Volume 2009, Article ID 832686,14pages doi:10.1155/2009/832686
Research Article
Composition Operator on Bergman-Orlicz Space
Zhijie Jiang
1, 2and Guangfu Cao
11College of Mathematics and Information Science, Guangzhou University,
Guangzhou, Guangdong, 510006, China
2Department of Mathematics, Sichuan University of Science and Engineering,
Zigong, Sichuan 643000, China
Correspondence should be addressed to Zhijie Jiang,[email protected]
Received 19 May 2009; Accepted 22 October 2009
Recommended by Shusen Ding
LetDdenote the open unit disk in the complex plane and letdAzdenote the normalized area measure onD. Forα >−1 andΦa twice differentiable, nonconstant, nondecreasing, nonnegative, and convex function on0,∞, the Bergman-Orlicz spaceLΦα is defined as followsLΦα {f ∈
HD:DΦlog|fz|1− |z|2αdAz<∞}.Letϕbe an analytic self-map ofD. The composition operatorCϕ induced by ϕ is defined byCϕf f◦ϕ forf analytic inD. We prove that the
composition operatorCϕis compact onLΦα if and only ifCϕis compact onA2α, andCϕhas closed
range onLΦα if and only ifCϕhas closed range onA2α.
Copyrightq2009 Z. Jiang and G. Cao. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
1. Introduction
LetDbe the open unit disk in the complex plane and letϕbe an analytic self-map ofD. The
composition operatorCϕinduced byϕis defined byCϕff◦ϕforfanalytic inD. The idea
of studying the general properties of composition operators originated from Nordgren1 .
As a sequence of Littlewood’s subordinate theorem, eachϕinduces a bounded composition
operator on the Hardy spacesHpDfor allp0< p <∞and the weighted Bergman spaces
ApαDfor allp0 < p < ∞and for allα−1 < α <∞. Thus, boundedness of composition
operators on these spaces becomes very clear. Nextly, a natural problem is how to characterize the compactness of composition operators on these spaces, which once was a central problem for mathematicians who were interested in the theory of composition operators. The study of compact composition operators was started by Schwartz, who obtained the first compactness
theorem in his thesis2 , showing that the integrability of1− |ϕ|−1 over∂D implied the
compactness ofCϕ onHp. The work was continued by Shapiro and Taylor3 , who showed
∂D. Moreover, MacCluer and Shapiro4 pointed out that nonexistence of the finite angular
derivatives ofϕwas a sufficient condition for the compactness ofCϕ onApαbut it failed on
Hp. So looking for an appropriate tool of characterizing the compactness ofC
ϕ onHpwas
difficult at that time. Fortunately, Shapiro5 developed relations between the essential norm
ofCϕonH2and the Nevanlinna counting function ofϕ, and he obtained a nice essential norm
formula ofCϕin 1987. As a result, he completely gave a characterization of the compactness
ofCϕin terms of the function properties ofϕ.
Another solution to the compactness of Cϕ on H2 was done by the Aleksandrov
measures which was introduced by Cima and Matheson 6 . It is well known that the
harmonic functionRλϕz/λ−ϕzcan be expressed by the Possion integral
Rλϕz
λ−ϕz
∂D
Pz, ζdmλζ 1.1
for eachλ∈∂D. Cima and Matheson appliedσλthe singular part ofmλto give the following
expression:
Cϕ2
esup
λ∈∂Dσλ. 1.2
They showed thatCϕwas compact onH2if and only if all the measuresmλwere absolutely
continuous.
The study of compactness of composition operators is also an important subject on other analytic function spaces, and we have chosen two typical examples above, and for more
related materials one can consult7,8 . Another natural interesting subject is the composition
operator with closed range. Considering angular derivatives of ϕ, it is known that Cϕ is
compact onA2if and only ifϕfails to have finite angular derivatives on∂D, in this case,C
ϕ
does not have closed range sinceCϕis not a finite rank operator. And ifϕhas finite angular
derivatives on∂D, thenϕis necessarily a finite Blaschke product and hence one can easily
verify thatCϕhas closed range onA2. Zorboska has given a necessary and sufficient condition
forCϕwith closed range onH2, and she also has done onApα9 . Luecking10 considered
the same question on Dirichlet space after Zorboska’s work. Recently, Kumar and Partington
11 have studied the weighted composition operators with closed range on Hardy spaces
and Bergman spaces.
