Math 115 (2006-2007) Yum-Tong Siu 1

Bessel Functions and Vibrating Circular Membrane

Method of Separation of Variables. For a linear partial dierential equation
Lu = 0, we can use the method of separation of variables when the linear
partial dierential operator L can be written as the sum of two linear partial
dierential operators P and Q so that P depends only on variables x
1
, , x
k
and Q depends only on the complementary set of variables x
k+1
, , x
n
.
Another condition which needs to be satised before the method of separation
of variables can be applied is that the domain D must be the product of a
domain G in the space of variables x
1
, , x
k
and a domain H in the space
of variables x
k+1
, , x
n
.
The idea of the method of separation of variables is to consider rst
unknown functions u of the form u = vw so that v depends on the variables
x
1
, , x
k
and w depends on the variables x
k+1
, , x
n
. The equation Lu =
0 now becomes wPv + vQw = 0 which can be rewritten as
Pv
v
=
Qw
w
so that the left-hand side depends only on the variables x
1
, , x
k
and the
right-hand side depends only on the variables x
k+1
, , x
n
and as a conse-
quence both sides must be equal to some constant . The logic is that if the
partial dierential equation Lu = 0 is satised on the domain D = G H
for some function u of the form
u(x
1
, , x
n
) = v (x
1
, , x
k
) w(x
k+1
, , x
n
) ,
then there exists a constant such that
Pv
w
=
Qw
w
= ,
which means that
Pv v = 0 on G,
Qw + w = 0 on H.
We then nd all such special product solutions u = v
j
w
j
(for j J) and use
yet-to-be-determined coecients c
j
to nd a solution u =

jJ
c
j
v
j
w
j
which
satises the prescribed boundary conditions.
Math 115 (2006-2007) Yum-Tong Siu 2
In practice, we already use some of the prescribed boundary conditions
when we seek special product solutions u = v
j
w
j
so as to limit the index
set J to a countable set. Then we use yet-to-be-determined coecients c
j
to nd a solution u =

jJ
c
j
v
j
w
j
which satises the remaining prescribed
boundary conditions.
Wave Equation for Vibrating Circular Membrane. To present the details of
the method of separation of variables, we choose to work out the example of
the wave equation for a vibrating circular membrane. The circular membrane
is given by the disk { 0 r c } of radius c > 0 in polar coordinates (r, ).
The displacement of the membrane at time t in the direction perpendicular to
the disk is given by u(t, r, ). The wave equation for the vibrating membrane
is given by
u
tt
= a
2
u
with boundary conditions
u = 0 at r = c and for all t 0,
u
t
= 0 at t = 0,
u = f(r, ) at t = 0,
where a > 0 is a constant determined by the surface tension of the membrane
and f (r, ) is a given function on the disk { 0 r c }. The Laplacian u
of u in polar coordinates is given by
u = u
rr
+
1
r
u
r
+
1
r
2
u

.
The wave equation u
tt
= a
2
u can be written as Lu = 0 with
L =
1
a
2

2
t
2
+

2
r
2
+
1
r

r
+
1
r
2

2
.
The partial dierential operator L can be written as P +Q so that the partial
dierential operator
P =
1
a
2

2
t
2
depends only on the variable t and the partial dierential operator
Q =

2
r
2
+
1
r

r
+
1
r
2

2
Math 115 (2006-2007) Yum-Tong Siu 3
depends only on the variables r, . The domain D in the space of the variables
t, r, can be written as the product of the domain { t 0 } in the space of the
variable t and the domain { 0 r c } of the space of the variables r, . We
apply the method of the separation of variables and consider special product
functions u = T(t)v (r, ). For the special product function u = T(t)v (r, )
we impose the boundary conditions
u = 0 at r = c and for all t 0,
u
t
= 0 at t = 0,
and save the remaining boundary condition
u = f (r, ) at t = 0
to be used later to determine the coecients c
j
. The method of separation
of variables gives us two equations
d
2
dt
2
T (t) + a
2
T (t) = 0,

