Accepted Manuscript

Some properties of a class of elliptic partial differential operators

Shaolin Chen, Matti Vuorinen

PII: S0022-247X(15)00580-6
DOI: http://dx.doi.org/10.1016/j.jmaa.2015.06.026
Reference: YJMAA 19579

To appear in: Journal of Mathematical Analysis and Applications

Received date: 12 April 2015

Please cite this article in press as: S. Chen, M. Vuorinen, Some properties of a class of elliptic partial differential operators, J.
Math. Anal. Appl. (2015), http://dx.doi.org/10.1016/j.jmaa.2015.06.026

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are
providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting
proof before it is published in its final form. Please note that during the production process errors may be discovered which could
affect the content, and all legal disclaimers that apply to the journal pertain.
 SOME PROPERTIES OF A CLASS OF ELLIPTIC PARTIAL
DIFFERENTIAL OPERATORS

SHAOLIN CHEN AND MATTI VUORINEN

Abstract. We prove Schwarz-Pick type estimates and coefficient estimates for

a class of functions induced by the elliptic partial differential operators. Then we
apply these results to obtain a Landau type theorem.

1. Introduction and main results

Let C be the complex plane. For a ∈ C, let r > 0 and D(a, r) = {z : |z − a| < r}.
In particular, we use Dr to denote the disk D(0, r) and D, the open unit disk D1 .
For a real 2 × 2 matrix, we will consider the matrix norm A = sup{|Az| : |z| =
1} and the matrix function l(A) = inf{|Az| : |z| = 1}. For z = x + iy ∈ C with x
and y real, we denote the complex differential operators
   
∂ 1 ∂ ∂ ∂ 1 ∂ ∂
= −i and = +i .
∂z 2 ∂x ∂y ∂z 2 ∂x ∂y
If we denote the formal derivative of f = u + iv by
 
ux uy
Df = ,
vx vy
 
then Df  = |fz | + |fz | and l(Df ) = |fz | − |fz |, where u, v are real functions,
fz = ∂f /∂z and fz = ∂f /∂z. Throughout this paper, we denote by C n (D) the
set of all n-times continuously differentiable complex-valued functions in D, where
n ∈ {1, 2, · · · }.
For α ∈ R and z ∈ D, let
 
α2 α ∂ ∂ 1
Tα = − (1 − |z|2 )−α−1 + (1 − |z|2 )−α−1 z +z + (1 − |z|2 )−α Δ
4 2 ∂z ∂z 4
be the second order elliptic partial differential operator, where Δ is the usual complex
Laplacian operator
∂2 ∂2 ∂2
Δ := 4 = + .
∂z∂z ∂x2 ∂y 2
Of particular interest to our analysis is the homogeneous partial differential equation
(1.1) Tα (f ) = 0 in D
File: YJMAA19579_source.tex, printed: 17-6-2015, 8.07
2000 Mathematics Subject Classification. Primary: 31A05; Secondary: 35J25.
Key words and phrases. Differential operator, coefficient estimate, Schwarz-Pick estimate.

