμ (h) ≤ C/Ψ2 AΨ−1 1 (1/h) with C > 1, then the convexity of Ψ2 gives
Ψ2 (t/C) ≤ Ψ2 (t)/C and hence:
1
ρμ (h) ≤  A −1 ·
Ψ2 C Ψ1 (1/h)
 
b) If ρμ (h) ≤ 1/Ψ2 AΨ−1 1 (1/h) only for h ≤ hA , one can find some C =
 
CA > 0 such that ρμ (h) ≤ C/Ψ2 AΨ−1 1  (1/h) for every h ∈ (0, 1]. In fact, ρμ (h) ≤
  
μ(D) and 1/Ψ2 AΨ−1 (1/h) ≥ 1/Ψ AΨ −1
(1/h ) for h ≥ hA ; hence ρμ (h) ≤
 −1
 1 2
 1−1 A

C/Ψ2 AΨ1 (1/h) with C = μ(D)/Ψ2 AΨ1 (1/hA ) .
The same remark applies for Kμ .    
2. a) In the case where Ψ1 = Ψ2 = Ψ, one has: Ψ AΨ−1 (t) ≤ Ψ Ψ−1 (t) = t
when A ≤ 1 and Ψ AΨ−1 (t) ≥ Ψ Ψ−1 (t) = t when A ≥ 1. On the other hand,
if Ψ ∈ Δ2 , one has, for some constant C = CA > 0: Ψ(Ax) ≤ CΨ(x), when A ≥ 1
and Ψ(Ax) ≥ (1/C)Ψ(x) when A ≤ 1. Hence, when Ψ ∈ Δ2 , one has, for every
A > 0:
1 1
 = ≈h
−1
Ψ AΨ (1/h) χ (1/h)
A
and Theorem 4.10 is nothing but Carleson’s Theorem.
b) If Ψ1 (x) = xp and Ψ2 (x) = xq with p < q < ∞, then:
 
Ψ2 AΨ−11 (t) = A t
q q/p
,
and condition (R) means that μ is a β-Carleson measure, with β = q/p (see [11],
Theorem 9.4).
Ψ1 (x)
c) If, for fixed A > 0, the function x → is non increasing, at least for
Ψ2 (Ax)
x large enough, conditions (R) and (K) (resp. conditions (R0 ) and (K0 ) below)
are clearly equivalent. This is the case in the framework of classical Hardy spaces:
Ψ1 (x) = xp and Ψ2 (x) = xq , with q ≥ p.
When Ψ1 = Ψ2 = Ψ, this is equivalent, if A > 1, to the convexity of the
function κ(x) = log Ψ(ex ) (see Proposition 4.7).
3. a) When Ψ1 = Ψ2 = Ψ, the condition μ(T) = 0 is automatically fulfilled (and
so can be removed from (K0 )). This follows on the one hand from the majorization
in (K0 ), which implies that Kμ (h) → 0 (when h → 0); and on the other hand from
the inequality:
  2π
μ D \ rD ≤ ρμ (1 − r) ≤ 2πKμ (1 − r).
1−r
Indeed, D \ rD can be covered by less than 1−r 2π
Carleson’s windows of “size” 1 − r.
b) Nevertheless, the condition μ(T) = 0 cannot be removed in full generality in
Theorem 4.11. Indeed, if we consider the identity j from H 4 into L2 (D, m̃), where
m̃ is 0 on D and its restriction to the torus is the normalized Lebesgue measure. It
1
is easily seen that K(h) is bounded and so less than 2 1/2 , for h small enough.
A h
But j is not compact.
4. In the case where Ψ1 = Ψ2 = Ψ and μ is a Carleson measure, then Kμ is
bounded, by say K ≥ 1, and condition  (K) is satisfied for A = 1/K,
 since A ≤ 1
implies, by the convexity of Ψ: Ψ AΨ−1 (1/h) ≤ AΨ Ψ−1 (1/h) = A/h. Hence
the canonical embedding H Ψ → LΨ (μ) is continuous. We get hence, by Carleson’s
Theorem ([11], Theorem 9.3):
3. GENERAL MEASURES 43

Proposition 4.12. Let μ be a positive finite measure on D. Assume that the
canonical embedding jμ : H p → Lp (μ) is continuous for some 0 < p < ∞. Then
jμ : H Ψ → LΨ (μ) is continuous.
Note that this is actually a consequence of the fact that H Ψ is an interpolation
space for H 1 and H ∞ (see [4], Theorem V.10.8).
When μ = μφ is the image of the Haar measure m under φ∗ , where φ is an
analytic self-map of D, we know (Proposition 3.12) that the composition operator
Cφ : H Ψ → H Ψ is always continuous. This can be read as the continuity of H Ψ →
LΨ (μφ ). Hence condition (R) must be satisfied, for some A > 0. Note that for
A ≤ 1, 1/χA (1/h) ≥ h, and so condition (R) is implied by the fact that μφ is a
Carleson measure.
5. To have the majorization in condition (K0 ), it suffices that: for every A > 0,
there exists hA ∈ (0, 1] such that
1/h
(4.10) Kμ (h) ≤   , for every h ∈ (0, hA ].
Ψ2 AΨ−1
1 (1/h)
In fact, fixing A > 0 and ε ∈ (0, 1), we have, by convexity:
Ψ2 (Ax) ≤ εΨ2 (Ãx),
with à = A/ε. Since we have (4.10) with Ã, when h is small enough (depending
on A and ε), we get, for x = Ψ−1
1 (1/h), condition (K0 ).

To prove both Theorem 4.10 and Theorem 4.11, we shall need some auxiliary
results. The following is actually the heart of the classical Theorem of Carleson,
though it is not usually stated in this form. The maximal (non-tangential) function
Mf (which is essentially the same as Nα f in the previous chapter) will be defined
by:  
Mf eiθ = sup{|f (z)|; z ∈ Gθ },
where
Gθ = {z ∈ D ; |eiθ − z| < 3(1 − |z|)}.
Theorem 4.13 (Carleson’s Theorem). For every f ∈ H 1 and every finite pos-
itive measure μ on the closed unit disk D, one has, for every h ∈ (0, 1] and every
t > 0:
 
μ {z ∈ D ; |z| > 1 − h and |f (z)| > t} ≤ 2πKμ (h) m({Mf > t}).
As this theorem is not usually stated in such a way, we shall give a few words
of explanations.
Proof. For convenience , we shall denote, when I is a subarc of T:
z
W (I) = {z ∈ D ; |z| > max(0, 1 − |I|/2) and ∈ I}.
|z|
Obviously, when |I| ≤ 2, we have W (I) = W (ξ, |I|/2), where ξ is the center of I.
Since it is clear that z ∈ Gθ when z = |z| eiθ , the condition |f (z)| > t implies
z
that Mf |z| > t. Hence it suffices to majorize the measure of the set
z 
M = {z ∈ D ; |z| > 1 − h and Mf > t}.
|z|
The open set {Mf > t} is the disjoint union of a countable family of open arcs
Ij of T. So every z ∈ M belongs to some W (Ij ).
44 PASCAL LEFÈVRE, DANIEL LI, HERVÉ QUEFFÉLEC, LUIS RODRÍGUEZ-PIAZZA

Now, when |Ij | > 2h, we may cover Ij in such a way that we can write Ij ⊂
Jj,1 ∪ · · · ∪ Jj,Nj with the arcs Jj,1 , . . . , Jj,Nj satisfying: |Jj,k | = 2h for every k =

Nj
1, . . . , Nj and 2|Ij | ≥ |Ij | + 2h ≥ |Jj,k |.
k=1
When |Ij | ≤ 2h, we simply define: Nj = 1 and J1,Nj = Ij .
We then notice that, for every j, we have by definition of Kμ (h):
  |Jj,k |
(4.11) μ W (Jj,k ) ≤ Kμ (h) ·
2
Given E ⊆ D, let us denote by Eh the set of points z ∈ E such that |z| > 1 − h.
Since, for every j: 
 
W (Ij ) h ⊆ W (Jj,k ),
1≤k≤Nj
we get, using (4.11):
   
N N
 j
 j
1
μ W (Ij )h ≤ μ W (Jj,k ) ≤ Kμ (h) |Jj,k | ≤ Kμ (h) |Ij |.
2
k=1 k=1
 
On the other hand, M ⊆ W (Ij ) h .
j
It follows that:
   
μ(M) ≤ μ W (Ij )h ≤ Kμ (h) |Ij |
j j

= 2πKμ (h) m(Ij ) = 2πKμ (h) m({Mf > t}),
j

remembering that {Mf > t} is the disjoint union of the Ij ’s.
We obtain
 
μ {z ∈ D ; |z| > 1 − h and |f (z)| > t} ≤ 2πKμ (h) m({Mf > t}) ,
as announced. 
The following estimation will be useful for the study both of boundedness and
compactness.
Lemma 4.14. Let μ be a finite Borel measure on D. Let Ψ1 and Ψ2 be two
Orlicz functions. Suppose that there exists A > 0 and hA ∈ (0, 1) such that
1/h
Kμ (h) ≤   , for every h ∈ (0, hA ).
Ψ2 AΨ−1
1 (1/h)

Then, for every f ∈ H Ψ1 such that f Ψ1 ≤ 1 and every Borel subset E of D, we
have: A
π
Ψ2 |f | dμ ≤ μ(E)Ψ2 (xA ) + Ψ1 (Mf ) dm
E 8 2 T
A
where xA = Ψ−1 (1/hA ).
2 1
Proof. For every s > 0 and every z ∈ D, the inequality |f (z)| > s implies that the
norm of the evaluation at z is greater than s; hence by Lemma 3.11:
1
s < 4Ψ−1 ,
1
1 − |z|
3. GENERAL MEASURES 45

i.e.:
1
|z| > 1 − ·
Ψ1 (s/4)
Obviously, this inequality still holds if z ∈ T.
Carleson’s Theorem (Theorem 4.13) gives:
  1
μ {|f (z)| > s} ≤ 2πKμ m({Mf > s})
Ψ1 (s/4)
when Ψ1 (s/4) ≥ 1. Hence:
A ∞
Ψ2 |f | dμ = Ψ2 (t) μ({|f | > 8t/A} ∩ E) dt .
E 8 0
But our hypothesis means that, when Ψ1 (s/4) > 1/hA :
1 Ψ1 (s/4)
Kμ ≤ ·
Ψ1 (s/4) Ψ2 (As/4)
We have then
  Ψ1 (2t/A)
μ {|f (z)| > 8t/A} ≤ 2π m({Mf > 8t/A}).
Ψ2 (2t)
So:
A xA
Ψ2 |f | dμ ≤ Ψ2 (t)μ(E) dt
E 8 0
+∞
Ψ1 (2t/A)
+ 2π Ψ2 (t) m({Mf > 8t/A}) dt
xA Ψ2 (2t)
≤ Ψ2 (xA )μ(E)
+∞ 
Ψ2 (t)
+ 2π Ψ1 (2t/A) m({Mf > 8t/A}) dt .
xA Ψ2 (2t)
For the second integral, note that one has Ψ(x) ≤ xΨ (x) ≤ Ψ(2x), for any
Orlicz function Ψ. This leads to:
∞ 
Ψ2 (t)
Ψ1 (2t/A) m({Mf > 8t/A}) dt
xA Ψ2 (2t)

Ψ1 (2t/A)
≤ m({Mf > 8t/A}) dt
t
0

2 ∞ 
≤ Ψ1 (2t/A) m({Mf > 8t/A}) dt
A 0
∞ 1
= Ψ1 (x) m({Mf > 4x}) dx = Ψ1 Mf dm
T 4
0

1
≤ Ψ1 (Mf ) dm.
4 T
which leads to the desired result. 

For the proofs of Theorem 4.10 and Theorem 4.11, we may restrict ourselves to
the case of functions Ψ1 and Ψ2 satisfying ∇2 . Indeed, suppose that Ψ1 and Ψ2 are
Orlicz functions and define Ψ j (t) = Ψj (t2 ), for j ∈ {1, 2}. The functions Ψ
1 and 

Ψ2 are Orlicz functions satisfying ∇2 since, with β = 2, we have for every t ≥ 0:
j (βt) = Ψj (4t2 ) ≥ 4Ψj (t2 ) = 2β Ψ
Ψ j (t).
46 PASCAL LEFÈVRE, DANIEL LI, HERVÉ QUEFFÉLEC, LUIS RODRÍGUEZ-PIAZZA

 
Now, we claim that μ satisfies (R), (K), (R0 ) or (K0 ) for the couple Ψ1 , Ψ2
 
if and only if μ satisfies it for the couple Ψ 2 . This is simply due to the fact
1 , Ψ
that for every A > 0 and t ≥ 0, we have

2 AΨ
Ψ 1 −1 (t) = Ψ2 A2 Ψ−1 (t) .
1

Moreover, notice that, writing f = Bg 2 (where B is a Blaschke product), we have
 
f ∈ H Ψ if and only if g ∈ H Ψ ; thus gLΨ  = f LΨ . It is then clear that
   2
     
Id : H Ψ1 −→ LΨ2 (μ) = Id : H Ψ1 −→ LΨ2 (μ) ,
 
so that the canonical embedding is bounded (resp. compact) for the couple Ψ1 , Ψ2
 
if and only if it is so for the couple Ψ1 , Ψ
2 , thanks to Proposition 4.9.

