Key words and phrases: Harmonic, Univalent, Subordination, Convex, Starlike, Close-to-convex

AN APPLICATION OF SUBORDINATION ON HARMONIC FUNCTION

H.A. AL-KHARSANI DEPARTMENT OFMATHEMATICS

GIRLSCOLLEGE

P.O. BOX838, DAMMAM, SAUDIARABIA

hakh73@hotmail.com

Received 28 June, 2006; accepted 15 May, 2007 Communicated by A. Sofo

ABSTRACT. The purpose of this paper is to obtain sufficient bound estimates for harmonic func- tions belonging to the classesS_H^∗[A, B], KH[A, B]defined by subordination, and we give some convolution conditions. Finally, we examine the closure properties of the operatorDⁿon these classes under the generalized Bernardi integral operator.

Key words and phrases: Harmonic, Univalent, Subordination, Convex, Starlike, Close-to-convex.

2000 Mathematics Subject Classification. Primary 30C45; Secondary 58E20.

1. INTRODUCTION

A continuous function f = u+iv is a complex-valued harmonic function in a complex domainC if both u andv are real harmonic in C. In any simply connected domain D ⊂ C, we can writef = h+g, where h andg are analytic in D. We call hthe analytic part and g the co-analytic part off. A necessary and sufficient condition forf to be locally univalent and orientation-preserving inDis that|g⁰(z)|<|h⁰(z)|inD[2].

We denote by S_H the family of functions f = h +g which are harmonic univalent and orientation-preserving in the open diskU = {z : |z| <1}so thatf = h+g is normalized by f(0) = h(0) = fz(0)−1 = 0. Therefore, forf = h+g ∈ SH, we can express the analytic functionshandg by the following power series expansion:

(1.1) h(z) =z+

∞

X

m=2

a_mz^m, g(z) =

∞

X

m=1

b_mz^m.

Note that the familyS_H of orientation-preserving, normalized harmonic univalent functions reduces to the classS of normalized analytic univalent functions if the co-analytic part off = h+gis identically zero.

LetK, S^∗, C, K_H, S_H^∗ andC_Hdenote the respective subclasses ofSandS_H where the images off(u)are convex, starlike and close-to-convex.

176-06

A functionf(z)is subordinate toF(z)in the diskU if there exists an analytic functionw(z) with w(0) = 0 and |w(z)| < 1 such that f(z) = F(w(z)) for |z| < 1. This is written as f(z)≺F(z).

LetK[A, B], S^∗[A, B]denote the subclasses ofS defined as follows:

S^∗[A, B] =

f ∈S, zf⁰(z)

f(z) ≺ 1 +Az

1 +Bz, −1≤B < A≤1

, K[A, B] =

f ∈S,(zf⁰(z))⁰

f⁰(z) ≺ 1 +Az

1 +Bz, −1≤B < A≤1

.

We now introduce the following subclasses of harmonic functions in terms of subordination.

Letf =h+g ∈S_H such that

ϕ(z) = h(z)−g(z) 1−b1

, (1.2)

ψ(z) = h(z)−e^iθg(z)

1−e^iθb₁ , 0≤θ <2π, (1.3)

and let −1 ≤ B < A ≤ 1, then we can construct the classes KH[A, B], S_H^∗[A, B] using subordination as follows:

K_H[A, B] =

f ∈S_H,(zψ⁰(z))⁰

ψ⁰(z) ≺ 1 +Az 1 +Bz

, S_H^∗[A, B] =

f ∈S_H,zϕ⁰(z)

ϕ(z) ≺ 1 +Az 1 +Bz

.

LetDⁿdenote then-th Ruscheweh derivative of a power seriest(z) =z+P∞

m=2t_mz^mwhich is given by

Dⁿt= z

(1−z)ⁿ⁺¹ ∗t(z)

=z+

∞

X

m=2

C(n, m)t_mz^m, where

C(n, m) = (n+ 1)m−1

(m−1)! = (n+ 1)(n+ 2)· · ·(n+m−1)

(m−1)! .

In [5], the operatorDⁿwas defined on the class of harmonic functionsS_H as follows:

Dⁿf =Dⁿh+Dⁿg.