2
r
2
v (r, ) +
1
r

r
v (r, ) +
1
r
2

2
v (r, ) + v (r, ) = 0,
where is a constant and must be nonnegative, because is a nonnegative
dierential operator in the sense that the integral of the product of g and
g is nonnegative for g with compact support as one can see by transforming
the integral to the integral of the gradient square of f by using integration
by parts.
The solution of the dierential equation
d
2
dt
2
T (t) + a
2
T (t) = 0
with the boundary condition T
t
= 0 for t = 0 gives (up to a nonzero constant
multiple) the solution
T(t) = cos
_
a

t
_
for each .
Though the partial dierential operator
Q =

2
r
2
+
1
r

r
+
1
r
2

2
Math 115 (2006-2007) Yum-Tong Siu 4
is not the sum of two partial dierential operators with each one depending
only on one of the two variables r, , yet we can change

2
r
2
v (r, ) +
1
r

r
v (r, ) +
1
r
2

2
v (r, ) + v (r, ) = 0
to the new partial dierential equation
r
2

2
r
2
v (r, ) + r

r
v (r, ) +

2

2
v (r, ) + r
2
v (r, ) = 0
so that the partial dierential operator
r
2

2
r
2
+ r

r
+

2

2
+ r
2
is the sum of the two partial dierential operator
r
2

2
r
2
+ r

r
+ r
2
and

2
with each one depending on only one of the two variables r, . We can
now apply again the method of separation of variables and consider special
product functions v (r, ) = R(r)(). We get two dierential equations
r
2
d
2
R
dr
2
+ r
dR
dr
+ r
2
R R = 0,
d
2

dl
2
+ = 0,
where is a constant and must be nonnegative, because the operator

d
2
d
2
is nonnegative as one can see by integration by parts. Recall that we still have
the boundary condition R(c) = 0. There are one other boundary condition
for R(r) and two boundary conditions for (), which are not as obvious.
The other boundary condition for R(r) is that R(r) is nite at r = 0. As we
Math 115 (2006-2007) Yum-Tong Siu 5
will see later that this is indeed an important boundary condition for R(r),
because the dierential equation
r
2
d
2
R
dr
2
+ r
dR
dr
+ r
2
R R = 0
is not regular at r = 0 in the sense that the coecient r
2
for the term
d
2
R
dr
2
of
the highest-order dierentiation is not nonzero at r = 0. The two boundary
conditions for () come from the periodicity of () with period 2. These
two boundary conditions are
(0) = (2),

(0) =

(2).
For the solution () of a second-order dierential equation these two con-
ditions are equivalent to the the periodicity of () with period 2.
Second-Order Ordinary Dierential Equations with Boundary Conditions In-
volving Both End-Points. Suppose we have a second-order ordinary dieren-
tial equation
a(x)y

+ b(x)y

+ c(x)y = 0
on a nite closed interval [, ] with a(x), b(x), c(x) smooth on [, ] and
a(x) nowhere zero on [, ]. For the boundary condition y () = 0 and
y

() = 0 involving only one single end-point , there is only one solution

y(x) of the dierential equation which is identically zero.
However, when we have two homogeneous boundary conditions involv-
ing both end-points and , for example, the two homogeneous boundary
conditions
y() = 0 and y() = 0
or the two homogeneous boundary conditions
y() = y() and y

() = y

(),
there may exist another solution y(x) other than the identically zero solution.
The existence of non identically zero solutions for certain homogeneous
linear second-order ordinary dierential equations subject to two homoge-
neous boundary conditions involving both end-points makes it possible to
apply the method of separation to use complete system of eigenfunctions for
dierential operators.
Math 115 (2006-2007) Yum-Tong Siu 6
In our analysis of the wave equation for a vibrating circular membrane,
for a nonnegative number we have the second order dierential equation
d
2

d
2
+ = 0
for the unknown function () on [0, 2] with the two homogeneous boundary
conditions
(0) = (2),

(0) =

(2).
involving both end-points 0 and 2 of the interval [0, 2]. A general solution
of the second-order dierential equation is of the form
sin (