Proof of Theorem 4.10. 1) Let C be the norm of the canonical embedding
j : H Ψ1 → LΨ2 (μ), and let ξ ∈ T and h ∈ (0, 1). It suffices to test the continuity of
j on f = Ψ−1 1 (1/h)uξ,1−h , which is in the unit ball of HM
Ψ1
, by Corollary 3.10.
But, when z ∈ W (ξ, h) one has, with a = (1 − h)ξ:

 z   z 

|1 − āz| ≤ |1 − āξ| + |āξ − āz| ≤ h + (1 − h) ξ − + − z
|z| |z|
≤ h + (1 − h)[h + (1 − |z|)] ≤ h + (1 − h)[h + h] ≤ 3h;
hence |uξ,1−h (z)| ≥ 1/9 and |f (z)| ≥ (1/9)Ψ−11 (1/h); therefore:
|f | 1  
1≥ Ψ2 dμ ≥ Ψ2 Ψ−1
1 (1/h) μ W (ξ, h) ,
D C 9C
which is (R).
2) By Proposition 3.5, the maximal (non-tangential) function M is bounded
on LΨ1 (T): there exists a constant C ≥ 1 such that Mf Ψ1 ≤ Cf Ψ1 for every
f ∈ LΨ1 (T). We fix f in the unit ball of H Ψ1 (note that f /CΨ1 remains ≤ 1)
and use Lemma 4.14, with E = D and f replaced by f /C (here hA = 1)). Writing
π
C̃ = + μ(D)Ψ2 (xA ), we get:
2
A A
1
Ψ2 |f | dμ ≤ Ψ2 |f | dμ
D 8C C̃ C̃ D 8C
 
1 π 1
≤ μ(D)Ψ2 (xA ) + Ψ1 Mf dm
C̃ 2 T C
1 π
≤ μ(D)Ψ2 (xA ) + = 1,
C̃ 2
8C C̃
which means that f LΨ2 (μ) ≤ · 
A

Proof of Theorem 4.11. 1) Suppose that the embedding is compact, but that
condition (R0 ) is not satisfied. Then there exist ε0 ∈ (0, 1), A > 0, a sequence of
positive numbers (hn )n decreasing to 0, and a sequence of ξn ∈ T, such that:
  ε0
μ W (ξn , hn ) ≥  ·
Ψ2 AΨ−1
1 (1/hn )
3. GENERAL MEASURES 47

Consider the sequence of functions
h2n
fn (z) = Ψ1 −1 (1/hn ) = Ψ1 −1 (1/hn ) uξn ,|an | ,
(1 − ān z)2
where an = (1−hn )ξn . By Corollary 3.10, fn is in the unit ball of HM Ψ1 . Moreover,
it is plain that (fn )n converges to 0 uniformly on every compact subset of D. By the
compactness criterion (Proposition 3.6), (fn )n is norm-converging to 0 in LΨ2 (μ).
But, as above (proof of Theorem 4.10), for every n ≥ 1, one has |fn (z)| ≥
(1/9)Ψ−1 1 (1/hn ) when z ∈ W (ξn , hn ); hence:
9A A  
Ψ2 |fn | dμ ≥ Ψ2 Ψ−1
1 (1/hn ) μ W (ξn , hn )
D ε0 ε0
A ε0
≥ Ψ2 Ψ−1 (1/h n )   ≥ 1,
ε0 1 Ψ2 AΨ−1
1 (1/hn )

by the convexity of Ψ2 . This implies that fn LΨ2 (μ) ≥ ε0 /9A and gives a contra-
diction.
2) We have to prove that for every ε > 0, there exists an r ∈ (0, 1) such
that the norm of the injection Ir : H Ψ1 −→ LΨ2 (D \ rD, μ) is smaller than ε (see
Proposition 4.9).
Let C ≥ 1 be the norm of the maximal operator, as in the proof of Theo-
rem 4.10: Mf Ψ1 ≤ Cf Ψ1 for every f ∈ LΨ1 (T), and set A = 16C/ε. Condition
(K0 ) gives us hA ∈ (0, 1) such that:
1 1/h
Kμ (h) ≤  
2 Ψ2 AΨ−1
1 (1/h)
when h ≤ hA .
Let f in the unit ball of H Ψ1 and r ∈ (0, 1). By Lemma 4.14:
|f | A A
1
Ψ2 dμ = Ψ2 |f | dμ ≤ Ψ2 |f | dμ
D\rD ε D\rD 16C 2 D\rD 8C
   
1   π Mf
≤ μ D \ rD Ψ2 (xA ) + Ψ1 dm
2 2 T C
π  
≤ + Ψ2 (xA )μ D \ rD .
4

 
As μ(T) = 0, there exists some r0 ∈ (0, 1) such that π4 + Ψ2 (xA )μ D \ rD ≤ 1, for
every r ∈ (r0 , 1). This ends the proof of 2).
3) Assume that Ψ satisfies condition ∇0 and that condition (R0 ) is fulfilled:
for every A > 0 and every h ∈ (0, hA ) (hA small enough), we have:
1
ρμ (h) ≤  ·
Ψ AΨ−1 (1/h)
This implies that:
ρμ (s) 1/s Ψ(x)
Kμ (h) = sup ≤ sup  = sup ·
s −1
0<s≤h 0<s≤h Ψ AΨ (1/s) x≥Ψ−1 (1/h) Ψ(A x)

Fix an arbitrary β > 1 and choose A = βCβ > 1, where Cβ is given by the
∇0 condition for Ψ and Proposition 4.6. We have, for h small enough and x ≥
48 PASCAL LEFÈVRE, DANIEL LI, HERVÉ QUEFFÉLEC, LUIS RODRÍGUEZ-PIAZZA

Ψ−1 (1/h):  
Ψ βΨ−1 (1/h) Ψ(βCβ x) Ψ(A x)
  ≤ = ;
−1
Ψ Ψ (1/h) Ψ(x) Ψ(x)
we get hence, for h small enough:
 
Ψ Ψ−1 (1/h) 1/h
Kμ (h) ≤  −1 =  ,
Ψ βΨ (1/h) Ψ βΨ−1 (1/h)
and condition (K0 ) is fulfilled. Keeping in mind that μ(T) = 0, by Remark 3.a.,
the two first part 1) and 2) imply the result. 

Remark. Actually, the proof of Theorem 4.10, 1) gives, for every measure μ on D
and every Orlicz function Ψ:

|uξ,1−h | 1  
1≥ Ψ dμ ≥ Ψ μ W (ξ, h) ,
D uξ,1−h LΨ (μ) 9 uξ,1−h LΨ (μ)
and hence:
  1
(4.12) μ W (ξ, h) ≤ ·
1
Ψ
9 uξ,1−h LΨ (μ)
In particular, if μ = μφ is the image of the Haar measure m under a self-map φ of
D, one has:
  1
(4.13) μ W (ξ, h) ≤ ·
1
Ψ
9 Cφ (uξ,1−h )Ψ
Condition (4.12) allows to have an upper control for the μ-measure of Carleson
windows, with uξ,1−h LΨ (μ) . It is possible, conversely, to majorize these norms.
Definition 4.15. We shall say that a measure μ on D is a Ψ-Carleson measure
if there exists some A > 0 such that
  1
μ W (ξ, h) ≤  −1
 , for every ξ ∈ T and every h ∈ (0, 1).
Ψ AΨ (1/h)
We shall say that a measure μ on D is a vanishing Ψ-Carleson measure if, for
every A > 0,  
lim Ψ AΨ−1 (1/h) .ρμ (h) = 0.
h→0
Equivalently if, for every A > 0, there exists hA ∈ (0, 1) such that
  1
μ W (ξ, h) ≤  −1
 , for every ξ ∈ T and every h ∈ (0, hA ).
Ψ AΨ (1/h)
Remarks. 1) Let us precise how the second assertion implies the first one: this
follows from the convexity of Ψ and the fact that the condition involves arbitrary
A > 0. Indeed, fixing A > 0, we have for every ε ∈ (0, 1):
1 ε
ρμ (h) ≤  A −1 ≤   , for every h ∈ (0, h A ).
Ψ ε Ψ (1/h) Ψ AΨ−1 (1/h) ε

2) Let us point out that the fact that μ is a Ψ-Carleson measure (resp. a
vanishing Ψ-Carleson measure) means that μ satisfies condition (R) in Theorem
4.10 (resp. condition (R0 ) in Theorem 4.11), with Ψ1 = Ψ2 = Ψ.
3. GENERAL MEASURES 49

We have the following characterizations:

Proposition 4.16. 1) μ is a Ψ-Carleson measure on D if and only if there
exists some constant C ≥ 1 such that :

C
uξ,1−h LΨ (μ) ≤ , for every ξ ∈ T and every h ∈ (0, 1).
Ψ−1 (1/h)

2) μ is a vanishing Ψ-Carleson measure on D if and only if

lim sup Ψ−1 (1/h)uξ,1−h LΨ (μ) = 0.
h→0 ξ∈T

Proof. The sufficiency (both for 1) and 2)) follows easily from (4.12) in the pre-
ceding remark.
The converse is an obvious consequence of the following lemma.

Lemma 4.17. Suppose that there exist A > 0 and h0 ∈ (0, 1) such that:

1
ρμ (h) ≤   , for every h ∈ (0, h0 ).
Ψ AΨ−1 (1/h)

Then there exists h1 ∈ (0, 1) such that:

24
uξ,1−h LΨ (μ) ≤ ,
AΨ−1 (1/h)

for every ξ ∈ T and every h ∈ (0, h1 ).

Proof of the lemma. This is inspired from [12], Chapter VI, Lemma 3.3, page
239. We may assume that h ≤ h0 /4 ≤ 1/4.
First, writing a = (1 − h)ξ (where ξ ∈ T), we observe that, when |z − ξ| ≥ bh
for a b > 0, we have:

¯
|1 − āz|2 = 1 + |a|2 |z|2 − 2|a|Re (ξz)
= |a||ξ − z|2 + (1 − |a|) + |a|2 |z|2 − |a||z|2
≥ |a|b2 h2 + (1 − |a|)2 ≥ (|a|b2 + 1)h2 .

1
So, we have |uξ,1−h (z)| ≤ ≤ min(1, 2/b2 ), when |z − ξ| ≥ bh.
|a|b2 + 1
Now, define, for every n ∈ N and ξ ∈ T:

Sn = S(ξ, 2n+1 h) = {z ∈ D ; |z − ξ| < 2n+1 h} ⊂ W (ξ, 2.2n+1 h).

Our observation implies that |uξ,1−h (z)| ≤ min(1, 2/4n ), for every z ∈ D \ Sn−1 .
For z ∈ S0 , one has simply |uξ,1−h (z)| ≤ 1.
There exists an integer N such that 2N +2 h ≤ h0 < 2N +3 h.
50 PASCAL LEFÈVRE, DANIEL LI, HERVÉ QUEFFÉLEC, LUIS RODRÍGUEZ-PIAZZA

Let us compute:

    N
 
Ψ M |uξ,1−h | dμ = Ψ M |uξ,1−h | dμ + Ψ M |uξ,1−h | dμ
D S0 n=1 Sn \Sn−1

 
+ Ψ M |uξ,1−h | dμ
D\SN

N 2M 2M
≤ Ψ(M )μ(S0 ) + Ψ n μ(Sn ) + Ψ N μ(D)
n=1
4 4

N
1 4M 2M
≤ Ψ n μ(Sn ) + Ψ N μ(D).
n=0
2n+1 2 4

But for n ≤ N , we have 2.2n+1 h ≤ 2N +2 h ≤ h0 , so the hypothesis gives:
1
μ(Sn ) ≤  ·
Ψ AΨ−1 (1/2n+2 h)
Take now:
A −1 1
(4.14) M= Ψ ·
24 h
 1   
We have, using that Ψ−1 2n+2 h ≥ 2n+2 Ψ
1 −1 1
h ,
  
  N
1 A −1 1
Ψ M |uξ,1−h | dμ ≤ + μ(D)Ψ Ψ
D n=0
2n+1 12.4N h
 2 
1 16A.h −1 1
≤ + μ(D)Ψ Ψ ,
2 3h20 h
 2
1 8h
because N ≤ .
4 h0
We can choose h1 small enough to have:
 
16A.h2 −1 1 1
μ(D)Ψ 2 Ψ ≤
3h0 h 2

for every h ∈ (0, h1 ), since lim h2 Ψ−1 h1 = 0.
h→0
We get for such h
 
Ψ M |uξ,1−h | dμ ≤ 1,
D
so that
1 24 1
uξ,1−h LΨ (μ) ≤ = −1
,
M A Ψ (1/h)
as it was announced. 

