MATH306 SUPPLEMENTARY MATERIAL

A BRIEF INTRODUCTION TO BESSEL and

RELATED SPECIAL FUNCTIONS
Edmund Y. M. Chiang
c Draft date December 1, 2008

Contents
Contents

1 Trigonometric and Gamma Functions

1.1

Trigonometric Functions . . . . . . . . . . . . . . . . . . . . . . . . .

1.2

Gamma Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1.3

Chebyshev Polynomials . . . . . . . . . . . . . . . . . . . . . . . . . .

2 Bessel Functions

2.1

Power Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2.2

Bessel Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2.3

Basic Properties of Bessel Functions . . . . . . . . . . . . . . . . . . . 12

2.3.1

Zeros of Bessel Functions . . . . . . . . . . . . . . . . . . . . . 12

2.3.2

Recurrence Formulae for J . . . . . . . . . . . . . . . . . . . 12

2.3.3

Generating Function for Jn . . . . . . . . . . . . . . . . . . . . 13

2.3.4

Lommels Polynomials . . . . . . . . . . . . . . . . . . . . . . 14

2.3.5

Bessel Functions of half-integer Orders . . . . . . . . . . . . . 14

2.3.6

Formulae for Lommels polynomials . . . . . . . . . . . . . . . 14

2.3.7

Pythagoras Theorem for Bessel Function . . . . . . . . . . 15

2.3.8

Orthogonality of the Lommel Polynomials . . . . . . . . . . . 15

2.4

Integral formaulae

2.5

Asymptotic Behaviours of Bessel Functions . . . . . . . . . . . . . . . 17

2.5.1

2.6

. . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

Addition formulae . . . . . . . . . . . . . . . . . . . . . . . . . 18

Fourier-Bessel Series . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
i

CONTENTS
3 An New Perspective

1
21

3.1

Hypergeometric Equations . . . . . . . . . . . . . . . . . . . . . . . . 21

3.2

Confluent Hypergeometric Equations . . . . . . . . . . . . . . . . . . 23

3.3

A Definition of Bessel Functions . . . . . . . . . . . . . . . . . . . . . 23

4 Physical Applications of Bessel Functions

27

5 Separation of Variables of Helmholtz Equations

31

5.1

What is not known? . . . . . . . . . . . . . . . . . . . . . . . . . . . 36

CONTENTS

Chapter 1
Trigonometric and Gamma
Functions
1.1

Trigonometric Functions

Pythagoras Theorem:
sin2 x + cos2 x = 1

(1.1)
holds for all real x.
Addition Theorem:
(1.2)

sin(A + B) = sin A cos B + sin B cos A

(1.3)

cos(A + B) = cos A cos B sin A sin B

hold for all angles A, B.

1.2

Gamma Function

We recall that
(1.1)

k! = k (k 1) 3 2 1.

Euler was able to give a correct definition to k! when k is not a positive integer. He
invented the Euler-Gamma function in the year 1729. That is,
Z
(1.2)
(x) =
tx1 et dt.
0

CHAPTER 1. TRIGONOMETRIC AND GAMMA FUNCTIONS

So the integral will converge for all positive real x. Since

Z
(1.3)
(1) =
et dt = 1,
0

so one can show

(1.4)

(x + 1) = x(x),

and so for each positive integer n

(1.5)

(n + 1) = n!.

In fact, the infinite integral will not only converge for all positive real x but can
be analytically continued into the whole complex plane C. We have
Z
(1.6)
(z) =
tz1 et dt, <(z) > 0,
0

which is analytic in C except at the simple poles at the negative integers including
0. So it is a meromorphic function.
Euler worked out that
(1.7)

1
2

!=

1
.
2

So
 1
11 9 7 5 3  1 
5 !=
, , ,
!.
2
2 2 2 2 2 2
One can even compute negative factorial:

(1.8)

(1.9)

1.3

(1/2) = (1/2)(3/2)(5/2)(19/2)(11/2)(11/2).

Chebyshev Polynomials

Theorem 1.3.1 (De Moivres Theorem). Let z = r(cos + i sin ) be a complex

number. Then for every non-negative integer n,
(1.1)

z n = rn (cos n + i sin n).

Proof It is clear that that the formula holds when n = 1. We assume it holds at the
integer n. Then the addition formulae for the trigonometric theorem imply that
z n+1 = rn+1 (cos + i sin )(cos n + i sin n)
= rn+1 [cos(n + 1) + i sin(n + 1)].
Thus, the result follows by applying the principle of induction.

1.3. CHEBYSHEV POLYNOMIALS

Exercise 1.3.2. Justify the formula for negative integers n.

Remark 1.3.1. De Moivres formula actually works for all real and complex n. See
MATH304.
Example 1.3.3 (Multiple angle formulae). Let n = 2 in De Movires formula yields
z 2 = cos 2 + i sin 2.

(1.2)
But
(1.3)

z 2 = (cos + i sin )2 = cos2 sin2 + i(2 cos sin ).