The purpose of this paper is to obtain sufficient bound estimates for harmonic functions be- longing to the classesS_H^∗[A, B], KH[A, B], and we give some convolution conditions. Finally, we examine the closure properties of the operatorDⁿon the above classes under the generalized Bernardi integral operator.

2. PRELIMINARYRESULTS

Cluni and Sheil-Small [2] proved the following results:

Lemma 2.1. Ifh, g are analytic inU with|h⁰(0)| > |g⁰(0)| andh+g is close-to-convex for each,||= 1, thenf =h+gis harmonic close-to-convex.

Lemma 2.2. Iff = h+g is locally univalent inU andh+g is convex for some, || ≤ 1, thenf is univalent close-to-convex.

A domainDis called convex in the directionγ (0≤γ < π)if every line parallel to the line through 0 ande^iγ has a connected intersection withD. Such a domain is close-to-convex. The convex domains are those that are convex in every direction.

We will make use of the following result which may be found in [2]:

Lemma 2.3. A function f = h+g is harmonic convex if and only if the analytic functions h(z)−e^iγg(z), 0≤γ <2π, are convex in the direction ^γ₂ andf is suitably normalized.

Necessary and sufficient conditions were found in [2, 1] and [4] for functions to be inK_H, S_H^∗ and C_H. We now give some sufficient conditions for functions in the classes S_H^∗[A, B] and K_H[A, B], but first we need the following results:

Lemma 2.4 ([7]). If q(z) = z +P∞

m=2Cmz^m is analytic in U, then q maps onto a starlike domain ifP∞

m=2m|C_m| ≤1and onto convex domains ifP∞

m=2m²|C_m| ≤1.

Lemma 2.5 ([4]). Iff =h+g with

∞

X

m=2

m|a_m|+

∞

X

m=1

m|b_m| ≤1, thenf ∈C_H. The result is sharp.

Lemma 2.6 ([4]). Iff =h+g with

∞

X

m=2

m²|a_m|+

∞

X

m=1

m²|b_m| ≤1, thenf ∈K_H. The result is sharp.

Lemma 2.7 ([6]). A functionf(z)∈Sis inS^∗[A, B]if

∞

X

m=2

{m(1 +A)−(1 +B)} |a_m| ≤A−B, where−1≤B < A≤1.

Lemma 2.8 ([6]). A functionf(z)∈Sis inK[A, B]if

∞

X

m=2

m{m(1 +A)−(1 +B)} |a_m| ≤A−B, where−1≤B < A≤1.

Lemma 2.9 ([3]). Lethbe convex univalent inU withh(0) = 1andRe(λh(z)+µ)>0 (λ, µ∈ C). Ifpis analytic inU withp(0) = 1, then

p(z) + zp⁰(z)

λp(z) +µ ≺h(z) (z ∈U) implies

p(z)≺h(z) (z ∈U).

3. MAINRESULTS

Theorem 3.1. If (3.1)

∞

X

m=2

{m(1 +A)−(1 +B)} |a_m|+

∞

X

m=1

{m(1 +A)−(1 +B)} |b_m| ≤A−B, thenf ∈S_H^∗[A, B]. The result is sharp.

Proof. From the definition ofS_H^∗[A, B], we need only to prove that ϕ(z) ∈ S^∗[A, B], where φ(z)is given by (1.2) such that

φ(z) = z+

∞

X

m=2

a_m−b_m 1−b₁

z^m. Using Lemma 2.7, we have

∞

X

m=2

{m(1 +A)−(1 +B)}

A−B

am−bm

1−b₁

≤

∞

X

m=2

{m(1 +A)−(1 +B)}

A−B

|am|+|bm| 1− |b₁|

≤1 if and only if (3.1) holds and hence we have the result.

The harmonic function f(z) = z+

∞

X

m=2

1

(A−B){m(1 +A)−(1 +B)}x_mz^m +

∞

X

m=1

1

(A−B){m(1 +A)−(1 +B)} y_mz^m where

∞

X

m=2

|x_m|+

∞

X

m=1

|y_m|=A−B−1

!

shows that the coefficient bound given by (3.1) is sharp.