1
2 Sh. Chen and M. Vuorinen

and its associated the Dirichlet boundary value problem as follows


Tα (f ) = 0 in D,
(1.2)
f = f∗ on ∂D.
Here, the boundary data f ∗ ∈ D (∂D) is a distribution on the boundary ∂D of D,
and the boundary condition in (1.2) is interpreted in the distributional sense that
fr → f ∗ in D (∂D) as r → 1−, where
(1.3) fr (eiθ ) = f (reiθ ), eiθ ∈ ∂D,
for r ∈ [0, 1). In [25], Olofsson proved that, for parameter values α > −1, if a
function f ∈ C 2 (D) satisfies (1.1) with limr→1− fr = f ∗ ∈ D (∂D), then it has the
form of a Poisson type integral
 2π
1
(1.4) f (z) = Kα (ze−iτ )f ∗ (eiτ )dτ, for z ∈ D,
2π 0
where
(1 − |z|2 )α+1
Kα (z) = cα ,
|1 − z|α+2
2 ∞
cα = Γ(α/2 + 1) /Γ(1 + α) and Γ(s) = 0 ts−1 e−t dt for s > 0 is the standard
Gamma function. In fact, by [25, Proposition 3.2], temperate growth is equivalent
to distributional boundary value for a solution of (1.1) when α > −1.
If we take α = 2(n − 1), then f is polyharmonic (or n-harmonic), where n ∈
{1, 2, . . .} (cf. [1, 2, 5, 16, 22, 23, 24, 27]). Furthermore, Borichev and Hedenmalm
[5] proved that
(1 − |z|2 )n Δn = 4(1 − |z|2 )T0 ◦ 4(1 − |z|2 )2 T2 ◦ · · · ◦ 4(1 − |z|2 )n T2(n−1) .
In particular, if α = 0, then f is harmonic (cf. [15]).
For a, b, c ∈ R with c = 0, −1, −2, . . . , the hypergeometric function is defined by
the power series
∞
(a)n (b)n xn
F (a, b; c; x) = , |x| < 1,
n=0
(c)n n!
where (a)0 = 1 and (a)n = a(a+1) · · · (a+n−1) for n = 1, 2, . . . are the Pochhammer
symbols. Obviously, for n = 0, 1, 2, . . ., (a)n = Γ(a + n)/Γ(a). In particular, for
a, b, c > 0 and a + b < c, we have (cf. [3, 4])
Γ(c)Γ(c − a − b)
(1.5) F (a, b; c; 1) = lim F (a, b; c; x) = < ∞.
x→1 Γ(c − a)Γ(c − b)
For α = 0, Heinz [17] and Colonna [14] proved the following Schwarz-Pick type
estimates on planar harmonic functions, which are the following. For the related
discussions on this topic, see [7, 9, 13, 19, 26].

Theorem A. ([17, Lemma]) Let f be a harmonic function of D into D with f (0) = 0.

Then, for z ∈ D,
4
|f (z)| ≤ arctan |z|.
π
 Differential operators 3

This estimate is sharp.

Theorem B. ([14, Theorems 3 and 4]) Let f be a harmonic function of D into D.

Then, for z ∈ D,
4 1
Df (z) ≤ .
π 1 − |z|2
This estimate is sharp, and all the extremal functions are
 
2γ 1 + ψ(z)
f (z) = arg ,
π 1 − ψ(z)
where |γ| = 1 and ψ is a conformal automorphism of D.
For α > −1, we establish the following Schwarz-Pick type estimate on the solu-
tions to (1.1).
Theorem 1. For α > −1, let f ∈ C 2 (D) satisfy (1.1) and supz∈D |f (z)| ≤ M , where
M is a positive constant. Then, for z ∈ D,
   2π 
 (1 − |z|)α+1  1 (1 − |z|) α+1
(1.6) f (z) − f (0) ≤ M −it
Kα (ze )dt − Kα (0)
1 + |z| 2π 0 1 + |z|
and

2 + α + |α||z| 2 + α + |α||z|
(1.7) Df (z) ≤ M Mα (|z|) ≤ M ,
1 − |z|2 1 − |z|2
where
  2  
1 2π
i Γ 1 + α2 α α
(1.8) Mα (r) = Kα (re )d = F − , − ; 1; r2 , r ∈ [0, 1).
2π 0 Γ(1 + α) 2 2
Let f be a harmonic mapping of D onto D with f (0) = 0. In [17], Heinz showed
that, for θ ∈ [0, 2π],
2
Df (eiθ ) ≥ .
π
For the extensive discussion on the Heinz’s inequality for real harmonic functions in
high dimension, see [18].
By using Theorem 1, we get a Heinz type inequality on ∂D as follows.
Theorem 2. For α ≥ 0, let f ∈ C 2 (D) ∩ C 1 (D) satisfying (1.1). Suppose that
f (0) = 0, f (D) = D and f (∂D) = ∂D.
(a) If α = 0, then, for θ ∈ [0, 2π],
2
Df (eiθ ) ≥ ;
π
(b) If α > 0, then, for θ ∈ [0, 2π],
d α
Df (eiθ ) ≥ lim Mα (r) = ,
r→1− dr 2
where Mα (r) is given by (1.8).
4 Sh. Chen and M. Vuorinen

The following result is the homogeneous expansion of solutions to (1.2).