) + cos (

) with , R.
The two homogeneous boundary conditions are equivalent to the condition
that the solution is periodic of period 2. In order to have a non identically
zero solution, it is necessary and sucient that

is an integer. Thus the
constant must be of the form n
2
for some nonnegative integer n. When
= n
2
, the two linearly independent solutions are sin n and cos n. Each
of the two functions sin n and cos n is an eigenfunction for the operator

d
2
d
2
corresponding to the eigenvalue n
2
, because

d
2
d
2
sin n = n
2
sin n,

d
2
d
2
cos n = n
2
cos n,
We know from the theory of Fourier series that the set of all eigenfunctions
1, cos nx, sin nx for n N
form a complete system of functions on [0, 2] in the sense that any L
2
function (i.e., square-integrable function) on [0, 2] is an innite linear com-
bination of them in the sense of L
2
.
Math 115 (2006-2007) Yum-Tong Siu 7
We now use = n
2
consider the dierential equation
r
2
d
2
R
dr
2
+ r
dR
dr
+ r
2
R n
2
R = 0
for the unknown function R(r) on [0, c] with the two boundary conditions
R(x) is nite at x = 0 and R(c) = 0.
We get rid of with the rescaling which replaces r by r as the variable. In
other words, we dene y = y(x) by R(r) = y
_

r
_
so that y(r) = R
_
1

r
_
.
Then y = y(x) satises the dierential equation
() x
2
y

+ xy

+
_
x
2
n
2
_
y = 0,
because
x
dy
dx
(x) =
_
1

x
_
R

_
1

x
_
,
x
2
d
2
y
dx
2
(x) =
_
1

x
_
2
R

_
1

x
_
and
x
2
y

+ xy

+
_
x
2
n
2
_
y
=
_
1

x
_
2
R

_
1

x
_
+
_
1

x
_
R

_
1

x
_
+
_

_
1

x
_
2
n
2
_
R

_
1

x
_
.
The boundary conditions become y(x) nite at x = 0 and y(c) = 0. The
dierential equation () is known as Bessels dierential equation for the
Bessel function of order n. We are going to solve this equation by using a
generating function which is a Laurent series in a new indeterminate t whose
n-th coecient is the Bessel function of order n. We are going to write down
the generating function and verify that its n-th coecient satises Bessels
dierential equation for the Bessel function of order n.
Math 115 (2006-2007) Yum-Tong Siu 8
Generating Function for Bessel Functions. We now introduce the generating
function for the Bessel functions. It is e
x
2
(t
1
t
)
and we denotes the n-th
coecient by J
n
(t) for n Z so that
() e
x
2
(t
1
t
)
=

n=
J
n
(x)t
n
.
First we are going to verify that the function J
n
(x) satises Bessels dier-
ential equation (). We will do the verication by dierentiating () with
respect to x twice and dierentiating () with respect to t twice. Dierenti-
ating () with respect to x once, we get
()

1
2
_
t
1
t
_
e
x
2
(t
1
t
)
=

n=
J

n
(x)t
n
.
Dierentiating () with respect to x one more time, we get
()

_
1
2
_
t
1
t
__
2
e
x
2
(t
1
t
)
=

n=
J

n
(x)t
n
.
Dierentiating () with respect to t once, we get
x
2
_
1 +
1
t
2
_
e
x
2
(t
1
t
)
=

n=
nJ
n
(x)t
n1
.
and multiplying it by t, we get
()

x
2
_
t +
1
t
_
e
x
2
(t
1
t
)
=

n=
nJ
n
(x)t
n
.
Dierentiating ()

with respect to t one more time, we get

_
x
2
_
1
1
t
2
_
+
x
2
_
t +
1
t
_
x
2
_
1 +
1
t
2
__
e
x
2
(t
1
t
)
=

n=
n
2
J

n
(x)t
n1
.
and multiplying it by t, we get
()

_
x
2
_
t
1
t
_
+
_
x
2
_
t +
1
t
__
2
_
e
x
2
(t
1
t
)
=

n=
n
2
J

n
(x)t
n
.
Math 115 (2006-2007) Yum-Tong Siu 9
We now form the equation
x
2
()

+ x()

+ x
2
() ()

whose left-hand side is e

x
2
(t
1
t
)
times
_
x
2
_
t
1
t
__
2
+
x
2
_
t
1
t
_

_
x
2
_
t
1
t
_
+
_
x
2
_
t +
1
t
__
2
_
+ x
2
=
_
x
2
_
t
1
t
__
2

_
x
2
_
t +
1
t
__
2
+ x
2
=
_
x
2
_
2
_
t
2
2 +
1
t
2
_

_
x
2
_
2
_
t
2
+ 2 +
1
t
2
_
+ x
2
= 0
and whose right-hand side is

n=
_
x
2
J

n
(x) + xJ

n
(x) +
_
x
2
n
2
_
J
n
(x)
_
t
n
.
This is nishes the verication of the dierential equation () for the Bessel
functions J
n
(x).
Relation Between Bessel Functions for an Integer and the Negative of the
Integer. The generating function e
x
2
(t
1
t
)
is unchanged when t is replaced by