Corollary 3.2. If A = 1, B = −1, then we have the coefficient bound given in [1] with a different approach.

Theorem 3.3. Iff =h+gwith

∞

X

m=2

{m(1 +A)−(1 +B)}|a_m|C(n, m) +

∞

X

m=1

{m(1 +A)−(1 +B)}|b_m|C(n, m)≤A−B, thenDⁿf =H+G∈S_H^∗[A, B]. The function

f(z) = z+ (1 +δ)(A−B)

{m(1 +A)−(1 +B)}C(n, m)z^m, δ >0 shows that the result is sharp.

Corollary 3.4. If A = 1, B = −1, then we have the coefficient bound given in Theorem 3.1, α= 0[5] with a different approach.

Theorem 3.5. If (3.2)

∞

X

m=2

m{m(1 +A)−(1 +B)}|a_m|+

∞

X

m=1

m{m(1 +A)−(1 +B)}|b_m| ≤A−B, thenf ∈K_H[A, B]. The result is sharp.

Proof. From the definition of the classK_H[A, B]and the coefficient bound ofK[A, B]given in Lemma 2.8, we have the result. The function

f(z) = z+ (1 +δ)(A−B)

m{m(1 +A)−(1 +B)}z^m, δ >0

shows that the upper bound in (3.2) cannot be improved.

Theorem 3.6. Iff =h+gwith

∞

X

m=2

m{m(1 +A)−(1 +B)}C(n, m)|am| +

∞

X

m=1

m{m(1 +A)−(1 +B)}C(n, m)|b_m| ≤A−B, thenDⁿf ∈K_H[A, B]. The function

f =z+ (1 +δ)(A−B)

m{m(1 +A)−(1 +B)}C(n, m)z^m, δ >0 shows that the result is sharp.

Corollary 3.7. Ifn= 0, A= 1, B =−1, we have Theorem 3 in [4] and ifA= 1, B =−1, we have Theorem 2 in [5].

In the next two theorems, we give necessary and sufficient convolution conditions for func- tions inS_H^∗[A, B]andK_H[A, B].

Theorem 3.8. Letf =h+g ∈S_H. Thenf ∈S_H^∗[A, B]if h(z)∗ z+^(ξ−A)_A−Bz²

(1−z)²

!

+B g(z) ξ z−^(−1−Aξ)_A−B z² (1−z)²

!

6= 0, |ξ|= 1, 0<|z|<1.

Proof. LetS(z) = ^h(z)−g(z)_1−b

1 , thenS ∈S^∗[A, B]if and only if zS⁰

S ≺ 1 +Az 1 +Bz or

zS⁰(z)

S(z) 6= 1 +Ae^iθ

1 +Be^iθ, 0≤θ <2π, z ∈U.

It follows that

zS⁰(z)−S(z)1 +Ae^iθ 1 +Be^iθ

6= 0.

SincezS⁰(z) = S(z)∗ _(1−z)^z 2, the above inequality is equivalent to 06=S(z)∗

z

(1−z)² − 1 +Ae^iθ 1 +Be^iθ

z 1−z

(3.3)

= 1 λe^it





 S(z)∗





z+ ^(−e_(A−B)^−iθ−A)z² (−e^−iθ−B)(1−z)²











, 1−b₁ =λe^it

= 1 λe^it

(

h(z)∗ z+ ^(−e_A−B^−iθ−A)z² (−e^−iθ−B)(1−z)²

!

−g(z)

∗ z+ (−e^−iθ−A)z² (A−B)(e^−iθ/B)

(1−z)²(−B−e^iθ)

= 1 λ

(

h(z)∗ z+^(−e_A−B^−iθ−A)z² (1−z)²e^it

!

−g(z)∗ Be^iθz+ ^B(−e_A−B^{−iθ−A)eiθ}z² e^it(−B−e^iθ)(1−z)²

!) . Now, ifz1−z2 6= 0and|z1| 6=|z2|, thenz1−z2 6= 0, ||= 1, i.e.,

= 1

λ(−B −e^−iθ)

"

h(z)∗ z+^(−e_A−B^−iθ−A)z² (1−z)²e^it

!#

− g(z)∗





Be^+iθz+^(−1−Ae_A−B^iθ)Bz² e^it(−B−e^iθ)(1−z)²





= 1

λ(−B −e^−iθ)

"

h(z)∗ z+^(−e_A−B^−iθ−A)z² (1−z)²e^it

!#

− g(z)∗ (−B)(−e^−iθz+^B(−1−Ae_A−B^−iθ)z² (1−z)²e^−it

! .

