2020

Unit V Special Functions

Contents

1. Gamma and Beta functions

a. Definition and properties
b. Relation between Beta and Gamma functions, Duplication formula
c. Problems
2. Bessel function
a. Series solution of Bessel differential equation
b. Recurrence relations
c. Generating functions, Jacobi series
d. Bessel’s integral formula and orthogonality of Bessel functions
e. Problems

1 Department of Science & Humanities (Mathematics) PESU

 2020

Unit V Special Functions

Beta Gamma Functions

Introduction:

• The Gamma and Beta functions are defined in terms of improper definite integrals.
• These functions belong to the category of the special transcendental functions
• These functions are very useful in many areas like asymptotic series, Riemann-Zeta function,
number theory, etc.
• They also have many applications in Engineering and Physics.
• Many integrals which cannot be expressed in terms of elementary functions can be evaluated
in terms of beta and gamma functions.

Beta Function:

Beta function is a definite integral whose integrand depends on two variables.

It is also known as Euler’s integral of First kind.

1
Beta function is defined as b (m, n) = òx m-1
(1- x)n-1dx where m, n > 0
0

Properties of beta functions:

1. Symmetry: b (m, n) = b (n, m)

1
Proof: By definition, b (m, n) = òx m-1
(1- x)n-1dx ------------ (1)
0

Put x = 1- t in (1)

Then,
0 1
b (m, n) = ò (1 - t ) m -1 n -1
t  -dt  = ò t n-1 (1 - t )m-1dt = b (n, m)
1 0

2. Beta function in terms of trigonometric functions:

p
2
b (m, n) = 2 ò sin 2 m-1 x cos2n-1 xdx
0

1
Proof: By definition, b (m, n) = òx m-1
(1- x)n-1dx ------------ (1)
0

Put x = sin 2 q , dx = 2sin q cosq dq in (1)

2 Department of Science & Humanities (Mathematics) PESU

 2020

Unit V Special Functions

Then,
p
2
b (m, n) = ò (sin 2 q ) m -1 (1 - sin 2 q ) n -1 2sin q cos q dq
0

p
2
= 2 ò sin 2 m -1 q cos 2 n -1 q dq
0

p 1 q 1
Note: put 2m - 1 = p, 2n - 1 = q  m = , n=
2 2
p
2
1 æ p 1 q 1 ö
\ ò sin p q cosq q dq = b ç , ÷
0 2 è 2 2 ø

¥
x m-1
3. As an improper integral, b (m, n) = ò (1 x) mn
dx
0
1 1
Proof: Put x = , dx = - dy in (1)
1 y 
2
1 y

0
1 æ 1 ö
n-1 æ ö
Then, b (m, n) = ò m-1 ç
1- ÷ ç - 1 2 ÷dy
ç 1 y ÷
¥ 1 y è 1 y ø è   ø
n-1
¥
1 æ y ö 1
= ò m-1 ç ÷ dy
1 y  è 1 y ø 1 y 
2
0

¥
y n-1
=ò dy
1 y
mn
0
¥
x n-1
=ò dx (dummy var iable)
1 x 
mn
0

Gamma function:

Gamma function is a definite integral whose integrand depends on one variable. It is also known as
Euler’s integral of second kind. It is the generalization of factorial notation from integer values to real
numbers.
¥
Gamma function is defined as G(n) = òe -x
x n-1dx where n > 0 ---------------------- (2)
0

G(n) is also defined for negative non-integers.

3 Department of Science & Humanities (Mathematics) PESU
 2020

Unit V Special Functions

Put x = t 2, dx = 2tdt in (2)

¥ ¥
Then, G (n) = ò e - t  t 2 
n -1
2tdt = 2 ò e- t t 2 n-1dt
2 2

0 0

¥
Therefore, another form of gamma function is G(n) = 2 ò e- x x 2n-1dx
2

Evaluation of gamma function:

1. G(1) =1
Proof: Put n = 1in (2)
¥
¥ é e- x ù ¥
Then, G(1) = ò e x dx = ò e dx = ê ú = 1
- x 1-1 -x

0 0 ë -1 û0

2. G(n 1) = nG(n) where n is a real number. (Reduction formula)

¥
Proof: From the definition, G(n) = òe -x
x n-1dx
0

¥ ¥
Therefore, G(n 1) = ò e- x x (n1)-1dx = ò e- x x n dx
0 0

¥
æ x n e- x ö ¥

÷  n ò e x dx =0  nG(n)
- x n-1
=ç
è -1 ø0 0

= nG(n)

3. G(n 1) = n! where n is a positive integer.

Proof: From Result 2,

G ( n  1) = nG (n)
= n(n - 1)G (n - 1)

= n(n - 1)( n - 2)G(n - 2)

.
.
= n(n - 1)( n - 2)( n = 3)..........2.1.G(1) = n !

Note: G(n) is not defined when n is zero or a negative integer.