Comparing the real and imaginary parts of (1.2) and (1.3) yields
cos 2 = cos2 sin2
= 2 cos2 1
= 1 2 sin2
and
(1.4)

sin 2 = 2 cos sin .

Similarly, we have
(1.5)

z 3 = cos 3 + i sin 3.

But

(1.6)

z 3 = (cos + i sin )3
= cos3 + 3 cos2 (i sin ) + 3 cos (i sin )2 + (i sin )3
= cos3 3 cos sin2 + i(3 cos2 sin sin3 ).

Comparing the real and imaginary parts of (1.5) and (1.6) yields
cos 3 = cos3 3 cos sin2
= 4 cos3 3 cos
and
sin 3 = 3 cos2 sin sin3
= 3 sin 4 sin3 .

CHAPTER 1. TRIGONOMETRIC AND GAMMA FUNCTIONS

More generally, we have

(1.7)

cos n + i sin n = z n = (cos + i sin )n

n  
X
n
=
cosnk (i sin )k
k
k=0
n  
X
n k
=
i cosnk sink .
k
k=0

Equating the real parts of the last equation and noting that i2l = (1)l give
 
 
n
n
n
n2
2
cos n = cos
cos
sin +
cosn4 sin4 +
2
4


n
[n/2]
+ (1)
(1.8)
cosn2[n/2] sin2[n/2] .
2[n/2]
But each of the sine function above has even power. Thus the substitution of
sin2 = 1 cos2 in (1.8) yields
[n/2] 

(1.9)

cos n =

X
l=0


 
l
X


n
n2l
k l
cos

(1)
cos2k .
2l
k
k=0

Finally, it is possible to re-write the summation formula as

(1.10)

cos n =

[n/2] 

X
k=0

(1)

[n/2] 

X
j=k

n
2j

 
j
cosn2k .
k

Writing = arccos x for 0 , and Tn (x) = cos(n arccos x). We deduce from
the above formula that Tn (x) is a polynomial in x of degree n. The polynomial Tn
is called the Chebyshev1 polynomial of the first kind of degree n. The first
few polynomials are given by
T0 (x) = 1;
T1 (x) = x;
T2 (x) = 2x2 1;
T3 (x) = 4x3 3x;
T4 (x) = 8x4 8x2 + 1;
T5 (x) = 16x5 20x3 + 5x;

P. L. Chebyshev (18211894) Russian mathematician. The collection of the polynomials

{Tn (x)}
0 is a complete orthogonal polynomial set on [1, 1]

1.3. CHEBYSHEV POLYNOMIALS

The set of Chebyshev polynomials {Tn (x)} can generate a L2 [1, 1] space. That is,
it is a Hilbert space.
Interested reader should consult the handbook [1] and the book [5] about the
Chebyshev polynomials by T. J. Rivlin. Our library has a copy of this book.

CHAPTER 1. TRIGONOMETRIC AND GAMMA FUNCTIONS

Chapter 2
Bessel Functions
2.1

Power Series

We define the Bessel function of first kind of order to be the complex function
represented by the power series
(2.1)

J (z) =

+
+
X
X
(1)k ( 21 z)+2k
(1)k ( 21 )+2k 2k
= z
z .
(
+
k
+
1)
k!
(
+
k
+
1)
k!
k=0
k=0

Here is an arbitrary complex constant and the notation () is the Euler Gamma
function defined by
Z
(2.2)
(z) =
tz1 et dt, <(z) > 0.
0

2.2

Bessel Equation

A common introduction to the series representation for Bessel functions of first kind
of order is to consider Frobenius method: Suppose
y(z) =

ck z +k ,

k=0

where is a fixed parameter and ck are coefficients to be determined. Substituting

the above series into the Bessel equation
(2.1)

z2

dy
d2 y
+ x + (z 2 2 )y = 0.
2
dz
dz
9

10

CHAPTER 2. BESSEL FUNCTIONS

and separating the coefficients yields:

0=
=

2
2d y
z
dz 2

+x

dy
+ (z 2 2 )y
dz

ck ( + k)( + k 1)z +k +

k=0

ck ( + k)z +k

k=0

+ (z 2 2 )

ck z +k

k=0

ck [( + k) ]z

k=0

+k

ck z +k+2

k=0

= c0 (2 2 )z +

{ck [( + k)2 2 ] + ck2 }z +k

k=0

The first term on the right side of the above expression is

c0 (2 2 )z
and the remaining are

(2.2)
(2.3)

c1 [( + 1)2 2 ] + 0
c2 [( + 2)2 2 ] + c0
c3 [( + 3)2 2 ] + c1

2
2
ck [( + k) ] + ck2

= 0
= 0
= 0
= 0

Hence a series solution could exist only if = . When k > 1, then we

require
ck [( + k)2 2 ] + ck2 = 0,
and this determines ck in terms of ck2 unless = 2 or + = 2 is an
integer. Suppose we discard these exceptional cases, then it follows from (2.2) that
c1 = c3 = c5 = = c2k+1 = = 0.