1
t
which from () means that
J
n
(x) = (1)
n
J
n
(x) for n Z.
Since Bessels dierential equation () is invariant under n n, the rela-
tion J
n
(x) = (1)
n
J
n
(x) for n Z precludes the easy way of using J
n
(x)
to get another solution which is not a scalar multiple of J
n
(x). For n Z it
will take a more involved procedure to get another solution which is not a
scalar multiple of J
n
(x). We will indicate how this additional solution is ob-
tained but for the problem of the vibrating circular membrane the additional
solution is not needed.
Power Series Expansion for Bessel Functions. We now compute the power
series expansion for Bessel functions directly from the power series expansion
Math 115 (2006-2007) Yum-Tong Siu 10
of the generating function e
x
2
(t
1
t
)
. We write the generating function as the
product two functions so that
e
x
2
(t
1
t
)
= e
xt
2
e

x
2t
.
The power series expansion of the rst factor is
e
xt
2
=

j=0
x
j
j!2
j
t
j
and the power series expansion of the second factor is
e

x
2t
=

k=0
(1)
k
x
k
k!2
k
t
k
.
We are interested in the coecient of t
n
in the product of these two power
series expansion. In order to get t
n
as a product we have to consider it as
the product of t
j
and t
k
with j k = n. These the coecient of t
n
in the
product of the two power series expansion is the sum of the coecient of t
j
in
the rst series times the coecient of t
k
in the second series with j k = n.
Thus
J
n
(x) =

jk=n
x
j
j!2
j
t
j
(1)
k
x
k
k!2
k
t
k
which can be rewritten as
(&) J
n
(x) =

k=0
(1)
k
k!(n + k)!
_
x
2
_
n+2k
when we remove the index j by using j = n + k. We would like to remark
that at x = 0 the value of J
0
is 1 but the value of J
n
is 0 for any positive
integer n.
Bessel Functions of Non-Integral Order and the Other Solution of the Bessel
Dierential Equation in the Case of Integral Order. When we substitute
the power series (&) into the Bessel dierential equation, we can readily see
that J
n
(x) given by (&) satises the Bessel dierential equation. The same
computation shows that if we allow n to be a non-integer and replace (n+k)!
by the Gamma function (n + k + 1), then Bessels dierential equation is
Math 115 (2006-2007) Yum-Tong Siu 11
satised also by J
n
(x) given by (&) even when n is not an integer. Recall
that the Gamma function is dened by
(x) =
_

t=0
e
t
t
x1
dt.
For any real number we dene the Bessel function by the power series
J

(x) =

k=0
(1)
k
k! ( + k + 1)
_
x
2
_
+2k
.
Then the Bessel function J

(x) satises the Bessel dierential equation

x
2
y

+ xy

+
_
x
2

2
_
y = 0.
When is not an integer, the two solutions J

(x) and J

(x) are linearly

independent. We would like to mention without going into any further details
that for an integer n we can use the following Hankels function
Y
n
(x) = lim
0
1

(J
n+
(x) (1)
n
J
n
(x))
as the second solution of Bessels dierential equation for n. This is the
standard technique of taking the limit, as 0, of a normalized linear
combination of two solutions for Bessels dierential equation for n + . It
turns out that another less obvious linear combination gives a more elegant
simpler expression.
1
2
Y
n
(x)+J
n
(x) (log 2 ) = J
0
(x) log x+2
_
J
2
(x)
1
2
J
4
(x) +
1
3
J
6
(x) +
_
,
where
= lim
k
_
1 +
1
2
+ +
1
k
log k
_
is Eulers constant. This formula shows that the singularity order of Y
n
(x)
is log x as x 0, which is to be expected, because Y
n
(x) is obtained by
dierentiating
J

