Some Integrals Involving Bessel Functions

arXiv:math.CA/9307213 v1 9 Jul 1993

M.L. Glasser
Department of Physics and Department of Mathematics and Computer
Science, Clarkson University, Potsdam, N.Y. 13699-5815 (USA)

E. Montaldi
Dipartimento di Fisica, Universita di Milano, Via Celoria 16, 20133
Milano, Italy

Abstract

A number of new definite integrals involving Bessel functions are pre-

sented. These have been derived by finding new integral representations for
the product of two Bessel functions of different order and argument in terms
of the generalized hypergeometric function with subsequent reduction to spe-
cial cases. Connection is made with Weber’s second exponential integral and
Laplace transforms of products of three Bessel functions.

Keywords: Bessel Function, Hypergeometric Function, Integral Repre-

sentation

AMS Classification No. 33A40, 33A30

1
 Some Integrals Involving Bessel Functions

M.L. Glasser* and E. Montaldi**

*Department of Physics and Department of Mathematics and Computer

Science, Clarkson University, Potsdam, N.Y. 13699-5815 (USA)

**Dipartimento di Fisica, Universita di Milano, Via Celoria 16, 20133

Milano, Italy

1. Introduction
The aim of this work is to derive a number of infinite integrals involving
Bessel functions which appear to be new. The approach is, beginning with
an expression for the product of two Bessel functions as a sum of Gauss
functions, to integrate and perform a resummation to obtain other hyper-
geometric functions, and then to reduce these to more familiar form. In
particular we exploit the relation between Bessel functions and the function
0 F3 [1]. Finally, we generalize Weber’s second exponential integral by express-
ing the Laplace transform of a product of three Bessel functions as an infinite
series of products of modified Bessel functions. We have not aimed at com-
plete rigor or generality; operations such as interchange of limits and Hankel
inversion are carried out formally, i.e. without verifying that the conditions
stated ensure their validity. In the majority of cases approriate conditions
can be supplied by appeal to convergence and analytic continuation.

2. On the product Jµ (ax)Jν (bx)

Our starting point is the familiar expansion [2]

Γ(µ + 1)Γ(ν + 1)Jµ (ax)Jν (bx) =

1 µ 1 ν
X∞
(−1)m ( 12 ax)2m b2
( ax) ( bx) 2 F1 (−m, −µ − m; ν + 1; 2 ) (2.1)
2 2 m=0 m!(µ + 1)m a

2
By making use of the standard transformation[3], we have
b2
2 F1 (−m, −µ − m; ν + 1; )=
a2
b2 µ+ν+2m+1 b2
= (1 − ) 2 F1 (ν + m + 1, µ + ν + m + 1; ν + 1; ) (2.2)
a2 a2
Next, we recall that [4]
b2
2 F1 (ν + m + 1, ν + µ + m + 1; ν + 1; )=
a2
4(a/b)ν
= ·
Γ(µ + ν + 1)(ν + 1)m (µ + ν + 1)m
b
Z ∞
· tµ+ν+2m+1 Kµ (2t)Iν (2 t)dt (2.3)
0 a
b
m = 0, 1, 2, . . . , Re ν > −1, Re (µ + ν) > −1, |Re( )| < 1.
a
By inserting (2.2) and (2.3) into (2.1), we get
1 1 b2
Γ(ν + 1)Γ(µ + 1)Γ(µ + ν + 1)Jµ (ax)Jν (bx) = 4( ax)µ ( bx)ν (1 − 2 )µ+ν+1 ·
2 2 a
a b
Z ∞
·( )ν tµ+ν+1 0 F3 (µ + 1, ν + 1, µ + ν + 1; −z 2 t2 )Kµ (2t)Iν (2 t)dt (2.4)
b 0 a
b
Re ν > −1, Re (µ + ν) > −1, |Re( )| < 1,
a
1 b2
where z = 2 ax(1 − a2 ).
It is interesting to observe that eq. (2.4) enables us to derive the Mellin
transform of Jµ (ax)Jν (bx), i.e. the Weber-Schafheitlin integral [5], in a simple
way. Indeed from [6],
Z ∞
xµ+ν−s 0 F3 (µ + 1, ν + 1, µ + ν + 1; −z 2 t2 )dx =
0