This paper will study the compactness of composition operator on Bergman-Orlicz space. We are mainly inspired by the following results.
iLiu et al. 12 showed that composition operator was bounded on Hardy-Orlicz
space. Lu and Cao13 also showed that composition operator was bounded on
Bergman-Orlicz space.
iiA composition operator was compact on the Nevanlinna classNif and only if it
was compact onH214 .
iiiIf a composition operator was compact onHpfor somep >0, then it was compact
on Hp for all p > 0 3 . Moreover, paper 15 compared the compactness of
CϕonA2αand on the Bergman-Orlicz space, and whether there is a equivalence for
the closed range of Cϕ onA2αand on the Bergman-Orlicz space. In this paper, we
are going to give affirmative answers for the proceeding questions.
2. Preliminaries
LetHDdenote the space of all analytic functions onD. LetdAzdenote the normalized
area measure onD, that is,AD 1. LetS denote the class of strongly convex functions
Φ:0,∞ → 0,∞, which satisfies
i Φ0 Φ0 0,Φt/t → ∞ast → ∞,
ii Φexists on0,∞,
iii Φ2t≤CΦtfor some positive constantCand for allt >0.
ForΦ∈ Sandα >−1 the Bergman-Orlicz spaceLΦα is defined as follows:
LΦα
f∈HD:fΦ
DΦ
logfz 1− |z|2αdAz<∞
, 2.1
where logx max{0,logx}. Although · Φ does not define a norm in LΦα, it holds that
thedf, g f−gΦdefines a metric on LΦα, and makesLΦα into a complete metric space.
Obviously, the inequalities
logx≤log1x≤1logx, x≥0,
2logx≤log 1x2≤12logx, x≥0,
2.2
and the fact thatΦis nondecreasing convex function imply that
Φlogx≤Φlog1x≤Φ1logx
≤ 1
2Φ2
1
2Φ
2 logx
≤ 1
2Φ2
1
2CΦ
logx,
Φlogx≤Φ2 logx≤Φ log 1x2
≤Φ12 logx
≤ 1
2Φ2
1
2Φ
4 logx
≤ 1
2Φ2
1
2CΦ
logx.
Thenf ∈LΦα if and only if
DΦ
log1fz 1− |z|2αdAz<∞ 2.4
or if and only if
DΦ log 1 fz2
1− |z|2αdAz<∞. 2.5
Throughout this paper, constants are denoted byC, they are positive and may differ
from one occurrence to the other. The notationabmeans that there is a positive constant
Csuch thata≤Cb. Moreover, if bothabandbahold, we writeaband say thatais
asymptotically equivalent tob.
In this section we will prove several auxiliary results which will be used in the proofs of the main results in this paper.
Lemma 2.1. Iff ∈LΦα, then
fΦΦ log 1f0|2
Df
Φz 1− |z|2αdAz, 2.6
whereΔis Laplacian andfΦz ΔΦlog1|fz|2.
Proof. By the Green Theorem, ifu, v∈C2Ω, whereΩis a domain in the plane with smooth
boundary, then
ΩuΔv−vΔudx dy
∂Ω
u∂v ∂n−v
∂u ∂n
ds. 2.7
Let 0< ε < r <1,uz logr/|z|,vz Φlog1|fz|2, andΩ {z∈D:ε <|z|< r}.
SinceΔuz 0, by2.7we have
ΩΔΦ log 1
fz2
log r
|z|dx dylog r ε
|z|ε
∂
∂nΦ log 1fz
2
ds
|z|r
Φ log 1fz2
r ds−
|z|ε
Φ log 1fz2
ε ds.
2.8
Since∂/∂nΦlog1|fz|2is bounded near to 0, we get
lim
ε→0log
r ε
|z|ε
∂
∂nΦ log 1fz
2
Letε → 0 in2.8, we have
|z|<r
ΔΦ log 1fz2log r
|z|dx dy
2π
0 Φ
log
1f reiθ2
dθ−2πΦ log 1f02.
2.10
Integrating equality2.10with respect torfrom 0 to 1, we obtain
1
0
|z|<r
ΔΦ log 1fz2log r
|z| 1−r
2αrdr dAzf
Φ−
2π
α1Φ log 1f0
2
.
