Bessel Functions

Series solution of differential equation:

Let the differential equation y// + p(x)y/ + q(x)y = 0.
If p(x) and q(x) are analytic at a point x = a (i.e. p(x) and q(x) has
Taylor's series expansion about the point x = a). Then the point
x = a is called an ordinary point of the equation. (If a point is not
an ordinary point then it is a singular point.)
If p(x) and q(x) are not analytic at x = a but (x – a)p(x) and
(x – a)2q(x) are analytic then x = a is called a regular singular point
of the equation. If the equation has a regular singular point at x = a
then there exists at least one series solution of the form

y = (x – a)s a
n 0
n ( x  a) n ; a0  0 .

Bessel's equation and its solution: The differential equation of

the form x2y// + xy/ + (x2 – n2)y = 0 ....... (1)
is called Bessel's equation of order n, n being a non-negative
constant. Now we solve the Bessel's equation in series by using the
method of Frobenius. Since x = 0 is a regular singular point of
equation (1), there exists at least one solution of the form

y = xs a
m 0
m x m ; a0  0 ...... (2)

Or, y= a
m 0
m x m  s ; a0  0

Or, y/ =  (m  s)a
m 0
m x m s 1

Or, y// =  (m  s)(m  s  1)a
m 0
m x m s 2

Substituting in equation (1), we have

1
  

 (m  s)(m  s  1)am x m s +  (m  s)am x ms

m 0 m 0
 
+ a
m 0
m x m s 2 – n
m 0
2
am x m s = 0
 
Or, {(m  s)(m  s  1)  (m  s)  n 2 }am x ms +  am x ms2 = 0
m 0 m 0
 
Or, {(m  s)
m 0
2
 n 2 }a m x m s +  a m x m s  2 = 0 ; a0  0 .... (3)
m 0
Equating the terms with lowest power of x to zero, we get the
indicial equation
(s2 – n2) a 0 = 0 is the indicial equation.
Since a0  0 so, s2 – n2 = 0 s =  n
Equating the coefficients of x s 1 to zero, we have
{(s + 1)2 – n2} a1 = 0
For s =  n , {(s + 1)2 – n2}  0  a1 = 0.
Equating the coefficients of x r  s where r  2
{(r  s) 2  n 2 }ar  ar 2  0
ar 2
Or, ar = – ...... (4)
(r  s) 2  n 2
Putting r = 3, 5, 7, ... in (4) and using a1 = 0, we get
a1 = a 3 = a 5 = a 7 = ... = 0
Taking s = n in (4), we have
ar 2 a r 2
ar = – =– ...... (5)
( r  n)  n
2 2
r ( r  2n)
Putting r = 2, 4, 6, ... in (4), we get
a0 (1)a0
a2 = – = 2
2(2  2n) 2 .1!(n  1)

2
 a2 a0 (1) 2 a0
a4 = – = 2 = 4
4(4  2n) 2 .1!(n  1)4(4  2n) 2 .2!(n  1)(n  2)
a4 (1) 3 a0
a6 = – = 6 and so on.
6(6  2n) 2 .3!(n  1)(n  2)(n  3)
Continuing we get, in general
(1) m a0
a2m = 2m
2 .m!(n  1)(n  2)(n  3)...(n  m)
1
If we particularly chose a 0 = n , then we write
2 (n  1)
1
a 2 m = (1) m 2 m n
2 .m!.(m  n  1)
Then (2) gives a particular solution as follows:
1 x2 x4
y = xn [ n – 2 n + 4 n – ...]
2 (n  1) 2 .1!.(n  2) 2 .2!.(n  3)

x 2 m n
Or, y =  (1) m
m 0 2 2 m n.m!.(m  n  1)
This particular solution of Bessel's equation is known as Bessel's
function J n (x) of order n of the first kind, and is usually expressed
 2 m n
1  x
as J n (x) =  (1) m
  .
m 0 m!.(m  n  1)  2 
Putting s = – n in (4) and proceeding as above gives another
particular solution
 2 mn
1  x
J n (x) =  (1) m
 
m 0 m!.(m  n  1)  2 
If n be an integer then J n (x) is a constant multiple of J n (x) . But
when n is not an integer, then J n (x) and J n (x) are the two

3
independent solutions of Bessel's equation. So, the general solution
of the Bessel's equation is
y = c1 . J n (x) + c 2 . J n (x)
where c1 , c 2 are arbitrary constants and n is not an integer.