Examples and counterexamples.
We are going to give some examples showing that we do not have the reverse
implications in general in Theorem 4.10 and Theorem 4.11.
1. Condition (R) is not sufficient in general to have a continuous embedding.
Let Ψ(x) = ex − 1 (note that this Orlicz function even fulfills the Δ1 condition !).
Note that Ψ(AΨ−1 (1/h)) ∼ h−A , when h → 0.
3. GENERAL MEASURES 51

a. Let ν be a probability measure on T, supported by a compact set L of
Lebesgue measure zero, such that ν(I) ≤ |I|1/2 , for each I. We can associate to ν
the measure on D defined by ν̃(E) = ν(E ∩ T). Then the identity map from H Ψ to
LΨ (ν̃) is not even defined. Nevertheless the condition (R) is clearly fulfilled with
A = 1/2.
b. Now, we exhibit a similar example (less artificial) on the open disk. Let ν be
as previously. By a standard argument: for every integer n, there exists a function
gn in the unit ball of the disk algebra such that |gn | = 1 on L and gn Ψ ≤ 4−n . As
L is compact, there exists some rn ∈ (1/2, 1) such that |gn (rn z)| ≥ 1/2 for every
z ∈ L. Now, define the measure μ by:
∞
1
μ(E) = ν (E),
n n
n=1
2

where:
 
νn (E) = ν {z ∈ T| rn z ∈ E} .
If W is a Carleson window of “size” h then, for each n ≥ 1, we have:
 
ν {z ∈ T| rn z ∈ W } ≤ ν(W ∩ T) ≤ (2h)1/2 .
Hence, μ(W ) ≤ (2h)1/2 and the condition (R) is fulfilled.
Nevertheless, the identity from H Ψ to L1 (μ) is not continuous: gn Ψ ≤ 4−n ;
but
1 1 1
gn 1 ≥ n |gn | dνn ≥ n |gn (rn w)| dν(w) ≥ n+1 ·
2 rn T 2 L 2
2. Condition (K) is not necessary in general to have a continuous embedding.
When Ψ satisfies Δ2 , the identity from H Ψ to LΨ (μ) is continuous if and only if μ
is a Carleson measure. So the conditions (R), (K) and the continuity are equivalent
in this case. Actually, when Ψ does not satisfy Δ2 , we construct below a measure
μ on D such that the identity from H Ψ to LΨ (μ) is continuous and order bounded,
but μ is not a Carleson measure (a fortiori does not verify (K)). Note that the
measure μ is then Ψ-Carleson but not Carleson. Here is the example:
We have assumed that Ψ does not satisfy Δ2 ; so there exists an increasing
Ψ(an ) Ψ(2an )
sequence (an )n≥1 such that is increasing and ≥ n2n . Now, define
n Ψ(an )
the discrete measure
∞  
n n+1
μ= − δxn ,
n=1
Ψ(2an ) Ψ(2an+1 )

where:
1
xn = 1 − ·
Ψ(2an )
  N ,
As μ [xN , 1] = the measure μ is not Carleson: it should be bounded
Ψ(2aN )
c
by c(1 − xN ) = , where c is some constant.
Ψ(2aN )
 forevery f in the unit ball of H and every x ∈ (0, 1), we have
Ψ
We know that
1
|f (x)| ≤ 4Ψ−1 : see Lemma 3.11. So we only have to see that g ∈ LΨ (μ),
1−x
52 PASCAL LEFÈVRE, DANIEL LI, HERVÉ QUEFFÉLEC, LUIS RODRÍGUEZ-PIAZZA

 
1
where g(x) = Ψ−1 . Indeed, we have:
1−x
  ∞
|g|  n n + 1 |g(xn )|
Ψ dμ = − Ψ
D 2 n=1
Ψ(2an ) Ψ(2an+1 ) 2

 n 1 1
≤ Ψ Ψ−1
n=1
Ψ(2an ) 2 1 − xn

 ∞

Ψ(an ) 1
≤ n ≤
n=1
Ψ(2an ) n=1 2n
≤1

so gΨ ≤ 2.
3. Condition (K0 ) is not necessary in general to have a compact embedding.
We can find, for every Orlicz function Ψ not satisfying ∇0 , a measure μ such
that the identity from H Ψ to LΨ (μ) is compact but (K0 ) is not satisfied. Indeed:
since Ψ ∈/ ∇0 , we can select two increasing sequences (xn )n≥1 and (yn )n≥1 , with
1 ≤ xn ≤ yn ≤ xn+1 and Ψ(xn ) > 1, and such that lim xn = +∞ and:

Ψ(2xn ) Ψ(2n yn )
≥ ·
Ψ(xn ) Ψ(yn )
1
Define rn = 1 − and the discrete measure:
Ψ(yn )

 1
μ= δr .
n=1
Ψ(2n yn ) n

The series converge since Ψ(2n yn ) ≥ 2n .
Thanks to Lemma 3.11, we have, for every f in the unit ball of H Ψ and every
n ≥ 1:
 
1
|f (rn )| ≤ 4Ψ−1 = 4yn .
1 − rn
Given r > r1 , there exists an integer N ≥ 1 such that rN < r ≤ rN +1 . Then, for
every f in the unit ball of H Ψ , we have f LΨ (D\rD,μ) ≤ 2−N +2 , since:
  ∞
  
|f | 1 |f (rn )|
Ψ −N +2 dμ = Ψ
D\rD 2 Ψ(2n yn ) 2−N +2
n=N +1
∞
Ψ(2N yn )

Ψ(2n yn )
n=N +1
∞
1
≤ = 1.
2n−N
n=N +1

This implies that lim sup f LΨ (D\rD,μ) = 0. By Proposition 4.9, the identity
r→1 f ≤1
Ψ

from H Ψ to LΨ (μ) is compact.
3. GENERAL MEASURES 53

1 1 ,
On the other hand, writing hn = and tn = we have:
Ψ(xn ) Ψ(yn )
∞
μ([1 − tn , 1]) 1 Ψ(yn )
Kμ (hn ) ≥ = Ψ(yn ) my )
≥ ny )
tn m=n
Ψ(2 m Ψ(2 n

Ψ(xn ) 1/hn
≥ =  −1 ,
Ψ(2xn ) Ψ 2Ψ (1/hn )
and this shows that (K0 ) is not satisfied.
This shows that condition Ψ ∈ ∇0 is necessary and sufficient in order to have:
the identity from H Ψ to LΨ (μ) is compact if and only if μ satisfies (K0 ).
4. Condition (R0 ) is not sufficient in general to have a compact embedding.
We can find an Orlicz function Ψ and a vanishing Ψ-Carleson measure (i.e. (R0 )
is satisfied) μ such that the identity from H Ψ to LΨ (μ) is not compact.
We shall use the Orlicz function introduced in [25]. The key properties of this
function Ψ are:
i) For every x > 0, Ψ(x) ≥ x3 /3.
ii) For every integer k ≥ 1, Ψ(k!) ≤ (k!)3 .
iii) For every integer k ≥ 1, Ψ(3(k!)) > k.(k!)3 .
Once again, the job is done by a discrete measure.
(k + 1)! 1 1
Define xk = k!; yk = 1/3
; rk = 1 − and ρk = 1 − · Of course,
k Ψ(yk ) Ψ(xk )
x2 < y2 < x3 < · · · .
Let ν be the discrete measure defined by:


ν= νk ,
k=2

where:
1 
νk =   δr k a .
Ψ (k + 1)! k2
a =1
2 2
k 3.k
Observe that νk  ≤   ≤ so that the series converges. Note
Ψ (k + 1)! (k + 1)!3
that ν is supported in the union of the circles of radii rk and not in a subset of the
segment [0, 1] as in the preceding counterexamples.
In order to show that (R0 ) is satisfied, it is clearly sufficient to prove that, when
1 1
≤h< (with k ≥ 3), we have:
Ψ(yk ) Ψ(yk−1 )
1
ρν (h) ≤ k1/3 ·
−1
Ψ Ψ (1/h)
2
1 1 , we have Ψ−1 (1/h) ≤ yk so
Supposing then ≤h<
Ψ(yk ) Ψ(yk−1 )
k1/3 1  
Ψ Ψ−1 (1/h) ≤ Ψ (k + 1)! .
2 2
2
Therefore, it is sufficient to establish that ρν (h) ≤  ·
Ψ (k + 1)!
54 PASCAL LEFÈVRE, DANIEL LI, HERVÉ QUEFFÉLEC, LUIS RODRÍGUEZ-PIAZZA

A Carleson window W (ξ, h) (where ξ ∈ T) can contain at most one k2 -root of
2 6 6(k − 1) 2π
the unity, since 2h < ≤ 3 ≤ ≤ 2 · This implies that
Ψ(yk−1 ) yk−1 (k!)3 k

  1
νk W (ξ, h) ≤  ·
Ψ (k + 1)!

Besides, when j < k, the window W(ξ, h) cannot meet any circle of radius rj
(centered at the origin), so νj W (ξ, h) = 0. We obtain:

     1  j2
ν W (ξ, h) = νj W (ξ, h) ≤  +  
j=k
Ψ (k + 1)! j>k
Ψ (j + 1)!
∞
1 3j 2
≤  +
Ψ (k + 1)! (j + 1)!3
j=k+1
∞  s
1 3 1
≤  +
Ψ (k + 1)! (k + 1)!3 s=1 k + 1
2
≤  ·
Ψ (k + 1)!

This proves that (R0 ) is satisfied.
 2
 2 1 − ρk
Let us introduce the function fk (z) = xk u1,ρk z k = xk .
1 − ρk z k 2
It lies in the unit ball of H Ψ by Corollary 3.10:
x
fk Ψ = xk u1,ρk Ψ ≤  k 1  = 1.
Ψ−1 1−ρk

2
An easy computation gives rkk ≥ ρk , for every k ≥ 2. So, for every a ∈ T with
2
ak = 1, we have:
 2
1 − ρk 1
fk (ark ) ≥ xk ≥ xk .
1 − ρ2k 4
So


k2
Ψ(12|fk |) dν ≥ Ψ(12|fk |) dνk ≥ Ψ(3xk )
D\rk−1 D D\rk−1 D Ψ((k + 1)!)
k 2  
> k.(k!)3 ≥ 1.
Ψ((k + 1)!)

1
Therefore, we conclude that sup f LΨ (D\rk D,μ) ≥ , though rk → 1. By
f Ψ ≤1 12
Proposition 4.9, the identity from H Ψ to LΨ (μ) is not compact.
4. CHARACTERIZATION OF THE COMPACTNESS OF COMPOSITION OPERATORS 55

4. Characterization of the compactness of composition operators
For composition operators, compactness can be characterized in terms of Ψ-
Carleson measures, as stated in the following result.
Theorem 4.18. For every analytic self-map φ : D → D and every Orlicz func-
tion Ψ, the composition operator Cφ : H Ψ → H Ψ is compact if and only if one
has:
1
(R0 ) ρμφ (h) = o   as h → 0, for every A > 0.
Ψ AΨ−1 (1/h)
In other words, if and only if μφ is a vanishing Ψ-Carleson measure.
In order to get this result, we shall show that in Theorem 4.11, conditions (R0 )
and (K0 ) are equivalent for the pull-back measure μφ induced by φ. This is the
object of the following theorem.
Theorem 4.19. There exists a constant k1 > 0 such that, for every analytic
self-map φ : D → D, one has:
   
(4.15) μφ S(ξ, εh) ≤ k1 ε μφ S(ξ, h) ,
for every h ∈ (0, 1 − |φ(0)|), and every ε ∈ (0, 1).
Note that we prefer here to work with the sets:
S(ξ, h) = {z ∈ D ; |z − ξ| < h} , ξ ∈ T, 0 < h < 1,
instead of the Carleson windows W (ξ, h). Recall also that the pull-back measure
μφ is defined by (4.3).
We are going to postpone the proof of Theorem 4.19, and shall give before some
consequences.
4.1. Some consequences. An immediate consequence of Proposition 4.16
and Theorem 4.18 is the following
Theorem 4.20. Let φ : D → D be analytic and Ψ be an Orlicz function.
The operator Cφ on H Ψ is compact if and only if
 
1
(W) sup Cφ (uξ,1−h )Ψ = o   , as h → 0.
ξ∈T Ψ−1 1/h
We deduce:
Theorem 4.21. Let φ : D → D be analytic.
1) Assume that the Orlicz function Ψ satisfies condition Δ0 . Then, the operator
Cφ on H Ψ is weakly compact if and only if it is compact.
2) Assume that the Orlicz function Ψ satisfies condition ∇2 . Then, the operator
Cφ on H Ψ is a Dunford-Pettis operator if and only if it is compact.
Recall that (see Theorem 3.24), under condition Δ2 for Ψ, the weak compact-
ness of the composition operator Cφ is equivalent to its compactness, and even
to Cφ being order bounded into M Ψ (T). However, we shall see below, in Theo-
rem 4.22, that there exist Orlicz functions Ψ ∈ Δ0 (and even Ψ ∈ Δ1 ) for which
Cφ is compact, but not order bounded into M Ψ (T).
Proof. In both cases, the result follows from Theorem 4.20, since condition (W)
is satisfied. Indeed, if Cφ : H Ψ → H Ψ is weakly compact and Ψ ∈ Δ0 , we use
56 PASCAL LEFÈVRE, DANIEL LI, HERVÉ QUEFFÉLEC, LUIS RODRÍGUEZ-PIAZZA

Theorem 3.20. If Cφ : H Ψ → H Ψ is a Dunford-Pettis operator, this is due to
Proposition 3.28. 
Now, we have:
Theorem 4.22. There exist an Orlicz function Ψ satisfying Δ1 , and an analytic
self-map φ : D → D such that the composition operator Cφ : H Ψ → H Ψ is not order
bounded into M Ψ (T), though it is compact.
Remark. It follows that our assumption that Ψ ∈ Δ2 in Theorem 3.24 is not only
a technical one, though it might perhaps be weakened.