Thus we could express the coefficients c2k in terms of

c2k =

(1)k c0
.
( + 2)( + 4) ( + 2k) ( + + 2)( + + 4) ( + + 2k)

2.2. BESSEL EQUATION

11

If we now choose = , then we obtain

h
X
c0 z 1 +

(2.4)

k=0

i
(1)k ( 12 z)2k
.
k!( + 1)( + 2) ( + k)

Alternatively, if we choose = , then we obtain

c00 z

(2.5)

1+

X
k=0

i
(1)k ( 12 z)2k
.
k!( + 1)( + 2) ( + k)

Since c0 and c00 are arbitrary, so we choose them to be

(2.6)

c0 =

1
2 (

+ 1)

c00 =

1
2 (

+ 1)

so that the two series (2.4) and (2.5) can be written in the forms:
(2.7)

J (z) =

+
X
(1)k ( 21 z)+2k
,
(
+
k
+
1)
k!
k=0

J (z) =

+
X
(1)k ( 21 z)+2k
( + k + 1) k!
k=0

and both are called the Bessel function of order if the first kind.
It can be shown that the Wronskian of J and J is given by (G. N. Watson
A Treatise On The Theory Of Bessel Functions, pp. 4243):
(2.8)

W (J , J ) =

2 sin
.
z

This shows that the J and J forms a fundamental set of solutions when is
not equal to an integer. In fact, when = n is an integer, then we can easily check
that
(2.9)

Jn (z) = (1)n Jn (z).

Thus it requires an extra effort to find another linearly independent solution.

It turns out a second linearly independent solution is given by
(2.10)

Y (z) =

J (z) cos J (z)

sin

when is not an integer. The case when is an integer n is defined by

(2.11)

J (z) cos J (z)

.
n
sin

Yn (z) = lim

12

CHAPTER 2. BESSEL FUNCTIONS

The Y so defined is linearly independent with J for all values of .

In particular, we obtain
n1

(2.12)
(2.13)

1 X (n k 1)!  z 2kn
Yn (z) =
k=0
k!
2

i
z
1 X (1)k (z/2)n+2k h
2 log (k + 1) (k + n + 1)
+
k=0 k!(n + k)!
2

for | arg z| < and n = 0, 1, 2, with the understanding that we set the sum
to be 0 when n = 0. Here the (z) = 0 (z)/(z). We note that the function is
unbounded when z = 0.

2.3

Basic Properties of Bessel Functions

The general reference for Bessel functions is G. N. Watsons classic: A Treatise on

the Theory of Bessel Functions, published by Cambridge University Press in 1922
[6].

2.3.1

Zeros of Bessel Functions

See A. Abramowitz and I. A. Stegun, Handbook of Mathematical Functions with

formulas, graphs and mathematical tables, 10th Edt., National Bureau of Standards, 1964. [1]. [3] and [6]

2.3.2

Recurrence Formulae for J

We consider arbitrary complex .

d
(1)k z 2+2k
d
z J (z) =
dz
dz 2+2k k!( + k + 1)
d (1)k z 21+2k
=
dz 21+2k k!( + k)
= z J1 (z).
But the left side can be expanded and this yields
(2.1)

zJ0 (z) + J (z) = zJ1 (z).

2.3. BASIC PROPERTIES OF BESSEL FUNCTIONS

Similarly,
d nu
z J (z) = z J+1 (z).
dz

(2.2)
and this yields

zJ0 (z) J (z) = zJ+1 (z).

(2.3)

Substracting and adding the above recurrence formulae yield

2
J (z)
z
J1 (z) J+1 (z) = 2J0 (z).
J1 (z) + J+1 (z) =

(2.4)
(2.5)

2.3.3

Generating Function for Jn

Jacobi in 1836 gave

(2.6)

1
z(t 1t )
2

+
X

tk Jk (z).

k=

Many of the forumulae derived above can be obtained from this expression.
(2.7)

1
z(t 1t )
2

+
X

ck (z)tk

k=

for 0 < |t| < . We multiply the power series

zt

(2.8)

e2 =1+

(z/2)2 2
(z/2)
t+
t +
1!
1!

and
zt

e 2 = 1

(2.9)

(z/2)
(z/2)2 2
t+
t
1!
1!

Multiplying the two series and comparing the coefficients of tk yield

(2.10)
(2.11)

cn (z) = Jn (z), n = 0, 1,
cn (z) = (1)n Jn (z), n = 1, 2, .

Thus
(2.12)

1
z(t 1t )
2

= J0 (z) +

+
X
k=1

Jk [tk + (1)k tk ].

13

14

CHAPTER 2. BESSEL FUNCTIONS

2.3.4

Lommels Polynomials

Iterating the recurrence formula

2
J (z) J1
z
with respect to a number of times give

(2.13)

J+1 (z) =

J+k (z) = P (1/z)J (z) Q(1/z)J1 .

(2.14)

Lommel (1871) [6, pp. 294295] found that

J+k (z) = Rk, (z)J (z) Rk1,+1 J1 .