Remark: Bessel's function of the first kind of order n is denoted

by J n (x) and is defined as
 2 m n
1  x
J n (x) =  (1) m
 
m 0 m!.(m  n  1)  2 
where n is any non-negative constant.
When n is an integer, (m  n  1) = (m + n)! and so
 2 mn
1  x
J n (x) =  (1) m
 
m 0 m!.(m  n)!  2 
Replacing n by 0 and 1, Bessel's functions of order 0 and 1 are
given by
x2 x4 x6
J 0 ( x) = 1 – 2 + 2 2 – 2 2 2 + ...
2 2 .4 2 .4 .6
3 3
x x x
J 1 ( x) = – 3 + 2 2 – ...
2 2 .3 2 .4 .6

Theorem (1): If n is an integer, then J n (x) = (1) n J n (x) .

Proof: Case-I: When n is a non-negative integer:
We know the Bessel's function
 2 m n
1  x
J n (x) =  (1) m  
m 0 m!.(m  n  1)  2 
 2 mn
1  x
and J n (x) =  (1) m
  .
m 0 m!.(m  n  1)  2 

4
Since  of a negative integer is  (infinity), therefore in that case
J n (x) = 0. Non-zero values of J n (x) exists only if m  n.
Let m = n + r ; r  0, then
 2r n
1  x
J n (x) =  (1) nr
 
r 0 (n  r )!.(r  1)  2 
 2r n
1  x
= (1) n  (1) r  
r 0 r!.(r  n  1)  2 
n
= (1) J n (x)
So, J n (x) = (1) n J n (x) for non-negative integer n.
Case-II: When n is a negative integer:
Let n = – p where p is a positive integer. So, p satisfies case-I.
i.e., J  p (x) = (1) p J p (x) so that J p (x) = (1)  p J  p (x) .
But p = – n, hence the above result becomes
J n (x) = (1) n J n (x)
Therefore, in both cases J n (x) = (1) n J n (x) holds. [Proved]

Remark: This theorem proves when n is an integer, J n (x) and

J n (x) are not independent solutions of the Bessel's equation.

Theorem (2): The two independent solutions of Bessel's equation

cos n J n ( x)  J n ( x)
may be taken to be J n (x) and Yn (x) = for
sin n
all values of n.
Proof: Case-I (When n is not an integer): Since n is not a
integer, sin n  0. Hence Yn (x) is a linear combination of
J n (x) and J n (x) . But we know that J n (x) and J n (x) are
independent solutions if n is not an integer. Hence J n (x) and a
linear combination of J n (x) and J n (x) are also two independent

5
solutions. Thus we find that J n (x) and Yn (x) are two independent
solutions of Bessel's equation.
Case-II (When in is an integer): As n is an integer, cos n =
(1) n , sin n = 0 and J n (x) = (1) n J n (x) . Using these values,
0
we find Yn (x) = is undefined. To make Yn (x) meaningful, we
0
define it as
Yn (x) = lt Yv (x)
vn

cos v J v ( x)  J v ( x)
= lt ........... (1)
vn sin v

{cos v J v ( x)  J v ( x)}
= lt v [By L'Hospital's rule]
vn 
(sin v )
v
 
[ sin v J v ( x)  cos v J v ( x)  J v ( x)]v n
= v v
[ cos v ]v n
 
(1) n [ J v ( x)]v n  (1) 2 n [ J v ( x)]v n
= v v
 (1) n

1  
 Yn (x) =  J n ( x)  (1) n J n ( x) ......... (2)
  n n 
We now establish the following two results about Yn (x) as given
by (1):-
(i) Yn (x) is a solution of Bessel's equation.
(ii) Yn (x) is a solution independent of J n (x) .
Proof of (i): Since J v (x) and J v (x) are solutions of Bessel's
equation of order v, we must have

6
 d 2Jv dJ
x2 2
 x v  ( x 2  v 2 ) J v = 0 ......... (3)
dx dx
2
d J v dJ
and x2 2
 x v  ( x 2  v 2 ) J v = 0 ....... (4)
dx dx
Differentiating (3) and (4) with respect to v gives
d 2 J d J J
x 2 2 ( v )  x ( v )  ( x 2  v 2 ) v  2vJ v = 0 ........ (5)
dx v dx v v
d 2
J d J J
x 2 2 ( v )  x ( v )  ( x 2  v 2 ) v  2vJ v = 0 ....... (6)
dx v dx v v
v
Multiplying (6) by (1) and subtracting from (5) gives
d 2 J v J d J J
x2 {  (1) v v }  x { v  (1) v v }
dx v
2
v dx v v
J J
+ ( x 2  v 2 ){ v  (1)v  v }  2v{J v  (1) v J v } = 0
v v
Taking v = n and using (2), we have
2
2 d d
x 2
{Yn ( x)}  x {Yn ( x)}  ( x 2  n 2 ){Yn ( x)}
dx dx
 2n{J n  (1) n J n } = 0
Since n is an integer, J n (x) = (1) n J n (x) and hence the last term
of the above equation vanishes. So the above equation reduces to
x2 Yn// + x Yn/ + (x2 – n2) Yn = 0 ....... (7)
It shows that Yn (x) is also a solution of Bessel's equation of order n.
Proof of (ii): We know that an explicit expression of Yn (x) for n
integral is given by
2 x 1 n 1
Yn (x) = log      J n ( x)
 2 2 r 1 r 
 n2m m
1 1  x 1 1 
–  (1) m
    r  r  n 
 m 0 m!(n  m)!  2  r 1

7
  n2m
1 (n  m  1)!  x 
n 1
– 
 m 0 m!
 