Proof. Let:    √
exp (log x)2 if x ≥ e,
Ψ(x) = √
e−1/4 x if 0 ≤ x ≤ e.
It is plain that Ψ ∈ Δ1 ∩ ∇0 .
Moreover, for every A > 0, one has, for h small enough:
1/h   
  = exp − (log A)2 − 2(log A) log(1/h) .
Ψ AΨ−1 (1/h)
Consider now φ = φ2 , the analytic self-map of D constructed in Theorem 4.1.
Then Cφ : H Ψ → H Ψ is not order bounded into M Ψ (T), by Theorem 3.15, since,
otherwise, Cφ1 : H Ψ → H Ψ would also be order bounded into M Ψ (T), which is
easily seen to be not the case (we may also argue as follows: Cφ1 : H Ψ → H Ψ
would be compact, and hence, by Theorem 4.3, Cφ1 would be compact from H 2
into H 2 , which is false).  
On the other hand, we have proved that ρμφ (h) = O h3/2 . So the conclusion
follows from Theorem 4.11, 3) and the fact that for every c > 0:
 1
−(log A)2 − 2(log A) log(1/h) ≥ −c log(1/h)
2
when h is small enough. 
Remark. We saw that a compact composition operator on H Ψ is always compact
on H 2 , and even nuclear when Ψ ∈ Δ2 (see Theorem 3.24). A natural question is:
when Ψ ∈ Δ0 , is a compact composition operator on H Ψ automatically a Hilbert-
Schmidt operator on H 2 ? The answer is “no” in general: the function φ2 in the
previous proof gives the counterexample. Indeed, Cφ2 is compact on H Ψ . If it
was a Hilbert-Schmidt operator on H 2 , Cφ1 would be Hilbert-Schmidt as well since
|φ∗1 | = |φ∗2 | but Cφ1 is not compact on H 2 .

4.2. Preliminary results. We shall use the radial maximal function N , de-
fined for every harmonic function u on D by:
(4.16) (N u)(ξ) = sup |u(rξ)| , ξ ∈ T.
0≤r<1

Recall that for every positive harmonic function u : D → C whose boundary
values u∗ are in L1 (T) (i.e. u ∈ h1 ), one has, for every ξ ∈ T:
(4.17) N u(ξ) ≤ M u∗ (ξ) ≤ πN u(ξ) ,
where M u∗ is the Hardy-Littlewood maximal function of u∗ (see [35], Theorem
11.20 and Exercise 19, Chapter 11).
4. CHARACTERIZATION OF THE COMPACTNESS OF COMPOSITION OPERATORS 57

We shall denote by Π the right half-plane:
Π = {z ∈ C ; Re z > 0}
and by C the cone:
C = {z ∈ C ; −π/6 < Arg z < π/6}.
The next result follows from Kolgomorov’s Theorem, saying that the Hilbert
transform is a weak (1, 1) operator (in applying this theorem to the positive har-
monic function 2 Re g = g + ḡ, noting that Re g1 = Re g(0) = g(0)).
Lemma 4.23. There exists a constant c > 0 such that, if g : D → Π is an
analytic function with g(0) > 0, then:
g(0)
m({|g ∗ | > λ}) ≤ c , for all λ > 0.
λ
 3
Applying Lemma 4.23 to g(z) = f (z) + f (0) , and taking into account that
|w1 + w2 | ≥ |w1 |, if w1 , w2 ∈ C , we get:
Lemma 4.24. Let f : D → C be an analytic function with values in the cone C ,
and write f = u + iv, with u, v real-valued. Then:
u(0) 3
m({|f ∗ | > λ}) ≤ 8c , for all λ > 0.
λ
The next proposition is one of the keys. We postpone its proof.
Proposition 4.25. There exists a constant k2 > 0 such that, for every analytic
function f = u + iv : D → C with values in the cone C , one has:
α 3
(4.18) m({|f ∗ | > λ} ∩ I) ≤ k2 m(I), for all λ > 0,
λ
where I is the arc I = {eit ; a < t < b}, with a, b ∈ R and α > 0 satisfying
b − a < π/2, α ≥ N u(eia ), and α ≥ N u(eib ).
As a corollary we obtain:
Proposition 4.26. Let f : D → C be an analytic function, and write f = u+iv,
as in Proposition 4.25. If α > 0 satisfies m({N u > α}) < 1/4, then:
α 3 2α
(4.19) m({|f ∗ | > λ}) ≤ k2 m({N u > α}), for all λ ≥ √ ·
λ 3
Proof. The set {N u > α} is open, and one can decompose it into a disjoint union
of open arcs {Ij }j . Each arc has measure m(Ij ) ≤ m({N u > α}) < 1/4, and so is
an arc of length less than π/2. We can then apply Proposition 4.25 and we obtain:
α 3
m({|f ∗ | > λ} ∩ Ij ) ≤ k2 m(Ij ), for every j.
λ
Summing up all these inequalities we get:
  α 3
m {|f ∗ | > λ} ∩ {N u > α} ≤ k2 m({N u > α}) .
λ
The proposition follows since |f ∗ | ≤ √23 u∗ ≤ √23 N u, and then {|f ∗ | > λ} is
contained in {N u > α}, for λ ≥ √ 2α
3
· 
We shall need one more result.
58 PASCAL LEFÈVRE, DANIEL LI, HERVÉ QUEFFÉLEC, LUIS RODRÍGUEZ-PIAZZA

Proposition 4.27. There exists a constant k3 > 0 such that for every analytic
function f : D → C with values in the cone C , one has, writing f = u + iv:
m({M u∗ > α} ≤ k3 m({u∗ > α/2}, for every α > 0.
In order to prove it, we shall first prove the following lemma.
Lemma 4.28. There exists a constant k4 > 0 such that for every analytic func-
tion f : D → C with values in the cone C , one has, writing f = u + iv:
   2
M (u∗ )2 (ξ) ≤ k4 M u∗ (ξ) , for all ξ ∈ T.
Proof. Observe that u2 − v 2 is a positive harmonic function since it is the real
part of f 2 . But f 2 belongs to H 1 (see [11], Theorem 3.2), so u2 − v 2 ∈ h1 and
we can use inequalities (4.17). We also have, since −π/3 < Arg (f 2 ) < π/3, that
u2 ≥ 3v 2 , and so u2 − v 2 ≥ 2u2 /3 and |f | ≤ √23 u. We get:
  3   3π
M (u∗ )2 ≤ M (u∗ )2 − (v ∗ )2 ≤ N (u2 − v 2 )
2 2
 2
3π 3π 3π 2
≤ 2
N (f ) = (N f ) ≤
2
√ Nu
2 2 2 3
≤ 2π (M u∗ )2 . 
Proof of Proposition 4.27. Write A = {M u∗ > α} and B = {u∗ > α/2}. For
every ξ ∈ A, there exists an open arc Iξ centered at ξ such that:

1
p= u∗ dm > M u∗ (ξ)/2 , and p > α .
m(Iξ ) Iξ
We have, using Lemma 4.28:

1    2
(u∗ )2 dm ≤ M (u∗ )2 (ξ) ≤ k4 M u∗ (ξ) ≤ 4k4 p2 .
m(Iξ ) Iξ
Let L be the set L = {u∗ > p/2} ∩ Iξ . Let P be the probability measure: P(ω) =
1
m(ω ∩ Iξ ). We apply the Paley-Zygmund inequality (see [28], Proposition
m(Iξ )
III.3. page 22; or [21], page 8) to the random variable u∗ , with the probability P
and we obtain
1 (Eu∗ )2 1
P(L) ≥ · ≥
4 E(u∗ )2 16k4
since Eu∗ = p and E(u∗ )2 ≤ 4k4 p2 .
Therefore, 16k4 m(L) ≥ m(Iξ ), and, since L ⊆ B ∩ Iξ , we have, for every
ξ ∈ A, an arc Iξ containing ξ such that 16k4 m(B ∩ Iξ ) ≥ m(Iξ ). Applying the
Hardy-Littlewood covering lemma, we then obtain:

n 
n
m(A) ≤ 3 m(Iξj ) ≤ 3 × 16 × k4 × m(Iξj ∩ B) ≤ k3 m(B). 
j=1 j=1

Proof of Proposition 4.25. Composing f with a suitable rotation, we can sup-
pose that a = −δ and b = δ, for 0 < δ < π/4. Let us call I − and I + the arcs
I − = {eit ; −δ < t < 0} , I + = {eit ; 0 < t < δ} .
4. CHARACTERIZATION OF THE COMPACTNESS OF COMPOSITION OPERATORS 59

We shall prove that:
α 3
(4.20) m({|f ∗ | > λ} ∩ I + ) ≤ k2 m(I + ) ,
λ
using just the fact that N u(eiδ ) ≤ α.
In the same way one can prove:
α 3
(4.21) m({|f ∗ | > λ} ∩ I − ) ≤ k2 m(I − ) ,
λ
using just that N u(e−iδ ) ≤ α.
Then, summing up (4.20) and (4.21), Proposition 4.25 will follow.
Let Q be the right half-disc
Q = {z ∈ D ; Re z > 0},
and denote by ψ the (unique) homeomorphism from D to Q, which is a conformal
mapping from D onto Q and sends 1 to 1, i to i, and −i to −i.
We can construct ψ as the composition of a Moebius transformation T , with
the square root function and then with T −1 . Namely, let:
z+i
T z = −i ;
z−i
T maps D onto the upper-half plane, sending −i into 0, −1 into −1, 1 into 1, and
also 0 into i, and i into ∞. The square root function maps the upper-half plane
into the first quadrant and T −1 :
z−i
T −1 z =
1 − iz
maps this quadrant onto the half-disk Q. √
It is not difficult to see that ψ(−1) = 0, ψ(0) = 2 − 1, and that there
exist ρ ∈ (0,√π/2) such that ψ(eiρ ) = e√πi/4 and ψ(e−iρ ) = e−πi/4 (we must have

e = 1/3 + ( 8)i/3; hence ρ = arctan( 8)).
Let J be the arc:
J = {eit ; −ρ < t < ρ}.
The map ψ is regular on J, and so there exist two constants γ1 and γ2 > 0 such
that for every Borel subset E of J, one has:
 
γ1 m(E) ≤ m ψ(E) ≤ γ2 m(E).
If now β ∈ (0, 1), we put:
ψβ (z) = (ψ(z))β .
Then, it is easy to see that for every Borel subset E of J, one has:
   
m ψβ (E) = β m ψ(E) ,
and so  
γ1 β m(E) ≤ m ψβ (E) ≤ γ2 β m(E).
In order to prove (4.20), consider the function F : D → C defined by:
 
F (z) = f eiδ ψβ (z) , where β = 4δ/π.
Then:   √ 
Re F (0) = u ( 2 − 1)β eiδ ≤ N u(eiδ ) ≤ α.
Let us call χ the map:
χ(z) = eiδ ψβ (z).
60 PASCAL LEFÈVRE, DANIEL LI, HERVÉ QUEFFÉLEC, LUIS RODRÍGUEZ-PIAZZA

It is clear that I + is contained in χ(J). If A = {|f ∗ | > λ} ∩ I + , then E = χ−1 (A)
is a Borel subset of J, and for every ξ ∈ E, one has |F ∗ (ξ)| > λ. Then:
   
m(A) = m χ(E) = m ψβ (E) ≤ γ2 β m(E)
δ
≤ 8γ2 m({|F ∗ | > λ})

= 8γ2 m(I + ) m({|F ∗ | > λ})
and using Lemma 4.24 for F ,
 3
Re F (0)
≤ 8γ2 m(I + ) × 8c3
λ
α 3 α 3
≤ 64γ2 c3 m(I + ) = k2 m(I + ) .
λ λ
The proof of (4.20) is finished, and Proposition 4.25 follows. 
 