(x) =

k=0
(1)
k
k! ( + k + 1)
_
x
2
_
+2k
.
Math 115 (2006-2007) Yum-Tong Siu 12
with respect to and
d
dt
t
x
= t
x
log x.
We conclude that the only solution of Bessels dierential equation () for
an integer n which has no singularity at x = 0 is J
n
(x) up to a constant
multiple. Thus, to specify that the solution of () without singularity has
the same eect specifying an initial condition to single out certain solutions
of dierential equations. The reason why the initial condition takes the form
of specifying no singularity at x = 0 is because the coecient of the higher-
order term in Bessels dierential equation vanishes at x = 0 and is thus a
dierential equation which is singular at x = 0.
Derivatives of Bessel Functions and Recurrent Relation of Bessel Functions.
We now discuss the derivatives of Bessel functions and recurrent relation of
Bessel functions. In two ways this discussion is needed for the problem of the
vibrating circular membrane. The rst is that we need some properties of
zeroes of J
n
(x) for x 0. The second is that we need the L
2
norm of J
n
(x)
with respect to the weight function x over [0, c].
Bessels dierential equation expresses the second-order derivative of the
Bessel function in terms of the function and its rst-order derivative, but
does not give immediately an explicit expression for the rst-order derivative
of the Bessel function. On the other hand, the generating function, when
its rst-order derivative with respect to x is compared with its rst-order
derivative with respect to t, can give formulas for rst-order derivatives of
the Bessel function as follows. Dierentiating the generating function in ()
with respect to x gives ()

and multiplying both sides by x gives

()
x
2
_
t
1
t
_
e
x
2
(t
1
t
)
=

n=
xJ

n
(x)t
n
.
Dierentiating the generating function in () with respect to t and multi-
plying both sides by x gives ()

. In order to compare () with ()

, we
add
x
t
e
x
2
(t
1
t
)
=

n=
x
t
J
n
(x)t
n
=

n=
xJ
n+1
(x)t
n
to () so that we have
()
x
2
_
t +
1
t
_
e
x
2
(t
1
t
)
=

n=
(xJ

n
(x) + xJ
n+1
(x)) t
n
.
Math 115 (2006-2007) Yum-Tong Siu 13
We now can equate the right-hand side of ()

with the right-hand side of

() and get
nJ
n
(x) = xJ

n
(x) + xJ
n+1
(x)
or equivalently
(%) xJ

n
(x) = nJ
n
(x) xJ
n+1
(x).
Replacing n by n, we get
xJ

n
(x) = nJ
n
(x) xJ
n+1
(x).
using J
n
= (1)
n
J
n
(x), we get
($) xJ

n
(x) = nJ
n
(x) + xJ
n1
(x).
A more compact way of writing these two formulas for the rst-order deriva-
tives of Bessel functions is
_
J
n
x
n
_

=
J
n+1
x
n
,
(x
n
J
n
)

= x
n
J
n1
.
Eliminating J

n
(x) from (%) and ($), we get the following algebraic recurrent
formula for Bessel functions.
xJ
n+1
(x) = 2nJ
n
(x) xJ
n1
(x).
Zeros of Bessel Function of Order Zero. Recall that we have the boundary
condition R(c) = 0. With R(c) = y
_

c
_
the boundary condition becomes
y
_

c
_
= 0. Now y(x) = J
n
(x). So we have to formulate the boundary
condition in terms of J
n
and it becomes J
n
_

c
_
= 0. In order to determine
the constant , we have to locate the zeroes of J
n
for each nonnegative integer
n. Let us rst consider the simplest case of J
0
(x). The Bessel dierential
equation for it is
x
2
y

+ xy

+
_
x
2
_
y = 0
which is the same as
xy

+ y

+ xy = 0.
Math 115 (2006-2007) Yum-Tong Siu 14
We can get rid of the term involving the rst-order derivative by introducing
a new dependent variable u(x) =

x y. Then u

=
1
2

x
y +

xy

and
u

=
1
4x

x
y +
1

x
y

xy

so that
u

=
1
4x

x
y +
1

x
(y

+ xy

) =
_

x +
1
4x

x
_
y
which is the same as
u

(x) =
_
1 +
1
4x
2
_
u(x).
When x is large, this can be compared with
v

(x) = v(x)
which admits v(x) = sin x as a solution. The key technique is to use the
derivative of the Wronskian