2.11
Thus
f
Φ
2π
α1Φ log 1f0
21
0
|z|<r
ΔΦ log 1fz2log r
|z| 1−r
2αrdr dAz
2π
α1Φ log 1f0
2
DΔΦ log 1
fz2dAz
1
|z|
log r
|z| 1−r
2αrdr.
2.12
Since1|z|logr/|z|1−r2αrdr 1−z|22α,
fΦ 2π
α1Φ log 1f0
2
DΔΦ log 1
fz2 1− |z|2αdAz
Φ log 1f02
Df
Φ 1− |z|2αdAz,
2.13
the proof is complete.
Letϕbe an analytic self-map ofD. The generalized Nevanlinna counting function ofϕ
is defined by
Nϕ,α2w
z∈ϕ−1w
log 1
|z|
α2
Lemma 2.2see9 . Ifϕis an analytic self-map ofDand g is a nonnegative measurable function in
D, then
Dg◦ϕz ϕz2
log 1
|z|
α2
dAz
DgzNϕ,α2zdAz.
2.15
Lemmas2.1and2.2see9 can lead to the following corollary.
Corollary 2.3. Letϕbe an analytic self-map ofDandf ∈HD, then
f◦ϕΦ 1logf◦ϕ02
Df
ΦwN
α2wdAw. 2.16
We will end this section with the following lemma, which illustrates that the counting
functionalδz:f→fzis continuous onLΦα.
Lemma 2.4. Letf ∈LΦα, then
fz≤exp ⎛ ⎜ ⎝Φ−1
⎛ ⎜
⎝ CfΦ
1− |z|2α2
⎞ ⎟ ⎠
⎞ ⎟
⎠ ∀ zinD. 2.17
Proof. By the subharmonicity of mapz→log1|fz|, we get
log1fz≤ 1
AαDz,1− |z|/2
Dz,1−|z|/2log
1fz 1−z|2αdAz. 2.18
SinceΦlog1|fz|is convex and increasing, we have
Φlog1fz≤ 1
AαDz,1− |z|/2
Dz,1−|z|/2Φ
log1fzdAαz
≤ C
1− |z|2α2
Dz,1−|z|/2Φ
log1fzdAαz
≤ C
1− |z|2α2
DΦ
log1fzdAαz
≤ CfΦ
1− |z|2α2 .
SinceΦlogx≤Φlog1x, we get
Φlogfz≤ CfΦ
1− |z|2α2
, 2.20
that is, log|fz| ≤ Φ−1Cf
Φ/1− |z|2α2. Thus|fz| ≤ expΦ−1CfΦ/1− |z|2α2.
3. Compactness
In this section, we are going to investigate the equivalence between compactness of
composition operator on the Bergman-Orlicz spaceLΦα and on the weighted Bergman space
A2
α. The following lemma characterizes the compactness ofCϕ onLΦα in terms of sequential
convergence, whose proof is similar to that in7, Proposition 3.11 .
Lemma 3.1. Letϕbe an analytic self-map ofD, bounded operatorCϕis compact onLΦα if and only if
whenever{fn}is bounded inLΦα andfn → 0uniformly on compact subsets ofD, thenCϕfnΦ → 0
asn → ∞.
In order to characterize the compactness ofCϕ, we need to introduce the notion of
Carleson measure. For|ξ| 1 andδ >0 we defineQδξ {z∈ D:|z−ξ|< δ}. A positive
Borel measureμonDis called a Carleson measure if sup|ξ|1μQδξ Oδα2. Moreover, if
μsatisfies the additional condition limδ→0μQδξ/δα20,μis called a vanishing Carleson
measuresee16 for the further information of Carleson measure. The following result for
the compactness ofCϕonA2αis useful in the proof ofTheorem 3.3.
Lemma 3.2see14,17 . Letϕ be an analytic self-map ofD. Then the following statements are equivalent:
iCϕis compact onA2α,iilim|z| →1Nϕ,α2z/1− |z|2 α2
0, andiiithe pull measure
μϕis a vanshing Carleson measure onD.
Theorem 3.3. Letϕ be an analytic self-map ofD, thenCϕ is compact onA2α if and only ifCϕ is
compact onLΦα.
Proof. First we assume thatCϕ is compact onA2α. Choose a sequence{fn}that is bounded
by a positive constantMinLΦα and converges to zero uniformly on compact subsets ofD.