Theorem C. ([25, Theorem 2.2]) Let α ∈ R and f ∈ C 2 (D). Then f satisfies (1.1)
if and only if it has a series expansion of the form
∞  α α 
(1.9) f (z) = ck F − , k − ; k + 1; |z|2 z k
k=0
2 2
∞  α α 
+ c−k F − , k − ; k + 1; |z| z k , z ∈ D,
2

k=1
2 2
for some sequence {ck }∞
k=−∞ of complex numbers satisfying
1
(1.10) lim sup |ck | |k| ≤ 1.
|k|→∞

In particular, the expansion (1.9), subject to (1.10), converges in C ∞ (D), and every
solution f of (1.1) is C ∞ -smooth in D.
For α = 0, there are numerous discussions on coefficient estimates of harmonic
mappings in the literature, see for example [8, 9, 10, 12, 15, 21, 28]. We investigate
the problem of coefficient estimates on the solutions to (1.2) as follows.
Theorem 3. For α > −1, let f ∈ C 2 (D) be a solution to (1.1) with the series
expansion of the form (1.9) and supz∈D |f (z)| ≤ M , where M is a positive constant.
Then, for k ∈ {1, 2, . . .},
  α    α  4M
 α   α 
(1.11) ck F − , k − ; k + 1; 1  + c−k F − , k − ; k + 1; 1  ≤
2 2 2 2 π
and   α α 
 
c0 F − , − ; 1; 1  ≤ M.
2 2
In particular, if α = 0, then the estimate of (1.11) is sharp and all the extreme
functions are  
2εM 1 + ϑz k
fk (z) = Im log ,
π 1 − ϑz k
where |ε| = |ϑ| = 1.
The following result easily follows from Theorem 3 and [25, Proposition 1.4].
Corollary 1.1. For α > −1, let f ∈ C 2 (D) be a solution to (1.1) with the series
expansion of the form (1.9) and supz∈D |f (z)| ≤ M , where M is a positive constant.
Then, for k ∈ {1, 2, . . .},
4M Γ 1 + α2 Γ k + 1 + α
2
|ck | + |c−k | ≤ .
k!Γ(α + 1)π
The classical theorem of Landau shows that there is a ρ = M +√1M 2 −1 such that
every function f , analytic in D with f (0) = f  (0) − 1 = 0 and |f (z)| < M in
D, is univalent in the disk Dρ and in addition, the range f (Dρ ) contains a disk
of radius M ρ2 (see [20]), where M ≥ 1 is a constant. Recently, many authors
 Differential operators 5

considered Landau type theorem for functions f satisfying (1.1) with α = 0 (see
[6, 8, 9, 10, 11]), α = 2 (see [1]) and α = 2(n − 1) (see [12]), respectively, where
n ∈ {1, 2, . . .}.
The last main aim of this paper is to establish a Landau type theorem for a more
general class of functions. Applying Theorems 1 and 3, we get the following Landau
type theorem for a class of functions f satisfying (1.1) with α ∈ (−1, 0].
Theorem 4. For α ∈ (−1, 0], let f ∈ C 2 (D) be a solution to (1.1) satisfying f (0) =
|Jf (0)| − β = 0 and supz∈D |f (z)| ≤ M , where M, β are positive constants and Jf
is the Jacobian of f . Then f is univalent in Dρ0 , where ρ0 satisfies the following
equation  
β 4M ρ0 2 − ρ0 2ρ0
− + = 0.
M (2 + α) π (1 − ρ0 )2 (1 − ρ0 )(1 − ρ20 )2
Moreover, f Dρ0 contains a univalent disk DR0 with

2ρ0 β M ρ0 (2 − ρ0 )
R0 ≥ − .
3 M (2 + α) π(1 − ρ0 )2
We remark that Theorem 4 is a generalization of [6, Theorem 2].
The proofs of Theorems 1, 2 and 3 will be presented in Section 2, and the proof
of Theorem 4 will be given in Section 3.

