NATIONAL OPEN UNIVERSITY OF NIGERIA

SCHOOL OF SCIENCE AND TECHNOLOGY

COURSE CODE: PHY 312

COURSE TITLE: MATHEMATICAL METHODS OF

PHYSICS
PHY312 COURSE GUIDE

COURSE
GUIDE

PHY312
MATHEMATICAL METHODS OF PHYSICS I

Course Team Oluwatoyin K. Ogunbamike (Developer/Writer) -

Wesley University of Science & Technology, Ondo
Dr. S.O. Ajibola (Editor/Programme Leader) - NOUN

ii
PHY312 COURSE GUIDE

NATIONAL OPEN UNIVERSITY OF NIGERIA

National Open University of Nigeria

Headquarters
14/16 Ahmadu Bello Way
Victoria Island
Lagos

Abuja Office
5, Dar es Salaam Street
Off Aminu Kano Crescent
Wuse II, Abuja

e-mail: centralinfo@noun.edu.ng
URL: www.noun.edu.ng

Published By:
National Open University of Nigeria

First Printed 2012

ISBN: 978-058-431-5

All Rights Reserved

iii
PHY312 COURSE GUIDE

iv
PHY312 COURSE GUIDE

CONTENTS PAGE

Introduction…………………………...................…………….…..... 1
What You Will Learn in This Course…………………..................... 1
Course Aim………………………………………………………….. 1
Course Objectives……………………………………........................ 1
Working through This Course…………………………..................... 2
Course Materials…………………………………….......................... 2
Study Units………………………………………….......................... 2
Textbooks and References……………………………....................... 3
Assessment…………………………………………........................... 3
Tutor-Marked Assignment ……………..……………........................ 3
Final Examinations and Grading……………………......................... 3
Presentation Schedule…………………………………...................... 4
Course Marking Scheme………………………………...................... 4
How to Get the Most from This Course………………....................... 4
Facilitators/Tutors and Tutorials………………………...................... 5
Summary…………………………………….………......................... 5

v
Introduction

The course Mathematical Method of Physics 1- is meant to provide

essential methods for solving mathematical problems.

In scientific problems, often times we discover that a factor depends

upon several other related factors. For instance, the area of solid depends
on its length and breadth. Potential energy of a body depends on gravity,
density and height of the body etc. Moreover, the strength of a material
depends on temperature, density, isotopy and softness etc.

What You Will Learn in This Course

This is a 3unit course, it is grouped into four (4) modules i.e. module1, 2,
3 and 4. Module 1 has 2units; module 2 also has 2units. Module 3 has
only one unit while module 4 has 3units. In summary we have four (4)
modules and 8 units.

The course guide gives a brief summary of the total contents contained
in the course material. Functions of several variables streamline the
relationship between function and variables, the application of Jacobian,
down to functional dependence and independence. Also discussed are
multiple, line and improper integrals.

Course Aim

The overall aim of this course is to provide you with the essential
methods for solving mathematical problems in physics.

Course Objectives

At the end of this course, you should be able to:

• define linear second-order partial differential equation in more

than one independent variable
• use the technique of separation of variables in solving important
second order linear partial differential equations in physics
• solve the exercises at the end of this unit
• identify whether a given function is even, odd or periodic
• evaluate the Fourier coefficients
• derive and apply Fourier series in forced vibration problems
• use Fourier Integral for treating various problems involving
periodic function
• apply half range expansion to solutions of some problems.
PHY312 MATHEMATICAL METHODS OF PHYSICS I

Working through This Course

This course involves that you would be required to spend lot of time to
read. The content of this material is very dense and require you spending
great time to study it. This accounts for the great effort put into its
development in the attempt to make it very readable and comprehensible.
Nevertheless, the effort required of you is still tremendous.

I would advice that you avail yourself the opportunity of attending the
tutorial sessions where you would have the opportunity of comparing
knowledge with your peers.

Course Materials

You will be provided with the following materials:

• Course guide
• Study units

In addition, the course comes with a list of recommended textbooks,

which though are not compulsory for you to acquire or indeed read, are
necessary as supplements to the course material.

Study Units

The following are the study units contained in this course. The units are
arranged into four identifiable but related modules.

Module 1 Partial Differential Equations with Applications in

Physics

Unit 1 Partial Differential Equations

Unit 2 Fourier Series

Module 2 Application of Fourier to PDEs (Legendre polynomials

and Bessel Functions)

Unit 1 Legendre Polynomials

Unit 2 Bessel Functions

ii
PHY312 MATHEMATICAL METHODS OF PHYSICS I

Module 3 Application of Fourier to PDEs (Hermite Polynomials

and Laguerre Polynomials)

Unit 1 Hermite Polynomials

Unit 2 Laguerre Polynomials

Textbook and References

The following editions of these books are recommended for further

reading.