1 1 b2 s−µ−ν−1 Γ( µ+ν−s+1
2
)Γ(µ + 1)Γ(ν + 1)Γ(µ + ν + 1)
= [ a(1 − 2 )t] µ−ν+s+1 , (2.5)
2 2 a Γ( 2 )Γ( ν−µ+s+1
2
)Γ( µ+ν+s+1
2
)
Re (µ + ν − s) > −1,
and [4]

3
 b (b/a)ν µ+ν +s+1 ν−µ+s+1
Z ∞
ts Kµ (2t)Iν (2 t)dt = Γ( )Γ( )·
0 a 4Γ(ν + 1) 2 2
ν +µ+s+1 ν −µ+s+1 b2
, 2 F1 ( ; ν + 1; 2 ) (2.6)
2 2 a
b
Re(ν ± µ + s) > −1, |Re( )| < 1
a
we immediately obtain
Z ∞
−s ν s−ν−1 Γ( µ+ν−s+1
2
)
x Jµ (ax)Jν (bx)dx = 2 b a
−s
µ−ν+s+1 ·
0 Γ(ν + 1)Γ( 2 )
b2 s µ+ν+s+1 ν −µ+s+1 b2
(1 − 2 ) 2 F1 ( , ; ν + 1; 2 )
a 2 2 a
µ+ν−s+1
Γ( 2 ) ν −µ−s+1 ν +µ−s+1 b2
= 2−s bν as−ν−1 2 F1 ( , ; ν+1; )
Γ(ν + 1)Γ( µ−ν+s+1
2
) 2 2 a2
(2.7)
Re(µ + ν − s) > −1, 0<b<a
With b → ib (b > 0),and x positive, eq. (2.4) becomes
Γ(µ + 1)Γ(ν + 1)Γ(µ + ν + 1)Jµ (ax)Iν (bx) =
1 1 b2 a
( ax)µ ( bx)ν (1 + 2 )µ+ν+1 ( )ν ·
4 4 a b
2
Z ∞ 1 2 2 b b
tµ+ν+1 0 F3 (µ+1, ν+1, µ+ν+1; − a x (1+ 2 )2 t2 )Kµ (t)Jν ( t)dt (2.8)
0 16 a a
Re ν > −1, Re(µ + ν) > −1
or, by writing b=ay and x=a/4
1 1 a2
Γ(µ + 1)Γ(ν + 1)Γ(µ + ν + 1)Jµ ( a2 )Iν ( a2 y) = ( )µ+ν (1 + y 2 )µ+ν+1 ·
4 4 16
a4
Z ∞
· tµ+ν+1 0 F3 (µ + 1, ν + 1, µ + ν + 1; − (1 + y 2)2 t2 )Kµ (t)Jν (yt)dt (2.9)
0 256
Re ν > −1, Re(µ + ν) > −1.
16a
Also, with a2 → 1+y 2
,

4
 4a 4ay
Γ(µ + 1)Γ(ν + 1)Γ(µ + ν + 1)Jµ ( 2
)Iν ( 2
) = (1 + y 2 )aµ+ν ·
1+y 1+y
Z ∞
· tµ+ν+1 0 F3 (µ + 1, ν + 1, µ + ν + 1; −a2 t2 )Kµ (t)Jν (yt)dt (2.10)
0

Re ν > −1, Re(µ + ν) > −1, a, y > 0.