2. Schwarz-Pick type estimates and coefficient estimates

Proof of Theorem 1. We first prove (1.6). By [25, Proposition 3.2], we know that
the temperate growth is equivalent to distributional boundary value for a solution of
(1.1) when α > −1. Without loss of generality, we assume that fr → f ∗ in D (∂D)
as r → 1−, where fr is given by (1.3) for r ∈ [0, 1). By (1.4), for z = reiθ ∈ D, we
have
 
 α+1 
f (z) − (1 − |z|) f (0) 
 1 + |z| 
 2π  
1  −it ∗ it (1 − |z|)α+1 2π ∗ it


= K (ze )f (e )dt − K (0)f (e )dt
2π  0 
α α
1 + |z| 0
 2π   
1  −it (1 − |z|)α+1 ∗ it


= K (ze ) − K (0) f (e )dt
2π  0 
α α
1 + |z|
 2π  
1 (1 − |z|)α+1
≤ Kα (ze−it ) − Kα (0) |f ∗ (eit )|dt
2π 0 1 + |z|
 2π 
1 −it (1 − |z|)α+1
≤ M Kα (ze )dt − Kα (0) .
2π 0 1 + |z|
Next we prove (1.7). By the proof of [25, Theorem 3.1], we observe that
 2π  2  
1 i Γ 1 + α2 α α
(2.1) Mα (r) = Kα (re )d = F − , − ; 1; r2
2π 0 Γ(1 + α) 2 2
6 Sh. Chen and M. Vuorinen

and Mα (r) is increasing on r ∈ [0, 1) with

lim Mα (r) = 1.
r→1−

From the proof of [25, Theorem 1.1], for z ∈ D, we have the formulas
 
2 α
∂ (1 − |z| ) α 2 + α 1 − ze −it
Kα (ze−it ) = cα − zeit +
∂z |1 − ze−it |2+α 2 2 1 − ze−it

and  
∂ (1 − |z|2 )α α 2 + α 1 − ze−it
Kα (ze−it ) = cα − ze−it + ,
∂z |1 − ze−it |2+α 2 2 1 − ze−it
which, together with (1.4) and (2.1), imply that
  2π    2π 
1 ∂  1 ∂ 
Df (z) =   −it ∗ it
Kα (ze )f (e )dt +   Kα (ze )f (e )dt
−it ∗ it
2π 0 ∂z 2π 0 ∂z
 2π 2 α
 
M cα (1 − |z| ) 2 + α |α|
≤ + |z| dt
2π 0 |1 − ze−it |2+α 2 2
  
M cα 2π (1 − |z|2 )α 2 + α |α|
+ + |z| dt
2π 0 |1 − ze−it |2+α 2 2
 2π
M 2 + α + |α||z|
= Kα (ze−it )dt
2π(1 − |z|2 ) 0
2 + α + |α||z|
= M Mα (|z|)
1 − |z|2
2 + α + |α||z|
≤ M .
1 − |z|2
The proof of this theorem is complete. 

Proof of Theorem 2. Since (a) easily follows from the inequality (15) in [17], we
only need to prove (b). Let α > 0. By Theorem 1 (1.6), we have

|f (eiθ ) − f (reiθ )| 1 − |f (reiθ )|

≥
1−r 1−r
1 2π (1−|z|)α+1
1 − 2π 0
Kα (ze−it )dt + 1+|z|
Kα (0)
(2.2) ≥ .
1−r
where z = reiθ ∈ D and θ ∈ [0, 2π).
Applying [25, Theorem 3.1], we get
 2π
1
lim Kα (ze−it )dt = lim Mα (r) = 1,
2π |z|→1− 0 r→1−

which, together with L’Hopital’s rule and (2.2), yield that

 Differential operators 7

 
 ∂f (reiθ 
)
Df (eiθ ) ≥  
 ∂r 
r=1
|f (e ) − f (reiθ )|
iθ
= lim
r→1− 1−r
2π α+1
1 − 2π 0 Kα rei(θ−t) dt + (1−r)
1
1+r
Kα (0)
≥ lim
r→1− 1−r
2π α+1
1 − 2π 0 Kα re dη + (1−r)
1 iη
1+r
Kα (0)
= lim
r→1− 1−r
 2π
1 d
= lim Kα reiη dη
2π r→1− dr 0
d
= lim Mα (r).
r→1− dr

where Mα (r) is given by (1.8). It follows from the proof of [25, Theorem 3.1] that