4.3. Proof of Theorem 4.19. Using the fact h → μφ S(ξ, h) is nondecreas-
ing, it is enough to prove that there exist h0 > 0, and ε0 > 0, such that (4.15) is
true for 0 < h < h0 (1 − |φ(0)|), and 0 < ε < ε0 , because changing the constant k1 ,
if necessary, the theorem will follow.
We can also suppose that ξ = 1.
The real part of 1/ 1 − φ(z) is positive, in fact greater than 1/2, for every
z ∈ D. Take 0 < h < h0 (1 − |φ(0)|), and consider the analytic function f defined
by:
 1/3
h
f (z) = ,
1 − φ(z)
where the cubic root is taken in order that, for every z ∈ D, f (z) belongs to the
cone
C = {z ∈ C : −π/6 < Arg (z) < π/6 }.
    √
Clearly μφ S(1, h) = m({|f ∗ | > 1}) and μφ S(1, εh) = m({|f ∗ | > 1/ 3 ε}). We
also have:  1/3 
h
|f (0)| ≤ < 3 h0 .
1 − |φ(0)|
We shall write f = u + iv where u and v are real-valued harmonic functions.
Observe that:
1
|v(z)| < √ u(z), for every z ∈ D.
3
It is known that f ∈ H , for every p < 3 (see [11], Theorem 3.2), and so u and v
p

are the Poisson integrals of u∗ and v ∗ (in particular √ u, v ∈ h1 ).
We are looking for a control of m({|f | > 1/ ε}) by ε times m({|f ∗ | > 1}).
∗ 3

Proposition 4.26 provides this control if we replace m({|f ∗ | > 1}) by m({N u > 2}):

m({|f ∗ | > 1/ 3 ε}) ≤ 8k2 εm({N u > 2}),
when m({N u > 2}) < 1/4.
As f is valued in C , |f ∗ | is controlled by u∗ . Then what we need in fact is a
control of the measure of level sets of N u by the measure of level sets of u∗ . This
will be done by using Proposition 4.27.
Indeed, by Proposition 4.27, we have:
m({N u > 2}) ≤ m({M u∗ > 2}) ≤ k4 m({u∗ > 1}).
4. CHARACTERIZATION OF THE COMPACTNESS OF COMPOSITION OPERATORS 61

We know that ||u∗ ||1 = u(0) ≤ |f (0)| ≤ h0 . Then, choosing h0 small enough,
1/3

1/3
we can have k4 h0 < 1/4, and so m({N u > 2}) < 1/4; therefore, we can use
Proposition
√ 4.26. Moreover we also have {u∗ > 1} ⊆ {|f ∗ | > 1}. Taking ε0 <
3 3/64, we have, for 0 < ε < ε0 , by Proposition 4.26:

m({|f ∗ | > 1/ 3 ε}) ≤ 8k2 ε m({N u > 2}) ≤ 8k4 k2 ε m({u∗ > 1})
≤ k1 ε m({|f ∗ | > 1}) ,
taking k1 = 8k4 k2 . 
CHAPTER 5

Bergman spaces

1. Bergman-Orlicz spaces
Definition 5.1. Let dA (z) = dxπdy (z = x + iy) be the normalized Lebesgue
measure on D. The Bergman-Orlicz space B Ψ denotes the space of analytic func-
tions f : D → C which are in the Orlicz space LΨ (D, dA ). The Bergman-Morse-
Transue space is the subspace BM Ψ = B Ψ ∩ M Ψ (D, dA ).
B Ψ , equipped with the induced norm of LΨ (D, dA ), is a Banach space, as an
obvious consequence of the following lemma, analogous to Lemma 3.11.
Lemma 5.2. For every a ∈ D, the norm of the evaluation functional δa , which
maps f ∈ B Ψ to f (a), is:
1
δa  ≈ Ψ−1 ·
(1 − |a|)2
Proof. For every analytic function g : D → C, the mean-value property gives:

g(0) = g(z) dA (z).
D

Hence if φa : D → D denotes the analytic automorphism
z−a
φa (z) = ,
1 − āz
whose inverse is φ−1
a = φ−a , one has, for every f ∈ B , using the change of variable
Ψ

formula:

f (a) = f ◦ φ−a (0) = f ◦ φ−a (z) dA (z) = f (w) |φa (w)|2 dA (w)
D D

= f (w)Ha (w) dA (w) ,
D

where:
(1 − |a|2 )2
Ha (w) = |φa (w)|2 = ·
|1 − āw|4
The kernel Ha plays for B Ψ the role that the Poisson kernel Pa plays for H Ψ : the
analytic reproducing kernel Ka for B 2 being Ka (z) = (1−āz)
1
2 ·

We therefore have (using [33], Proposition 4, page 61, or [4], Theorem 8.14):
|f (a)| ≤ 2f Ψ Ha Φ ,
which proves the continuity of δa . To estimate its norm, we are going to majorize 
Ha Φ , with the help of Lemma 3.9. Let us notice that, on the one hand, Ha 1 = 1
63
64 PASCAL LEFÈVRE, DANIEL LI, HERVÉ QUEFFÉLEC, LUIS RODRÍGUEZ-PIAZZA

(take f = I in the above identity); and on the other hand:
(1 − |a|2 )2 (1 + |a|)2
Ha ∞ = = ·
(1 − |a|) 4 (1 − |a|)2
We get, setting b = Ha ∞ , and using Lemma 3.9 for  Φ :
b
Ha Φ ≤ ·
Φ−1 (b)
But b ≤ Φ−1 (b)Ψ−1 (b) (see [33], Proposition 1 (ii), page 14). Hence Ha Φ ≤
Ψ−1 (b). Now:
4
b≤ ·
(1 − |a|)2
We have Ψ−1 (4t) ≤ 4Ψ−1 (t) for all t > 0. It follows that
1
Ha Φ ≤ CΨ−1 ·
(1 − |a|)2
Since δa  ≤ 2Ha Φ , we get the upper bound in Lemma 5.2.
For the lower bound, we simply observe that Ha = |Ga |, where Ga (z) =
(1 − |a|2 )2
, and by Lemma 3.9:
(1 − az)4
1
|Ga (a)| |Ha (a)| (1 − |a|2 )2 Ψ−1 (b)
δa  ≥ = ≥ −1
=
Ga Ψ Ha Ψ b/Ψ (b) b(1 − |a|2 )2
Ψ−1 (b) 1 −1 1 −1 1
= ≥ Ψ (b) ≥ Ψ ,
(1 + |a|)4 16 16 (1 − |a|)2
since b ≥ 1/(1 − |a|)2 . 
Proposition 5.3. We have the following properties
i) BM Ψ is the closure of H ∞ (D) in LΨ (D, A ) and actually the algebraic
polynomials are dense in BM Ψ .  
ii) On the unit ball of B Ψ , the weak-star topology σ LΨ (D, A ), M Φ (D, A )
coincides with the topology of convergence on compact subsets of D.
iii) B Ψ is closed in LΨ (D, A ) for the weak-star topology.
iv) If Ψ ∈ ∇2 , B Ψ is (isometric to) the bidual of BM Ψ .
Proof. For the first point, let us fix f ∈ BM Ψ . Setting fr (z) = f (rz) for z ∈ D
and 0 ≤ r < 1, it suffices to show that fr − f Ψ −→ 0, since, being analytic in the
r→1
disk rD ⊂ D, fr can be uniformly approximated on D by its Taylor series. But the
norm of M Ψ is absolutely continuous (see [33], Theorem 14, page 84) and therefore,
for every ε > 0, there is some R > 0, with 1/3 ≤ R < 1, such that f ID\RD Ψ ≤ ε;
hence:
|f | 1 |f | 1
Ψ dA ≤ Ψ dA ≤ ·
D\RD 4ε D\RD 4 ε 4
When r ≥ 2R
R+1 ≥ 1/2, we therefore have:
|f |
r
Ψ dA ≤ 1,
D\ 2 D
1+R 4ε
1. BERGMAN-ORLICZ SPACES 65

and, by convexity of Ψ:

|fr − f | |f − f |
r
Ψ dA ≤ Ψ dA
D 8ε 2 D
1+R 8ε

1  |fr | |f | 
+ Ψ +Ψ dA
D\ 1+R
2 D
2 4ε 4ε
|f − f |
r
≤ Ψ dA
2 D
1+R 8ε
|f | |f |
1 r 1
+ Ψ dA + Ψ dA
2 D\ 1+R2 D
4ε 2 D\RD 4ε
≤ 1,
for r close enough to 1 since fr − f tends to 0 uniformly on 1+R2 D. Hence, for some
r0 < 1, one has fr − f Ψ ≤ 8ε for r0 ≤ r < 1. This was the claim.
ii) It suffices to use a sequential argument since the topologies are metrizable on
balls (the space M Φ (D, A ) is separable). Assume that f ∈ B Ψ (with f Ψ ≤ 1) is
the weak-star limit of a sequence of analytic functions fn ∈ B Ψ (with fn Ψ ≤ 1).
Testing this with the function hk (z) = (k + 1)z k , we obtain that the kth Taylor
coefficient ak (n) of fn converges to the kth Taylor coefficient ak of f :

ak (n) = fn hk dA −→ f hk dA = ak , for every k ≥ 0.
D D
Fix a compact K ⊂ D, there exists an r ∈ (0, 1) such that K ⊂ rD. We have:

sup |fn (z) − f (z)| ≤ |ak (n) − ak |r k −→ 0
z∈K
k≥0

by the dominated convergence Theorem (observe that |ak − ak (n)| ≤ 2(k + 1) for
every k ≥ 0 and every n ≥ 0).
For the converse, suppose now that fn ∈ B Ψ (with fn Ψ ≤ 1) converges
uniformly on every compact subsets of D to f ∈ B Ψ (with f Ψ ≤ 1). Fixing
g ∈ M Φ (D, A ) and ε > 0, there exists an r ∈ (0, 1) such that gID\rD Φ ≤ ε. We
have:
   
   
 (fn − f )gdA  ≤  (fn − f )gdA  + 2ε ≤ sup |fn (z) − f (z)|.g1 + 2ε.
  
D rD z∈rD

By hypothesis sup |fn (z) − f (z)| −→ 0. The conclusion follows.
z∈rD
iii) By the classical Theorem of Banach-Dieudonné, it is sufficient to prove that
the balls are weak-star closed (equivalently weak-star compact) and by separability
of M Φ (D, A ), the weak-star topology is metrizable on balls. The previous fact ii)
shows that it is equivalent to prove that the unit ball of B Ψ is compact for the
topology of convergence on compact subsets. But this is easy: indeed, if fn in the
unit ball of B Ψ . This is a normal family thanks to Lemma 5.2. A subsequence
converges to an analytic function f on compact subsets of D and the Fatou Lemma
implies that f actually lies in the unit ball.
iv) Assume now that Ψ satisfies ∇2 . Since (M Ψ )∗∗ = LΨ (D, A ), we have
w∗
(BM Ψ )∗∗ = BM Ψ , in the space LΨ (D, A ). Hence it suffices to show that
w∗
BM Ψ = B Ψ . We already know that B Ψ is weak-star closed. Now, let f ∈ B Ψ .
Obviously, fr ∈ BM Ψ for every r ∈ (0, 1), where fr (z) = f (rz). Moreover fr Ψ ≤
66 PASCAL LEFÈVRE, DANIEL LI, HERVÉ QUEFFÉLEC, LUIS RODRÍGUEZ-PIAZZA 

f Ψ . But it is clear that fr is uniformly convergent to f on compact subsets of
D, when r → 1− . By ii), the conclusion follows. 
In the previous proof, we can see the points ii) and iii) in a slighlty different
way: the unit ball B of B Ψ is compact for the topology τ of uniform convergence on
compact subsets. Then observe that the identity from B, equipped with the topol-
ogy τ , to B, equipped with the weak-star topology, is continuous (the sequential
argument is sufficient by metrizability). This implies, since the weak-star topology
is separated, that B is weak-star compact (hence closed) and that the topologies
coincide. Now, by Banach-Dieudonné’s theorem, the space B Ψ is weak-star closed.

2. Compact composition operators on Bergman-Orlicz spaces
We shall begin with showing that, as in the Hardy-Orlicz case, every symbol φ
defines a bounded composition operator.
Proposition 5.4. Every analytic self-map φ : D → D induces a bounded com-
position operator Cφ : B Ψ → B Ψ . Moreover, Cφ f ∈ BM Ψ for every f ∈ BM Ψ ;
hence, when Ψ ∈ ∇2 , the former operator is the bi-transposed of Cφ : BM Ψ →
BM Ψ .
Proof. It suffices to follow the lines of Proposition 3.12, and to integrate the
integrals written there between 0 and 1, with respect to the measure 2rdr. 