u v
u

.
In general the Wronskian of n functions f
1
, , f
n
is

f
1
f
2
f
n
f

1
f

2
f

n



f
(n1)
1
f
(n1)
2
f
(n1)
n

and is used to investigate the linear independence of f

1
, , f
n
. The deriva-
tive of the Wronskian

u v
u

is
(uv

vu

= uv

vu

,
Math 115 (2006-2007) Yum-Tong Siu 15
which is equal to
uv v
_
1 +
1
4x
2
_
u =
uv
4x
2
.
Integrating over [a, b] yields
_
b
a
uv
4x
2
dx = vu

uv

x=b
x=a
.
We choose a = 2k and b = (2k + 1) for k Z. Since
v(a) = v(b) = 0, v

(a) = 1, v

(b) = 1
for our choice of v(x) = sin x, it follows that
_
(2k+1)
2k
u(x)
sin x
x
2
dx = 4 (u(2k) + u((2k + 1))) .
Since
sin x
x
2
> 0 for 2k < x < (2k + 1),
it follows by sign consideration that u(x) has at least one zero in the closed
interval [2k, (2k + 1)] of length . Since u(x) = J
0
(x) is a convergent
power series on [2k, (2k +1)], there can only be a nite number of zeroes
for u(x) in [2k, (2k + 1)]. We are only interested in the zeroes of J
0
(x)
which are positive numbers. All the zeroes of J
0
(x) must be simple, because
it satises a linear homogeneous second-order dierential equation, otherwise
the uniqueness statement for the second-order dierential equation will force
J
0
(x) to be identically zero. Thus the zeroes

0,1
,
0,2
,
0,3
, ,
0,

of J
0
(x) for x 0 can be arranged in a strictly increasing sequence which
approaches innity. Remember that J
0
(0) = 1 and as a result all the non-
negative zeroes of J
0
(x) are positive. Let
0,
=

0,
c
so that J
0
(
0,
c) = 0.
Let
0,
=
0,
2
. Then R(r) = J
0
(
0,
r) satises the dierential equation
r
2
d
2
R
dr
2
+ r
dR
dr
+
0,
r
2
R = 0
Math 115 (2006-2007) Yum-Tong Siu 16
and the two boundary conditions that R(r) is nite at r = 0 and R(c) = 0.
After we rewrite the dierential equation as

d
2
R
dr
2
r
dR
dr
=
0,
R
and conclude that R(r) = J
0
(
0,
r) is an eigenfunction of the dierential
operator

d
2
dr
2

1
r
2
d
dr
for the eigenvalue
0,
=
0,
2
with the two boundary conditions that R(r) is
nite at r = 0 and R(c) = 0.
By the Sturm-Liouville theorem which we will introduce later, the family
of functions J
0
(
0,
x) for N is a complete orthogonal family of functions
on [0, c] with respect to the weight function x. This completeness property
will be used to express our sought-after solution of the vibrating circular
membrane as an innite sum of special product solutions obtained by the
method of separation of variables.
Zeroes of Bessels Function for Higher Integral Order. We now use Rolles
theorem and induction on n to investigate the zeroes of J
n
(x). The key is
the formula for the rst-order derivative of J
n
(x), which is
()
_
J
n
x
n
_

=
J
n+1
x
n
.
By Rolles theorem, there is at least one zero of J

n
between two consecutive
zeroes of J
n
. The above formula () tells us that there is at least zero of
J
n+1
between two consecutive zeroes of J
n
. Note that from the power series
expansion (&) of J
n
(x) centered at x = 0 we know that the vanishing order
of J
n
(x) at x = 0 is precisely n. Recall that for the problem of the vibrating
circular member only the nonnegative zeroes of J
n
(x) are of interest.
Since J
n
is the solution of a linear homogeneous second-order dierential
equation, its derivative cannot vanish at any one of its zeros. This means
that J

n
must alternate its sign between two consecutive zeroes of J
n
for
the following reason. For example, if after the zero of J
n
(x) at x = and
before the next zero at = > , the sign of J
n
(x) is positive, then from the
consideration of the dierence quotient the derivative J

n
(x) at x = must
Math 115 (2006-2007) Yum-Tong Siu 17
be nonnegative and the derivative J

