Professional Documents
Culture Documents
The Second Hankel Determinant For A Class of Analytic Functions Associated With The Carlson-Shaffer Operator
The Second Hankel Determinant For A Class of Analytic Functions Associated With The Carlson-Shaffer Operator
The Second Hankel Determinant For A Class of Analytic Functions Associated With The Carlson-Shaffer Operator
n=2
a
n
z
n
(z U). (1.1)
For the functions f and g in A given by the series expansion:
f (z) =
n=0
a
n
z
n
, g(z) =
n=0
b
n
z
n
(z U),
the Hadamard product (or Convolution) f g, is dened by
(f g)(z) =
n=0
a
n
b
n
z
n
=(g f )(z) (z U).
The function f g A. We recall that the Carlson- Shaffer operator [3]
L(, ) : A
0
A
0
( C, C\Z
0
; Z
0
: {0, 1, 2, . . . })
Corresponding author: A. K. Mishra.
2010 Mathematics Subject Classication. Primary 30C45, Secondary 33C15.
Key words and phrases. Univalent, Starlike, Hadamard product, Carlson-Shaffer operator, Hankel de-
terminant..
73
74 A. K. MISHRA ANDS. N. KUND
is dened by:
L(, )f (z) =(, ; z) f (z) (z U, f A), (1.2)
where
(, ; z) =
k=0
()
k
()
k
z
k+1
(z U) (1.3)
and()
k
is the Pochhammer symbol (or shiftedfactorial) denedin terms of the Gamma func-
tion by
()
k
=
(+k)
()
:=
_
_
_
1 (k =0)
(+1). . . (+k 1) (k N={1, 2, . . . }).
It can be readily veried that L(, ) is the identity operator; the operators L(, ) and
L(, ) commute, that is
L(, )L(, )f (z) =L(, )L(, )f (z) (f A
0
)
and the following transitive property holds true:
L(, )L(, )f =L(, )f (, C\Z
0
, f A
0
).
In the particular case =2, =1, the operator L(, ) reduces to the Alexanders transform:
L(2, 1)f (z) =z f
(z) (f A
0
).
Moreover, the popular Owa-Srivastava fractional differential operator
z
: A
0
A
0
(0 <1, z U)
is related to the Carlson-Shaffer operator by the formula:
z
f (z) =L(2, 2)f (z)
(see [23, 24, 25], also see [18, 19]). By using the Carlson-Shaffer operator we introduce the
following class of functions:
Denition 1. The function f A
0
is said to be in the class R
,
(, ) (
2
< <
2
, 0 <
1, C, C\Z
0
) if
_
e
i
L(, )f (z)
z
_
>cos (z U). (1.4)
THE SECOND HANKEL DETERMINANT FOR A CLASS OF ANALYTIC FUNCTIONS 75
The class R
,
(, ) generalizes several well known subclasses of A
0
. For example, tak-
ing = ; = 2, = 1 and = 2, = 2 (0 < 1) respectively, we get the following
interesting classes:
R
,
(, ) =
_
f A
0
:
_
e
i
f (z)
z
_
> cos
_
:=R
0
(, ) (1.5)
R
0
(, ) =R
0
(),
R
2,1
(, ) =
_
f A
0
:
_
e
i
f
(z)
_
> cos
_
:=R
1
(, )
R
1
(, ) =R
1
()
and
R
2,2
(, ) =
_
f A
0
:
_
e
i
z
f (z)
z
_
> cos
_
:=R
(, )
(, ) =R
().
It is well known that the functions in the class R
1
() are univalent close-to-convex [4]. More-
over, if 0 <<1 then
R
1
() R
() R
() R
0
()
(cf. [16, 20]). For initial seminal work on the class R
1
(0) := R
1
one may see the classical
paper of Macgregor [17]. The family of functions R
,
(, ) is characterized by the following
function class:
P :={p A : p(0) =1, (p(z)) >0, z U}.
Infact, it follows from (1.4) that the function f A
0
is in the class R
,
(, ) if and only if
e
i
L(, )f (z)
z
=[(1)p(z) +] cos+i sin (1.6)
for some function p P.