G(n 1)
4. When n is a negative non integer, G(n) =
n
Proof: we have G(n 1) = nG(n)

4 Department of Science & Humanities (Mathematics) PESU

 2020

Unit V Special Functions

G(n 1) G(n  2) G(n  3)

\ G(n) = = =
n n(n 1) n(n 1)(n  2)
G(n  k 1)
= ............ =
n(n 1)(n  2).....(n  k)

This formula is used to compute G(n) when n is a negative non-integer.

Some standard results:

æ1ö
1. Gç ÷ = p
è2ø
¥
Proof: we have, G(n) = 2 ò e- x x 2n-1dx
2

1 æ1ö ¥ ¥
Put n =  G ç ÷ = 2 ò e dx = 2 ò e- y dy
- x2 2

2 è2ø 0 0

2
æ æ 1 öö ¥ ¥ ¥ ¥

ç G ç ÷÷ = 2 ò e dx ´ 2 ò e dy = 4 ò òe  dxdy
- x2 - y2 - x 2 y 2

è è 2 øø 0 0 0 0

Put x = r cos q and y = r sin q , then x 2  y 2 = r 2 and dxdy = rdrdq

Since both x and y vary from 0 to ∞ , the region of integration is first quadrant and in first quadrant r
varies from 0 to ∞ and 𝜃 varies from 0 to .

Therefore,
p
2 ¥ ¥ ¥
æ æ 1 öö 2
p éæ 1 ö 2 ù
ò ò e- r rdrdq = 4. ò e- r rdr = 2p êç - ÷ e- r ú = p
2 2
G
ç ç 2 ÷÷ = 4
è è øø 0 0
20 ëè 2 ø û0

æ1ö
 Gç ÷ = p
è2ø

2. Relationship between beta and gamma functions:

G(m)G(n)
b (m, n) =
G(m  n)
¥ ¥
Proof: By definition, G(m) = 2 ò e - x2
dx and G(n) = 2 ò e- y y 2n-1dy
2
2m-1
x
0 0

¥ ¥
Therefore, G(m)G(n) = 2 ò e - x x 2 m-1dx ´ 2 ò e - y y 2 n -1dy
2 2

0 0

5 Department of Science & Humanities (Mathematics) PESU

 2020

Unit V Special Functions

¥¥
= 4ò ò e

- x2  y2  x 2 m -1 y 2 n -1dxdy
0 0

Put x = r cosq y = r sinq . Then dxdy = rdrdq

p
2 ¥
Then, G(m)G(n) = 4 ò ò e - r  r cos q   r sin q 
2 2 m -1 2 n -1
rdrdq
0 0

p
2 ¥
= 2 ò cos 2 m -1 q sin 2 n -1 q dq ´ 2 ò e - r r 2 m -1r 2 n -1rdr
2

0 0

dt
Put r 2 = t , rdr = in second integral, we get,
2
¥
1 - t m  n-1
dt = b (m, n) G  m  n 
2 ò0
G(m)G(n) = b (m, n).2. e t

G(m)G(n)
Thus, b (m, n) =
G(m  n)

æ p 1 ö p1 ¥

÷ = qa ò x e dx, p and q are positive constants.

p -ax q
3. G ç q

è q ø 0
¥
Proof: Consider, òx e p -ax q
dx
0

Put, ax q = y  qax q-1dx = dy

¥ ¥ ¥
dy dy
\ ò x p e-ax dx = ò òx e
q
x p e- y = p -y
q-1
0 0 qax q-1 0 æ yö q
qa ç ÷
èaø
p p1-q
æ y öq
¥
dy 1 ¥ æ yö
= ò ç ÷ e- y òç ÷
q
= e- y dy
è ø
0 a æ yö
1-
1
q
aq 0 è a ø
qa ç ÷
èaø
q- p-1 ¥ p-q1 - p1 ¥ p1
1 1 -1
=
aq
a q
ò y q -y
e dy = a
q
q
ò e- y y q
dy
0 0

1 ¥ p1
-1 1 æ p 1 ö
=  p1 òe -y
y q
dy =  p1
Gç
è q ø
÷
q 0 q
a q a q

6 Department of Science & Humanities (Mathematics) PESU

 2020

Unit V Special Functions

p
4. G  n  G 1- n  = ,0 < n <1
sin np
Proof: We know that, G  n  G  m = b (n, m)G(m  n)

Put m = 1- n  G  n  G 1- n  = b (n,1- n)G(1)

 G  n  G 1- n  = b (n,1- n)