2
........... (8)

where  is Euler's constant.

From (8) we find that Yn (x) is infinite when x = 0, whereas J n (x)
is finite when x = 0. So, Yn (x) and J n (x) are two independent
solutions of Bessel's equation of order n. Hence the theorem.

Remark-1: The above theorem shows that the general solution of

Bessel's equation is y = A J n (x) + B Yn (x) when n is an integer.
Remark-2: Bessel's function of the second kind of order n is
denoted by Yn (x) and is defined as
cos n J n ( x)  J n ( x)
Yn (x) = , n  integer
sin n
cos v J v ( x)  J v ( x)
and Yn (x) = lt , n is an integer.
vn sin v
Yn (x) is also called the Neumann function of order n and is
denoted by N n (x) .
Remark-3: Equation reducible to Bessel's equation: Consider
x2y// + xy/ + (a2x2 – n2)y = 0 ......... (*)
2 2
dy dy d y 2d y
Let z = ax so that =a and = a .
dx dz dx 2 dz 2
2
2d y dy
Then (*) becomes z 2
+ z + (z2 – n2)y = 0 which is
dz dz
Bessel's equation of order n. So, the general solution of (*) is
y = A J n (z ) + B Yn (z )
or, y = A J n (ax) + B Yn (ax) . Thus J n (ax) and Yn (ax) are
solutions of (*), which is the modified Bessel's equation.

8
  2
Example (1): Show that (i) J 1 ( x) =   sin x ;
2  x 
 2 2
  cos x and (iii) [ J 1 ( x)] + [ J 1 ( x)] =
2 2
(ii) J 1 ( x) =

2  x  2

2
x
Solution: We know that
 2 m n
1  x
J n (x) =  (1) m
  ....... (1)
m 0 m!.(m  n  1)  2 
1
(i) Putting n = in (1), we get
2
 2m 1
1  x
J 1 ( x) =  (1)
2
 
m

2 m 0 m!.(m  2 )  2 
3

1
 x 2 1 x2 x4 x6
=  [ 3 – 2 + – + ...]
 2  ( 2 ) 2 .1!.( 52 ) 2 4.2!.( 72 ) 2 6.3!.( 92 )
1
 x 2 1 x2 x4 x6
=   . 3 [1 – 2 3 + 4 5 3 – 6 7 5 3 + ...]
 2  ( 2 ) 2 .1!. 2 2 .2!. 2 . 2 2 .3!. 2 . 2 . 2
x 1 x2 x4 x6
= . [1 – + – + ...]
2 12 .  3! 5! 7!
x 2 x3 x5 x7
= . [x – + – + ...]
2 x  3! 5! 7!
 2
=   sin x
 x 
1
(ii) Putting n = – in (1), we get
2

9
  2m 1
1  x
J 1 ( x) =  (1)
2
 
m

2 m 0 m!.(m  2 )  2 
1

1
 x 2 1 x2 x4 x6
=  [ 1 – 2 + – + ...]
2 ( 2 ) 2 .1!.( 32 ) 2 4.2!.( 52 ) 2 6.3!.( 72 )
1
 2 2 1 x2 x4 x6
=   . 1 [1 – 2 1 + 4 3 1 – 6 5 3 1 + ...]
 x  ( 2 ) 2 .1!. 2 2 .2!. 2 . 2 2 .3!. 2 . 2 . 2
2 1 x2 x4 x6
= . [1 – + – + ...]
x  2! 4! 6!
 2
=   cos x
 x 
 2  2
(iii) Since J 1 ( x) =   sin x and J 1 ( x) =   cos x ,
2  x  
2  x 
2 2
hence [ J 1 ( x)]2 + [ J 1 ( x)]2 = sin 2 x + cos 2 x
2

2
x x
2
= ( sin 2 x  cos 2 x )
x
2
=
x

Example (2): Find the value of (i) J 3 ( x) and (ii) J 3 ( x) .