n
(x) at x = must be nonpositive, and
since they are both nonzero, the derivative J

n
(x) at x = must be positive
and the derivative J

n
(x) at x = must be negative. By the above formula
() we conclude that J
n+1
must alternate its sign between two consecutive
zeroes of J
n
.
Just like the case of J
0
(x), we have the following similar situation for
J
n
(x) for any nonnegative integer n. The zeroes

n,1
,
n,2
,
n,3
, ,
n,

of J
n
(x) for x 0 can be arranged in a strictly increasing sequence which
approaches innity. For n 1 the origin x = 0 is a zero of J
n
(x) of order
n and
n,1
= 0 for n 1. Let
n,
=

n,
c
so that J
n
(
0,
c) = 0. Let

n,
= (
n,
)
2
. Then R(r) = J
n
(
n,
r) satises the dierential equation
r
2
d
2
R
r
2
+ r
dR
r
+
n,
r
2
R n
2
R = 0
and the two boundary conditions that R(r) is nite at r = 0 and R(c) = 0.
After we rewrite the dierential equation as

d
2
R
r
2
r
dR
r

n
2
r
2
R =
n,
R
and conclude that R(r) = J
n
(
n,
r) is an eigenfunction of the dierential
operator

d
2
dr
2

1
r
2
d
dr

n
2
r
2
for the eigenvalue
n,
= (
n,
)
2
with the two boundary conditions that R(r)
is nite at r = 0 and R(c) = 0.
Again, by the Sturm-Liouville theorem which we will introduce later, the
family of functions J
n
(
n,
x) for N is a complete orthogonal family of
functions on [0, c] with respect to the weight function x and this completeness
property will be used to express our sought-after solution of the vibrating
circular membrane as an innite sum of special product solutions obtained
by the method of separation of variables.
Math 115 (2006-2007) Yum-Tong Siu 18
Orthogonality of the Family J
n
(
n,
x) for N. While the completeness
of the family J
n
(
n,
x) for N over [0, c] with weight function x depends
on the theorem of Sturm-Liouville with some sophisticated arguments, the
orthogonality property of the family comes from the orthogonality of two
eigenfunctions of a self-adjoint operator for two distinct eigenvalues. We
now verify the orthogonality property of the family.
Fix a nonnegative integer n. Take < m. Let y(x) = J
n
(
n,
x) and
z(x) = J
n
(
n,m
x) on [0, c]. These two functions satisfy the following two
dierential equations
x
2
d
2
y
x
2
+ x
dy
dx
+
n,
x
2
y n
2
y = 0,
x
2
d
2
z
dx
2
+ x
dz
dx
+
n,m
x
2
z n
2
z = 0,
which we can put in divergence form
(xy

n
2
x
y =
n,
xy.
(xz

n
2
x
z =
n,m
xy.
Multiplying the rst equation by z and the second one by y and taking their
dierence, we get
(xy

(x))

z(x) (xz

(x))

y(x) = (
n,

n,m
) xy(x)z(x).
Integrating over 0 x c and using integration by parts, we get
xy

(x)z(x) xz

(x)y(x)

x=c
x=0
= (
n,

n,m
)
_
x=c
x=0
xy(x)z(x) dx.
From the vanishing of y(x) and z(x) at x = c and their niteness at x = 0 it
follows that the left-hand side is zero and
_
x=c
x=0
xy(x)z(x) dx = 0,
which is the orthogonality of J
n
(
n,
x) and J
n
(
n,m
x) over [0, c] with respect
to the weight function x for 1 < m.
Math 115 (2006-2007) Yum-Tong Siu 19
Norm of Eigenfunctions of Bessels Dierential Equation for Integral Order.
In order to determine the coecients in the innite sum which expresses
a given function in terms of the family J
n
(
n,
x) for N over [0, c] with
weight function x, we need to have the L
2
norm of each member of the family
over [0, c] with weight function x. We now compute this L
2
norm.
Fix a nonnegative integer n and x a positive integer . Let y(x) =
J
n
(
n,
x). The function y(x) satises the dierential equation
x
2
y

+ xy

+
_
(
n,
)
2
x
2
n
2
_
y = 0.
After multiplying the equation by 2y

, we can rewrite it as
d
dx
(xy

)
2
+
_
(
n,
)
2
x
2
n
2
_
d
dx
y
2
= 0.
Integrating from x = 0 to x = c yields
(xy