¥
x m -1
Also, we know that, b (m, n) = ò dx
0
(1  x) m  n

¥
x n-1 p
Therefore, b (n,1 - n) = ò dx =
0
(1  x) sin np

p
2
1 æ 1 n 1 ö
5. ò sin q dq = 2 b çè 2 ,
n

2 ø
÷
0
p
2
Proof: We know that, b ( p, q) = 2 ò cos 2 p-1 q sin 2q-1 q dq
0

p
1 æ1 ö 2
For p = , b ç , q ÷ = 2 ò sin 2 q-1 q dq
2 è2 ø 0

n 1
Let 2q -1 = n  q = .
2
p
æ 1 n 1 ö 2
Then, b ç , ÷ = 2 ò sin q dq
n

è2 2 ø 0

p p
2
1 æ 1 n 1 ö 2
Therefore, ò sin n q dq = b ç , ÷= ò cos qdq
n

0 2 è2 2 ø 0

p
2
1 æ p 1 q 1 ö
Also, ò sin q cos q dq = 2 b çè
p q

2
,
2 ø
÷
0

7 Department of Science & Humanities (Mathematics) PESU

 2020

Unit V Special Functions

æ 1ö
6. Legendre duplication formula for gamma function: p G  2m  = 2 2 m-1 G  m G ç m  ÷
è 2ø
p
2
Proof: We know that, b (m, n) = 2 ò sin 2 m-1 q cos2 n-1 q dq
0

p
2
Therefore, b (m, m) = 2 ò sin 2 m -1 q cos 2 m -1 q dq
0

p p

sin 2 m -1 2q
2
2 2 2 m -1
= 2ò
22 m -1 ò0
d q = sin 2q dq
0
22 m -1

Let 2q = t  2dq = dt .
p
p
2 dt 2 2 1 æ1 ö
Then, b (m, m) = 2 m-1 ò sin 2 m -1 t = = 2 m -1 ò sin 2 m -1 tdt = 2 m -1 b ç , m ÷
2 0
2 2 0
2 è2 ø

G ( m )G ( m ) 1 G ( 1 )G ( m)
i.e. = 2 m-1 2
G(2m) 2 æ 1ö
Gçm  ÷
è 2ø

G ( m) 1 p
 = 2 m -1
G(2m) 2 æ 1ö
Gçm  ÷
è 2ø
æ 1ö
 p G(2m) = 22 m -1 G( m)G ç m  ÷
è 2ø

7. b (m, n) = b (m, n 1)  b (m 1, n)

1 1
Proof: RHS= ò x m-1 1- x  òx 1- x 
n1-1 n-1
dx  m1-1
dx
0 0

1 1 1
= ò [x m-1 1- x  dx  ò x 1- x  ò 1- x  [1- x  x ] dx
n n-1 n-1 m-1
m
]dx = x
0 0 0
1

ò x 1- x  dx = b  m, n = LHS
n-1
= m-1

(m -1)!(n -1)!
8. b (m, n) =
 m  n -1!
Proof: We know that, b (m, n) = b (m, n 1)  b (m 1, n)

8 Department of Science & Humanities (Mathematics) PESU

 2020

Unit V Special Functions

G(m 1)G(n) G(m)G(n 1)

= 
G(m  n 1) G(m  n 1)
m! n -1!  m -1!n! m! n -1!  m -1!n!
=  =
 m  n !  m  n !  m  n !
=
 m -1! n -1 {m  n} =  m -1! n -1
 m  n !  m  n -1!
1
1 æ p 1 ö
ò x 1 - x 
q r
9. p
dx = bç , r  1÷
0
q è q ø
Proof: Take x q = y  qx q -1dx = dy

1 p 1 p-q1
dy 1
ò y q 1- y  ò 1- y
r q r
LHS = q-1
= y dy
0 q
q 0
qy

1 æ p 1 ö
= bç , r 1÷ = RHS
q è q ø

SOLVED PROBLEMS:
2p
1. Evaluate: ò sin q dq
8

0
2
p æ9ö æ1ö 7 5 3 1 æ æ 1 öö
G ç ÷ G ç ÷ ç G ç ÷÷
2p 2
1 æ9 1ö è2ø è2ø 2 2 2 2 è è 2 øø 35p
Ans: ò sin q dq = 4 ò sin q dq = 4. b ç , ÷ = 2 = 2. =
8 8

0 0 2 è2 2ø G  5 4! 2

p
2
2. Evaluate: ò sin q cos q dq
4 5

0
æ 5ö 131
p
G ç ÷ G  3 p2
2
1 æ5 ö 1 è2ø 8
Ans: ò sin 4 q cos5 q dq = b ç , 3÷ = = 222 =
2 è 2 ø 2 G æ 11 ö 97531
p 315
0
ç ÷
è2ø 22222

p
2
3. Evaluate: ò cot q dq
0
p p p 1 p

æ cos q ö
2 2 1 2 2 1 1
- 2
Ans: ò
0
cot q dq = ò cot q dq = ò ç
0 0è
2
÷ dq = ò sin q cos q dq
sin q ø 0
2 2