Before stating and proving the main theorem, we are going to prove the follow-
ing auxiliary result, interesting in itself, and which reinforces an example of J. H.
Shapiro ([39], Example, page 185).
Proposition 5.5. There exists a Blaschke product B having angular derivative
at no point of T = ∂D, in the following sense:
(5.1) (∀ε > 0) (∃cε > 0) 1 − |B(z)| ≥ cε (1 − |z|)ε , ∀z ∈ D.
Proof. We shall take:

+∞
|zn | zn − z
B(z) = ,
z 1 − z̄n z
n=1 n
where: 
{zn ; n ≥ 1} = An ,
n≥1

with:
An = {rn ωnj ; ωn = e2πi/pn , 0 ≤ j ≤ pn − 1} ,
where (rn )n≥1 is a (strictly) increasing sequence with 0 < rn < 1, and the integers
pn will have to be adjusted, satisfying the Blaschke condition:

+∞ 
+∞
(1 − |zn |) = pn (1 − rn ) < +∞.
n=1 n=1

One has:

+∞  z − z 2 +∞
 n   (1 − |z|2 )(1 − |zn |2 )

 
|B(z)|2 =   = 1− ≤ exp − S(z) ,
n=1
1 − z̄n z n=1
|1 − z̄n z| 2
2. COMPACT COMPOSITION OPERATORS ON BERGMAN-ORLICZ SPACES 67

where:

+∞
(1 − |z|2 )(1 − |zn |2 )
S(z) = ·
n=1
|1 − z̄n z|2
We now proceed to minorize S(z). For this purpose, we shall need the following
simple lemma, whose proof will be temporarily postponed.
Lemma 5.6. For every positive integer p and every a ∈ D, one has, setting
ω = e2πi/p :
1
p−1
1 1 − |a|2p 1 1
= ≥ p |a|p .
p |1 − aω |
k 2 1 − |a| |1 − a |
2 p 2 4
k=0

Then, setting r = |z|, we have:

+∞ n −1
p
1
S(z) ≥ (1 − r) (1 − rn )
n=1 k=0
|1 − rn ωn−k z|2

1−r  2
+∞
≥ p (1 − rn )(rn r)pn .
4 n=1 n
We shall take:  
pn = (1 − rn )εn −1 + 1 ,
with:
1 1
εn = √ and rn = 1 − n
n 2
and where [ ] stands for the integer part. More explicitly:
 √ 
pn = 2n− n + 1.
If r ≥ 1/2, let N ≥ 1 be such that rN < r ≤ rN +1 . One has:
1−r 1
S(z) ≥ (1 − rN )2εN −1 rN2pN ≥ (1 − rN )2εN rN2pN ,
4 8
since 1 − r ≥ 1 − rN +1 = (1 − rN )/2. Moreover:
pN ≤ 2(1 − rN )εN −1 = 2.2N (1−εN ) ≤ 2.2N ,
so that:
1
(1 − r)2εN (1 − 2−N )2
N +1
S(z) ≥ ≥ c(1 − r)2εN ,
8
where c is a positive numerical constant.
Hence, setting:
ε(z) = 2εN for |z| ≥ 1/2 and rN < r ≤ rN +1 ,
one has:
ε(z) −→− 0 ,
|z|→1
and we get:
1 − |B(z)|2 ≥ 1 − e−S(z) ≥ 1 − e−c(1−|z|) ≥ c (1 − |z|)ε(z) ,
ε(z)

where c is another positive numerical constant. This gives condition (5.1), since
1 − |B(z)|2
1 − |B(z)| ≥ ·
2
68 PASCAL LEFÈVRE, DANIEL LI, HERVÉ QUEFFÉLEC, LUIS RODRÍGUEZ-PIAZZA

Finally, the Blaschke condition is satisfied, since:

pn (1 − rn ) ≤ 2(1 − rn )εn = 2.2− n
.
This ends the proof of Proposition 5.5. 
Proof of Lemma 5.6. Let Gp be the finite group of pth roots of unity, equipped
with its normalized Haar measure. For u : Gp → C and 0 ≤ k ≤ p − 1, we denote
by û(k) the kth Fourier coefficient of u, i.e.:
1 
û(k) = u(z)z −k .
p
z∈Gp

Then, the Plancherel-Parseval formula for Gp reads:

p−1
1 
|û(k)|2 = |u(z)|2 .
p
k=0 z∈Gp

Applying this to

1  
p−1
lp+k k
u(z) = = a z = û(k)z k ,
1 − az l≥0 k=0
0≤k≤p−1

with
 ak
û(k) = alp+k = ,
1 − ap
l≥0
we get:
1  |a|2k
p−1 p−1
1 1 − |a|2p
= = ·
p |1 − aω |
k 2 |1 − a |
p 2 (1 − |a|2 )|1 − ap |2
k=0 k=0
To finish, we note that |1 − a | ≤ 2, and that, by the arithmetico-geometric in-
p

equality, we have, with x = |a|2 :
1 − |a|2p
= 1 + x + · · · + xp−1
1 − |a|2
 1/p p−1
≥ p x1+2+···+(p−1) = p x 2 ≥ p xp/2 = p |a|p .

Theorem 5.7. If the composition operator Cφ : B Ψ → B Ψ is compact, then

1
Ψ−1
(1 − |φ(a)|)2
(5.2)  −→ 0.
−1
1 |a|→1−
Ψ
(1 − |a|)2
This condition is sufficient if Ψ ∈ Δ2 .
Before giving the proof of this theorem, let us note that in case where Ψ = Ψ2 ,
2
with Ψ2 (x) = ex − 1, it reads: the composition operator Cφ : B Ψ2 → B Ψ2 is
compact if and only if:
(5.3) (∀ε > 0) (∃cε > 0) 1 − |φ(z)| ≥ cε (1 − |z|)ε , ∀z ∈ D.
2. COMPACT COMPOSITION OPERATORS ON BERGMAN-ORLICZ SPACES 69

Indeed, Ψ2 ∈ Δ2 , and:
 
1 1
Ψ−1 log
2
(1 − |φ(a)|)2 (1 − |φ(a)|)2
 =  ;
−1 1 1
Ψ2 log
(1 − |a|)2 (1 − |a|)2
hence Cφ is compact if and only if
1
log
(1 − |φ(a)|)2
−→ 0.
1 |a|→1
log
(1 − |a|)2
Then, for every ε > 0, we can find some cε > 0 such that, for all a ∈ D:
1 1
log ≤ ε log + cε ;
1 − |φ(a)| (1 − |a|)
which is equivalent to (5.3). 
That allows to have compact composition operators on B Ψ2
which are not
compact on H Ψ2 . However it is very likely that this is the case for every Ψ ∈ Δ2 ,
but we have not tried to see this full generality.
Theorem 5.8. There exist symbols φ : D → D such that the composition op-
erators Cφ is compact from B Ψ2 into itself, but not compact from H Ψ2 into itself,
and even Cφ is an isometry onto its image.
A similar example is well-known for the Hilbert spaces B 2 and H 2 (see [39],
pages 180–186).
Proof. Let B be a Blaschke product verifying the condition of Proposition 5.5.
We introduce φ(z) = zB(z). The function still verifies 1 − |φ(z)| ≥ Cε (1 − |z|)ε ,
since |φ(z)| ≤ |B(z)| on D. From Theorem 5.7, it follows, since Ψ2 ∈ Δ2 , that
Cφ : B Ψ2 → B Ψ2 is compact.
We are now going to see that Cφ : H Ψ2 → H Ψ2 is an isometry. Indeed, recall
the following well-known fact, that we already used (see [32], Theorem 1): since
φ is an inner function, the image φ(m) of the Haar measure m of T under φ is
equal to Pa .m, where a = φ(0) and Pa is the Poisson kernel at a. Here φ(0) = 0 so
φ(m) = m. It follows that for every f ∈ H Ψ2 , one has, for C > 0:
|f ◦ φ| |f |
Ψ2 dm = Ψ2 dm
T C T C
so that f Ψ2 = f ◦ φΨ2 . 

We shall need also the following lemma, which completes Lemma 5.2.
Lemma 5.9. For every f ∈ BM Ψ , one has:
 
−1 1
f (a) = o Ψ as |a| −→ 1− .
(1 − |a|)2
Proof.
 This is obvious
 for the monomials en : z → z n since |en (a)| ≤ 1, whereas
−1
Ψ 1/(1 − |a|) −→ +∞. Since the evaluation δa is bounded on BM Ψ and
2
|a|→1
70 PASCAL LEFÈVRE, DANIEL LI, HERVÉ QUEFFÉLEC, LUIS RODRÍGUEZ-PIAZZA 

 −1  
δa / Ψ (1/1 − |a|2 )  = O (1), it suffices to use that the polynomials are dense
in BM Ψ ; but this was already proved in Proposition 5.3. 

Proof of Theorem 5.7. If Cφ : B Ψ → B Ψ is compact, then so is the restriction
 ∗  ∗
Cφ : BM Ψ → BM Ψ and its adjoint Cφ∗ = Cφ : BM Ψ → BM Ψ . Since
w∗
Cφ∗ (δa ) = δφ(a) , Lemma 5.9 gives δa /δa  −→ 0. Compactness of Cφ∗ now leads us
|a|→1
to
δ
Cφ∗
a
−→ 0.
δa  |a|→1
That gives (5.2), in view of Lemma 5.2.
Conversely, assume that (5.2) is verified. Observe first that, since Ψ ∈ Δ2 , one
has:

(5.4) Ψ−1 (x2 ) ≤ αΨ−1 (x) for x large enough.
 2
Indeed, let x0 > 0 be such that Ψ(αx) ≥ Ψ(x) for x ≥ x0 . For x ≥ y0 = 
 2
Ψ(αx0 ), with y = Ψ−1 (x2 ), one has x2 = Ψ(y) ≥ Ψ(y/α) , i.e. x ≥ Ψ(y/α),
and hence Ψ−1 (x2 ) = y ≤ αΨ−1 (x).
Therefore condition (5.2), which reads:
 
−1 1 −1 1
Ψ =o Ψ ,
(1 − |φ(z)|)2 (1 − |z|)2
reads as well, because of (5.4):
 1 
1
(5.5) Ψ−1 = o Ψ−1 , as |z| → 1.
1 − |φ(z)| 1 − |z|

We have to prove that (5.5) implies the compactness of Cφ : B Ψ → B Ψ . So, by
Proposition 3.8, we have to prove that: for every sequence (fn )n in the unit ball of
B Ψ which converges uniformly to 0 on compact sets of D, one has fn ◦ φΨ −→ 0.
n→∞
But Lemma 5.2 and (5.4) imply that, for some C > 0:
1
(5.6) |fn (z)| ≤ C Ψ−1 , ∀z ∈ D .
1 − |z|
Let ε > 0 and set ε0 = ε/αC. Due to (5.5), we can find some r with 0 < r < 1
such that:
⎧ √ 1
⎪ 1

⎪ 1−r ≤ ; Ψ−1 ≥ αx0 ;
⎨ 8 1−r
(5.7) 1

⎪ 1

⎩ Ψ−1 ≤ ε0 Ψ−1 if z ∈ D \ rD .
1 − |φ(z)| 1 − |z|
Then, since (fn )n converges uniformly to 0 on φ(rD), we have, for n large
enough (n ≥ n0 ):
|f ◦ φ|
n 1
Ψ dA (z) ≤ ·
rD ε 2
ACKNOWLEDGEMENT 71

On the other hand, by (5.7):
|f ◦ φ|  
n C −1 1
Ψ dA (z) ≤ Ψ Ψ dA (z)
D\rD ε D\rD ε 1 − |φ(z)|
 
ε0 C −1 1
≤ Ψ Ψ dA (z)
D\rD ε 1 − |z|
 
1 −1 1
= Ψ Ψ dA (z)
D\rD α 1 − |z|
$ 
1 
≤ Ψ Ψ −1 dA (z)
D\rD 1 − |z|
1
since Ψ−1 ≥ αx0 ,
1−r
1
1 ρ dρ
=  dA (z) = 2 √
D\rD 1 − |z| r 1−ρ
1
dρ √ 1
≤2 √ =4 1−r ≤ ·
r 1 − ρ 2
Putting together these two inequalities, we get, for n ≥ n0 :

|fn ◦ φ| 1 1
Ψ dA (z) ≤ + = 1,
D ε 2 2
and hence: fn ◦ φΨ ≤ ε, which ends the proof of Theorem 5.7. 

