Jim Lambers

MAT 415/515
Fall Semester 2013-14
Lecture 15 Notes
These notes correspond to Section 14.2 in the text.
Orthogonality of Bessel Functions
Since Bessel functions often appear in solutions of PDE, it is necessary to be able to compute
coecients of series whose terms include Bessel functions. Therefore, we need to understand their
orthogonality properties.
Consider the Bessel equation

2
d
2
J

(k)
d
2
+
dJ

(k)
d
+ (k
2

2
)J

(k) = 0,
where 1. Rearranging yields

_
d
2
d
2
+
1

d
d


2

2
_
J

(k) = k
2
J

(k).
Thus J

(k) is an eigenfunction of the linear dierential operator

d
d


2

2
_
with eigenvalue k
2
.
The operator L is not self-adjoint with respect to the standard scalar product, as the coecients
p
0
() = 1 and p
1
() = 1/ do not satisfy the condition p

1
() = p
0
(), so we use the weight
function
w() = p
0
()e

p
1
()
p
0
()
d
= e

1/ d
= e
ln
= .
It follows from the relation
(
u

v
)
_
b
a
v

(x)u(x)w(x) dx =
_
w(x)p
0
(x)(v

(x)u

(x) (v

(x)u(x)
_

b
a
that
_
a
0
J

(k)J

(k

) d =
a[k

(ka)J

(k

a) kJ

(k

a)J

(ka)]
k
2
k

2
.
Therefore, in order to ensure orthogonality, we must have ka and k

a be zeros of J

. Thus we have
the orthogonality relation
_
a
0
J

a
_
J

a
_
d = 0, i = j,
where
j
is the jth zero of J

.
It is worth noting that because of the weight function being the Jacobian of the change of
variable to polar coordinates, Bessel functions that are scaled as in the above orthogonality relation
are also orthogonal with respect to the unweighted scalar product over a circle of radius a.
1
Normalization
Now that we have orthogonal Bessel functions, we seek orthonormal Bessel functions. From
_
a
0
[J

(k)]
2
d = lim
k

k
a[k

(ka)J

(k

a) kJ

(k

a)J

(ka)]
k
2
k

2
.
Substituting ka =
i
and applying lHospitals Rule yields
_
a
0
[J

(k)]
2
d = lim
k

kaJ

(k

a)J

(
i
)
k
2
k

2
=
a
2
2
[J

(
i
)]
2
.
Using the recurrence relation
J

(x) = J

1
(x) +
1
x
J
1
(x),
we then obtain
_
a
0
[J

(k)]
2
d =
a
2
2
[J
+1
(
i
)]
2
.
Bessel Series
Now we can easily describe functions as series of Bessel functions. If f() has the expansion
f() =

j=1
c
j
J

a
_
, 0 a, > 1,
Then, the coecients c
j
are given by
c
j
=

a
_
|f()
_

a
_
|J

a
__ =
2
a
2
[J
+1
(
j
)]
2
_
a
0
J

a
_
f() d.
It is worth noting that orthonormal sets of Bessel functions can also be obtained by imposing
Neumann boundary conditions J

(k) = 0 at = a, in which case ka =

.
We now consider an example in which a Bessel series is used to describe a solution of a PDE.
Example Consider Laplaces equation in a hollow cylinder of radius a with endcaps at z = 0 and
z = h,

2
V = 0,
with boundary conditions
V (a, , z) = 0, V (, , 0) = 0, V (, , h) = f(, )
for a given potential function f(, ). In cylindrical coordinates, Laplaces equation becomes
1

_
+
1

2
V

2
+

2
V
z
2
= 0.
Using separation of variables, we assume a solution of the form
V (, , z) = P()()Z(z).
2
Substituting this form into the PDE and dividing by V yields
1
P
d
d
_

dP
d
_
+
1

d
2

d
2
+
1
Z
d
2
Z
dz
2
= 0.
Due to periodicity, we require that satisfy
d
2

d
2
= m
2
,
where m is an integer, and therefore () = e
im
. We also have
d
2
Z
dz
2
=
2
Z,
which has solutions e
z
. Because of the boundary condition at z = 0, we take linear combinations
of solutions so that
Z(z) = sinh lz.
Then, we have

2
d
2
P
d
2
+
dP
d
+ (
2

2
m
2
)P = 0.
This is the Bessel equation of order m, which has solutions J
m
(). To satisfy the boundary
condition at = a, we set =
mj
/a, where
mj
is the jth zero of J
m
. We then have
V (, , z) = J
m
_

mj

a
_
e
im
sinh
_

mj
z
a
_
.
To satisfy the boundary condition at z = h, we take a linear combination of solutions of this
and seek the coecients of the expansion
f(, ) =

m=

j=1
c
mj
J
m
_

mj

a
_
e
im
sinh
_

mj
h
a
_
.
From the orthogonality relation for Bessel functions, as well as the orthogonality relation
_
2
0
e
im
e
im

d = 2
mm
,
we obtain
c
mj
=
1
a
2
sinh
_

mj
h
a
_
J
2
m+1
(
mj
)
_
2
0
_
a
0
e
im
J
m
_

mj

a
_
f(, ) d d.
Thus the PDE and all boundary conditions are satised. 2
3