9 Department of Science & Humanities (Mathematics) PESU

 2020

Unit V Special Functions

æ1ö æ 3ö
Gç ÷Gç ÷
1 æ1 3ö 1 è 4ø è 4ø 1 p p
= bç , ÷= = =
2 è 4 4ø 2 G 1 2 æp ö 2
sin ç ÷
è4ø
1
1- x
4. Evaluate: ò0
x
dx
2
æ1ö æ3ö æ 1 ö 1 æ 1 ö 1 æGæ 1 öö
Gç ÷Gç ÷ Gç ÷ Gç ÷ ç ç ÷÷
1
1- x
1 1 1
æ1 3ö 2 2 è 2 ø 2 è è 2 øø p
dx = ò x 2 1 - x  2 dx = b ç , ÷ = è ø è ø = è ø
- 2 2
Ans: ò
0
x 0 è2 2ø G  2 G  2
=
1
=
2

¥
dx
5. Evaluate: ò 1 x 4
0
2 2
Ans: take x 2 = tan q , 2xdx = sec 2 q dq  dx = sec2qxdq = sec q dq
2 tan q

p p p p
¥
dx 2
1 se c 2 q dq 2 1 se c 2 q dq 2 dq 12 1
 
-
ò0 1  x 4 ò0 1  tan 2 q 2 tan q ò0 se c2 q 2 tan q ò0 2 tan q 2 ò0
= = = = tan q 2 dq

p p æ1ö æ3ö
-
1 Gç ÷Gç ÷
1 æ sin q ö
2 2 1 2
-
1 1
1 1 æ1 3ö 1 è4ø è4ø
2 ò0 è cos q ø
= ç ÷ dq = ò sin q cos q dq =
2
bç , ÷=
2
20 2 2 è4 4ø 4 G 1
1 p p
= =
4 sin p 2 2
4

BESSEL FUNCTIONS

One of the most important differential equations in applied mathematics is,

x y  xy    x - n
2 2 2
 y = 0- - - - (1) which is known as Bessel’s differential equation of order n.

We employ generalized power series method to find solution in the form of power series of the above
differential equation.
Let the power series solution of (1) is of the form
¥
y =  ar x r  k - - - - - (2) where k is a positive real number and ar ' s are constants and a0  0 .
r =0

Differentiating (2) wrt x, we get

dy ¥ d2y ¥
=  ar  r  k x r  k -1 - - - - - (3) and 2
=  ar  r  k  r  k - 1 x r  k - 2 - - - - - (4)
dx r =0 dx r =0

Substituting equation (2), (3) and (4) in equation (1) and simplifying, we get

 r  k  
¥ ¥

a x - n 2   ar x r  k = 0
r k 2
r
r =0 r =0

10 Department of Science & Humanities (Mathematics) PESU

 2020

Unit V Special Functions

Equating the co-efficients of x k and x k 1 to zero, we get

a0  k 2 - n 2  = 0
Since a0  0 , k 2 - n 2 = 0  k =  n


Also, a1  k  1 - n 2 = 0  a1 = 0
2

Now, by equating the higher powers of x to zero, we get
ar - 2
ar = - - - - - - - - -(5)
 r  k  - n2
2

Now Let us discuss two cases.

Case (i ) k = n
Put k = n in equation (5)
a
Then we get, ar = - 2 r - 2 - - - - - - - -(6)
r  2nr
Put r = 1, 2,3, 4,........ equation (6), we get
 -1
1
a0
a2 = - - - - - - - - -(7)
2 (n  1)1!
2

 -1
2
a0
a4 = - - - - - - - - -(8)
24 (n  2)(n  1)2!
And ar = 0 for r = 1,3, 5, …
1
Writing the values of ar 's in eqn (2) and standardizing the result by taking a0 = , we get,
2 G(n  1)
n

¥ n 2 r
æ xö 1 1
y1 =  (-1) ç ÷
r
---------(9)
r =0 è2ø G  n  r  1 r !
(Since this is one solution of equation (1) it is called y1 ).
Equation (9) is a series solution of Bessel differential equation and it is called Bessel function of order
n. It is denoted as J n ( x) .
Particular solutions of Bessel’s equation of order n are called Bessel functions or cylindrical functions

of order n denoted by , J n ( x ) = 
¥
- 1r
æ xö
n2r

ç ÷
r = 0 r!G ( n  r  1) è 2 ø

Case(2): Put k = - n we get J - n ( x) .

 -1
r - n2r
¥
æxö
Therefore, J - n ( x ) =  ç ÷
r = 0 r ! G ( - n  r  1) è 2 ø

The complete solution of Bessel differential equation is

y = C1 J n ( x)  C2 J - n ( x) , provided n is a non- integer.
When n is an integer we get a complete solution of (1) in the form,
y = C1Yn ( x)  C2Y- n ( x )
Where Yn ( x) is called Bessel function of second kind of order n and is given by,

11 Department of Science & Humanities (Mathematics) PESU

 2020

Unit V Special Functions

dx
Yn ( x ) = J n ( x ) ò
x  J n ( x) 
2

Note: J -n ( x) = (-1) n J n ( x) when ‘n’ is an integer. i.e., when n is an integer, J - n ( x) and J n ( x ) are not
linearly independent.