)
2
+
_
(
n,
)
2
x
2
n
2
_
y
2

x=c
x=0
= 2 (
n,
)
2
_
c
x=0
xy
2
dx.
Since y(c) = 0 and since both y(x) and y

(x) is nite at x = 0 and y(0) = 0

when n > 0, it follows that
_
(
n,
)
2
x
2
n
2
_
y
2

x=c
x=0
= 0
and
(xy

)
2

x=c
x=0
= (cy

(c))
2
.
Thus we have
(cy

(c))
2
= 2 (
n,
)
2
_
c
x=0
xy
2
dx
and the formula for the L
2
is given by
_
c
x=0
x(J
n
(
n,
x))
2
dx =
c
2
2
(J

n
(
n,
c))
2
.
We can get an alternative form by using the formula (%)
xJ

n
(x) = nJ
n
(x) xJ
n+1
(x)
Math 115 (2006-2007) Yum-Tong Siu 20
at x =
n,
c so that

n,
c J

n
(
n,
c) = nJ
n
(
n,
c) (
n,
c) J
n+1
(
n,
c)
and
J

n
(
n,
c) = J
n+1
(
n,
c) ,
because J
n
(
n,
c) = 0. Thus
_
c
x=0
x(J
n
(
n,
x))
2
dx =
c
2
2
(J
n+1
(
n,
c))
2
.
Final Answer of Problem of Vibrating Circular Membrane. We now have the
special product solutions
T(t)R(r)() = cos (a
n,
t) J
n
(
n,
r)
_
_
_
1 if n = 0
cos n if n N
sin n if n N
for N. To get to this point, we have already used up the following ve
boundary conditions
T

(0) = 0,
(0) = (2),

(0) =

(2),
R(r) nite at r = 0,
R(c) = 0.
There is the following boundary condition u(t, r, ) = f(r, ) at t = 0, which
has not been used. Observe that cos (a
n,
t) = 1 at t = 0. Now we form the
R-linear combination of all the special product functions given above with
yet-to-be-determined coecients A
n,
(for integers n 0 and 1) and B
n,
(for integers n 1 and 1) so that
f(r, ) =

=1
A
0,
2
J
0
(
0,
r)+

n=1

=1
J
n
(
n,
r) (A
n,
cos (n) + B
n,
sin (n)) .
Note that here the case of n = 0 is singled out, because the Fourier series
expansion for f() is of the form
A
0
2
+

n=1
(A
n
cos (n) + B
n
sin (n)) ,
Math 115 (2006-2007) Yum-Tong Siu 21
where
A
n
=
1

_
2
=0
f () cos (n) d for n 0,
B
n
=
1

_
2
=0
f () sin (n) d for n 1,
and the case for n = 0 has to be treated separately. Here for the vibrating
circular membrane, we determine the coecients A
n,
(for integers n 0 and
1) and B
n,
(for integers n 1 and 1) by rst using the formula for
the Fourier series coecients to get
g
n
(r) =
1

_
2
=0
f (r, ) cos (n) d for n 0,
h
n
(r) =
1

_
2
=0
f (r, ) sin (n) d for n 1,
and then for any xed n using the expansion in terms of the complete family
J
n
(
n,
x) for N to get A
n,
from g
n
(r) and to get B
n,
from h
n
(r). We
use the formula
_
c
x=0
x(J
n
(
n,
x))
2
dx =
c
2
2
(J
n+1
(
n,
c))
2
for the L
2
norm of J
n
(
n,
x) to get
A
n,
=
1
c
2
2
(J
n+1
(
n,
c))
2
_
c
r=0
r g
n
(r)J
n
(
n,
r) dr,
B
n,
=
1
c
2
2
(J
n+1
(
n,
c))
2
_
c
r=0
r h
n
(r)J
n
(
n,
r) dr
to get the following nal answer u(t, r, ) for the problem of the vibrating
circular membrane.
u(t, r, ) =

=1
A
0,
2
cos (a
0,
t) J
0
(
0,
r)
+

n=1

=1
cos (a
n,
t) J
n
(
n,
r) (A
n,
cos (n) + B
n,
sin (n)) .