2 2
Solution: We know the Bessel's function
 2 m n
1  x
J n (x) =  (1) m
  ....... (a)
m 0 m!.(m  n  1)  2 
3
(i) Putting n = in (a), we get
2

10
  2m 3
1  x
J 3 ( x) =  (1)
2
 
m

2 m 0 m!.(m  2 )  2 
5

3
 x 2 1 x2 x4 x6
=  [ 5 – 2 + – + ...]
 2  ( 2 ) 2 .1!.( 72 ) 2 4.2!.( 92 ) 2 6.3!.( 112 )
3
 x 2 1 x2 x4 x6
=   . 5 [1 – 2 5 + 4 7 5 – 6 9 7 5 + ...]
 2  ( 2 ) 2 .1!. 2 2 .2!. 2 . 2 2 .3!. 2 . 2 . 2
x x 1 x2 x4 x6
= . [1 – + – + ...]
2 2 . . 
3 1
2 2
2 .5 2.4.5.7 2.4.5.6.7.9
 2 x
2
x4 x6 x8
=  [ – + – + ...]
 x  3 2.3.5 2.3.4.5.7 2.3.4.5.6.7.9
 2  2x
2
4x 4 6x6 8x 8
=  [ – + – + ...]
 x  3! 5! 7! 9!
 2  (3  1) x (5  1) x 4 (7  1) x 6 (9  1) x 8
2
=  [ – + – + ...]
 x  3! 5! 7! 9!
 2 x2 x4 x6 x8
=   [{– + – + – ...}
 x  3! 5! 7! 9!
x2 x4 x6 x8
+{ – + – + ...}]
2! 4! 6! 8!
 2 x2 x4 x6 x8
=   [{1 – + – + – ...}
 x  3! 5! 7! 9!
x2 x4 x6 x8
+ {–1 + – + – + ...}]
2! 4! 6! 8!
 2 1 x3 x5 x7 x9
=   [ {x – + – + – ...}
 x  x 3! 5! 7! 9!

11
 x2 x4 x6 x8
– {1 – + – + – ...}]
2! 4! 6! 8!
 2 1
=  [ sin x  cos x] [Answer]
 x  x
3
(ii) If we put n = – in (a), then for m = 0 the denominator
2
 1
contains    =  . To avoid this situation, we multiply the
 2
numerator and the denominator of (a) by (m + n + 1), we get
2 m n

(m  n  1)  x 
J n (x) =  (1) m
  ..... (b)
m 0 m!.(m  n  2)  2 
3
Putting n = – in (b), we get
2
2m 3

(m  1 / 2)  x 
J 3 ( x) =  (1)
2
 
m

2 m 0 m!.(m  12 )  2 
3
 x
2 3 4 5 6
2 1 1 12 .x 2 .x 2 .x
=  [  – + – + ...]
2 ( 12 ) 2 2 2.1!. 12 2 4.2!. 32 . 12 2 6.3!. 52 . 32 . 12
2 2 1 x2 x4 x6
= (– )[1 + – + – ...]
x x.  2 2 8 144
 2 1 1 1 1 1 1
= –   [1 + (1  ) x 2 – (  ) x 4 + (  ) x 6 – ...]
 x  x 2 3! 4! 5! 6!
 2 1 1 1 1 4
= –   [ {1  x 2 + x 4 – x + ... }
 x  x 2! 4! 6!
1 1 1 6
+ { x2 – x4 + x – ... }]
x 3! 5!

12
  2 1 1 1
=   [ cos x + ( x – x 3 + x 5 – ... )]
 x  x 3! 5!
 2 1
=   [ cos x  sin x] [Answer]
 x  x
Theorem (3): (Generating function for Jn(x)) J n (x) is the
x ( t 1 )
coefficient of t n in the expression of the function e 2 t
, that is
x 1
(t  ) 
e 2 t
= J
n  
n ( x)t n .
xt
Proof: Expanding the exponential function e 2 in Maclaurent's
series, we get
2 3
 xt   xt   xt 
xt      
e 2 = 1 +   +   +   + ...
2 2 2
1! 2! 3!
 r
1  x
=    tr ............................ (1)
r  0 r!  2 
Similarly,
2 3
 x  x  x
x      
e

2t
=1+  2t 
+  2t 
+ 
2t 
+ ...
1! 2! 3!
 s
s 1  x  s
=  (1)   t ............................. (2)
s 0 s!  2 
The product of (1) and (2) gives,
 r s
1  x r 
x 1
s 1  x  s
(t  )
e2 t
=    t .  (1)   t
r  0 r!  2  s 0 s!  2 
  r s
1 1 x
=  (1) .   s
t r s ................. (3)
r 0 s 0 r! s!  2 
13
We want to find the coefficient of tn, therefore we put r – s = n
where n can be positive, negative or zero.
First we consider the case n  0, i.e., r – s  0. So, we vary s from
0 to  , and we assign such value to r so that r – s  0.
Since r – s = n  r = s + n. Then we get,
x 1  2sn
(t  ) 1 1  x
e 2 t
=  (1) . s
  tn
s 0 s! ( s  n)!  2 
 2sn
1 1  x
 the coefficient of tn =
s 0
.   (1) s
s! ( s  n  1)  2 
= J n (x) ; where n  0
Again we consider the case n < 0 ; i.e., r – s < 0. So, we vary r
from 0 to  and we assign such values to s so that r – s < 0.
Since r – s = n  s = r – n (n < 0). In this case, we have
x 1  2r n
(t  ) 1 1  x
e2 t
=  (1) r n
r 0
.  
r! (r  n)!  2 
tn