Proof: We know that J n ( x) = 

æ xö ¥
- 1r n 2 r

ç ÷
r =0 r!G ( n  r  1) è 2 ø

Replacing n by –n, we get,

J -n ( x) = 
¥
- 1r æ xö
- n 2 r

ç ÷
r =0 r!G ( - n  r  1) è 2 ø

Because n is an integer and r is an integer, - n  r  1 is also an integer.

G ( - n  r  1) = r - n !

 - 1 æ x ö
r - n 2r
¥
J - n ( x) =  ç ÷ r - n  0
r = n r ! r - n  ! è 2 ø

Put r - n = s then, r = n  s and -n  2r = n  2s

Therefore, J -n ( x) = 
¥
æ xö- 1n s n 2 s

ç ÷
s =0 ( n  s )! s!è 2 ø

 -1
s n 2 s
¥
æ xö
= (-1)  n
ç ÷
s = 0 G( n  s  1) s ! è 2 ø

Thus, J -n ( x) = (-1) n J n ( x) when n is an integer.

Recurrence relations:

We will now learn some important properties of Bessel functions which are called recurrence relations.
Bessel Functions of higher order be expressed by Bessel functions of lower order for all real values of
‘n’.

d
Property (1): x J n ( x ) =n J n ( x) - x J n 1 ( x )
dx

Proof: Consider, J n ( x ) = 
¥
æ xö - 1r n 2 r

-------- (1)
ç ÷
r = 0 r!G ( n  r  1) è 2 ø

12 Department of Science & Humanities (Mathematics) PESU

 2020

Unit V Special Functions

Differentiating wrt x on both sides of equation (1) and then multiplying by x and splitting the
summation into two summations, we get

 -1  -1
r n2r r n 1 2 r
¥ ¥
d æxö æxö
x. J n ( x) = n ç ÷ - x. ç ÷
dx r = 0 r ! G ( n  r  1) è 2 ø r = 0 r ! G ( n  1  r  1) è 2 ø

d
i.e. x. J n ( x ) = n J n ( x ) - x J n 1 ( x )
dx

d
Property (2): x J n ( x) = x J n -1 ( x ) - n J n ( x )
dx

Proof: Consider, J n ( x) = 
¥
- 1r
æ xö
n 2r

-------- (1)
ç ÷
r = 0 r! G ( n  r  1) è 2 ø

Differentiate wrt x and multiply with x , we get,

d ¥
- 1r n  2r æ x ö n  2 r -1 . x
x.
dx
J n ( x) =  r!G(n  r  1) çè 2 ÷ø
r =0 2

Now, writing n+2r =2(n+r) – n and splitting the summation, we get

 -1  -1
r n  2 r -1 r n2r
¥ ¥
d æxö æxö
x. J n ( x) = x  ç ÷ - n. ç ÷
dx r = 0 r !G( n - 1  r  1) è 2 ø r = 0 r !G( n  r  1) è 2 ø

d
i.e. x J n ( x ) = x J n -1 ( x) - n J n ( x )
dx

The remaining four results can be proved using these results.

Property (3): 2 n J n ( x) = xJ n -1 ( x)  J n 1 ( x) 

Proof: consider property (1) and (2)

d d
x J n ( x ) =n J n ( x) - x J n 1 ( x ) and x J n ( x ) = x J n -1 ( x ) - n J n ( x )
dx dx

Subtracting one equation from the other, we get,

0 = - x( J n -1 ( x)  J n 1 ( x))  2n J n ( x)

i.e. 2 n J n ( x) = xJ n -1 ( x)  J n 1 ( x) 

13 Department of Science & Humanities (Mathematics) PESU

 2020

Unit V Special Functions

d 1
Property (4): J n ( x ) =  J n -1 ( x) - J n 1 ( x) 
dx 2

Proof: consider property (1) and (2)

d d
x J n ( x) =n J n ( x ) - x J n1 ( x ) and x J n ( x ) = x J n-1 ( x) - n J n ( x )
dx dx

Adding the above equations, we get

d
2x J n (x) =x( J n-1 ( x) - J n1 ( x )
dx

d 1
This implies J n ( x ) =  J n -1 ( x) - J n 1 ( x) 
dx 2

Property 5:
d n
dx
 
x J n ( x) = x n J n-1 ( x)

Proof: Consider,
d n
dx

x J n (x) = x n
d

dx
( J n (x) ) + J n (x) n x n-1

d
Using property (2), x J n ( x) = x J n-1 ( x) - n J n ( x) in the above equation,
dx

We get
d n
dx
  n
x J n ( x) = x n ( J n-1 ( x) - J n ( x) )  J n ( x) nx n-1
x

d n
Therefore,
dx
 x Jn (x) = xn Jn-1(x)

d -n
Propery (6):
dx
 x J n ( x )  = - x - n J n1 ( x )