ByLemma 3.1, it is enough to show thatfn◦ϕΦ → 0 asn → ∞. Letε > 0, we can find
0 < r < 1 such thatNϕ,α2z < ε1− |z|2α2 for all|z| > r.Since fn → 0 uniformly on
compact subsets ofDasn → ∞, so isfn. Thus we can chooseN > 0 such that|fn| < εand
|fn|< εonrD, whenevern > N. Hence for suchnwe have
Cϕfn
Φ≤Φ log 1fn◦ϕ02
Df
Φ
As|fn◦ϕ0| → 0 asn → ∞andΦlog1|fn◦ϕ0|2 → 0 asn → ∞, we only need to
verify thatDfnΦNϕ,α2dAw → 0 asn → ∞. Now
Df
Φ
nNϕ,α2dAw
rDf
Φ
nNϕ,α2dAw
D\rDf
Φ
nNϕ,α2dAw III. 3.2
We first prove that the first term in previous equality is bounded by a constant multiple ofε
I
rDf
Φ
nNϕ,α2dAw
rD
Φ log 1f nw2
fnw2 Φ log 1fnw2
× fnw
2
1fnw2
2Nϕ,α2wdAw
≤Φlog1εε Φlog1εε rD
Nϕ,α2wdAw
≤Φlog1εε Φlog1εε.
3.3
Now, we show that the previous second term above is also bounded by a constant multiple ofε
II
D\rDf
Φ
nNϕ,α2dAw≤εfnΦw 1− |w|2
α
≤εfnΦ≤Mε. 3.4
Conversely, we assume thatCϕ is compact onLΦα. ByLemma 3.2, we need to verify
thatμϕ is a vanishing Carleson measure. For 0 < δ < 1 andξ ∈ ∂Dwe write a 1−δξ
andgaz 1− |a|2 α2
/1−az2α4. Then|ga| ∈L1D, dA
α. PutGz Φ|gaz|.Gis well
defined, beacuseGis nondecreasing on range of|ga|. SinceΦ−1is concave, there is a constant
C > 0 such that Φ−1t ≤ Ctfor enough bigt. Thus we getGz ∈ L1D, dA
α. Sethz
exp20πeitz/eit−zGeitdt. SinceΦlog|hz| ΦGz |g
az| ∈ L1D, dAα, it
means thath ∈ LΦα. Let faz 1− |a|hz. Then clearlyfa → 0 uniformly on compact
subsets ofDas|a| → 1. Moreover,
fa
Φ
DΦ
logfazdAαz≤
DΦ
log|hz|dAαz
D
1− |a|2α2
|1−az|2α4 dAαz 1.
On the other hand, if|1−zζ|/1− |a|< γfor some fixed 0< γ <1/4, whereζa/|a|, that is,
z∈Qγδζ, we have
1− |a|2
|1−az|2
1− |a|2
1− |a|2
1− |a|2
|1−az|2
1− |a|2
1− |a|2
⎛ ⎜
⎝1|a| 1−zζ
1− |a|
⎞ ⎟ ⎠ −2
≥ 1
1γ2
1− |a|2
1− |a|2 >
1− |a|2
41− |a|2 ≥
1 4δ.
3.6
Hence, forz∈Qγδζwe have
Φ−1
1− |a|2
|1−az|2
α2
≥Φ−1
1 4δ
α2
. 3.7
Thus, forz∈Qγδζwe obtain
Φlogfaz Φ
log1− |a|hz Φlog1− |a| log|hz|
≥Φlog|hz|gaz
1− |a|2
|1−az|2
α2
≥
1 4δ
α2
.
3.8
So, for allξ∈∂Dand 0< δ <1 we get
1 4δ
α2
μϕ
Qγδξ
≤
QγδξΦ
logfazdμϕ ≤
DΦ
logfazdμϕ
DΦ
logfa◦ϕzdAαz CϕfaΦ.
3.9
For the compactness of Cϕ, we know that CϕfaΦ → 0 as |a| → 1, which means that
limδ→1μϕQγδξ/δα2 0 uniformly for ξ ∈ ∂D. This means that μϕ is a vanishing
Carleson measure. ByLemma 3.2,Cϕis compact onA2α.
For special caseΦt tpp > 1, the Bergman-Orlicz spaceLΦ
α is called the area-type
Nevanlinna class and we writeNpα.