(2.15)

2.3.5

Bessel Functions of half-integer Orders

One can check that

(2.16)

J 1 (z) =
2

 2  12
cos z,
z

J 1 (z) =
2

 2  12
sin z.
z

Moreover,
r
(2.17)

Jn+ 1 (z) = (1)

2

2 n+ 1  d n sin z
z 2
,

z dz
z

n = 0, 1, 2, .

Thus applying a recurrence formula and using the Lommel polynomials yield
Jn+ 1 (z) = Rn, (z)J 1 (z) Rn1,+1 J 1 (z)

(2.18)

That is, we have

(2.19)

2.3.6

 2  12
 2  12
sin z Rn1,+1
sin z.
Jn+ 1 (z) = Rn, (z)
2
z
z

Formulae for Lommels polynomials

For each fixed , the Lommel polynomials are given by

[n/2]
X (1)k (n k)!()nk  2 n2k
(2.20)
Rn (z) =
k!(n 2k)!()k
z
k=0
where the [x] means the largest integer not exceeding x.
Lommel is a German mathematician who made a major contribution to Bessel
functions around 1870s.

2.3. BASIC PROPERTIES OF BESSEL FUNCTIONS

2.3.7

15

Pythagoras Theorem for Bessel Function

These Lommel polynomials have remarkable properties. Since

 2  12
 2  12
cos z,
J 1 (z) =
sin z
(2.21)
J 1 (z) =
2
2
z
z
and sin2 x + cos2 x = 1; we now have
2
2
Jn+
(z) = 2(1)n
1 (z) + J
n 1

(2.22)

R2n, 1 n (z)
2

That is, we have

2
Jn+
1 (z)
2

(2.23)

2
Jn
1 (z)
2

n
2 X (2z)2n2k (2n k)!(2n 2k)!
.
=
z k=0
[(n k)!]2 k!

A few special cases are

1. J 21 (z) + J2 1 (z) =
2

2. J 23 (z) + J2 3 (z) =
2

2
;
z
2 
z

1+

2 
1+
2
2
z
2 
4. J 27 (z) + J2 7 (z) =
1+
2
2
z
3. J 25 (z) + J2 5 (z) =

2.3.8

1
;
z2
3
9
+
;
z2 z4
6
45 225 
+
+ 6
z2 z4
z

Orthogonality of the Lommel Polynomials

Let us set
[n/2]

(2.24)

hn, (z) = Rn, (1/z) =

X (1)k (n k)!()nk  z n2k

,
k!(n

2k)!()
2
k
k=0

then the set {hn (z) is called the modified Lommel polynomials. Since the Bessel
functions J (z) satisfies a three-term recurrence relation, so the Lommel polynomials
inherit this property:
(2.25)

2z(n + )hn (z) = hn+1, (z) + hn1, (z)

16

CHAPTER 2. BESSEL FUNCTIONS

with initial conditions

(2.26)

h0, (z) = 1,

h1, (z) = 2z.

If one start with a different set of initial conditions

(2.27)

h0, (z) = 0,

h1, (z) = 2,

then the sequence {hn, (z)} generated by the (2.25) is called the associated Lommel polynomials. It is known that a three-term recurrence relation for polynomials
with the coefficients of as in (2.25) will generate a set of orthogonal polynomials
on
R +(, +). That is, there is a probability measure onb (, +) with
d = 1

Theorem 2.3.1 (A. A. Markov, 1895). Suppose the set of {pn (z)}of orthogonal
polynomials with its measure supported on a bounded internal [a, b], then
Z b
d(t)
pn (z)
=
(2.28)
lim
n pn (z)
a z t
holds uniformly for z 6 [a, b].
Since we know
Theorem 2.3.2 (Hurwitz). The limit
(2.29)

(z/2)+n Rn, +1 (z)

= J (z),
n
(n + + 1)
lim

holds uniformly on compact subsets of C.

So we have
Theorem 2.3.3. For > 0, the polynomials
R +{hn (z) are orthogonal with respect to
a discrete measure normalized to have d(t) = 1, and
Z
d(t)
J (1/z)
(2.30)
= 2
.
J1 (1/z)
R z t
One can use this formula to recover the measure (t). Thus the Lommel polynomials was found to have very distinct properties to be complete in L2 (1, +1).

2.4. INTEGRAL FORMAULAE

2.4

17

Integral formaulae

(2.1)
(z/2)
J (z) = 1
( 2 )( + 12 )

(1 t2 )1/2 cos zt dt,

<() > 1/2,

| arg z| < .

<() > 1/2,

| arg z| < ,

Or equivalently,
(2.2)
Z
(z/2)
cos(z cos ) sin2 d,
J (z) = 1
1
( 2 )( + 2 ) 0
where t = cos .
(1)

Writing H (z) = J (z) + iY (z). Then

Theorem 2.4.1 (Hankel 1869).
(2.3)

H(1) (z)

 2  12 ei[z/2/4] Z exp(i)

iu  12
u 12
=
e u
du,
1+
z
( + 1/2) 0
2z

where || < /2.