Γ 1 + α2
2 
α α   Γ 1 + α 2 ∞  − α 2
2 2 n
Mα (r) = F − , − ; 1; r = 2
r2n ,
Γ(1 + α) 2 2 Γ(1 + α) n=0 (n!)2
which yields that
 2  
d α2 Γ 1 + α2 α α
Mα (r) = rF 1 − , 1 − ; 2; r2 ,
dr 2 Γ(1 + α) 2 2
where r ∈ (0, 1). By (1.5), for α > 0, we see that
 2  
d α2 Γ 1 + α2 α α
lim Mα (r) = F 1 − , 1 − ; 2; 1
r→1− dr 2 Γ(1 + α) 2 2
  2
α2 Γ 1 + α2 Γ(2)Γ(α)
=  
2 Γ(1 + α) Γ 1 + α 2
2
2
α Γ(α) α
= = .
2 Γ(1 + α) 2
Therefore, for θ ∈ [0, 2π],
d α
Df (eiθ ) ≥ lim
Mα (r) = ,
r→1− dr 2
where α > 0. The proof of this theorem is complete. 
Proof of Theorem 3. For r ∈ [0, 1), let
 α α 
Ak (r, α) = ck F − , k − ; k + 1; r2
2 2
and  α α 
Bk (r, α) = c−k F − , k − ; k + 1; r2 ,
2 2
8 Sh. Chen and M. Vuorinen

where r = |z|. Then

 2π
k 1
Ak (r, α)r = f (z)e−ikθ dθ
2π 0

and
 2π
k 1
Bk (r, α)r = f (z)eikθ dθ,
2π 0

which imply that

 2π
k 1
(2.3) |Ak (r, α)|r = f (z)e−ikθ e−i arg Ak (r,α) dθ
2π 0

and
 2π
k 1
(2.4) |Bk (r, α)|r = f (z)eikθ e−i arg Bk (r,α) dθ ,
2π 0

where Ak (r, α) = |Ak (r, α)|ei arg Ak (r,α) , Bk (r, α) = |Bk (r, α)|ei arg Bk (r,α) and z = reiθ .
By (2.3), (2.4) and [13, Lemma 1], we have

 
(2.5)  |Ak (r, α)| + |Bk (r, α)| rk 
  2π   
1
=  f (z) e −i kθ+arg A k (r,α)
+e i kθ−arg B k (r,α)
dθ
 2π
0
 2π  
1  
≤ |f (z)|e−i kθ+arg Ak (r,α) + ei kθ−arg Bk (r,α) dθ
2π 0

M 2π  −i kθ+arg Ak (r,α) 

≤ e + ei kθ−arg Bk (r,α) dθ
2π 0

M 2π  
i 2kθ+arg Ak (r,α)−arg Bk (r,α) 
= 1 + e dθ
2π 0
   
M 2π  arg Ak (r, α) − arg Bk (r, α) 
= cos kθ +  dθ
π 0  2 
4M
= .
π

By letting r → 1− on (2.5), we obtain

4M
|Ak (1, α)| + |Bk (1, α)| ≤ .
π
 Differential operators 9

On the other hand, for k = 0, we have

 2π   α α  2
1  
(2.6) lim |f (reiθ )|2 dθ = c0 F − , − ; 1; 1 
2π r→1− 0 2 2

∞
  α α  2
+ ck F − , k − ; k + 1; 1 
 2 2 
k=1
 
  α α  2

+c−k F − , k − ; k + 1; 1 
2 2
≤ M 2,
where r ∈ [0, 1). It follows from (2.6) that
  α α 
 
c0 F − , − ; 1; 1  ≤ M.
2 2
If α = 0, then the sharpness part follows from [12, Lemma 1]. The proof of this
theorem is complete. 

3. The Landau type theorem

Lemma 1. For x ∈ [0, 1), let
 
δ 4M x (2 − x) 2x
ϕ(x) = − + ,
M (2 + α) π (1 − x)2 (1 − x)(1 − x2 )2
where α > −2, δ > 0 and M > 0 are constant. Then ϕ is strictly decreasing and
there is an unique x0 ∈ (0, 1) such that ϕ(x0 ) = 0.