Corollary 3.4. Letϕ be an analytic self-map ofD, thenCϕ is compact onA2α if and only if Cϕ is
compact onNpα.
Remark 3.5. Theorem 3.3may be not true ifΦdoes not satisfy the given conditions in this
Φis nondecreasing butΦx>0 for somex /0. Then the compactness ofCϕon the Bergman
spaceA2i.e.,α0is different from that onLΦ
α. HereLΦα is defined as follows:
LΦα
f∈HD:∃t >0, s.t.
DΦ
logtfzdAz<∞
. 3.10
If we takeΦx 0 forx ≤ 1, and Φx ∞forx > 1, thenLΦα isH∞D. We know that
Cϕ is compact onH∞Dif and only ifϕ∞ < 1 consult2 . But MacCluer and Shapiro
constructed an inner functionϕin4 such thatCϕwas compact onA2.
4. Closed Range
In this section we will develop a relatively tractable if and only if condition for the
composition operator onLΦα with closed range. Considering that any analytic automorphism
ofDhas the formϕaz cz−a/1−az, where|c| 1 anda∈D. By13 , we have the
following lemma.
Lemma 4.1. If one ofCϕ,Cϕ◦ϕa,Cϕa◦ϕhas closed range onLΦα, so have the other two.
Now thatLΦα,0{f ∈LΦα :f0 0}is a closed subspace ofLΦα and dimLΦα/LΦα,0 1,
the following lemma is easily proved.
Lemma 4.2. Letϕbe an analytic self-map ofD, thenCϕhas closed range onLΦα if and only ifCϕhas
closed range onLΦα,0.
Recall that the pseudohyperbolic metricρz, w,z, w∈Dis given by
ρz, w w−z
1−wz
. 4.1
For z ∈ D and 0 < r < 1 we define Dz, r {w ∈ D : ρz, w < r}. Forε > 0 we put
Ωε {z ∈ D : 1 − |z|2/1− |ϕz|2 ≥ ε}and Gε ϕΩε. We say that Gε satisfies the
Φ-reverse Carleson measure condition if there exists a positive constantηsuch that
Gε
Φ 1logfz2 1− |z|2α2dAz≥η
DΦ 1log
fz2 1− |z|2α2dAz,
4.2
wherefis analytic inDandDΦ1log|fz|21− |z|2α2dAz<∞.
Theorem 4.3. Letϕbe an analytic self-map ofD. ThenCϕhas closed range onLΦα if and only if there
Proof. We first assume that there existsε > 0 such thatGε satisfies theΦ-reverse Carleson
measure condition. Iff∈LΦα, then
f◦ϕ
ΦΦ log 1f◦ϕ02
D
f◦ϕΦz 1− |z|2α2dAz
≥Φ log 1f◦ϕ02
Ωε
f◦ϕΦz 1− |z|2α2dAz
≥εα2
Φ log 1f◦ϕ02
Ωε
f◦ϕΦz 1−ϕz2α2dAz
εα2
n
Ωε∩Rn
f◦ϕΦz 1−ϕz2α2dAz εα2Φ log 1f◦ϕ02,
4.3
whereZis the zero point set ofϕand{Rn}is a partition ofD\Zinto at most countably many
semiclosed polar rectangles such thatϕis univalent on eachRn. LetSnϕRn∩Ωε. Then by
the change of variables involvingϕ, the last line above becomes
εα2
n
Gε
fΦ 1− |w|2α2χSnwdAz εα2Φ log 1f◦ϕ02
≥εα2
Gε
fΦ 1− |w|2α2dAz εα2Φ log 1f◦ϕ02
≥ηεα2
Df
Φ 1− |w|2α2dAz ηεα2Φ log 1f◦ϕ02
.
4.4
So we show thatCϕhas closed range onLΦα.
Conversely, byLemma 4.2, we need to prove thatCϕhas closed range onLΦα,0. Suppose
that there does not existε >0 such that GεsatisfiesΦ-reverse Carleson measure condition.
We can choose a sequence{fn}inLΦα,0such thatDΦ1log|fnz|2dAα2z 1 for allnand
yetGnfΦ1−w|2α2dA → 0 asn → ∞, whereGn ϕΩnandΩn {z∈D: 1− |ϕz|2 ≤
n1− |z|2}. Now
fn◦ϕΦΦ log 1fn◦ϕ02
D
fn◦ϕ
ΦzdA α2z
Φ log 1fn◦ϕ02
Ωn
fn◦ϕ
Φ
zdAα2z
D\Ωn
fn◦ϕ
ΦzdA α2z.