 2 r n
1 1 1  x
 the coefficient of t =
(1) n
n

r 0
(1) . r
 
r! (r  n  1)  2 
= (1) n J n ( x)
= J n (x) ; where n < 0
Therefore in the both cases n  0 and n < 0, we find that J n (x) is
x 1
(t  )
n 2 t
the coefficient of t in the expansion of e . Thus
x 1
(t  ) 
e 2 t
= J
n  
n ( x)t n ................... (4)
x 1
(t  )
2 t
Therefore, e is called the generating function for J n (x) .
Corollary-1: From (4), we have
x 1
(t  )
e2 t
= J0+ t.J1+ t–1 J–1+ t2 J2 + t–2 J–2 + ... + tnJn + t–n J–n + ...

14
 = J0+ t.J1– t–1 J1+ t2 J2 + t–2 J2 + ... + tnJn + (–1)n t–n Jn + ...
[Since J1 = (–1)J–1, ..., Jn = (–1)n J–n]
= J0+ J1 (t – t–1) + J2 (t2 + t–2) + ... + Jn (tn + (–1)n t–n) + ...
Corollary-2: Putting t = e i in the corollary-1, we have
x  i  i 
 e e 
e 2
= J0 + J1 ( e i – e i ) + J2 ( e 2i + e 2i ) + ...


+ Jn ( e ni + (–1)n e  ni ) + ...

Or, e ix sin  = J0 + 2iJ1 sin  + 2J2 cos2  + 2iJ3sin3  + ...
Since L.H.S is cos(xsin  ) + isin(xsin  ), therefore equating real
and imaginary parts, we get,
cos(x sin  ) = J0 + 2J2 cos2  + 2J4 cos 4  + ... ..... (a)
sin(x sin  ) = 2J1 sin  + 2J3 sin3  + 2J5 sin5  + ... .... (b)
These are expansion of sine and cosine functions in terms of
Bessel's function. Above (a) and (b) series can be written in the

another form, taking  = 
2
cos(x cos  ) = J0 – 2J2 cos2  + 2J4 cos 4  – ... ....... (c)
sin(x cos  ) = 2J1 cos  – 2J3 cos3  + 2J5 cos5  – ... ...... (d)
Putting  = 0 in (c) and (d), we have,
cos x = J0 – 2J2 + 2J4 – ...
sin x = 2J1 – 2J3 + 2J5 – ...
Theorem (4): (Orthogonality of Bessel's function) If  1 and  2
a
are different roots of J n (a) = 0, then J
0
n ( 1 r ) J n ( 2 r )rdr = 0.

This relationship is called the orthogonality of Bessel's function.

Proof: Since  1 and  2 are different roots of J n (a) = 0,
therefore J n (1a) = 0 and J n ( 2 a) = 0.
Also J n (x) is a solution of the Bessel's equation,
d2y dy
x2 2
+x + (x2 – n2)y = 0 .............. (1)
dx dx
15
 dr 1
Let, x =  1 r and y = u 
=
dx  1
dy du du dr 1 du d 1 d
So, = = . =  
dx dx dr dx  1 dr dx  1 dr
d2y d du 1 d 1 du 1 d 2u
And, = [ ] = [ ] =
dx 2 dx dx  1 dr  1 dr  12 dr 2
Substituting these in (1), we get
2
2d u du
r 2
+r + (  12 r2 – n2)u = 0 .............. (2)
dr dr
Similarly by putting x =  2 r and y = v, we get
d 2v dv
r2 2
+r + (  22 r2 – n2)v = 0 .............. (3)
dr dr
v u
Multiplying equation (2) by and equation (3) by and then
r r
subtracting we get,
d 2u d 2v du dv
r(v 2 – u 2 ) + (v – u ) + r(  12 –  22 )uv = 0
dr dr dr dr
d du dv
Or, [r(v – u )] + r(  12 –  22 )uv = 0
dr dr dr
Integrating over [0, a], we get
a a
du dv
r (v  u ) + ( 12   22 )  ruvdr = 0;
dr dr 0 0

where u = J n (1r ) and v = J n ( 2 r )

a
Or, r{J n ( 2 r ) J n/ (1r )  J n (1r ) J n/ ( 2 r )}
0
a
+ ( 12   22 )  J n ( 1r ) J n ( 2 r )rdr = 0
0

16
Since the first term is 0 because J n (1a) = 0 and J n ( 2 a) = 0 and
1   2 , so
a

J
0
n ( 1 r ) J n ( 2 r )rdr = 0

Hence Jn's are orthogonal with respect to weight function r.