For Reν ≤ −1/2 we may use the Hankel inversion formula [7], and eq. (2.10)
gives

(at)µ+ν 0 F3 (µ + 1, ν + 1, µ + ν + 1; −a2t2 )Kµ (t) = Γ(µ + 1)Γ(ν + 1)Γ(µ + ν + 1)·

Z ∞ y 4a 4ay
· J ν (ty)J µ ( )Iν ( )dy (2.11)
0 1 + y2 1 + y2 1 + y2
This is an interesting addition to the class of Sonine-Gegenbauer integrals
[8]. In particular, by dividing by aµ+ν and taking the limit a → 0, which
may be taken under the integral sign, we have

t Z ∞
y ν+1
( )µ+ν Kµ (t) = Γ(µ + ν + 1) Jν (ty)dy (2.12)
2 0 (1 + y 2 )µ+ν+1
The four particular cases
1 1 1 1
µ = 0, ν = − ; µ = 0, ν = ; ν = 0, µ = − ; ν = 0, µ =
2 2 2 2
of eq.(2.9) are of some interest. By using

1 1 x4
0 F3 ( , , 1; − ) = ber(x) (2.13)
2 2 256
and
x2 3 3 x4
0 F3 ( , , 1 : − ) = bei(x) (2.14)
4 2 2 256
we obtain the four integral representations

π a2 a2 y
Z ∞ q
(1 + y 2 )−1/2 J0 ( ) cosh( )= K0 (t)ber(a (1 + y 2 )t) cos(yt)dt
2 4 4 0
(2.15)

5
 π 2 −1/2 a2 a2 y Z ∞ q
(1 + y ) J0 ( ) sinh( )= K0 (t)bei(a (1 + y 2 )t) sin(yt)dt (2.16)
2 4 4 0
2 2
ay a Z ∞ q
(1 + y 2)−1/2 I0 ( ) cos( ) = e−t ber(a (1 + y 2 )t)J0 (yt)dt (2.17)
4 4 0

a2 y a2
Z ∞ q
2 −1/2
(1 + y ) I0 ( ) sin( ) = e bei(a (1 + y 2)t)J0 (yt)dt
−t
(2.18).
4 4 0
Therefore,
√ Z ∞ 1 a2 1 a2 y
K0 (t)ber(a t) = (1 + y 2 )−1/2 J0 ( ) cosh( ) cos(ty)dy
0 4 1 + y2 4 1 + y2
(2.19)
√ Z ∞
1 a 2
1 a2
y
K0 (t)bei(a t) = (1+y 2 )−1/2 J0 ( ) sinh( ) sin(ty)dy (2.20)
0 4 1 + y2 4 1 + y2
1 −t √ Z ∞
1 a2 y 1 a2
e ber(a t) = y(1 + y 2 )−1/2 I0 ( ) cos( )J0 (ty)dy (2.21)
t 0 4 1 + y2 4 1 + y2
1 −t √ Z ∞
1 a2 y 1 a2
e bei(a t) = y(1 + y 2 )−1/2 I0 ( ) sin( )J0 (ty)dy (2.22)
t 0 4 1 + y2 4 1 + y2
1
A curious formula arising from (2.4) may be mentioned. We set µ = 2
,
replace Iν (2 ab t) by Poisson’s integral [9]

b 1 bt ν Z 1 −2 b tu 1
Iν (2 t) = 1 ( ) e a (1 − u2 )ν− 2 du (2.23)
a Γ(ν + 2 ) a −1

1
Re ν > −
2
interchange the order of integration, and then use the Laplace transform
√
3 3 Γ(2ν + 2) −(2ν+1) α
Z ∞
−βt 2ν+1 2
e t 0 F3 ( , ν + 1, ν + ; −αt )dt = √ β sin(4 )
0 2 2 4 α β
(2.24)
Re ν > −1, Re β > 0
which is readily established by expressing the 0 F3 as its power series and
integrating term-by-term. This leads to