2.5

Asymptotic Behaviours of Bessel Functions

Expanding the integrand of Henkels contour integral by binomial expansion and

after some careful analysis, we have
Theorem 2.5.1. For < arg z < 2,
(2.1)

H(1) (x)


=

2
x

 21
e

i(x 12 41 )

X
p1


( 12 )m ( 12 + )m
(1)
+ Rp (x) ,
m m!
(2ix)
m=0

and for 2 < arg x < ,

(2.2)

H(2) (x)


=

2
x

 12
e

i(x 12 14 )

X
p1


( 12 )m ( 12 + )m
(2)
+ Rp (x) ,
(2ix)m m!
m=0

where
(2.3)

Rp(1) (x) = O(xp )

and

Rp(2) (x) = O(xp ),

as x +, uniformly in + < arg x < 2 and 2 + < arg x <

respectively.

18

CHAPTER 2. BESSEL FUNCTIONS

The two expansions are valid simultaneously in < arg x < .
(1)

(2)

We thus have J (z) = 12 (H (z) + H (z)). So

Theorem 2.5.2. For | arg z| < , we have

J (z)

2
z

 21 h

1
X (1)k (, 2k)
1
cos z
2
4 k=0
(2z)2k

1
X (1)k (, 2k + 1) i
1
sin z
2
4 k=0
(2z)2k+1

2.5.1

Addition formulae

Schlafli (1871) derived

(2.4)

J (z + t) =

Jk (t)Jk (z).

k=

Theorem 2.5.3 (Neumann (1867)). Let z, Z, R forms a triangle and let be the
angle opposite to the side R, then
(2.5)

p
X
2
2
J0
k Jk (Z)Jk (z) cos k,
(Z + z 2Zz cos ) =
k=0

where 0 = 1, k = 2 for k 1.
We note that Z, z can assume complex values. There are generalizations to
6= 0.

2.6

Fourier-Bessel Series

Theorem 2.6.1. Suppose the real function f (r) is piecewise continuous in (0, a)
and of bounded variation in every [r1 , r2 ] (0, a). Then if
Z a

(2.1)
r|f (r)| dr < ,
0

2.6. FOURIER-BESSEL SERIES

19

then the Fouier-Bessel series

(2.2)

f (r) =

ck J xk

k=1

r

a

converges to f (r) at every continuity point of f (r) and to

(2.3)

1
[f (r + 0) + f (r 0)]
2

at every discontinuity point of f (r). Here

Z 1
2
xf (x)J (k x) dx
(2.4)
ck = 2
J (k ) 0
is the Fourier-Bessel coefficients of f .

The xk are the zero of J (x) k = 0, 1, . The discussion of the orthogonality

of the Bessel functions is omitted. You may consult G. P. Tolstovs Fourier Series,
Dover 1962.
The situation here is very much like the classical Fourier series in terms of sine
and cosine functions.

20

CHAPTER 2. BESSEL FUNCTIONS

Chapter 3
An New Perspective
3.1

Hypergeometric Equations

We shall base our consideration of functions on the finite complex plane C in this
chapter. Let p(z) and q(z) be meromorphic functions defined in C. A point z0
belongs to C is called a regular singular point of the second order differential equation
d2 f (z)
df (z)
+ p(z)
+ q(z)f (z) = 0
2
dz
dz

(3.1)
if
(3.2)

lim (z z0 )2 q(z) = B

lim (z z0 )p(z) = A,

zz0

zz0

both exist and being finite. We start by recalling that any second order linear
equation with three regular singular points in the C can always be transformed ([2,
pp. 7375]) to the hypergeometric differential equation (due to Euler, Gauss
and Riemann)
(3.3)

z(1 z)

dy
d2 y
+ [c (a + b + 1)z] ab y = 0
2
dz
dz

which has regular singular points at 0, 1, .

It is known that the hypergeometric equation has a power series solution

(3.4)

2 F1

 
+
X
(a)k (b)k k
a, b 
z
=
F
[a,
b;
c;
z]
=
z
2 1

c
(c)k k!
k=0

21

22

CHAPTER 3. AN NEW PERSPECTIVE

which converges in |z| < 1. It is called the Gauss hypermetric series which
depends on three parameters a, b, c. Here the notation is the standard Pochhammer
notation (1891)
(a)k = a (a + 1) (z + 2) (a + k 1),

()0 = 1

for each integer k.

Theorem 3.1.1 (Euler 1769). Suppose <(c) > <(b) > 0, then we have

(3.5)

2 F1

 
Z 1
(c)
a, b 
z =
tb1 (1 t)cb1 (1 xt)a dt,
c 
(b)(c b) 0

defined in the cut-plane C\[1, +), arg t = arg(1 t) = 0, and that (1 xt)a
takes values in the principal branch (, ).
Thus Eulers integal representation of 2 F1 thus continues the function analytically into the cut-plane.
Special cases of (3.4) include: ...
In 1812 Gauss gave a complete set of 15 contiguous relations ...
We note that one can write familiar functions in terms of p Fq :
1.

2.

3.

4.

5.