Sinceϕis an analytic self-map ofD. The Nevanlinna counting functionNϕ,α2satisfies
Nϕ,α2z O
log 1
|z|
α2
4.6
as|z| → 1. Using4.6and decompositions of the disk into polar rectangles8 , one can find
a positive constantc1such that
Ωn
fn◦ϕ
Φ
dAα2z
Ωn
fnΦNϕ,α2dAz≤c1
Gn
fnΦz 1− |z|2α2dAz−→0 4.7
asn → ∞, and
D\Ωn
fn◦ϕ
Φ
dAα2z≤ 1
nα2
D\Ωn
fn◦ϕ
Φ
1−ϕz2α2dAz
≤ 1
nα2
Df
Φ
nNϕ,α2dAz
≤ 1
nα2fnΦ−→0
4.8
asn → ∞. Evidently,fn◦ϕΦ → 0 asn → ∞, thoughfnΦ1 for alln. It follows thatCϕ
does not have closed range onLΦα.
We have offered a criterion for the composition operator with closed range onLΦα, but
it seems that it is difficult to check whether or notGεsatisfies theΦ-reverse Carleson measure
condition.
Theorem 4.4. The composition operatorCϕ has closed range onLΦα if and only if there are positive
constantsε, c, andrsuch thatAαGε∩Dz, r≥c|Dz, r|α2for allz∈D.
Proof. We first assume thatCϕ has closed range onLΦα. Then there is a constant ε > 0 such
thatGε satisfies theΦ-reverse Carleson measure condition. Thus, applying the proceeding
constructed functionhzto theΦ-reverse Carleson condition gives
Gε
1− |a|2
|1−az|2
α2
dAαz≥η
D
1− |a|2
|1−az|2
α2
dAαz η. 4.9
SinceD1− |z|2αdAz<∞, it allows to choose a fixed constantr >0 such that
D\Db,r 1− |z|
2αdAz≤ 1
2C
D 1− |z|
2αdAz.
Changingw b−a 1−abz/1−ba b−azin4.10gives
D\Da,r
1− |a|2
|1−az|2
α2
dAαz≤
1 2C
D
1− |a|2
|1−az|2
α2
dAαz. 4.11
Combing4.11gives
Gε∩Da,r
1− |a|2
|1−az|2
α2
dAαz≥ 1
2C
D 1− |z|
2αdAz. 4.12
The integral in the left of4.12is dominated by
AαGε∩Da, rsup
⎧ ⎨ ⎩
1− |a|2
|1−az|2 α2
:z∈Da, r
⎫ ⎬
⎭. 4.13
Since1−a|2/|1−az|21/|Da, r|forz∈Da, r, we get
AαGε∩Da, r |Da, r|α2 ≥
1 2C
D 1− |z|
2αdAz.
4.14
The converse can be derived from modification of18 , so we omit it here.
Remark 4.5. From 18 , we find that the composition operator has closed range on the
weighted Bergman space Apαif and only if there are positive constants ε, cand r such that
AαGε∩Dz, r≥c|Dz, r|α2for allz∈D. Thus, we have the following fact.
The composition operatorCϕhas closed range onLΦα if and only ifCϕhas closed range
onApα.
Let us further investigate the Φ-reverse Carleson measure condition, which can be
formulated as follows.
The space{f|Gε :f∈LΦα}is a closed subspace ofLΦα if and only if there exists a constant
ε >0 such thatGεsatisfiesΦ-reverse Carleson measure condition.