Integral representation of Bessel's function: We know that,

cos(x sin  ) = J0 + 2J2 cos2  + 2J4 cos 4 
+ ... + 2J2m cos2m  + ... ..... (1)
sin(x sin  ) = 2J1 sin  + 2J3 sin3  + 2J5 sin5 
+ ... + 2J2m+1 sin(2m+1)  + ... . .... (2)
We multiply equation (1) by cos2m  and integrating over [0,  ]
 

 cos( x sin  ) cos 2m d

0
= J0  cos 2m d
0
 
+ J2  2 cos 2 cos 2m d + J4  2 cos 4 cos 2m d + ...
0 0

+ J2m  2 cos 2 2m d + ... ................... (3)
0
 
Since  cos 2m d = 0,  2 cos p cos q d = 0 ; (p  q)
0 0
  
sin 4m
and  2 cos 2m d =  (1  cos 4m ) d =  
2
=
0 0
4m 0

So from (3), we have



 cos( x sin  ) cos 2m d

0
= J2m. 


1
Or, J2m(x) =
  cos( x sin  ) cos 2m d
0
.......... (4)

17
Again multiplying (2) by sin(2m+1)  and integrating over [0,  ],
we get
 

 sin( x sin  ) sin(2m  1) d

0
= J1  2 sin  sin(2m  1) d
0

+ J3  2 sin 3 sin(2m  1) d + ...
0

+ J2m+1  2 sin 2 (2m  1) d + ... ........... (5)
0

Since  2 sin p sin q d = 0 ; (p  q) and
0
 

 2 sin
2
(2m  1) d =  {1  cos(4m  2) }d
0 0

sin(4m  1)
=  =
4m  2 0
From (5), we have


 sin( x sin  ) sin(2m  1) d

0
= J2m+1. 


1
 0
Or, J2m+1(x) = sin( x sin  ) sin(2m  1) d .......... (6)

Again multiplying (1) by cos n  and integrating over [0,  ] with

respect to  , we have
 

 cos( x sin  ) cos n d = J0  cos n d

0 0
 
+ J2  2 cos 2 cos n d + J4  2 cos 4 cos n d + ... (7)
0 0

18
 
Since  cos n d = 0,
0

0; when p  q
and  2 cos p cos q d  
0  ; when p  q
So, from (7), we have

0; when n is odd int eger
0 cos( x sin  ) cos n d  J n ( x); when n is even int eger .... (8)
Again multiplying (2) by sin n  and integrating over [0,  ] with
respect to  , we have
 

 sin( x sin  ) sin n d

0
= J1  2 sin  sin n d
0

+ J3  2 sin 3 sin n d + ... ........ (9)
0

0; when p  q
Since  2 sin p sin q d  
0  ; when p  q
So, from (9), we have

J n ( x); when n is odd int eger
0 sin( x sin  ) sin n d  
0; when n is even int eger
.... (10)

Adding (8) and (10), we have for any integer n,



 [cos( x sin  ) cos n  sin( x sin  ) sin n ]d

0
=  Jn(x)


Or,  cos(n  x sin  )d =  Jn(x)
0

1
 Jn(x) =
  cos(n  x sin  ) d ; for any integer n.
0

19
Therefore, for all positive integers m and n

1
J2m(x) =
  cos( x sin  ) cos 2m d
0

1
J2m+1(x) =
  sin( x sin  ) sin(2m  1) d
0

1
Jn(x) =
  cos(n  x sin  ) d
0

These are the integral representation of Bessel's function Jn(x).

2 2 2
Example (3): Prove that J 0 + 2 J 1 + 2 J 2 + ... = 1.
Solution: We know that,
cos(x sin  ) = J0 + 2J2 cos2  + 2J4 cos 4  + ... (1)
sin(x sin  ) = 2J1 sin  + 2J3 sin3  + 2J5 sin5  + ... (2)
Squaring and adding (1) and (2) and integrating over [0,  ] with
respect to  , we have
  

 [cos ( x sin  )  sin ( x sin  )]d = [ J 0  d  cos 2 d

2 2 2 2 2
+ 4 J2
0 0 0
 

 cos 4 d + ...] + [4 J 1  sin  d

2 2 2 2
+ 4 J4
0 0


 sin 3 d + ...]
2 2
+ 4 J3
0
 
[Because 4J0 Jm  cos m d = 0; 4Jp Jq  cos p cos q d = 0; and
0 0

4Jp Jq  sin p sin q d = 0]
0

  = [ J 0  + 2 J 2  + 2 J 4  + ...] + [2 J 1  + 2 J 3  + ...]
2 2 2 2 2

20
 J 0 + 2 J 1 + 2 J 2 + 2 J 3 + ... = 1
2 2 2 2
[Proved]

Example (4): Show that

n
 x 1
   cos( x sin  ) cos  d
2n
J n (x) =
1 2
 ( n  ) 0
2
Solution: We know that,
( x sin  ) 2 ( x sin  ) 4
cos(x sin  ) = 1 – + – ...
2! 4!