Proof: Consider,
dx

d -n
x J n (x) = x - nd
dx
( J n (x) ) + J n (x) (-n x -n-1 )

d
Using property (1), x J n ( x) =n J n ( x) - x J n1 ( x)
dx

d -n
dx
 n
x J n (x) = x - n ( J n ( x) - J n1 ( x) ) + J n (x) (-n x - n-1 )
x

Hence,
d -n
dx
 
x J n ( x) = - x - n J n 1 ( x)

14 Department of Science & Humanities (Mathematics) PESU

 2020

Unit V Special Functions

Generating function:

Generating function of a sequence of functions f n (x) is,

¥
G ( x, t ) =  f n ( x )t n which generates f n (x) i.e., f n (x) appear as coefficients of various powers of t.
-¥

¥ x
t -1t 
Prove that  J n ( x)t n = e 2
n = -¥
or Prove that Generating function for Bessel function of integral order

x t - 1t 
1
is e 2 .
xt -x
x
t - 1t 
Proof: e 2 = e 2 . e 2t

Expanding the exponential functions,

æ æ xt ö æ xt ö 2 æ xt ö3 n
æ xt ö æ xt ö
n 1
ö
1
x t - 1t 
ç ç ÷ ç ÷ ç ÷ ç ÷ ç ÷ ÷
e 2 ç
= 1 è 2ø è2ø è2ø
  ------ è 2ø è2ø
  - - - - -÷´
ç 1! 2! 3! n! (n  1)! ÷
ç ÷
è ø

æ æ x ö æ x ö 2 æ x ö3 n
æ xö æ xö
n 1
ö
ç ç ÷ ç ÷ ç ÷ ç ÷ ç ÷ ÷
ç 1 - è 2t ø  è 2t ø - è 2t ø  - - - - - - -  è 2t ø - è 2t ø  - - - - - - ÷
ç 1! 2! 3! n! (n  1)! ÷
ç ÷
è ø

Case (i): positive powers of t

Collecting only co-efficient of t n in the expansion, we get

æ æ x ö n æ x ö n2 æ xö
n4
æxö
n6
ö
çç ÷ ç ÷ ç ÷ ç ÷ ÷
çè2ø è 2ø è 2ø è 2ø ÷
Co-efficient of t = ç
n
-  -  - - - - -÷
n! 1!(n  1)! 2!(n  2)! 3!(n  3)!
ç ÷
ç ÷
è ø
n2 r
æ xö
¥
( -1) r ç ÷
= è2ø
r =0 r! G ( n  r  1)

i.e. Co-efficient t n = J n (x)

Case (ii): negative powers of t

15 Department of Science & Humanities (Mathematics) PESU
 2020

Unit V Special Functions

Collecting coefficients of t - n in the expansion

We get, coefficients of t - n = (-1) n J n ( x)

Case (iii): collecting terms independent of t (n=0)

æ æ x ö2 æ x ö4 æ xö
2n
ö
ç ç ÷ ç ÷ ç ÷ ÷
ç è2ø è 2ø è 2ø ÷
Terms independent of t = ç1 - 2
- 2
---- 2
- - - - -÷
(1!) (2! ) (n! )
ç ÷
ç ÷
è ø
0 2 r
æxö
¥
(-1) r ç ÷
= è2ø = J 0 ( x)
r = 0 r! G (0  r  1)

Combining the results of three cases,

1
 
x t - 1t ¥
We get e 2
=  J n ( x)t n
n = -¥

Bessel functions of various orders can be obtained as co-efficient of various powers of t on the power
x t - 1t 
1
series expansion of e 2
.

Important outcomes of Generating Function:

1. cos( x sin q ) = J 0  2  J 2 cos 2q  J 4 cos 4q  ..........

2. sin( x sin q ) = 2  J1 sin q  J 3 sin 3q  .J 5 sin 5q .........

p
1
cos  nq - x sin q  dq (integral form of Bessel function)
p ò0
3. J n ( x ) =

Let us first derive results (1) and (2) which are called Jacobi Series.

Jacobi Series:

cos( x sin q ) = J 0  2  J 2 cos 2q  J 4 cos 4q  ..........

sin( x sin q ) = 2  J1 sin q  J 3 sin 3q  .J 5 sin 5q .........

1
 
x t - 1t ¥
Proof: we know that e 2 =  J n ( x)t n
n = -¥

16 Department of Science & Humanities (Mathematics) PESU

 2020

Unit V Special Functions

Expanding the summation,

1
x  t - 1t 
e2 = J 0 ( x )  tJ1 ( x )  t 2 J 2 ( x )  t 3 J 3 ( x )  - - - - - - -  t -1 J -1 ( x )  t -2 J -2 ( x )  t -3 J -3 ( x )  - - - -
= J 0 ( x )  (t - t -1 ) J1 ( x )  (t 2  t -2 ) J 2 ( x )  ( t 3 - t -3 ) J 3 ( x )  - - - - - - (1)
( when n is an i nteger , J - n ( x ) = ( -1) n J n ( x ))