sin(ax)Jν (bx) =

6
 1
1 1 ν 2 2 ν+ 21
Z 1
(1 − u2 )ν− 2 a2 − b2
=√ ( bx) (a − b ) sin( x)du (2.25)
πΓ(ν + 12 ) 2 −1 (a + bu)2ν+1 a + bu
1
Re ν > − , |a| > |b| > 0.
2
π
Therefore, with x = 2a
and b = ay
πy
Jν ( )
2
1 1 1 (1 − u2 )ν−1/2 π 1 − y2
Z
ν 2 ν+ 21
=√ 1 ( πy) (1 − y ) sin( )du (2.26)
πΓ(ν + 2 ) 4 −1 (1 + uy)2ν+1 2 1 + uy
1
Re ν > − , |y| < 1.
2
By taking advantage of other known reductions for the 0 F3 functions in
(2.4) many additional new integrals can be derived. For example from [23]
we obtain
Z ∞ 3πν √ 3πν √
e−x Iν (x sin θ)[sin( )ber2ν (2 cos θ ux)+cos( )bei2ν (2 cos θ ux)]dx
0 2 2
= sec θ sin u Jν (u sin θ),
and in particular
Z ∞ √
e−x I2n (x sin θ)bei4n (2 cos θ ux)dx = (−1)n sec θ sin u J2n (u sin θ)
0
Z ∞ √
e−x I2n+1 (x sin θ)ber4n+2 (2 cos θ ux)dx = (−1)n+1 sec θ sin u J2n+1 (u sin θ).
0
We conclude this section by deriving a further integral representation for the
product Jν (ax)Jν (bx) in terms of a 0 F3 , but different from (2.4).
Let us consider eq.(2.1) with µ = ν. According to the quadratic transfor-
mation [10]

b2 b2 m m 1−m 2ab 2
2 F1 (−m, −ν − m; ν + 1; ) = (1 + ) 2 F1 (− , ; ν + 1; ( ) )=
a2 a2 2 2 a2 + b2
[m]
b2 2
1 ab 2r
= (1 + 2 )m m!
X
( 2 2
) (2.27)
a r=0 (m − 2r)!r!(ν + 1)r a + b

7
and (2.1) becomes
m 1
√
1 ∞
(−1) ( x a2 + b2 )2m
Jν (ax)Jν (bx) = ( abx2 )ν 2
X
·
4 m=0 Γ(ν + m + 1)
[m]
2
X ( a2ab
+b2
)2r
· . (2.28)
r=0 (m − 2r)!r!Γ(ν + r + 1)

By using [11]
∞ X
2
[m] ∞
X X
c(m, r) = c(m + 2r, r) (2.29)
m=0 r=0 m,r=0

it follows that
Jν (ax)Jν (bx) =
√
∞
X 1 abx ν+2r
X∞
(−1)m ( 21 x a2 + b2 )ν+2r+2m
( √ 2 ) =
r=0 r!Γ(ν + r + 1) 2 a + b2 m=0 m!Γ(ν + 2r + m + 1)

1 abx ν
∞
X ( 2√abx
a2 +b2
)2r √
=( √ ) J ν+2r (x a2 + b2 ) (2.30)
2 a2 + b2 r=0 r!Γ(ν + r + 1)
We note, in passing, that (2.30) provides a quick derivation of Weber’s second
integral [12]. Indeed, by using
Z ∞ 2 √ (a2 + b2 )ν/2+r − a24p
+b2
e−px xν+2r+1 Jν+2r (x a2 + b2 )dx = e (2.31)
0 (2p)ν+2r+1
we have
1 − a24p
+b2 X
∞
(ab/4p)ν+2r
Z ∞
−px2
xe Jν (ax)Jν (bx)dx = e =
0 2p r=0 r!Γ(ν + r + 1)

1 − a24p
+b2 ab
= e Iν ( ) (2.32)
2p 2p
We also observe that (2.30) can be obtained from the formula [13]

2 F1 (a, b; c; x)2 F1 (a, b; c; y) =

∞
(a)r (b)r (c − a)r (c − b)r
(xy)r 2 F1 (a + r, b + r; c + 2r; x + y − xy) (2.33)
X
=
r=0 r!(c) r (c) 2r