From the perspective of closed subspace, we will see the following special setting. Let
F {zn : n 1,2, . . .} be aδ-sequence inD. That is, there isk withρz, zk < δfor every
z∈D. We also assume thatFisγseparated for some fixedγ >0, that is,ρzm, zn≥γfor all
m /n. Using the subharmonicity of log|fz|for analytic functionfz, it is easy to see that
logfzk≤ C
Aα
Dzk, γ/2
Dzk,γ/2
logfzdAαz. 4.15
SinceΦlog|fz|is convex and increasing, we have
Φlogfzk≤
C Aα
Dzk, γ/2
Dzk,γ/2
Moreover, the formulaAαDzk, γ/21− |zk|2 α2
allows us to write
Φlogfzk≤C 1− |zk|2
−α−2
Dzk,γ/2
ΦlogfzdAαz. 4.17
SinceDzk, γ/2are disjoint, we obtain
k
Φlogfzk 1− |zk|2
α2
≤
DΦ
logfzdAαz. 4.18
Hence, the mapσ:f →f|FtakesLΦα intoLΦF, μ, whereμis a measure onFthat assignszk
to the mass1− |zk|2α2and spaceLΦF, μ {f:F → C|FΦlog|f|dμ <∞}. Of course,
the mapσmay be one to one. If the mapσis one to one, the mapσhas closed range if and
only if
DΦ
logfzdAαz≤
FΦ
logfdμ
k
Φlogfzk 1− |zk|2
α2
. 4.19
Acknowledgments
The authors are extremely thankful to the editor for pointing out several errors. This work
was supported by the Science Foundation of Sichuan Province no. 20072A04 and the
Scientific Research Fund of School of Science SUSE.
References
1 E. A. Nordgren, “Composition operators,”Canadian Journal of Mathematics, vol. 20, pp. 442–449, 1968. 2 H. J. Schwartz,Composition operators onHp, Ph.D. thesis, University of Toledo, Toledo, Ohio, USA,
1968.
3 J. H. Shapiro and P. D. Taylor, “Compact, nuclear, and Hilbert-Schmidt composition operators onH2,”
Indiana University Mathematics Journal, vol. 23, pp. 471–496, 1973.
4 B. D. MacCluer and J. H. Shapiro, “Angular derivatives and compact composition operators on the Hardy and Bergman spaces,”Canadian Journal of Mathematics, vol. 38, no. 4, pp. 878–906, 1986. 5 J. H. Shapiro, “The essential norm of a composition operator,”Annals of Mathematics, vol. 125, no. 2,
pp. 375–404, 1987.
6 J. A. Cima and A. L. Matheson, “Essential norms of composition operators and Aleksandrov measures,”Pacific Journal of Mathematics, vol. 179, no. 1, pp. 59–64, 1997.
7 C. C. Cowen and B. D. MacCluer,Composition Operators on Spaces of Analytic Functions, Studies in Advanced Mathematics, CRC Press, Boca Raton, Fla, USA, 1995.
8 J. H. Shapiro,Composition Operators and Classical Function Theory, Universitext: Tracts in Mathematics, Springer, New York, NY, USA, 1993.
9 K. H. Zhu,Operator Theory in Function Spaces, vol. 139 ofMonographs and Textbooks in Pure and Applied Mathematics, Marcel Dekker, New York, NY, USA, 1990.
10 D. H. Luecking, “Bounded composition operators with closed range on the Dirichlet space,”
Proceedings of the American Mathematical Society, vol. 128, no. 4, pp. 1109–1116, 2000.
11 R. Kumar and J. R. Partington, “Weighted composition operators on Hardy and Bergman spaces,” in
Recent Advances in Operator Theory, Operator Algebras, and Their Applications, vol. 153 ofOperator Theory Advance and Application, pp. 157–167, Birkh¨auser, Basel, Switzerland, 2005.
13 Q. Lu and G. F. Cao, “Weighted Orlicz-Bergman spaces and their composition operators,” Acta Analysis Functionalis Applicata, vol. 7, no. 4, pp. 366–369, 2005Chinese.
14 J. S. Choa and H. O. Kim, “Compact composition operators on the Nevanlinna class,”Proceedings of the American Mathematical Society, vol. 125, no. 1, pp. 145–151, 1997.
15 P. Lef`evre, D. Li, H. Queff`elec, and L. Rodr´ıguez-Piazza, “Compact composition operators onH2and
Hardy-Orlicz spaces,”Journal of Mathematical Analysis and Applications, vol. 354, no. 1, pp. 360–371, 2009.
16 J. B. Garnett,Bounded Analytic Functions, vol. 96 ofPure and Applied Mathematics, Academic Press, New York, NY, USA, 1981.
17 J. S. Choa and S. Ohno, “Products of composition and analytic Toeplitz operators,” Journal of Mathematical Analysis and Applications, vol. 281, no. 1, pp. 320–331, 2003.