For the complex sequence a
n
, a
n+1
, a
n+2
, . . . , the Hankel matrix, named after Herman
Hankel (1839-1873), is the innite matrix whose (i , j )
t h
entry a
i j
is dened by
a
i j
=a
n+i +j 2
(i , j , n N).
The q
t h
Hankel matrix (q N\{1}), is by denition, the following q q square sub matrix:
_
_
_
_
_
_
_
_
_
_
a
n
a
n+1
a
n+2
. . . a
n+q1
a
n+1
a
n+2
a
n+3
. . . a
n+q
a
n+2
a
n+3
a
n+4
. . . a
n+q+1
.
.
.
.
.
.
.
.
.
.
.
.
a
n+q1
a
n+q
a
n+q+1
. . . a
n+2q1
_
_
_
_
_
_
_
_
_
_
76 A. K. MISHRA ANDS. N. KUND
We observe that the Hankel matrix has constant positive slopping diagonals whose entries
also satisfy:
a
i j
=a
i 1, j +1
(i N\{1}; j N).
This also describes the Hankel matrix without reference to a particular sequence. The deter-
minant of the q
t h
Hankel matrix, usually denoted by H
q
(n), is called the q
t h
Hankel deter-
minant (cf. [22]). In the particular cases q = 2, n = 1, a
1
= 1 and q = 2, n = 2, the Hankel
determinant simplies respectively to
H
2
(1) =a
3
a
2
2
and H
2
(2) =a
2
a
4
a
2
3
.
We refer to H
2
(2) as the Second Hankel determinant.
It is fairly well known that for the univalent function of the form (1.1) the sharp inequal-
ity |H
2
(1)| = |a
3
a
2
2
| 1 holds true [4]. For a family of functions in A
0
, the more general
problem of nding sharp estimates for the functional |a
2
2
a
3
| ( R or C) is popu-
larly known as the Fekete-Szeg problem for . The Fekete-Szeg problem for the families of
univalent functions, starlike functions, convex functions, close-to-convex functions has been
completely settled in [5, 11, 12, 13]. For related results also see [19].
Recently Janteng et.al.[8] and the rst author and Gochhayat [20] obtained sharp esti-
mates on the Second Hankel determinant for the families R
1
() and R
(, ) respectively.
For some more recent work see [1, 6, 7, 9, 10, 21]. In this paper we generalize the results of [8]
and [20] by nding sharp bounds for |H
2
(2)| =|a
2
a
4
a
2
3
| for f in R
,
(, ). We also obtain
here some basic properties such as class preserving transforms for the class R
,
(, ).
2. Preliminaries
Each of the following results will be required in our present investigation:
Lemma 2.1. (cf. [4]) Let the function p P be given by the series
p(z) =1+c
1
z +c
2
z
2
+. . . (z U). (2.1)
Then,
|c
k
| 2 (k N). (2.2)
The estimate (2.2) is sharp.
Lemma 2.2. (cf.[15], p.254, also see [14]) Let the function p P be given by the power series
(2.1). Then,
2c
2
=c
2
1
+x(4c
2
1
) (2.3)
and
4c
3
=c
3
1
+2(4c
2
1
)c
1
x (4c
2
1
)c
1
x
2
+2(4c
2
1
)(1|x|
2
)z (2.4)
for some complex numbers x, z satisfying |x| 1 and |z| 1.
THE SECOND HANKEL DETERMINANT FOR A CLASS OF ANALYTIC FUNCTIONS 77
Lemma 2.3. (cf. [26]) Let f and g be univalent convex functions in U. Then, f g is also a
univalent convex function in U.
Lemma 2.4. (cf. [26], also see [16]) Let f and g be starlike of order 1/2. Then, for each function
F(z) satisfying (F(z)) > (0 <1, z U),
_
f (z) F(z)g(z)
f (z) g(z)
_
> (z U). (2.5)
3. Main results
We state and prove the following:
Theorem3.1. Let the function f , given by (1.1), be in the class R
,
(, ) (0 <1,
2
< <
2
). If 0 <<2, 0 <<
2+5
2
, then
|a
2
a
4
a
2
3
|
4
2
(+1)
2
(1)
2
cos
2
2
(+1)
2
. (3.1)
The estimate (3.1) is sharp.