 

e = 0 F0
x ;

  
a 
a
(1 x) = 1 F0
x ;




1, 1 
log(1 + z) = x 2 F1
x ;
2 



 x2
cos x = 0 F1
4 ;
1/2 
 

1/2, 1/2  2
1
sin x = x 2 F1
x .
3/2 
x

3.2. CONFLUENT HYPERGEOMETRIC EQUATIONS

3.2

23

Confluent Hypergeometric Equations

Writing the (3.1) in the variable x = z/b results in the equation

(3.1)

x(1 x/b)

d2 y
dy
+ [c (a + b + 1)x/b] a y = 0
2
dx
dx

which has regular singular points at 0, b, . Letting b tends to results in

(3.2)

d2 y
dy
+ (c x) a y = 0
2
dx
dx

which is called the confluent hypergeometric equation. But the equation has
a regular singular point at 0 and an irregular singular point at .
Substituting z/b in (3.4) and letting b to infinity results in a series
(3.3)

+
X
(a)k k
(a, c; z) := 1 F1 [a; c; z] =
z .
(c)k k!
k=0

This series is called a Kummer series. We note that the Gauss hypergeometric
series has radius of convergence 1, and after the above substitution of z/b results in
a series that has radius of convergence equal to |b|. So the Kummer series has radius
of convergence after taking the limit b . We call all solutions to the equation
(3.2) confluent hypergeometric functions because they are obtained from the
confluence of singularities of the hypergeometric equations. Further details can be
found in [4], better known as the Bateman project volumn one.

3.3

A Definition of Bessel Functions

We define the Bessel function of first kind of order to be the complex function
represented by the power series
(3.1)

+
+
X
X
(1)k ( 21 z)+2k
(1)k ( 21 )+2k 2k

=z
z .
J (z) =
(
+
k
+
1)
k!
(
+
k
+
1)
k!
k=0
k=0

Here is an arbitrary complex constant and the notation () is the Euler Gamma
function defined by
Z
(3.2)
(z) =
tz1 et dt, <(z) > 0.
0

24

CHAPTER 3. AN NEW PERSPECTIVE

It can be analytically continued onto the whole complex plane C, and has simple
poles at the negative integers including 0. So it is a meromorphic function. In
particular, (k + 1) = k! = k(k 1) 2 1. So the Euler-Gamma function is a
generalization of the factorial notation k!.
Let us set a = + 12 , c = 2 + 1 and replace z by 2iz in the Kummer series.
That is,

X
( + 21 )k
( + 1/2, 2 + 1; 2iz) =
(2iz)k
(2 + 1)k k!
k=0
 

+ 12 
2iz
= 1 F2
2 + 1 



z
2

iz
= e 0 F1

+ 1
2

X (1)k z 2k
= eiz
( + 1)k k! 22k
k=0

= eiz

X
(1)k ( + 1) z
2k
( + k + 1) k! 2
k=0

z
z
X
(1)k
z
2k

= e ( + 1)
2
2 k=0 ( + k + 1) k! 2
z

= eiz ( + 1)
J (z),
2
iz

where we have applied Kummers second transformation

 
  

a 
 2
2x
x
(3.3)
4x = e 0 F1
1 F1
2a 
a + 21 
in the third step above, and the identity ()k = ( + k)/() in the fourth step.
Since the confluent hypergeometric function ( + 1/2, 2 + 1; 2iz) satisfies
the hypergeometric equation
(3.4)

d2 y
dy
+ (c z) a y = 0
2
dz
dz

with a = + 1/2 and c = 2 + 1. Substituting eiz ( + 1)

and simplfying lead to the equation
(3.5)

z2

d2 y
dy
+ x + (z 2 2 )y = 0.
2
dz
dz

z
2

J (z) into the (3.4)

3.3. A DEFINITION OF BESSEL FUNCTIONS

25

The above derivation shows that the Bessel function of the first kind with order
is a special case of the confluent hypergeometric functions with specifically chosen
parameters.

26

CHAPTER 3. AN NEW PERSPECTIVE

Chapter 4
Physical Applications of Bessel
Functions
We consider radial vibration of circular membrane. We assume that an elastic
circular membrane (radius `) can vibrate and that the material has a uniform density.
Let u(x, y, t) denote the displacement of the membrane at time t from its equilibrium
position. We use polar coordinate in the xyplane by the change of variables:
(4.1)

x = r cos ,

y = r sin .

Then the corresponding equation

(4.2)

 2
2u
2u 
2 u
=
c
+
t2
x2 y 2

can be transformed to the form:

(4.3)

 2
2u
1 2u
1 2u 
2 u
=
c
+
+
.
t2
r2
r r
r2 2

Since the membrane has uniform density, so u = u(r, t) that is, it is independent of
the . Thus we have
 2
2u
1 2u 
2 u
(4.4)
.
=
c
+
t2
r2
r r
The boundary condition is u(`, t) = 0, and the initial conditions take the form
(4.5)

u(r, 0)
= g(r).
t

u(r, t) = f (r),
Separation of variables method yields

(4.6)

u(r, t) = R(r)T (t),

27

28

CHAPTER 4. PHYSICAL APPLICATIONS OF BESSEL FUNCTIONS

which satisfies the boundary condition u(`, t) = 0. Thus



1
RT 00 = c2 R00 T + R0 T .
r

(4.7)
Hence

T 00
R00 + (1/r)R0
+ 2 ; = 2 ,
R
cT

(4.8)

where the is a constant. Thus

1
R00 + R0 + 2 R = 0
r
T 00 + c2 2 T = 0.

(4.9)
(4.10)

We notice that the first equation is Bessel equation with = 0. Thus its general
solution is given by
(4.11)

R(r) = C1 J0 (r) + C2 Y0 (r).