x 2 r sin 2 r 
=  (1) r
r 0 (2r )!
  
x 2r
  cos( x sin  ) cos 2 n  d =  (1) r
r 0

(2r )! 0
sin 2 r  cos 2 n  d
0
 2r
x ( ).( )
2 r 1 2 n 1
=  (1)
r 0
r

(2r )! 2(
2

) 2r 2n2
2

x 2r n (n  12 ) 2 . 2 ..... 2 .( 2 )

 2 r 1 2 r 3 1 1
=  (1) r
. .
r 0 x n 2(r  n  1) 2r (2r  1)(2r  2)...2.1
1
. 

r x
2r n
( n  2 )
1 r 1
=  (1) . . 2
r 0 x n 2(r  n  1) 2r (2r  2)(2r  4)...2
1
. 

r x
2r n
( n  2 )
1 r 1
=  (1) . . 2
r 0 x n 2(r  n  1) 2 r .r!

x 2 r  n (n  12 ) 
=  (1) .r
. 2r
r 0 x n
(r  n  1) 2 .r!
2r n
2 n  .(n  12 )  1  x
= .  (1) r . 
x n
r 0 r!.(r  n  1)  2 

21
 2 n  .(n  12 )
= . J n (x)
xn
n 
1  x
 J n (x)    cos( x sin  ) cos  d [Proved]
2n
=
 ( n  2 )  2  0
1

Some recurrence relations of Bessel's function: Prove the

following recurrence relations of Bessel's function:
(i) J 0 ( x)   J 1 ( x)
/

(ii) x J n ( x)  nJ n ( x) – x J n1 ( x)
/

(iii) x J n ( x)  nJ n ( x) + x J n1 ( x)

/

(iv)
dx

d n

x J n (x) = – x  n J n1 ( x)

(v) x J n ( x)  nJ n ( x) – x J n1 ( x) ; x J n ( x)  nJ n ( x) + x J n1 ( x)

/ /

using generating function.

 2m
1  x
Proof of (i): We know that J 0 ( x) =  (1)  
m

m 0 m!.(m  1)  2 
 2 m 1
2m  x 1
 J 0 ( x) =  (1) m
/
  .
m 0 m!.(m  1)  2  2
 2 m 1
1  x
=  (1) m
 
m 0 (m  1)!.(m  1)  2 
Let (m – 1) = r  m = r + 1
 2 r 1
1  x
 (1) r 1
/
So, J 0 ( x) =  
r  1 r!.(r  1  1)  2 
Since (–1)! =  , therefore,
 2 r 1
1  x
J 0 ( x) = 0 –  (1)
/
 
r

r 0 r!.(r  1  1)  2 

22
 J 0 ( x)   J 1 ( x)
/
So, [Proved]
Proof of (ii): We know that
 2 m n
1  x
J n (x) =  (1) m  
m 0 m!.(m  n  1)  2 
Differentiating with respect to x, we have
2 m  n 1

( 2m  n)  x  1
J n ( x) =  (1)
/
 
m
.
m 0 m!.(m  n  1)  2  2
 2 m  n 1
2m  x 1
=  (1) m
 
m 0 m!.(m  n  1)  2  2
2 m n
n  1  x
+ 
x m 0
(1) m  
m!.(m  n  1)  2 
 2 m  n 1
1  x n
=  (1) m
  + J n (x)
m 0 (m  1)!.(m  n  1)  2  x
 2 ( m 1)  ( n 1)
1  x n
=  (1) m
  + J n (x)
m 0 (m  1)!.{(m  1)  (n  1)  1}  2  x
/
Let m – 1 = r and for r = –1, J n ( x) has zero in the first term.
 2 r  ( n 1)
1  x n
 J n ( x) = 0 +  (1) r 1
/
  + J n (x)
r 0 r!.{r  (n  1)  1}  2  x
/ n
Or, J n ( x) = – J n1 ( x) + J n (x)
x
x J n ( x)  nJ n ( x) – x J n1 ( x)
/
Or, [Proved]

Which may also be written as

d n
dx
 
x J n =  x  n J n1 .
Proof of (iii): We know that
 2 m n
1  x
J n (x) =  (1) m  
m 0 m!.(m  n  1)  2 