8
by applying the confluence principle [14] twice. We first replace x by x/b, y
by y/b, and let b→ ∞. This gives

(a)r (c − a)r
∞
(−xy)r 1 F1 (a + r; c + 2r; x + y)
X
1 F1 (a; c; x)1 F1 (a; c; y) =
r=0 r!(c) r (c) 2r
(2.34).
Next, we replace x by x/a, y by y/a and let a→ ∞; then
∞
X (xy)r
0 F1 (c; x) 0 F1 (c; x) = 0 F1 (c + 2r; x + y) (2.35)
r=0 r!(c)r (c)2r

which is equivalent to the desired result. If on the right hand side of (2.30)
we write [9] √
Jν+2r (x a2 + b2 ) =
√
2 ((1/2)x a2 + b2 )ν+2r Z 1 2 ν− 21 +2r
√
√ (1 − t ) cos(xt a2 + b2 )dt,
πΓ(ν + 12 ) ( ν2 + 14 )r ( ν2 + 34 )r 22r 0
(2.36)
we get (Re ν > − 21 )

2
Z 1 1 √
Jν (ax)Jν (bx) = (abx2 )ν (1 − t2 )ν− 2 cos(xt a2 + b2 )·
πΓ(2ν + 1) 0

ν 1 ν 3 1 2 2 4
+ , + ; a b x (1 − t2 )2 )dt
· 0 F3 (ν + 1, (2.37)
2 4 2 4 64
√ √
and with b = 1, x = a−1/2 , and a = 14 ( u2 + 2 + u2 − 2)2 ,
√ √ √ √
u2 + 2 + u2 − 2 u2 + 2 − u2 − 2
Jν ( )Jν ( )=
2 2
2 1
Z
(1 − t2 )ν−1/2 cos(ut)·
πΓ(2ν + 1) 0

ν 1 ν 3 1
+ , + ; (1 − t2 )2 )dt
· 0 F3 (ν + 1, (2.38)
2 4 2 4 64
Finally, with ν = 1/2 (2.37) yields [23]
s
Z 1 a2 + b2 √ √ dt
cos(ut )[I1 (u 1 − t2 ) + J1 (u 1 − t2 )] √
0 2ab 1 − t2

9
 q q
= (2/u) sin(u a/2b) sin(u b/2a).
Further, more complex, evaluations are possible by the same procedure.
√ √ √
3. The integral 0∞ e−αx J0 (β1 x)J0 (β2 x)J0 (β3 x)dx.
R

The integral in this section heading, which we denote I(β1 , β2 , β3 ) is an

extension of the ν = 0 case of Weber’s second integral, eq. (2.32), to which
it reduces when one of the three parameters vanishes.
By writing (see eq.(2.1))
√
√ √ X∞
(−1)n ( 12 β3 x)2n β22
J0 (β2 x)J0 (β3 x) = F
2 1 (−n, −n; 1; ) (3.1)
n=0 n!2 β32
and [15]
√ n! β2
Z ∞ 2
e−αx xn J0 (β1 x)dx = n+1 e−β1 /4α Ln ( 1 ), (3.2)
0 α 4α
Ln (x) being a Laguerre polynomial, we first have
1 − β12 X∞
(−1)n β32 n β2 β2
I(β1 , β2 , β3 ) = e 4α ( ) 2 F1 (−n, −n; 1; 22 )Ln ( 1 ) (3.3)
α n=0 n! 4α β3 4α
Now [16]
β22 β22 n β22
2 F1 (−n, −n; 1; ) = (1 − ) F
2 1 (−n, n + 1; 1; )
β32 β32 β22 − β32
β2 β 2 + β22 1 π Z
= (1 − 22 )n Pn ( 32 ) = (β 2 + β32 − 2β2 β3 cos θ)n dθ. (3.4)
β3 β3 − β22 πβ32n 0 2
Therefore, in terms of a well known generating function[17], (3.3) becomes
β 2
1
e− 4α Z π X
∞
(1/4α)n 2 β2
I(β1 , β2 , β3 ) = (β2 + β32 − 2β2 β3 cos θ)n Ln ( 1 )dθ
πα 0 n=0 n! 4α
2
e−β1 /4α Z π −R2 /4α β1 R
= e I0 ( )dθ (3.5)
πα 0 2α
where R = (β22 +β32 −2β2 β3 cos θ)1/2 . Next, by using Graf’s addition theorem
[18]
β1 R X∞
β1 β2 β1 β2
I0 ( )= (2 − δno )In ( )In ( ) cos(nθ) (3.6)
2α n=0 2α 2α