Proof. For x ∈ [0, 1), let

4M x(2 − x) 4M 2x2
f1 (x) = and f 2 (x) = .
π (1 − x)2 π (1 − x)(1 − x2 )2
Since, for x ∈ [0, 1),
8M 1
f1 (x) = > 0,
π (1 − x)3
we see that f1 is continuous and strictly increasing in [0, 1). We observe that f2 is
also continuous and strictly increasing in [0, 1). Then
δ
ϕ(x) = − f1 (x) − f2 (x)
M (2 + α)
is continuous and strictly decreasing in [0, 1), which, together with
δ
lim ϕ(x) = > 0 and lim ϕ(x) = −∞,
x→0 M (2 + α) x→1−

imply that there is an unique x0 ∈ (0, 1) such that ϕ(x0 ) = 0. 

10 Sh. Chen and M. Vuorinen

Proof of Theorem 4. By Theorem C, we can assume that

∞  α α 
f (z) = ck F − , k − ; k + 1; |z|2 z k
k=0
2 2
∞  α α 
+ c−k F − , k − ; k + 1; |z|2 z k , z ∈ D,
k=1
2 2

for some sequence {ck }∞

k=−∞ of complex numbers satisfying

1
lim sup |ck | |k| ≤ 1.
|k|→∞

For α ∈ (−1, 0], by [25, Proposition 1.4], we observe that

 α α  ∞
− α2 n k − α2 r2n
F − , k − ; k + 1; r2 = n
≥0
2 2 n=0
(k + 1)n n!

is bounded and increasing on r ∈ [0, 1), which imply that

 α α 
(3.1) |ck | + |c−k | F − , k − ; k + 1; r2
2
 α 2 
α
≤ |ck | + |c−k | F − , k − ; k + 1; 1
2 2
4M
≤ ,
π
where r = |z| and k ∈ {1, 2, . . .}.
By (3.1) and Theorem 3, we see that, for each k ∈ {1, 2, . . .},

− α2 n k − α2 n 1 4M
(3.2) (|ck | + |c−k |) ≤ ,
(k + 1)n n! π

where n ∈ {1, 2, . . .}.

Since c0 = f (0) = 0, we see that

∞  α α 
(3.3) fz (z) − fz (0) = kck F − , k − ; k + 1; w z k−1
k=2
2 2
d  α 
∞
α
+ ck F − , k − ; k + 1; w z k z
k=1
dw 2 2
d  α 
∞
α
+ c−k F − , k − ; k + 1; w z k+1
k=1
dw 2 2
 Differential operators 11

and
∞  α α 
(3.4) fz (z) − fz (0) = kc−k F − , k − ; k + 1; w z k−1
k=2
2 2
d  α 
∞
α
+ ck F − , k − ; k + 1; w z k+1
k=1
dw 2 2
d  α 
∞
α
+ c−k F − , k − ; k + 1; w z k z,
k=1
dw 2 2

where w = |z|2 .
Applying (3.2), (3.3) and (3.4), we obtain

(3.5) |fz (z) − fz (0)| + |fz (z) − fz (0)|

∞  α α 
≤ k |ck | + |c−k | F − , k − ; k + 1; w |z|k−1
k=2
2 2
d  α 
∞
α
+2 |ck | + |c−k | F − , k − ; k + 1; w |z|k+1
k=1
dw 2 2
∞ ∞
 ∞

4M 4M
≤ k|z|k−1 + 2 n|z|2(n−1) |z|k+1
π k=2 k=1
π n=1
∞
4M |z|(2 − |z|) 8M |z|k+1
= +
π (1 − |z|)2 π k=1
(1 − |z|2 )2
4M |z|(2 − |z|) 8M |z|2
= + .
π (1 − |z|)2 π (1 − |z|)(1 − |z|2 )2
Applying Theorem 1 (1.7), we get

β = |Jf (0)| = | det Df (0)| = Df (0)l Df (0) ≤ M (2 + α)l Df (0) ,

which gives that

β
(3.6) l Df (0) ≥ .
M (2 + α)
In order to prove the univalence of f in Dρ0 , we choose two distinct points z1 ,
z2 ∈ Dρ0 and let [z1 , z2 ] denote the segment from z1 to z2 with the endpoints z1 and
z2 , where ρ0 satisfies the following equation
 
