Bessels function(B.
F)
These are canonical solution of the Bessels differential equations (BDF)
d2 y
dy
2
2
x
+ x + ( x ) y=0
2
dx
dx
2
for an arbitrary complex number
Although and produce same BDF for real , it is conventional to
define different BF for these two values.
Bessels equation arises when finding separable solution to Laplaces equation and
Helmoltz equation in cylindrical and spherical coordinates.
.
=i
is named as the Laplacian
+j
+k
x y
z
. =
2
2
2
+
+
x 2 y2 z 2
. =
2 2 2
+
+
x2 y2 z2
If is electrostatic potential
2
. =0 =0 Leplaces equation
The general form of this equation is
2 + K 2 =0
If K2 = 0 the Laplaces equation
K= + constant then Helmoltz equation
K= - constant then Diffusion
= constant kinetic energy Schrodinger
Bessels functions are important for man problems related to propagation of wave
and static potential.
In solving problems in cylindrical coordinates. We obtain BF of inter order(
=n
1
In spherical coordinates we obtain BF of half integer order ( =n+ 2 )
2
2
Consider + K =0
Let K2 be a constant
Then we can write
Substitute
YZ
d2 x
d x2
(x , y, z )
(x , y, z )=X (x) Y (y ) Z(z )
in the equation 1
d2 y
d y2
+ XZ
d2 z
d z2
+ XY
+ K2 = 0
Divided by XYZ
1 d2 x
x d x2
1 d2 x
x d x2
=- K
d2 y
2
dy
1
y
2
= - l (say)
1
y
1 d2 z
z d z2
d2 y
d y2
+ K2 = 0
1 d2 z
z d z2
1 d2 x
x d x2
2
=- l
(2)
2
1
y
d y
2
dy
1
y
d2 y
2
dy
=-
1 d2 z
z d z2
1
y
d2 y
d y2
2
=- m
Take
1d z
z d z2
2
= l K2
2
+ l K2
1 d2 z
z d z2
2
2
2
= m + k - l
1 d2 z
z d z2
2
=- n
d2 x
2
dx
2
=- l x
d2 y
2
dy
2
=- m y
d2 z
d z2
2
=- n z
We have divided the DE into three simple DE, their solution are familiar for us.
The solution should be labeled according to our choice of constants l,m and n.
That is lmn ( xyz ) =X l ( x ) Y m ( y ) Z n (z)
2
2
Q) + k ( x ) =0 BH
Determine the three differential equation .
Circular cylindrical coordinates( z
2
2
( z + k ( z = 0
or
1 2
2 2
) +
2
z2
2
+ k ( z = 0
2
If k is a constant
( z =
( ) ( )Z(z)
d
d
1
p
d 2
d 2
d
dp
1
+ 2
d d
1
p
d 2
d 2
d
dp
1
+ 2
d d
PZ
2
dp
( d ) +
( )
( )
1
z
k2
d 2
d 2
d2 z
+ P* d z 2
d2 z
d z2
2
+ k =0
=-
1
z
2
= l
d2 z
2
dz
2
+ l z=0
1
p
d
2
d
d
dp
1
+ 2
d d
( )
2
2
+ k =- l
d2 z
d z2
2
+ k p z = 0
1
p
d
2
d
d
dp
1
d d 2
1
p
d
2
d
d
dp
1
+ 2
d d
( )
2
2
+ k + l =0
( )
2
+ n =0
2
Multiply by ,
d
dp
d d
( )
d 2
d 2
1
2 2
+ n =0
d
dp
d d
( )
d 2
2
d
1
2
= m
d
dp
d d
( )
d
dp
d d
( )
2 2
+ n =-
d
2
d
1
d2 p
2
d
2 2
2
+ n - m =0
2
- m p=0
dp
d ) +
2
- m p=0
The above equation is Bessel differential equation.
The solution the Helmoltz equation is
P Z
( z = ( ) () (z)
The general solution
a mn pmn () m () Z n(z )
( z =
mn
2
2
- n = l
amn are chosen to permit to satisfy boundary conditions.
Spherical polar coordinate(r )
2
+ K =0
1
2
1
sin
r
+
sin
+
=K 2 r )
2
2
r
r
sin
r sin
) (
R
Substitute r ) = (r ) () ( )
1
Rr 2
d
dR
r2
dr (
dr ) +
1
d
d
sin
2
(
d
d ) +
r sin
1
2
r sin 2
d 2
d 2
2
=- K
2
2
Multiply by r sin
sin
d 2
d 2
d
d )}
2
2
2
= - r sin { - K - Rr 2
d
2 dR
r
(
dr
dr ) -
sin
d
d (
d 2
d 2
1
R
d
dR
r2
(
dr
dr ) +
2
=- m
r k
=Q
2
Divided by r and multiply R(r)
1
r2
d
dR
r2
(
dr
dr ) +
1
=> sin
R k
= r2 R
d
d
sin
d (
d ) -
m2
sin
+Q =0
(2)
(3)
2
If k is constant the equation(2) is called Spherical Bessel equation.
If Q can be written as Q= l(l+1)
Where l is an integer
Then,
d2 R
d r2
+ r
d2 R
d r2
dR
dr
dR
+ 2r dr
Z (kr )
Put R(kr) =
To show,
(kr )
1
2
2
+( k -
l(l+1)
R=0
r2
2 2
+ ( k r - l(l+1) )R = 0
d2 Z
d r2
dZ
+ r dr
2 2
+{ k r
1 2
(l+ ) } Z = 0
2
1
This is a Bessels equation in Z with order (l+ 2 ) , where l is an integer.