. Since Y0 is unbounded when r = 0, so C2 = 0. Thus the boundary condition

implies that
(4.12)

J0 (`) = 0,

implying that = ` is a zero of J0 (). Setting C1 = 1, we obtain for each integer

n = 1, 2, 3, ,
(4.13)

Rn (r) = J0 (n r) = J0 (

n
r),
`

where n = n ` is the nth zero of J0 (). Thus we have

un (r, t) = (An cos cn t + Bn sin cn t) J0 (n r),

(4.14)

for n = 1, 2, 3, . Thus the general solution is given by

(4.15)

X
n=1

un (r, t) =

(An cos cn t + Bn sin cn t) J0 (n r),

n=1

and by
(4.16)

f (r) = u(r, 0) =

X
n=1

An J0 (n r),

29
and
(4.17)

X
u(r, 0) 
g(r) =
Bn cn J0 (n r).
 =
t
t=0
n=1

Fourier-Bessel Series theory implies that

Z `
2
(4.18)
An = 2 2
rf (r)J0 (n r) dr,
` J1 (n ) 0

(4.19)

2
Bn =
2
cn ` J12 (n )

rg(r)J0 (n r) dr.
0

30

CHAPTER 4. PHYSICAL APPLICATIONS OF BESSEL FUNCTIONS

Chapter 5
Separation of Variables of
Helmholtz Equations
The Helmholtz equation named after the German physicst Hermann von Helmholtz
refers to second order (elliptic) partial differential equations of the form:
(5.1)

(2 + k 2 ) = 0,

where k is a constant. If k = 0, then it reduces to the Laplace equations.

In this discussion, we shall restrict ourselves in the Euclidean space R3 . One
of the most powerful theories developed in solving linear PDEs is the the method
of separation of variables. For example, the wave equation

1 2 
(5.2)
2 2 2 (r, t) = 0,
c t
can be solved by assuming (r, t) = (r) T (t) where T (t) = eit . This yields

2 
(5.3)
2 2 (r) = 0,
c
which is a Helmholtz equation. The questions now is under what 3dimensional
coordinate system (u1 , u2 , u3 ) do we have a solution that is in the separation of
variables form
(5.4)

(r) = 1 (u1 ) 2 (u2 ) 3 (u3 ) ?

Eisenhart, L. P. (Separable Systems of Stackel. Ann. Math. 35, 284-305,

1934) determines via a certain St
ackel determinant is fulfilled (see e.g. Morse,
P. M. and Feshbach, H. Methods of Theoretical Physics, Part I. New York:
McGraw-Hill, pp. 125126, 271, and 509510, 1953).
31

32CHAPTER 5. SEPARATION OF VARIABLES OF HELMHOLTZ EQUATIONS

Theorem 5.0.1 (Eisenhart 1934). There are a total of eleven curvilinear coordinate systems in which the Helmholtz equation separates.
Each of the curvilinear coordinate is characterized by quadrics. That is,
surfaces defined by
(5.5)

Ax2 + By 2 + Cz 2 + Dxy + Exz + F yz + Gx + Hy + lz + J = 0.

One can visit http://en.wikipedia.org/wiki/Quadric for some of the quadric

surfaces. Curvilinear coordinate systems are formed by putting relevant orthogonal
quadric surfaces. Wikipedia contains quite a few of these pictures. We list the eleven
coordinate systems here:

33

Name of system
(1) Cartesian

Co-ordinate Systems
Transformation formulae
x = x, y = y, z = z

Degenerate Surfaces

(2) Cylindrical

x = cos , y = sin , z = z
0,
<

(3) Spherical polar

x = r sin cos ,
y = r sin sin ,
z = r cos
r 0, 0 ,

Parabolic
(4)
cylinder

x = u2 v 2 , y = 2uv,
z=z
u 0,
< v < +

Elliptic
(5)
cylinder

x = f cosh cos , y = f sinh sin ,

z=z
0, < < +

Rotation
(6)
paraboloidal

x = 2uv cos , 2uv sin ,

z = u2 v 2
u, v 0,
< <

Prolate
(7)
spheroidal

x = ` sinh u sin v cos ,

y = ` sinh u sin v sin ,
z = ` cosh u cos v,
u 0, 0 v , <

Finite line ;
segment
Two half-lines

Oblate
(8)
spheroidal

x = ` cosh u sin v cos ,

y = ` cosh u sin v sin ,
z = ` sinh u cos v,
u 0, 0 v , <

Circular plate (disc);