Acknowledgement
Part of this work was made when the fourth-named author was Professeur
invité de l’Université d’Artois in May-June 2005 and during a visit of this author
in Lille and in Lens in March 2006. He wishes to thank both departements for their
hospitality. He was also partially supported by Spanish project MTM2006-05622. A
part was made too during a stay of the first named author in the University of Sevilla
in September 2006. He is grateful to the departamento de Análisis Matemático.
References

[1] J. Arazy, S. D. Fisher, and J. Peetre, Möbius invariant function spaces, J. Reine Angew.
Math. 363 (1985), 110–145. MR814017 (87f:30104)
[2] F. Bayart, Hardy spaces of Dirichlet series and their composition operators, Monatsh. Math.
136, no. 3 (2002), 203–236. MR1919645 (2003i:42032)
[3] F. Bayart, Compact composition operators on a Hilbert space of Dirichlet series, Illinois J.
Math. 47, no. 3 (2003), 725–743. MR2007233 (2004j:47049)
[4] C. Bennett and R. Sharpley, Interpolation of operators, Pure and Applied Math. 129, Aca-
demic Press (1988). MR928802 (89e:46001)
[5] O. Blasco de la Cruz and H. Jarchow, A note on Carleson measures for Hardy spaces, Acta
Sci. Math. (Szeged) 71, no. 1-2 (2005), 371–389. MR2160373 (2006d:30048)
[6] J. S. Choa and H. O. Kim, On function-theoretic conditions characterizing compact com-
position operators on H 2 , Proc. Japan Acad. Ser. A Math. Sci. 75, no. 7 (1999), 109–112.
MR1729855 (2000k:47027)
[7] J. S. Choa and H. O. Kim, Composition Operators between Nevanlinna-Type Spaces, J.
Math. Anal. Appl. 257, no. 2 (2001), 378–402. MR1827329 (2002c:47045)
[8] J. S. Choa, H. O. Kim, Hong Oh and J. H. Shapiro, Compact composition operators on
the Smirnov class, Proc. Amer. Math. Soc. 128, no. 8 (2000), 2297–2308. MR1653449
(2000k:47034)
[9] J. A. Cima and A. L. Matheson, Essential norms of composition operators and Aleksandrov
measures, Pacific J. Math. 179, no. 1 (1997), 59–64. MR1452525 (98e:47047)
[10] C. C. Cowen and B. D. MacCluer, Composition operators on spaces of analytic func-
tions, Studies in Advanced Mathematics, CRC Press, Boca Raton, FL (1995). MR1397026
(97i:47056)
[11] P. L. Duren, Theory of H p spaces, Second edition, Dover Publications (2000).
[12] J. B. Garnett, Bounded Analytic Functions, Pure and Applied Math. 96, Academic Press
(1981). MR628971 (83g:30037)
[13] J. Gordon and H. Hedenmalm, The composition operators on the space of Dirichlet series
with square summable coefficients, Michigan Math. J. 46, no. 2 (1999), 313–329. MR1704209
(2000f:47046)
[14] A. Hartmann, Interpolation and harmonic majorants in big Hardy-Orlicz spaces, J. Anal.
Math. 103 (2007), 197–219. Arxiv.org/abs/math.cv/0603624. MR2373268 (2008j:30051)
[15] H. Hunziker and H. Jarchow, Composition operators which improve integrability, Math.
Nachr. 152 (1991), 83–99. MR1121226 (93d:47061)
[16] N. Jaoua, Order bounded composition operators on the Hardy spaces and the Nevanlinna
class, Studia Math. 134, No.1 (1999), 35–55. MR1688214 (2000b:47069)
[17] N. Jaoua, Similarity to a contraction and hypercontractivity of composition operators, Proc.
Am. Math. Soc. 129, No.7 (2001), 2085–2092. MR1825921 (2002d:47038)
[18] H. Jarchow, Some functional analytic properties of composition operators, First International
Conference in Abstract Algebra (Kruger Park, 1993), Quaestiones Math. 18, no. 1-3 (1995),
229–256. MR1340480 (96f:47058)
[19] H. Jarchow, Compactness properties of composition operators, International Workshop on
Operator Theory (Cefalù, 1997), Rend. Circ. Mat. Palermo (2) Suppl., no. 56 (1998), 91–97.
MR1710825 (2000e:47047)
[20] H. Jarchow and J. Xiao, Composition operators between Nevanlinna classes and Bergman
spaces with weights, J. Operator Theory 46, no. 3, suppl. (2001), 605–618. MR1897157
(2003d:47034)

73
74 REFERENCES

[21] J.-P. Kahane, Some random series of functions, Cambridge Studies in Adv. Math. 5. Cam-
bridge University Press, Cambridge, 1985. MR0833073 (87m:60119)
[22] M. A. Krasnosel’skiı̆ and Ja. B. Rutickiı̆, Convex functions and Orlicz spaces, P. Noordhoff
Ltd., Groningen (1961). MR0126722 (23:A4016)
[23] P. Lefèvre, Some Characterizations of weakly compact operators on H ∞ and on the disk
algebra. Application to composition operators, J. Operator Theory 54:2 (2005), 229–238.
MR2186350 (2006k:47070)
[24] P. Lefèvre, D. Li, H. Queffélec, and L. Rodrı́guez-Piazza, Some translation-invariant Ba-
nach function spaces which contain c0 , Studia Math. 163 (2) (2004), 137–155. MR2047376
(2005b:46062)
[25] P. Lefèvre, D. Li, H. Queffélec, and L. Rodrı́guez-Piazza, A criterion of weak compactness for
operators on subspaces of Orlicz spaces, J. Funct. Spaces Appl. Vol. 6, No 3 (2008), 277–292.
MR2444526
[26] P. Lefèvre, D. Li, H. Queffélec, and L. Rodrı́guez-Piazza, Some examples of compact compo-
sition operators on H 2 , J. Funct. Analysis 255, No 11 (2008), 3098–3124. MR2464571
[27] P. Lefèvre, D. Li, H. Queffélec, and L. Rodrı́guez-Piazza, Compact composition operators on
H 2 and Hardy-Orlicz spaces, J. Math. Anal. Appl. 354 (2009), 360–371. MR2510446
[28] D. Li and H. Queffélec, Introduction à l’étude des espaces de Banach –Analyse et Probabilités,
Cours Spécialisés 12, Société Mathématique de France (2004). MR2124356 (2006c:46012)
[29] L. Liu, G. Cao, and X. Wang, Composition operators on Hardy-Orlicz spaces, Acta Math.
Sci. Ser. B Engl. Ed. 25, no. 1 (2005), 105–111. MR2119342 (2005k:47053)
[30] B. D. MacCluer, Compact composition operators on Hp (B N ), Michigan Math. J. 32 (1985),
no. 2, 237–248. MR783578 (86g:47037)
[31] V. Matache, A short proof of a characterization of inner functions in terms of the composition
operators they induce, Rocky Mountain J. Math. 35, no. 5 (2005), 1723–1726. MR2206032
(2007m:47052)
[32] E. A. Nordgren, Composition operators, Canad. J. Math. 20 (1968), 442–449. MR0223914
(36:6961)
[33] M. M. Rao and Z. D. Ren, Theory of Orlicz spaces, Pure and Applied Mathematics 146,
Marcel Dekker, Inc. (1991). MR1113700 (92e:46059)
[34] M. Rosenblum and J. Rovnyak, Topics in Hardy classes and univalent functions, Birkhäuser
Advanced Texts: Basler Lehrbücher, Birkhäuser Verlag, Basel (1994). MR1307384
(97a:30047)
[35] W. Rudin, Real and complex analysis, Third edition, McGraw-Hill Book Co., New York
(1987). MR924157 (88k:00002)
[36] D. Sarason, Composition operators as integral operators, Analysis and partial differential
equations, 545–565, Lecture Notes in Pure and Appl. Math., 122, Dekker, New York (1990).
MR1044808 (92a:47040)
[37] H. J. Schwartz, Composition operators on H p , Thesis, University of Toledo (1969).
[38] J. H. Shapiro, The essential norm of a composition operator, Annals of Math. 125 (1987),
375–404. MR881273 (88c:47058)
[39] J. H. Shapiro, Composition operators and classical function theory, Universitext. Tracts in
Mathematics, Springer-Verlag, New York (1993). MR1237406 (94k:47049)
[40] J. H. Shapiro, What do composition operators know about inner functions?, Monatsh. Math.
130, no. 1 (2000), 57–70. MR1762064 (2001a:47029)
[41] J. H. Shapiro, W. Smith, and D. A. Stegenga, Geometric models and compactness of com-
position operators, J. Funct. Anal. 127, no. 1 (1995), 21–62. MR1308616 (95m:47051)
[42] J. H. Shapiro and P. D. Taylor, Compact, Nuclear, and Hilbert-Schmidt Composition Oper-
ators on H 2 , Indiana Univ. Math. Journal 23 (6) (1973), 471–496. MR0326472 (48:4816)
[43] J.-O. Strömberg, Bounded mean oscillation with Orlicz norms and duality of Hardy spaces,
Bull. Amer. Math. Soc. 82, no. 6 (1976), 953–955. MR0419772 (54:7790)
[44] J.-O. Strömberg, Bounded mean oscillation with Orlicz norms and duality of Hardy spaces,
Indiana Univ. Math. J. 28, no. 3 (1979), 511–544. MR529683 (81f:42021)
[45] N. Zorboska, Compact composition operators on some weighted Hardy spaces, J. Operator
Theory 22, no. 2 (1989), 233–241. MR1043725 (91c:47062)
Editorial Information
To be published in the Memoirs, a paper must be correct, new, nontrivial, and sig-
nificant. Further, it must be well written and of interest to a substantial number of
mathematicians. Piecemeal results, such as an inconclusive step toward an unproved ma-
jor theorem or a minor variation on a known result, are in general not acceptable for
publication.
Papers appearing in Memoirs are generally at least 80 and not more than 200 published
pages in length. Papers less than 80 or more than 200 published pages require the approval
of the Managing Editor of the Transactions/Memoirs Editorial Board. Published pages are
the same size as those generated in the style files provided for AMS-LATEX or AMS-TEX.
Information on the backlog for this journal can be found on the AMS website starting
from http://www.ams.org/memo.
A Consent to Publish and Copyright Agreement is required before a paper will be
published in the Memoirs. After a paper is accepted for publication, the Providence
office will send a Consent to Publish and Copyright Agreement to all authors of the
paper. By submitting a paper to the Memoirs, authors certify that the results have not
been submitted to nor are they under consideration for publication by another journal,
conference proceedings, or similar publication.