Consider Bessel differential equation of order n
2
x2
d y
2
dx
dy
2
2
+ x dx + ( x n y = 0
This is a second order differential equation. Therefore, there are two linear
independent solutions.
Dependent on the situation, we can use various linear combination of these two
solutions (or Bessel function)
Example:Y=AJn(x) + BYn(x)
Where Jn(x) is called Bessel function of first kind.
Y(n) is the second kind,
This is sometimes called Neumann function or weber function.
Bessel function of third kind, sometimes called as Hankel function, which is given
as
H 1n (x) = Jn(x) + iYn(x)
H 2n (x) = Jn(x) - iYn(x)
for x>0
Let x=1x then Bessel differential equation takes the form
x2
d2 y
d x2
dy
2
2
+ x dx - ( x + n y = 0
The solution of this modified equation gives modified, Bessel function In(x) and
Kn(x)
Y=( In(x) + D Kn(x))
For fixed value of n the Bessel equation has two laniary solutions one of this called
first kind Bessel function and is given by
Jn(x) =
(1)
x
2
(n+ s)! s !
s=0
also J-n(x) =
s+n
(1)
( n+ s)! s !
s=0
n
= (1) Jn(x)
The second kind is given by
Yn(x) =
Y n( x) cos nY n ( x)
sin n
n
Y(x) = (1) Yn(x)
AS dr 0
for n>0
x
2
for n<0
d d T 2 hr
r
dr dr
tk
( )
d2 T
d r2
r2
dT
dr
2 hr
kT
dT
+ r dr
d T
2
dr
(T- T = 0
2 hr
kT
(T- T = 0
(T- T = 0
This is modified Bessels equation of order o
Since the third term is negative
But
2h
T
=i
Tkt r
Then the above equation takes the form
d2
d 2
2
d
2
+ d + = 0
This is Bessels equation of zeroth order.
r
d2 r
d r2
dT
+ r dr
2 hr
kt
=0
= iMr
= iMdr
dT
dr
= dr
d d
= d dr
where M=
d
= iM d
2h
kt
dT
dr
= iM d
d2T
d r2
similarly,
d2
d 2
d2
d r2
d
2
+ d + = 0
= C1 J 0 ( ) + C2 J 0 ( )
C1 J 0
C Y
( iMr ) + 2 0 ( iMr )
J0
( Available
C1 and C2 can be determined from the boundary condition
100 =
20 =
C1 J 0
C1 J 0
C Y
( r 1 ) + 2 0 ( r 1 )
C Y
( r 0 ) + 2 0 ( r 0 )
C1 and C2 can be determined
100 Y 0 ( i M r 0 )20 Y 0 ( r 0)
C1 = J 0 ( r 1 ) Y 0 ( r 0 )J 0 ( r 0 ) Y 0 ( r 1 )
C2 =
Legendres polynomials
The ordinary differential equation to as Legendres equation is frequently in
counted in physics. In particular it occurs when solving laplaces equation in
spherical coordinates.
2 +K 2
The second order differential equation of the form
2
(1x )
d2 y
2
dx
dy
2x dx +n ( n+1 ) y=0
where n>0(0,1,2) and |x| < 1, is called Legendres equation
The general solution of this equation is given by
Y = APn(x) + BQn(x)
where n = 0,1,2,3
Pn(x) are polynomials called Legendres polynomials function of kind.* are called
Legendre function of second kind.
The function Pn(x) defined by
Pn(x) =
( 2 n1 )( 2 n3 ) ..
n!
n(n1)
n
{ x - 2( 2n1)
n+ ..
x
2.4 (2 n1)(2 n3) }
n ( n1 )( n2 )( n3 )
First few Legendres polynomials are as follows
x n2 +
P0 ( x )=1
P1 (x )=x
P
1
2 ( x ) = (2 x2 1)
2
1
3 ( x ) = (5 x 33 x)
2
1
4 ( x ) = (35 x 430 x 2+ 3)
8
1
5 ( x ) = ( 63 x 570 x 2+15 x )
8
2
Eg:- The laplaces equation u=0 can be split into equations
d2 R
d r2
dR
2
+ 2 r dr + R
d
d
sin
(
d
d ) -
sin
= 0 and
=0
=
u=
But,
u
r
)+
r2
1
r2
2 u
2
Put u = R(r)
=0
( )
1
2
r sin
if U = U( r 1 )
R(r ) ()
u
sin
(
) +
u is not a function of
1
2 u
2
2
2
r sin
=0
1
R
d
dR
r2
(
dr
dr ) = -
1 d
d
sin
(
sin d
d ) = + n(n+1)
2
=-
d
d
sin
d (
d ) + n(n+1)
Put
cos
d
d
dx
d
=x
d
=sin
d
d
= - sin
dx
d
= dx
sin
=0
d
dx
2 d
= - sin dx
d
d
sin
d (
d ) =
2
= ( cos 1) dx
d
x 21 )
d (
d
dx
d
d
2
{
x
1
) dx }
dx (
dx
d
d
d
2
{
x
1
) dx }(- sin
dx (