Plane with
circular aperture

(9) Paraboloidal

x = 12 `(cosh 2 + 2 cos 2 cosh 2,

y = 2` cosh cos sinh ,
z = 2` sinh sin cosh ,
, 0, <

Parabolic plate;
Plane with
parabolic aperture

(10) Elliptic conal

x = krsn sn ;
y = (ik/k 0 )rcn cn ,
z = (1/k 0 )rdn dn ;
r 0, 2K < 2K,
= K + iu, 0 u 2K 0

Plane sector;
Including quarter
plane

Half-plane

Infinite strip;
Plane with
straight aperture

Half-line

34CHAPTER 5. SEPARATION OF VARIABLES OF HELMHOLTZ EQUATIONS

Coordinate system

Laplace & Helmholtz

Laplace Equation

Helmholtz equation

(1) Cartesian

(Trivial)

(Trivial)

(2) Cylinderical

(Trivial)

Bessel

(3) Spherical polar

Associated Legender

Associated Legender

(4) Parabolic cylinder

(Trivial)

Weber

(5) Elliptic cylinder

(Trivial)

Mathieu

(6) Rotation-paraboloidal

Bessel

Confluent hypergeometric

(7) Prolate spheroidal

Associated Legender

Spheroidal wave

(8) Prolate spheroidal

Associated Legender

Spheroidal wave

(9) Paraboloidal

Mathieu

Whittaker-Hill

(10) Elliptic conal

Lame

Spherical Bessel, Lame

(11) Ellipsoidal

Lame

Ellipsoidal

35
1. Associated Legender:
(5.6)

d2 y
m2 o
dy n
(1 x ) 2 2x + n(n + 1)
y=0
dx
dx
(1 x2 )
2

2. Bessel:
x2

(5.7)

d2 y
dy
+ (x2 n2 )y = 0
+
x
2
dx
dx

3. Spherical Bessel:
x2

(5.8)

d2 y
dy
+ 2x + (x2 n(n + 1))y = 0
2
dx
dx

4. Weber:
d2 y
1
+ ( x2 )y = 0
2
dx
4

(5.9)

5. Confluent hypergeometric:
(5.10)

dy
d2 y
+ ( x) y = 0
2
dx
dx

6. Mathieu:
d2 y
+ ( 2q cos 2x)y = 0
dx2

(5.11)
7. Spheroidal wave:
(5.12)

o
d2 y
dy n
2
2
2
(1 x ) 2 2x +
+ (1 x ) y = 0
dx
dx
(1 x2 )
2

8. Lame:
(5.13)

d2 y
+ (h n(n + 1)k 2 sn 2 x)y = 0
2
dx

9. Whittaker-Hill:
(5.14)

d2 y
+ (a + b cos 2x + c cos 4x)y = 0
dx2

10. Ellipsoidal wave:

(5.15)

d2 y
+ (a + bk 2 sn 2 x + qk 4 sn 4 x)y = 0
dx2

Remark 5.0.1. The spheroidal wave and the Whittaker-Hill do not belong to the
hypergeometric equation regime, but to the Heun equation regime (which has four
regular singular points).

36CHAPTER 5. SEPARATION OF VARIABLES OF HELMHOLTZ EQUATIONS

5.1

What is not known?

It is generally regarded that the Bessel functions, Weber functions, Legendre functions are better understood, but the remaining equations/functions are not so well
understood.
1. Bessel functions. OK! Still some unknowns.
2. Confluent hypergeometric equations/functions. NOT OK.
3. Spheroidal wave, Mathieu, Lame, Whittaker-Hill, Ellipsoidal wave are poorly
understood. Some of them are relatd to the Heun equation which has four
regular singular points. Its research has barely started despite the fact that
it has been around since 1910.
4. Mathematicans are separating variables of Laplace/Helmholtz equations, but
in more complicated setting (such as in Riemannian spaces, etc)

Bibliography
[1] Milton Abramowitz and Irene A. Stegun, editors. Handbook of mathematical
functions, with formulas, graphs, and mathematical tables, volume 55 of National
Bureau of Standards Applied Mathematics Series. Superintendent of Documents,
US Government Printing Office, Washington, DC, 1965. Third printing, with
corrections.
[2] G. E. Andrews, R. A. Askey, and R. Roy. Special Functions. Cambridge University Press, Cambridge, 1999.
[3] A. Erdelyi, W. Magnus, F. Oberhettinger, and F. G. Tricomi. Higher Transcendental Functions, volume 2. McGraw-Hill, New York, 1953.
[4] A. Erdelyi, W. Magnus, F. Oberhettinger, and F. G. Tricomi. Higher Transcendental Functions, volume 1. McGraw-Hill, New York, 1953.
[5] T. J. Rivlin and M. W. Wilson. An optimal property of Chebyshev expansions.
J. Approximation Theory, 2:312317, 1969.
[6] G. N. Watson. A Treatise on the Theory of Bessel Functions. Cambridge University Press, Cambridge, second edition, 1944.

37