Information for Authors
Memoirs is an author-prepared publication. Once formatted for print and on-line
publication, articles will be published as is with the addition of AMS-prepared frontmatter
and backmatter. Articles are not copyedited; however, confirmation copy will be sent to
the authors.
Initial submission. The AMS uses Centralized Manuscript Processing for initial sub-
missions. Authors should submit a PDF file using the Initial Manuscript Submission form
found at www.ams.org/submission/memo, or send one copy of the manuscript to the follow-
ing address: Centralized Manuscript Processing, MEMOIRS OF THE AMS, 201 Charles
Street, Providence, RI 02904-2294 USA. If a paper copy is being forwarded to the AMS,
indicate that it is for Memoirs and include the name of the corresponding author, contact
information such as email address or mailing address, and the name of an appropriate
Editor to review the paper (see the list of Editors below).
The paper must contain a descriptive title and an abstract that summarizes the article
in language suitable for workers in the general field (algebra, analysis, etc.). The descrip-
tive title should be short, but informative; useless or vague phrases such as “some remarks
about” or “concerning” should be avoided. The abstract should be at least one com-
plete sentence, and at most 300 words. Included with the footnotes to the paper should
be the 2010 Mathematics Subject Classification representing the primary and secondary
subjects of the article. The classifications are accessible from www.ams.org/msc/. The
Mathematics Subject Classification footnote may be followed by a list of key words and
phrases describing the subject matter of the article and taken from it. Journal abbrevi-
ations used in bibliographies are listed in the latest Mathematical Reviews annual index.
The series abbreviations are also accessible from www.ams.org/msnhtml/serials.pdf. To
help in preparing and verifying references, the AMS offers MR Lookup, a Reference Tool
for Linking, at www.ams.org/mrlookup/.
Electronically prepared manuscripts. The AMS encourages electronically pre-
pared manuscripts, with a strong preference for AMS-LATEX. To this end, the Society
has prepared AMS-LATEX author packages for each AMS publication. Author packages
include instructions for preparing electronic manuscripts, samples, and a style file that gen-
erates the particular design specifications of that publication series. Though AMS-LATEX
is the highly preferred format of TEX, author packages are also available in AMS-TEX.
Authors may retrieve an author package for Memoirs of the AMS from www.ams.org/
journals/memo/memoauthorpac.html or via FTP to ftp.ams.org (login as anonymous,
enter your complete email address as password, and type cd pub/author-info). The
AMS Author Handbook and the Instruction Manual are available in PDF format from the
author package link. The author package can also be obtained free of charge by sending
email to tech-support@ams.org or from the Publication Division, American Mathematical
Society, 201 Charles St., Providence, RI 02904-2294, USA. When requesting an author
package, please specify AMS-LATEX or AMS-TEX and the publication in which your paper
will appear. Please be sure to include your complete mailing address.
After acceptance. The source files for the final version of the electronic manuscript
should be sent to the Providence office immediately after the paper has been accepted for
publication. The author should also submit a PDF of the final version of the paper to the
editor, who will forward a copy to the Providence office.
Accepted electronically prepared files can be submitted via the web at www.ams.org/
submit-book-journal/, sent via FTP, or sent on CD to the Electronic Prepress Depart-
ment, American Mathematical Society, 201 Charles Street, Providence, RI 02904-2294
USA. TEX source files and graphic files can be transferred over the Internet by FTP to
the Internet node ftp.ams.org (130.44.1.100). When sending a manuscript electronically
via CD, please be sure to include a message indicating that the paper is for the Memoirs.
Electronic graphics. Comprehensive instructions on preparing graphics are available
at www.ams.org/authors/journals.html. A few of the major requirements are given
here.
Submit files for graphics as EPS (Encapsulated PostScript) files. This includes graphics
originated via a graphics application as well as scanned photographs or other computer-
generated images. If this is not possible, TIFF files are acceptable as long as they can be
opened in Adobe Photoshop or Illustrator.
Authors using graphics packages for the creation of electronic art should also avoid the
use of any lines thinner than 0.5 points in width. Many graphics packages allow the user
to specify a “hairline” for a very thin line. Hairlines often look acceptable when proofed
on a typical laser printer. However, when produced on a high-resolution laser imagesetter,
hairlines become nearly invisible and will be lost entirely in the final printing process.
Screens should be set to values between 15% and 85%. Screens which fall outside of this
range are too light or too dark to print correctly. Variations of screens within a graphic
should be no less than 10%.
Inquiries. Any inquiries concerning a paper that has been accepted for publication
should be sent to memo-query@ams.org or directly to the Electronic Prepress Department,
American Mathematical Society, 201 Charles St., Providence, RI 02904-2294 USA.
Editors
This journal is designed particularly for long research papers, normally at least 80 pages in
length, and groups of cognate papers in pure and applied mathematics. Papers intended for
publication in the Memoirs should be addressed to one of the following editors. The AMS uses
Centralized Manuscript Processing for initial submissions to AMS journals. Authors should follow
instructions listed on the Initial Submission page found at www.ams.org/memo/memosubmit.html.
Algebra, to ALEXANDER KLESHCHEV, Department of Mathematics, University of Oregon, Eu-
gene, OR 97403-1222; e-mail: ams@noether.uoregon.edu
Algebraic geometry, to DAN ABRAMOVICH, Department of Mathematics, Brown University,
Box 1917, Providence, RI 02912; e-mail: amsedit@math.brown.edu
Algebraic geometry and its applications, to MINA TEICHER, Emmy Noether Research Insti-
tute for Mathematics, Bar-Ilan University, Ramat-Gan 52900, Israel; e-mail: teicher@macs.biu.ac.il
Algebraic topology, to ALEJANDRO ADEM, Department of Mathematics, University of British
Columbia, Room 121, 1984 Mathematics Road, Vancouver, British Columbia, Canada V6T 1Z2; e-mail:
adem@math.ubc.ca
Combinatorics, to JOHN R. STEMBRIDGE, Department of Mathematics, University of Michigan,
Ann Arbor, Michigan 48109-1109; e-mail: JRS@umich.edu
Commutative and homological algebra, to LUCHEZAR L. AVRAMOV, Department of Math-
ematics, University of Nebraska, Lincoln, NE 68588-0130; e-mail: avramov@math.unl.edu
Complex analysis and harmonic analysis, to ALEXANDER NAGEL, Department of Mathemat-
ics, University of Wisconsin, 480 Lincoln Drive, Madison, WI 53706-1313; e-mail: nagel@math.wisc.edu
Differential geometry and global analysis, to CHRIS WOODWARD, Department of Mathemat-
ics, Rutgers University, 110 Frelinghuysen Road, Piscataway, NJ 08854; e-mail: ctw@math.rutgers.edu
Dynamical systems and ergodic theory and complex analysis, to YUNPING JIANG, Depart-
ment of Mathematics, CUNY Queens College and Graduate Center, 65-30 Kissena Blvd., Flushing, NY
11367; e-mail: Yunping.Jiang@qc.cuny.edu
Functional analysis and operator algebras, to DIMITRI SHLYAKHTENKO, Department of
Mathematics, University of California, Los Angeles, CA 90095; e-mail: shlyakht@math.ucla.edu
Geometric analysis, to WILLIAM P. MINICOZZI II, Department of Mathematics, Johns Hopkins
University, 3400 N. Charles St., Baltimore, MD 21218; e-mail: trans@math.jhu.edu
Geometric topology, to MARK FEIGHN, Math Department, Rutgers University, Newark, NJ
07102; e-mail: feighn@andromeda.rutgers.edu
Harmonic analysis, representation theory, and Lie theory, to E. P. VAN DEN BAN, De-
partment of Mathematics, Utrecht University, P.O. Box 80 010, 3508 TA Utrecht, The Netherlands;
e-mail: E.P.vandenBan@uu.nl
Logic, to STEFFEN LEMPP, Department of Mathematics, University of Wisconsin, 480 Lincoln
Drive, Madison, Wisconsin 53706-1388; e-mail: lempp@math.wisc.edu
Number theory, to JONATHAN ROGAWSKI, Department of Mathematics, University of Califor-
nia, Los Angeles, CA 90095; e-mail: jonr@math.ucla.edu
Number theory, to SHANKAR SEN, Department of Mathematics, 505 Malott Hall, Cornell Uni-
versity, Ithaca, NY 14853; e-mail: ss70@cornell.edu
Partial differential equations, to GUSTAVO PONCE, Department of Mathematics, South Hall,
Room 6607, University of California, Santa Barbara, CA 93106; e-mail: ponce@math.ucsb.edu
Partial differential equations and dynamical systems, to PETER POLACIK, School of Math-
ematics, University of Minnesota, Minneapolis, MN 55455; e-mail: polacik@math.umn.edu
Probability and statistics, to RICHARD BASS, Department of Mathematics, University of Con-
necticut, Storrs, CT 06269-3009; e-mail: bass@math.uconn.edu
Real analysis and partial differential equations, to DANIEL TATARU, Department of Mathe-
matics, University of California, Berkeley, Berkeley, CA 94720; e-mail: tataru@math.berkeley.edu
All other communications to the editors, should be addressed to the Managing Editor, ROBERT
GURALNICK, Department of Mathematics, University of Southern California, Los Angeles, CA 90089-
1113; e-mail: guralnic@math.usc.edu.
Titles in This Series
975 Javier Ribón, Topological classification of families of diffeomorphisms without small
divisors, 2010
974 Pascal Lefèvre, Daniel Li, Hervé Queffélec, and Luis Rodrı́guez-Piazza,
Composition operators on Hardy-Orlicz spaces, 2010
973 Peter O’Sullivan, The generalised Jacobson-Morosov theorem, 2010
972 Patrick Iglesias-Zemmour, The moment maps in diffeology, 2010
971 Mark D. Hamilton, Locally toric manifolds and singular Bohr-Sommerfeld leaves, 2010
970 Klaus Thomsen, C ∗ -algebras of homoclinic and heteroclinic structure in expansive
dynamics, 2010
969 Makoto Sakai, Small modifications of quadrature domains, 2010
968 L. Nguyen Van Thé, Structural Ramsey theory of metric spaces and topological
dynamics of isometry groups, 2010
967 Zeng Lian and Kening Lu, Lyapunov exponents and invariant manifolds for random
dynamical systems in a Banach space, 2010
966 H. G. Dales, A. T.-M. Lau, and D. Strauss, Banach algebras on semigroups and on
their compactifications, 2010
965 Michael Lacey and Xiaochun Li, On a conjecture of E. M. Stein on the Hilbert
transform on vector fields, 2010
964 Gelu Popescu, Operator theory on noncommutative domains, 2010
963 Huaxin Lin, Approximate homotopy of homomorphisms from C(X) into a simple
C ∗ -algebra, 2010
962 Adam Coffman, Unfolding CR singularities, 2010
961 Marco Bramanti, Luca Brandolini, Ermanno Lanconelli, and Francesco
Uguzzoni, Non-divergence equations structured on Hörmander vector fields: Heat kernels
and Harnack inequalities, 2010
960 Olivier Alvarez and Martino Bardi, Ergodicity, stabilization, and singular
perturbations for Bellman-Isaacs equations, 2010
959 Alvaro Pelayo, Symplectic actions of 2-tori on 4-manifolds, 2010
958 Mark Behrens and Tyler Lawson, Topological automorphic forms, 2010
957 Ping-Shun Chan, Invariant representations of GSp(2) under tensor product with a
quadratic character, 2010
956 Richard Montgomery and Michail Zhitomirskii, Points and curves in the Monster
tower, 2010
955 Martin R. Bridson and Daniel Groves, The quadratic isoperimetric inequality for
mapping tori of free group automorphisms, 2010
954 Volker Mayer and Mariusz Urbański, Thermodynamical formalism and multifractal
analysis for meromorphic functions of finite order, 2010
953 Marius Junge and Javier Parcet, Mixed-norm inequalities and operator space Lp
embedding theory, 2010
952 Martin W. Liebeck, Cheryl E. Praeger, and Jan Saxl, Regular subgroups of
primitive permutation groups, 2010
951 Pierre Magal and Shigui Ruan, Center manifolds for semilinear equations with
non-dense domain and applications to Hopf bifurcation in age structured models, 2009
950 Cédric Villani, Hypocoercivity, 2009
949 Drew Armstrong, Generalized noncrossing partitions and combinatorics of Coxeter
groups, 2009
948 Nan-Kuo Ho and Chiu-Chu Melissa Liu, Yang-Mills connections on orientable and
nonorientable surfaces, 2009
947 W. Turner, Rock blocks, 2009
946 Jay Jorgenson and Serge Lang, Heat Eisenstein series on SLn (C), 2009
945 Tobias H. Jäger, The creation of strange non-chaotic attractors in non-smooth
saddle-node bifurcations, 2009
TITLES IN THIS SERIES

944 Yuri Kifer, Large deviations and adiabatic transitions for dynamical systems and Markov
processes in fully coupled averaging, 2009

943 István Berkes and Michel Weber, On the convergence of ck f (nk x), 2009
942 Dirk Kussin, Noncommutative curves of genus zero: Related to finite dimensional
algebras, 2009
941 Gelu Popescu, Unitary invariants in multivariable operator theory, 2009
940 Gérard Iooss and Pavel I. Plotnikov, Small divisor problem in the theory of
three-dimensional water gravity waves, 2009
939 I. D. Suprunenko, The minimal polynomials of unipotent elements in irreducible
representations of the classical groups in odd characteristic, 2009
938 Antonino Morassi and Edi Rosset, Uniqueness and stability in determining a rigid
inclusion in an elastic body, 2009
937 Skip Garibaldi, Cohomological invariants: Exceptional groups and spin groups, 2009
936 André Martinez and Vania Sordoni, Twisted pseudodifferential calculus and
application to the quantum evolution of molecules, 2009
935 Mihai Ciucu, The scaling limit of the correlation of holes on the triangular lattice with
periodic boundary conditions, 2009
934 Arjen Doelman, Björn Sandstede, Arnd Scheel, and Guido Schneider, The
dynamics of modulated wave trains, 2009
933 Luchezar Stoyanov, Scattering resonances for several small convex bodies and the
Lax-Phillips conjuecture, 2009
932 Jun Kigami, Volume doubling measures and heat kernel estimates of self-similar sets,
2009
931 Robert C. Dalang and Marta Sanz-Solé, Hölder-Sobolv regularity of the solution to
the stochastic wave equation in dimension three, 2009
930 Volkmar Liebscher, Random sets and invariants for (type II) continuous tensor product
systems of Hilbert spaces, 2009
929 Richard F. Bass, Xia Chen, and Jay Rosen, Moderate deviations for the range of
planar random walks, 2009
928 Ulrich Bunke, Index theory, eta forms, and Deligne cohomology, 2009
927 N. Chernov and D. Dolgopyat, Brownian Brownian motion-I, 2009
926 Riccardo Benedetti and Francesco Bonsante, Canonical wick rotations in
3-dimensional gravity, 2009
925 Sergey Zelik and Alexander Mielke, Multi-pulse evolution and space-time chaos in
dissipative systems, 2009
924 Pierre-Emmanuel Caprace, “Abstract” homomorphisms of split Kac-Moody groups,
2009
923 Michael Jöllenbeck and Volkmar Welker, Minimal resolutions via algebraic discrete
Morse theory, 2009
922 Ph. Barbe and W. P. McCormick, Asymptotic expansions for infinite weighted
convolutions of heavy tail distributions and applications, 2009
921 Thomas Lehmkuhl, Compactification of the Drinfeld modular surfaces, 2009
920 Georgia Benkart, Thomas Gregory, and Alexander Premet, The recognition
theorem for graded Lie algebras in prime characteristic, 2009
919 Roelof W. Bruggeman and Roberto J. Miatello, Sum formula for SL2 over a totally
real number field, 2009
918 Jonathan Brundan and Alexander Kleshchev, Representations of shifted Yangians
and finite W -algebras, 2008

For a complete list of titles in this series, visit the
AMS Bookstore at www.ams.org/bookstore/.
ISBN 978-0-8218-4637-7

9 780821 846377
MEMO/207/974