1
Let t = cos q  i sin q , then = cos q - i sin q
t
1 1
t  = 2 cos q ; t - = 2i sin q
t t
1 1
t n  n = 2 cos nq ; t n - n = 2i sin nq
t t

Using in eqn (1), we get

1
x (2 i sin q )
e2 = J 0 ( x )  2( J 2 cos 2q  J 4 cos 4q  - - - - - )  2i ( J1 sin q  J 3 sin 3q  - - - - - )

i.e. cos( x sin q )  i sin( x sin q ) = J 0 ( x )  2( J 2 cos 2q  J 4 cos 4q  --)  2i( J1 sin q  J 3 sin 3q  - - -)

Equating the real and imaginary parts, we get

cos( x sin q ) = J 0  2J 2 cos 2q  J 4 cos 4q  ..........

sin( x sin q ) = 2J 1 sin q  J 3 sin 3q  .J 5 sin 5q .........

Integral form of Bessel Function:

p
1
J n ( x) =
p ò cosnq - x sinq dq
0

Proof: Consider Jacobi Series

cos( x sin q ) = J 0  2  J 2 cos 2q  J 4 cos 4q  .......... - - - -(1)

sin( x sin q ) = 2  J1 sin q  J 3 sin 3q  .J 5 sin 5q ......... - - - -(2)

Let n be an even number.

Multiply equation (1) with cos nq and integrate wrt q between the limits 0 to p

Multiply equation (2) with sin nq and integrate wrt q between the limits 0 to p

17 Department of Science & Humanities (Mathematics) PESU

 2020

Unit V Special Functions

p p
Adding the equations, we get LHS = ò cos( x sin q ) cos nq  sin( x sin q ) sin nq = ò cos(nq - x sin q ) dq
0 0
p p
RHS = ò J 0 cos nq dq  2ò  J 2 cos 2q  J 4 cos 4q  ........ cos nq dq
0 0
p
 2ò  J1 sin q  J 3 sin 3q  .J 5 sin 5q ......... sin nq dq
0

p æp ö
ç when n = m, ÷
Using important result: ò cos nq cos mq dq = 2
ç ÷
è otherwise = 0 ø
0

ò cos nq sin mq dq = 0 always

0

p
Therefore, RHS = J 0 .0  2 J n .  0 = J np
2
p
1
Equating LHS and RHS, we get, J n ( x) =
p ò cosnq - x sin q dq
0

Orthogonality of Bessel Functions:

If  and b are the roots of the equation Jn (ax) = 0 , then prove that

a 0 when   b

ò0 J n ( x) J n (b x) dx =  a 2 J 2  a  when  = b
 n 1
2

Proof: Since  and b are the roots of the equation J n (ax) = 0 , Jn ( a) = 0 and J n ( b a ) = 0 ---- (1)

Consider the Bessel’s equations,

x 2 u   x u   ( 2 x 2 - n 2 )u = 0 - - - -(2)

x 2 v  xv  ( b 2 x 2 - n 2 )v = 0 - - - -(3)

The solutions of equations (2) and (3) are,

u = J n (x) and v = J n ( b x) (Equations reducible to Bessel form)

18 Department of Science & Humanities (Mathematics) PESU

 2020

Unit V Special Functions

Differentiating wrt x,

u = J n ( x). and v = J n ( b x).b

u = J n( x). 2 and v = J n(b x).b 2

v u
Multiplying equation (2) by and equation (3) by we get,
x x

uv
x u v  u v  ( 2 x 2 - n 2 ) = 0 - - - -(4)
x
uv
xvu  uv  ( b 2 x 2 - n 2 ) = 0 - - - -(5)
x

(4) - (5)  x  uv - uv   (uv - uv)  ( 2 - b 2 )uvx = 0

d
  x(uv - uv)  = (b 2 -  2 )uvx
dx
1 d
 uvx =  x(uv - uv) 
( b -  ) dx
2 2

Integrating wrt x between the limits 0 to a , we get,

a a
1 d
ò xJ n ( x) J n ( b x) dx = 2 ò  x(uv - uv)  dx
0
( b -  ) 0 dx
2

1
2 
x (u v - uv) 0
a
=
(b -  )
2

a( J n ( a). .J n ( b a ) - J n ( b a).J n ( a ).b )

=
(b 2 -  2 )
=0  when   b  because J n ( a ) = 0 and J n ( b a) = 0 

0
However, when  = b , the value of the integral is which is indeterminate.
0

To evaluate the integral, consider  as constant and b as a variable which approaches 

Applying limits on both sides

lim a lim æ a ö
2 ÷ n
ò xJ n (x) J n ( b x) dx = ç 2 J  ( a ).J n ( b a ) - 0 
b  0 b   è (b -  ) ø

Applying L’Hospital rule on RHS

19 Department of Science & Humanities (Mathematics) PESU

 2020

Unit V Special Functions

a
lim a 2 ( J n ( a ). .J n ( b a )
ò xJ n (x) dx =
2

0
b  2b

a 2 ( J n (a )) 2
= ---------(2)
2

d
But x J n ( x) =n J n ( x) - x J n1 ( x)
dx

Therefore, a J n ( a ) = n J n (a ) - a J n 1 (a )