β 4M ρ0 2 − ρ0 2ρ0
− + = 0.
M (2 + α) π (1 − ρ0 )2 (1 − ρ0 )(1 − ρ20 )2

By (3.5), (3.6) and Lemma 1, we have

12 Sh. Chen and M. Vuorinen

 
 
|f (z2 ) − f (z1 )| =  fz (z)dz + fz (z)dz 
[z ,z ]
 1 2 
 
=  fz (0)dz + fz (0)dz 
[z1 ,z2 ]
 
 
−  fz (z) − fz (0) dz + fz (z) − fz (0) dz 
[z1 ,z2 ]
≥ l(Df )(0)|z2 − z1 |

− |fz (z) − fz (0)| + |fz (z) − fz (0)| |dz|
[z1 ,z2 ]

β
> |z2 − z1 |
M (2 + α)
 
4M ρ0 2 − ρ0 2ρ0
− +
π (1 − ρ0 )2 (1 − ρ0 )(1 − ρ20 )2
= 0.
Thus, f (z2 ) = f (z1 ). The univalence of f follows from the arbitrariness of z1 and z2 .
This implies that f is univalent in Dρ0 .
Now, for any ζ  = ρ0 eiθ ∈ ∂Dρ0 , we obtain that
 
 
|f (ζ ) − f (0)| = 

fz (z)dz + fz (z)dz 
[0,ζ  ]
 
 
=  fz (0)dz + fz (0)dz 
[0,ζ  ]
 
 
−  fz (z) − fz (0) dz + fz (z) − fz (0) dz 
[0,ζ  ]

≥ l(Df )(0)ρ0 − |fz (z) − fz (0)| + |fz (z) − fz (0)| |dz|
[0,ζ  ]
 
4M ρ20 1 t(2 − ρ0 t) 2ρ0 t2
≥ l(Df )(0)ρ0 − + dt
π 0 (1 − ρ0 t)2 (1 − ρ0 t)(1 − ρ20 t2 )2
 
βρ0 4M ρ20 (2 − ρ0 ) 1
≥ − tdt
M (2 + α) π (1 − ρ0 )2 0
 1 
2ρ0
+ t2 dt
(1 − ρ0 )(1 − ρ20 )2 0
 
β 4M ρ0 2 − ρ0 2ρ0
= ρ0 − +
M (2 + α) π 2(1 − ρ0 )2 3(1 − ρ0 )(1 − ρ20 )2

2ρ0 β M ρ0 (2 − ρ0 )
= − .
3 M (2 + α) π(1 − ρ0 )2
 Differential operators 13

Hence f Dρ0 contains a univalent disk DR0 with


2ρ0 β M ρ0 (2 − ρ0 )
R0 ≥ − .
3 M (2 + α) π(1 − ρ0 )2
The proof of this theorem is complete. 

Acknowledgements: The authors would like to thank the referee for the use-
ful suggestions that helped to improve the manuscript. This research was partly
supported by the National Natural Science Foundation of China (No. 11401184
and No. 11326081), the Hunan Province Natural Science Foundation of China (No.
2015JJ3025), the Excellent Doctoral Dissertation of Special Foundation of Hunan
Province (higher education 2050205), the Construct Program of the Key Discipline
in Hunan Province, the Väisälä Foundation of The Finnish Academy of Sciences
and Letters.