Proof. Let f R
,
(, ) (0 <1,
2
< <
2
). Then using (1.2), (1.3), and (1.6) we write
e
i
L(, )f (z)
z
= e
i
_
1+
n=2
()
n1
()
n1
a
n
z
n1
_
=
_
(1)p(z) +
_
cos+i sin (3.2)
where p P and is given by (2.1).
A comparison of the coefcients, in (3.2) gives
e
i
a
2
=c
1
(1)cos,
()
2
e
i
()
2
a
3
=c
2
(1)cos, (3.3)
()
3
e
i
()
3
a
4
=c
3
(1)cos.
Therefore, (3.3) yields the following:
|a
2
a
4
a
2
3
| =
2
(+1)(1)
2
cos
2
2
(+1)
+2
+2
c
1
c
3
+1
+1
c
2
2
.
Since the functions p(z) and p(e
i
z) ( R) are members of the class P simultaneously, we
assume without loss of generality that c
1
>0. For convenience of notation, we take c
1
=c (c
[0, 2]).
78 A. K. MISHRA ANDS. N. KUND
Using Lemma 2.2, we get
|a
2
a
4
a
2
3
| =
2
(+1)(1)
2
cos
2
2
(+1)
+2
4(+2)
c{c
3
+2(4c
2
)cx c(4c
2
)x
2
+2(4c
2
)(1|x|
2
)z}
+1
4(+1)
{c
2
+x(4c
2
)}
2
2
(+1)(1)
2
cos
2
2
(+1)
+2
4(+2)
c
4
+
(+2)(4c
2
)c
2
2(+2)
x
(+2)(4c
2
)c
2
4(+2)
x
2
+
(+2)(4c
2
)c(1|x|
2
)z
2(+2)
+1
4(+1)
c
4
(+1)(4c
2
)c
2
2(+1)
x
(+1)(4c
2
)
2
4(+1)
x
2
.
=
2
(+1)(1)
2
cos
2
2
(+1)
()c
4
4(+1)
2
+
(4c
2
)c
2
()
2(+1)
2
x
4c
2
4
_
c
2
(+2)
(+2)
+
(+1)(4c
2
)
(+1)
_
x
2
+
c(4c
2
)(+2)(1|x|
2
)
2(+2)
z
.
An application of triangle inequality and replacement of |x| by , give
|a
2
a
4
a
2
3
|
2
(+1)(1)
2
cos
2
2
(+1)
_
()c
4
4(+1)
2
+
(4c
2
)c
2
()
2(+1)
2
+
4c
2
4
_
c
2
(+2)
(+2)
+
(+1)(4c
2
)
(+1)
_
2
+
c(4c
2
)(+2)
2(+2)
c(4c
2
)(+2)
2(+2)
2
_
.
=
2
(+1)(1)
2
cos
2
2
(+1)
_
()c
4
4(+1)
2
+
c(4c
2
)(+2)
2(+2)
+
(4c
2
)c
2
()
2(+1)
2
+
(4c
2
)(2c){2(+2)(+1) c()}
4(+1)
2
2
_
= F(c, ) (say) (3.4)
where 0 c 2 and 0 1.
We next maximize the function F(c, ) on the closed rectangle [0, 2] [0, 1]. We rst as-
sume that 0 <<. A routine calculation gives
F
2
(+1)(1)
2
(4c
2
)()cos
2
2
2
(+1)
2
(+2)
_
c
2
+(c 2)
_
c
2(+2)(+1)
()
_
_
=
2
(+1)(1)
2
(4c
2
)()cos
2
2
2
(+1)
2
(+2)
_
c
2
+(c 2)
_
(c 2) (c +4) 24
()
_
_
.
Therefore, for 0 <c <2 and 0 < <1, we have
F
(c) =
8c()
2
(+1)(1)
2
cos
2
_
c
2
+
(2)+5+2
_
4
2
(+1)
2
(+2)
.