But J n (a ) = 0  J n ( a ) = - J n 1 (a )

Using in equation (2)

a 2 ( J n1 (a )) 2
a

ò xJ n (x) dx = when  = b
2

0
2

Combining both the results, we get

a 0 when  b

ò0 xJ n (x)J n (bx)dx =  a 2 J 2 a when = b
 n 1
2
PROBLEMS

1. Using the values of J ( x ) and J - ( x ) find J ( x ) and J - ( x )

1
2
1
2
3
2
3
2

2 2
. Ans: We know that J 1 ( x) = sin( x) and J 1 ( x) = cos( x)
2
px -
2 p x

We have recurrence property, 2nJn (x) =x Jn-1(x) J n1(x)  ---------- (1)

1 æ ö
Put n = in equation (1), we get, J 1 (x) =x çç J -1 (x) J 3 (x) ÷÷
2 2 è 2 2 ø

Using the expressions for J 1 ( x) and J -1 ( x) , we get

2 2

2 æ sin x ö
J 3 ( x) = ç - cos x ÷
2 px è x ø

20 Department of Science & Humanities (Mathematics) PESU

 2020

Unit V Special Functions

1 2 æ cos x ö
Similarly, put n = - in equation (1), we get, J -3 ( x) = - ç sin x  ÷
2 2
px è x ø

2. Express J5 in terms of J0 and J1 .

Ans: We have recurrence property, 2nJ n (x) =xJ n-1(x) J n1(x)  ---------- (1)

1
Put n = 1 in equation (1), we get J2 (x) = 2J1(x) - xJ0 (x)  ---------------- (2)
x

1
Put n = 2 in equation (1), we get J3 (x) = 4J 2 (x) - xJ1(x)  ----------------- (3)
x

1
Put n=3 in eqn (1), we get J 4 (x) = 6J3 (x) - xJ2 (x)  --------------- (4)
x

Put n=4 in eqn (1), we get xJ5 (x) =8J4 (x) - xJ3 (x)  ------------------ (5)

æ 1 ö
Using eqn (4) in eqn (5), we get, xJ5 (x) =ç8 6J3 (x) - xJ 2 (x)  - xJ3 (x) ÷
è x ø

æ 48 ö
Therefore, J5 (x) =ç - x ÷J3 (x) - 8 J 2
è x ø

Using eqn (3),

æ 48 ö 1
J5 (x) = ç - x ÷  4J2 (x) - xJ1(x)  -8 J2
è x øx

æ 384 72 ö æ 12 192 ö
Using eqn (2) and simplifying, we get J 5 ( x) = ç 4 - 2 - 1÷ J1  ç - 3 ÷ J 0
è x x ø è x x ø

2
3. Prove that, ò J 3 ( x)dx = c - J 2 ( x) - J1 ( x)
x

d -n
Ans: We know that,
dx
 x J n ( x )  = - x - n J n 1 ( x )

\ ò x - n J n1 ( x)dx = -x - n J n ( x)

ò J ( x)dx = ò x x
-2
3
2
J 3 ( x) dx  c

21 Department of Science & Humanities (Mathematics) PESU

 2020

Unit V Special Functions

= x 2 ò x -2 J 3 ( x) dx - ò 2 x é ò x -2 J 3 ( x )dx ù dx  c
ë û

= x 2  - x -2 J 2 ( x )  - ò 2 x éë - x -2 J 2 ( x) ùû dx  c

2
= c - J 2 ( x)  ò J 2 ( x)dx
x

2
= c - J 2 ( x) - J1 ( x )
x

4. Prove that,∫ 𝑥 𝐽 (𝑥)𝑑𝑥 = 𝑥 𝐽 (𝑥) + 𝐽 (𝑥)

Ans: ∫ 𝑥𝐽 (𝑥)𝑑𝑥 = 𝐽 (𝑥) 𝑥 − ∫ 2𝐽 (𝑥)𝐽 (𝑥) 𝑥 𝑑𝑥

1 2 2
= x J 0 ( x) - ò x 2 J 0 ( x) J 0 ( x)dx
2

1 2 2
= x J 0 ( x)  ò x 2 J 0 ( x) J1 ( x)dx
2

= 𝑥 𝐽 (𝑥) + ∫ 𝑥𝐽 (𝑥)𝑥𝐽 (𝑥)𝑑𝑥

= 𝑥 𝐽 (𝑥) + ∫ (𝑥𝐽 (𝑥))𝑥𝐽 (𝑥)𝑑𝑥
= 𝑥 𝐽 (𝑥) + 𝑥𝐽 (𝑥)
= 𝑥 𝐽 (𝑥) + 𝐽 (𝑥)

********************************************

22 Department of Science & Humanities (Mathematics) PESU