References
1. Z. Abdulhadi and Y. Abu Muhanna, Landau’s theorem for biharmonic mappings, J. Math.
Anal. Appl., 338(2008), 705–709.
2. A. Alemam and J. A. Peláez, Spectra of integration operators and weighted square func-
tions, Indiana Univ. Math. J., 61(2012), 775–793.
3. G. D. Anderson, R. W. Barnard, K. C. Richards, M. K. Vamanamurthy and M.
Vuorinen, Inequalities for zero-balanced hypergeometric functions, Trans. Amer. Math. Soc.,
126(1995), 1713–1723.
4. G. D. Anderson, S. L. Qiu, M. K. Vamanamurthy and M. Vuorinen, Generalized
elliptic integrals and modular equations, Pacific J. Math., 192(2000), 1–37.
5. A. Borichev and H. Hedenmalm, Weighted integrability of polyharmonic functions, Adv.
Math., 264(2014), 464–505.
6. H. Chen, P. M. Gauthier and W. Hengartner, Bloch constants for planar harmonic
mappings, Proc. Amer. Math. Soc., 128(2000), 3231–3240.
7. H. H. Chen, The Schwarz-Pick lemma for planar harmonic mappings, Sci. China Math.,
54(2011), 1101–1118.
8. Sh. Chen, S. Ponnusamy and A. Rasila, Coefficient estimates, Landau’s theorem and
Lipschitz-type spaces on planar harmonic mappings, J. Aust. Math. Soc., 96(2014), 198–215.
9. Sh. Chen, S. Ponnusamy and X. Wang, Harmonic mappings in Bergman spaces, Monatsh.
Math., 170(2013), 325–342.
10. Sh. Chen, S. Ponnusamy and X. Wang, Integral means and coefficient estimates on planar
harmonic mappings, Ann. Acad. Sci. Fenn. Math., 37(2012), 69–79.
11. Sh. Chen, S. Ponnusamy and X. Wang, On planar harmonic Lipschitz and planar harmonic
Hardy classes, Ann. Acad. Sci. Fenn. Math., 36(2011), 567–576.
12. Sh. Chen, S. Ponnusamy and X. Wang, Bloch constant and Landau’s theorems for planar
p-harmonic mappings, J. Math. Anal. Appl., 373(2011), 102–110.
13. Sh. Chen and A. Rasila, Schwarz-Pick type estimates of pluriharmonic mappings in the
unit polydisk, arXiv:1409.7897, to appear in Illinois J. Math.
14. F. Colonna, The Bloch constant of bounded harmonic mappings, Indiana Univ. Math. J.,
38 (1989), 829–840.
15. P. Duren, Harmonic Mappings in the Plane, Cambridge Univ. Press, 2004.
16. M. Englis̆, A Loewner-type lemma for weighted biharmonic operators, Pacific J. Math.,
179(1997), 343–353.
17. E. Heinz, On one-to-one harmonic mappings, Pacific J. Math., 9(1959), 101–105.
18. D. Kalaj, Heinz inequality for the unit ball, arXiv: 1504.01686 [math. AP].
14 Sh. Chen and M. Vuorinen

19. D. Kalaj and M. Vuorinen, On harmonic functions and the Schwarz lemma, Proc. Amer.
Math. Soc., 140(2012), 161–165.
20. E. Landau, Über die Bloch’sche konstante und zwei verwandte weltkonstanten, Math. Z.,
30(1929), 608–634.
21. M. S. Liu, Z. W. Liu and Y. C. Zhu, Landau’s theorems for certain biharmonic mappings,
Acta Math. Sinica, Chinese Series, 54(2011), 1–12.
22. A. Olofsson, A representation formula for radially weighted biharmonic functions in the unit
disc, Publ. Math., 49 (2005), 393–415.
23. A. Olofsson, Regularity in a singular biharmonic Dirichlet problem, Monatsh. Math., 148
(2006), 229–239.
24. A. Olofsson and J. Wittsten, Poisson integrals for standard weighted Laplacians in the
unit disc, J. Math. Soc. Japan., 65 (2013), 447–486.
25. A. Olofsson, Differential operators for a scale of Poisson type kernels in the unit disc, J.
Anal. Math., 123 (2014), 227–249.
26. M. Pavlović, Harmonic Schwarz lemmas: Chen. Kalaj-Vuorinen. Pavlović. Heinz., Preprint.
27. M. Pavlović, Decompositions of Lp and Hardy spaces of polyharmonic functions, J. Math.
Anal. Appl., 216 (1997), 499–509.
28. W. Szapiel, Bounded harmonic mappings, J. Anal. Math., 111(2010), 47–76.

Shaolin Chen, College of Mathematics and Statistics, Hengyang Normal Univer-

sity, Hengyang, Hunan 421008, People’s Republic of China.
E-mail address: [email protected]

Matti Vuorinen, Department of Mathematics and Statistics, University of Turku,

Turku 20014, Finland.
E-mail address: [email protected]