23
Differentiating with respect to x, we have
2 m  n 1

( 2m  n)  x  1
J n ( x) =  (1) m
/
  .
m 0 m!.(m  n  1)  2  2
2 m  n 1

2(m  n)  n  x  1
=  (1) m
m 0
 
m!.(m  n  1)  2  2
2 m  n 1

( m  n)  x
=  (1) m
 
m 0 m!.(m  n  1)  2 
 2 m  n 1
n  x 1
–  (1) m
 
m 0 m!.(m  n  1)  2  2
 2 m  ( n 1)
1  x
=  (1) m
 
m 0 m!.(m  n)  2 
 2 m n
n 1  x
–
x m 0
 (1) m  
m!.(m  n  1)  2 
 2 m  ( n 1)
1  x
=  (1) m
 
m 0 m!.{m  (n  1)  1}  2 
 2 m n
n 1  x
–x
m 0
(1) m
 
m!.(m  n  1)  2 
n
 J n ( x) = J n1 ( x) – J n (x)
/

x
So, x J n ( x)  nJ n ( x) + x J n1 ( x) [Proved]
/

Which may also be written as

dx
d n

x J n = x n J n 1 . 
 2 m n
1  x
Proof of (iv): We have, J n (x) =  (1)  
m

m 0 m!.(m  n  1)  2 
 2m
1  x
Or, x n
J n (x) =  (1) m
  .2
n

m 0 m!. ( m  n  1)  
2
24
Differentiating with respect to x, we have
2 m 1
2  n.2m
   (1)

d n  x 1
 
m
x J n (x) = .
dx m 0 m!.(m  n  1)  2  2
2 ( m 1) 1
2 n

 x
m 0
=  (1) m  
(m  1)!.(m  n  1)  2 
Let m – 1 = r, then (m – 1)! in the denominator produces zero for
the first term.
2 r 1
2 n
 

d n  x

dx
x J n (x) = 0 +  (1)
r 0
r 1
 
r!.{r  (n  1)  1}  2 
 2 r  ( n 1)
1  x
= – x  (1)
n r
 
r 0 r!.{r  (n  1)  1}  2 
Therefore,
d n
dx
 
x J n (x) = – x  n J n1 ( x) [Proved]
Proof of (v): We know the generating function
x 1
(t  ) 
e2 t
= J
n  
n ( x)t n .................. (*)

Differentiating with respect to t, we get

x 1
(t  ) x 
1
e 2 t . (1  2 ) =  nJ n ( x)t n 1
2 t n  
 
x 1
Or, (1  2 ) .  J n ( x)t n =  nJ n ( x)t n 1
2 t n   n  
n 1
Equating the coefficients of t
x x
J n1 ( x) + J n1 ( x) = n J n (x)
2 2
Or, 2 n J n (x) = x{ J n1 ( x) + J n1 ( x) } ........ (a)
Again differentiating (*) with respect to x, we have
x 1
(t  ) 1 
1
e t . (t  ) =  J n ( x)t n
2 /

2 t n  

25
 1  
(t  ) .  J n ( x)t n = 2  J n ( x)t n
/
Or,
t n  n  
Equating the coefficients of tn from both sides, we have
/
J n1 ( x) – J n1 ( x) = 2 J n ( x) ................... (b)
Multiplying (b) by x then subtracting from (a), we get
/
2x J n1 ( x) = 2{ n J n (x) – x J n ( x) }
x J n ( x)  nJ n ( x) – x J n1 ( x)
/
Or, [Proved]
Multiplying (b) by x then adding with (a), we get
/
2x J n1 ( x) = 2{ n J n (x) + x J n ( x) }
/
Or, x J n ( x) = – n J n (x) + x J n1 ( x) [Proved]

Exercises:
1. Write the Bessel's equation and then find the general solution
of it when the order is not an integer.
2. Prove that the solutions of Bessel's equation J n (x) and
J n (x) are not independent when n is an integer. [Hint: prove
J n (x) = (1) n J n (x) ]
3. Find the general solution of Bessel's equation.
4. Define Bessel's functions of first and second kinds.
 x  1
5. Show that   J 3 ( x) = sin x  cos x
 2 2 x
6. State and prove the orthogonality of Bessel's function.
7. Define the generating function for Bessel's function and
establish the condition.
8. Prove that
(i) cos x = J0(x) – 2J2(x) + 2J4(x) – ...
(ii) sin x = 2J1(x) – 2J3(x) + 2J5(x) – ...
9. For all positive integer m and n, prove that

26
 
1
(i) J2m(x) =
  cos( x sin  ) cos 2m d
0

1
(ii) J2m+1(x) =
  sin( x sin  ) sin(2m  1) d
0

1
(iii) Jn(x) =
  cos(n  x sin  ) d
0
n
1  x
   cos( x cos  ) sin  d
2n
10. Prove that J n (x) =
 ( n  )  2 
1 0
2

27