Sincearg(1−b₁) =t 6=π, we obtain the result and the proof is thus completed.

Corollary 3.9. If A = 1, B −1and = 1, then we have Theorem 2.6 in [1] with a different approach.

Theorem 3.10. Letf =h+g ∈S_H. Thenf ∈K_H[A, B]if and only if h(z)∗

"

z+ ^2ξ−A−B_A−B z² (1−z)³

#

+g(z)∗

"

ξz− ^−2+(A+B)ξ_A−B z² (1−z)³

# 6= 0

||= 1, |ξ|= 1, 0<|z|<1

Proof. Letψ(z) = ^h(z)−e_1−eiγ^iγb^g(z)1 , 0 ≤γ < 2π and1−e^iγb₁ =λ e^it, then from (1.3) and (3.3), zψ⁰(z)∈S_H^∗[A, B]if and only if

zψ⁰(z)∗

"

z+ ^(−e_A−B^−iθ−A)z² (−e^−iθ −B)(1−z)²

# 6= 0

i.e., 06= 1

λe^it

"

zh⁰∗

( z+^(−e_A−B^−iθ−A)z² (−e^iθ−B)(1−z)²

)

−zg⁰∗

( z+^(−e_A−B^−iθ−A)z² (−e^−iθ−B)(1−z)²

)#

.

= 1 λe^it



h(z)∗

( z+^(−e_A−B^−iθ−A)z² (1−z)²(−e^−iθ −B)

)⁰

−g(z)∗

( z+ ^(−e_A−B^−iθ−A)z² (1−z)²(−e^−iθ−B)

)⁰



= 1 λe^it

"

h(z)∗ z+^−2e−iθ_A−B^−A−Bz² (1−z)³(−e^−iθ −B)

!

−g(z)∗ z+^−2e−iθ_A−B^−A−Bz² (1−z)³(−e^−iθ−B)

!#

= 1 λ

"

h(z)∗ z+^−2e−iθ_A−B^−A−Bz² e^it(1−z)³(−e^−iθ −B)

!

−g(z)∗ z+^−2e−iθ_A−B^−A−Bz² e^it(1−z)³(−B −e^−iθ)^e−iθ_B

!#

= 1 λ



h(z)∗ z+^−2e−iθ_A−B^−A−Bz²

e^it(1−z)³(−e^−iθ−B) −g(z)∗





Be^iθz+−2B−(A+B)Be^iθ A−B z² e^it(1−z)³(−B−e^iθ)









= 1 λ

"

h(z)∗ z+ ^−2e−iθ_A−B^−A−B

e^it(1−z)³(−e^−iθ −B)−g(z)∗ (−B)(−e^−iθ)z+−2B−(A+B)Be^−iθ A−B z² e^−it(−B−e^−iθ)(1−z)³

!#

= 1 λ

"

h(z)∗ z+^−2e−iθ_A−B^−A−B

e^it(1−z)³(e^−iθ −B)+Bg(z)∗ (−e^−iθ)z− −2+(A+B)(−e^−iθ) A−B z² e^−it(−B−e^−iθ)(1−z)³

!#

,

and we have the result.

Corollary 3.11. IfA= 1, B =−1, =−1, then we have Theorem 2.7 of [1].

Theorem 3.12. Iff =h+g ∈S_H with (3.4)

∞

X

m=2

mC(n, m)|a_m|+

∞

X

m=1

mC(n, m)|b_m| ≤1, thenDⁿf =H+G∈C_H. The result is sharp.

Proof. The result follows immediately. Using Lemma 2.5, the function f(z) =z+ 1 +δ

mC(n, m)z^m, δ >0

shows that the upper bound in (3.4) cannot be improved.