10
and the familiar integral representation
Z π β1 β2
cos θ β1 β2
e 2α cos(nθ)dθ = πIn ( ), (3.7)
0 2α
we finally have
Z ∞ √ √ √
e−αx J0 (β1 x)J0 (β2 x)J0 (β3 x)dx
0

1 − 1 (β12 +β22 +β32 ) X∞

= e 4α (2 − δ0n )Fn (β1 , β2 , β3 ) (3.8)
α n=0

where
β1 β2 β1 β3 β2 β3
Fn (β1 , β2 , β3 ) = In (
)In ( )In ( ) (3.9)
2α 2α 2α
We conclude by sketching the evaluation of the more general integral
Z ∞ √ √ √
Im (β1 , β2 , β3 ) = e−αx J0 (β1 x)Jm (β2 x)Jm (β3 x)dx (3.10)
0

where m is a positive integer. In the first place, (3.4) takes the form

1 β2 β3 m − β12
Im (β1 , β2 , β3 ) = ( ) e 4α ·
m!α 4α
∞
(−1)n β32 n β2 β2
( ) 2 F1 (−n, −n − m; m + 1; 22 )Lm+n ( 1 )
X
(3.11)
n=0 n! 4α β3 4α
By observing that [19]

β22 β22 n n!(2m)! m+1/2 β32 + β22

2 F1 (−n, −n − m; m + 1; ) = (1 − ) C ( 2 )
β32 β32 (2m + n)! n β3 − β22

22m m!2 Z π 2m
= sin θ (β22 + β32 − 2β2 β3 cos θ)n dθ (3.12)
(2m)!πβ32n 0
eq. (3.11) becomes

1 m! β2 β3 m −β12 /4α Z π 2m X ∞
(−1/4α)n
Im (β1 , β2 , β3 ) = ( ) e sin θ ·
πα (2m)! α 0 n=0 n!

11
 β12
(β22 + β32 − 2β2 β3 cos θ)n Lm+n ( )dθ. (3.13)
4α
Now, from [20],
n! d
Lm+n (x) = ex ( )m e−x xm Lm
n (x) (3.14)
(m + n)! dx
and [17]
∞
tn √
Lm −m/2 t
X
n (x) = (xt) e Jm (2 xt) (3.15)
n=0 (m + n)!

it follows that (where R has the same meaning as before)

∞
1 R2 β2 2 2 d
(− )n Lm+n ( 1 ) = e(β1 −R )/4α 2m ( )m e−x xm ·
X

n=0 n! 4α 4α dx
x −m x
r r
(R ) Im (R )| 2 . (3.16)
α α x=β1 /4α
By inserting (3.16) into (3.13) and using [21]
x −m x β2 β3 x −m ∞
r r
(m + n)Cnm (cos θ)·
X
(R ) Im (R )=( ) (m − 1)!
α α 2α n=0

x x
r r
Im+n (β2 )Im+n (β3 ) (3.18)
α α
and [22] Z π β2 β3
sin2m θ e 2α
cos θ
Cnm (cos θ)dθ
0
1−m
π2 (2m + n − 1)! 2α m β2 β3
= ( ) Im+n ( ), (3.18)
n!(m − 1)! β2 β3 2α
we finally obtain
Z ∞ √ √ √
e−αx J0 (β1 x)Jm (β2 x)Jm (β3 x)dx
0