Since 0 < < <
2+5
2
, we get
_
c
2
+
(2)+5+2
_
> 0, so that H
=(1)
2
(4c
2
)(2c)cos
2
>0
and
H(c) =(1)
2
(4c
2
)cos
2
.
Hence, the maximum of F(c, ) occurs at c =0 and =1.
Therefore,
|a
2
a
4
a
2
3
|
4
2
(+1)
2
(1)
2
cos
2
2
(+1)
2
.
This is the assertion of (3.1). Equality holds for the function
f (z) =(, ; z) e
i
_
z
_
1+(12)z
2
1z
2
cos+i sin
__
.
The proof of the theorem is completed.
Taking 0 <= in Theorem 3.1 we have the following:
Corollary 3.2. Let the function f given by (1.1), be a member of the class R
0
() (0 < 1).
Then
|a
2
a
4
a
2
3
| 4(1)
2
.
Equality holds for the function
z
_
1+(12)z
2
1z
2
_
.
Taking =2 and =1 Theorem3.1 we get the following:
80 A. K. MISHRA ANDS. N. KUND
Corollary 3.3. Let the function f , given by (1.1) be in the class R
1
() (0 <1). Then
|a
2
a
4
a
2
3
|
4(1)
2
9
.
Equality holds for the function
f (z) =(1, 2; z)
_
z
_
1+(12)z
2
1z
2
__
.
The choice =0 in Corollary 3.3 gives a result of Janteng et. al. [8] for the class R
1
.
Similarly the choice =2, =2, ( 0 1), in Theorem3.1 gives the following recent result
of Mishra and Gochhayat [20] for the class R
(, ).
Corollary 3.4. Let the function f , given by (1.1), be in the class R
(, ), (0 1, 0 <
1,
2
< <
2
). Then
|a
2
a
4
a
2
3
|
(1)
2
(2)
2
(3)
2
cos
2
9
.
The estimate is sharp, for the function
f (z) =(2, 2; z) e
i
_
z
_
1+(12)z
2
1z
2
_
cos+i sin
_
.
Corollary 3.5. Let the function f , given by (1.1), be a member of the class R
,
() (0 < < 2,
0 <<<
2+5
2
). Then
|a
2
a
4
a
2
3
|
4
2
(+1)
2
(1)
2
2
(+1)
2
.
Equality holds for the function
f (z) =(, ; z)
_
z
_
1+(12)z
2
1z
2
__
.
Theorem3.6. Let f S
(1/2) and g R
,
(, )
_
0 <1,
2
< <
2
_
. Then
f g R
,
(, ).
Proof. Since Hadamard product is associative and commutative, we have
L(, )(f g)(z) = f (z) L(, )g(z).
Therefore,
e
i
L(, )(f g)(z)
z
=
f (z)
e
i
L(,)g(z)
z
z
f (z) z
.
Now applying Lemma 2.4, we get
_
e
i
L(, )(f g)(z)
z
_
>cos.
Hence f g R
,
(, ) and the proof of Theorem 3.6 is completed.
THE SECOND HANKEL DETERMINANT FOR A CLASS OF ANALYTIC FUNCTIONS 81
Theorem 3.7. Let f R
,
(, )
_
0 <1,
2
< <
2
_
. Then, the function J(f ) dened
by the integral transform.
J(f )(z) =
+1
z
_
z
0
t
1
f (t )dt (z U, >1)
is also in R
,
(, ).
Proof. The integral transformJ(f ) can be written in terms of the Carlson-Shaffer operator
as
J(f )(z) =(L(+1, +2)f )(z).
Hence
(L(, )J(f ))(z) =L(+1, +2)L(, )f (z) =(+1, +2; z) L(, )f (z).
Therefore,
e
i
(L(, )J(f ))(z)
z
=
(+1, +2; z)
_
e
i
L(, )f (z)/z
_
z
(+1, +2; z) z
.
Using a result of Bernardi [2], it can be veried that (+1, +2; z) S