Theorem 3.13. Iff =h+gis locally univalent withP∞

m=2m²C(n, m)|a_m| ≤1, thenDⁿf ∈ C_H.

Proof. Take = 0in Lemma 2.2 and apply Lemma 2.4.

Corollary 3.14. Dⁿf =H+G∈C_H if|G⁰(z)| ≤ ¹₂ andP∞

m=2m²C(n, m)|a_m| ≤1.

Proof. The functionDⁿf is locally univalent if|H⁰(z)|>|G⁰(z)|forz ∈U. Since 2

∞

X

m=2

mC(n, m)|a_m| ≤

∞

X

m=2

m²C(n, m)|a_m| ≤1, we have

|H⁰(z)|>1−

∞

X

m=2

m|a_m|C(n, m)| ≥ 1 2.

Corollary 3.15. Ifh(z)∈K andw(z)is analytic with|w(z)|<1, then

f(z) = Dⁿh(z) + Z z

0

w(t)(Dⁿh(t))⁰dt ∈C_H.

Theorem 3.16. Letf =h+g ∈S_H. IfDⁿ⁺¹f ∈R, thenDⁿf ∈R, whereRcan beS_H^∗[A, B]

orK_H[A, B]orC_H.

Proof. We can prove the result whenR ≡S_H^∗[A, B]. IfDⁿ⁺¹f ∈S_H^∗[A, B], thenDⁿ⁺¹h

h−g 1−b1

i ∈ S^∗[A, B]and|Dⁿ⁺¹h|>|Dⁿ⁺¹g|. Using Lemma 2.9, we have

Dⁿ

h−g 1−b₁

∈S^∗[A, B].

Since

|Dⁿ⁺¹h|= z

z

(1−z)ⁿ⁺¹ ∗h 0

= z

1 z

z

(1−z)ⁿ⁺¹ ∗h⁰

,

this implies|Dⁿh|>|Dⁿg|, orDⁿ(h) +Dⁿg ∈S_H^∗[A, B]and we have the result.

Theorem 3.17. Let f = h+g ∈ S_H and let F_c(f) = ^1+c_zc

Rz

0 t^c−1f(t)dt. IfDⁿf ∈ R, then DⁿF_c(f)∈R, whereRcan beS_H^∗[A, B]orK_H[A, B]orC_H.

Proof. IfDⁿf ∈S_H^∗[A, B], thenDⁿ

h−g 1−b1

∈S^∗[A, B]. Using Lemma 2.9, we haveDⁿF_c(f)∈ S^∗[A, B]. That is, DⁿF_c(h−g)

1−b₁

∈ S^∗[A, B] or DⁿF_c(h) −DⁿF_c(g) ∈ S^∗[A, B]. Since

|DⁿF_c(n)|>|DⁿF_c(g)|, thenDⁿF_c(f)∈S_H^∗[A, B].

REFERENCES

[1] O.P. AHUJA, J.M. JAHANGIRI AND H. SILVERMAN, Convolutions for special classes of har- monic univalent functions, Appl. Math. Lett., 16 (2003), 905–909.

[2] J. CLUNI ANDT. SHEIL-SMALL, Harmonic univalent functions, Ann. Acad. Sci. Fenn. Ser. A.I.

Math., 9 (1984), 3–25.

[3] P. ENIGENBERG, S.S. MILLER, P.T. MOCANUANDM.O. READE, On a Briot-Bouquet differ- ential subordination, General Inequalities, Birkhäuser, Basel, 3 (1983), 339–348.

[4] J. JAHANGIRIANDH. SILVERMAN, Harmonic close-to-convex mappings, J. of Applied Mathe- matics and Stochastic Analysis, 15(1) (2002), 23–28.

[5] G. MURUGUSUNDARAMOORTHY, A class of Ruscheweyh-type harmonic univalent functions with varying arguments, South West J. of Pure and Applied Mathematics, 2 (2003), 90–95.

[6] H. SILVERMANANDE.M. SILVIA, Subclasses of starlike functions subordinate to convex func- tions, Canad. J. Math., 37 (1985), 48–61.

[7] H. SILVERMAN, Univalent functions with negative coefficients, Proc. Amer. Math. Soc., 51 (1975), 109–116.