1 22m+1 m! α m − β22 +β32 d m −x

= ( ) e 4α ( ) e · (3.19)
α (2m)! β2 β3 dx
∞
(n + m)(2m + n − 1)! x x β2 β3
r r
X
Im+n (β2 )Im+n (β3 )Im+n ( )| 2 .
n=0 n! α α 2α x=β1 /4α

12
In particular, by letting β3 → 0, after dividing by β3m ,
Z ∞ √ √
e−αx J0 (β1 x)Jm (β2 x)xm/2 dx
0

1 d x x
r
2
= e−β2 /4α ( )m e−x ( )m/2 Im (β2 )| 2 (3.20)
α dx α α x=β1 /4α
m
!
1 β2 m − β12 +β22 X n m β1 n β1 β2
= ( ) e 4α (−1) ( ) In ( ).
α 2α n=0 n β2 2α
which is of interest in connection with formulas (39)-(42) on page 186 of
reference [15].
In conclusion, we point out that the derivation of (2.24) can be extended
to give

Z ∞ 1
e−βx 0 F3 (µ, ν, ν + ; −a2 x2 )dx = (2a)1−µ Γ(µ)Γ(2ν)β µ−2ν−1 Jµ−1 (4a/β)
0 2
(3.21)
Therefore, the results of section 2, for example, are capable of extension
in a variety of directions. We leave this for the future and merely quote
one example, interesting because it contains each of the four types of Bessel
functions:
Z ∞
xJ1 (ax)I1 (ax)Y0 (x)K0 (x) dx = −(2πa2 )−1 ln(1 − a4 ) (3.22)
0

where 0 < a < 1.

References

[1] T.J. Lardner, Siam Review 11, 69 (1969).

[2] Bateman Manuscript Project, Higher Transcendental Functions (McGraw-
Hill, 1953), Vol. 2, p.11, eq. (47).
[3] Ibid. Vol. 1, p.64, eq. (23).
[4] Ibid. Vol. 2, p.52, eq. (31).
[5] Ibid. Vol. 2, p.51, eq. (29).
[6] L.J. Slater, Generalized Hypergeometric Functions (Cambridge 1966),
Sec. 4.8.
[7] I.N. Sneddon, The Use of Integral Transforms (McGraw-Hill 1972), Chap.
5, p.309.

13
[8] See Ref.[2] Vol. 2, p.96, eq. (59).
[9] Ibid. Vol. 2, p.81, eq. (10).
[10] See Ref. [1], p. 110, eq. (1).
[11] H.M. Srivastava and H.L. Manocha, A treatise on Generating Functions
(John Wiley 1983), p. 101, eq. (6).
[12] See Ref. [2], Vol. 2, p.50, eq. (25).
[13] J.L. Burchnall and T.W. Chaundy, Quart. J. Math. Oxford Ser. 11,
249-270 (1940).
[14] See Ref. [11], p.36.
[15] Bateman Manuscript Project, Tables of Integral Transforms (McGraw-
Hill 1953), Vol. 2, p. 30, eq. (13).
[16] See Ref. [1], p. 157, eq. (15).
[17] See Ref. [2], p. 189, eq. (18).
[18] Ibid. Vol. 2, p.44, eq.(5).
[19] Ibid. Vol. 2, p.177, eq.(31).
[20] Ibid. Vol. 2, p.190, eq.(28).
[21] Ibid. Vol. 2, p.101, eq.(30).
[22] See Ref. [15], p.281, eq.(7).
[23] A.P. Prudnikov, Ya. A. Brychkov and O.I. Marichev, Integrals and Series
(Gordon and Breach, N.Y. 1989) Vol. 3 Sec. 7.16.3 Eq. (6).
[24] Ibid. Sec. 7.16.2 Eq.(15).

14