Hildraban, F. B.(nd). Advanced Calculus for Application.

Murray, R. S.(1974). Schaums Outline Series or Theory and Problem of

Advanced Calculus. Great Britain: McGraw–Hill Inc.

Stephenor, G. (1977). Mathematical Methods for Science Students.

London: Longman, Group Limited.

Stroud, K.A. (1995). Engineering Maths (5th ed.). Palgraw.

Verma, P.D.S. (1995). Engineering Mathematics. New Delhi: Vikas

Publishing House PVT Ltd.

Assessment

There are two components of assessment for this course. The Tutor-
Marked Assignment (TMA) and the end of course examination.

Tutor-Marked Assignment

The (TMA) is the continuous assessment component of your course. It

accounts for 30% of the total score. You will be given four (4) TMAs’ to
answer. Three of these must be answered before you are allowed to sit
for the end of course examination. The TMAs’ would be given to you by
your facilitator and returned after you have done the assignment.

Final Examinations and Grading

This examination concludes the assessment for the course. It constitutes

70% of the whole course. You will be informed of the time of the
examination. It may or may not coincide with the University Semester
Examination.

iii
PHY312 MATHEMATICAL METHODS OF PHYSICS I

Presentation Schedule

Your course materials have important dates for the early and timely
completion and submission of your TMAs and attending tutorials. You
should remember that you are required to submit all your assignments
by the stipulated time and date. You should guard against falling behind
in your work.

Course Marking Scheme

Assignment Marks
Assignments 1-4 Four TMAs, best three marks of
the four count at 10% each – 30%
of course marks.
End of course examination 70% of overall course marks.
Total 100% of course materials.

At the end of each unit, assignments are given to assist you to assess
your understanding of the topics that have been discussed.
Course Overview

Each study unit consists of three hours work. Each study unit includes
introduction, specific objectives, directions for study, reading materials,
conclusions, and summary, Tutor -Marked Assignments (TMAs),
references / further reading. The units direct you to work on exercises
related to the required readings. In general, these exercises test you on
the materials you have just covered or require you to apply it in some
way and thereby assist you to evaluate your progress and to reinforce
your comprehension of the material. Together with TMAs, these
exercises will help you in achieving the stated learning objectives of the
individual units and of the course as a whole.

How to Get the Most Out of This Course

Implicit interest and regular culture of reading are of utmost

requirements for getting the best out of this course. It is paramount that
you should at least purchase one of the textbooks that are recommended
for you. More importantly, attending tutorials sessions and completing
your assignments on time will certainly assist you to get the best out of
this course.

iv
PHY312 MATHEMATICAL METHODS OF PHYSICS I

Facilitators/Tutors and Tutorials

There are 16 hours of tutorials provided in support of this course. You

will be notified of the dates, times and locations of these tutorials as well
as the name and phone number of your facilitator, as soon as you are
allocated a tutorial group.

Your facilitator will mark and comment on your assignments, keep a

close watch on your progress and any difficulties you might face and
provide assistance to you during the course. You are expected to mail
your Tutor -Marked Assignment to your facilitator before the scheduled
date (at least two working days are required). They will be marked by
your tutor and returned to you as soon as possible.

Do not delay to contact your facilitator by telephone or e-mail if you

need assistance.

The following might be circumstances in which you would find

assistance necessary. You would have to contact your facilitator if:

• you do not understand any part of the study or the assigned

readings
• you have difficulty with the self-tests
• you have a question or problem with assignments or with the
grading of assignments.

You should endeavour to attend the tutorials. This is the only chance to
have face to face contact with your course facilitator and to ask
questions which are answered instantly. You can raise any problem
encountered in the course of your study.

To gain much benefit from course tutorials prepare a question list before
attending them. You will learn a lot from participating actively in
discussions.

Summary

It is expected that, going through this course, you have learnt how to use
Method of Separation of Variables to Solve Heat Conduction Equation
and Wave Equation respectively.

The use of Fourier transforms to solve some differential Equation,

Boundary values problems etc. You should also have learnt the use of
Laplace transformation to solve some initial and Boundary value
problems, which are difficult to handle in addition to the application of
convolution theory in solving problems.

v
PHY312 MATHEMATICAL METHODS OF PHYSICS I

Course Code PHY312

Course Title Mathematical Methods of Physics I

Course Team Oluwatoyin K. Ogunbamike (Developer/Writer) -

Wesley University of Science & Technology, Ondo
Dr. S.O. Ajibola (Editor/Programme Leader) - NOUN

vi
PHY312 MATHEMATICAL METHODS OF PHYSICS I

NATIONAL OPEN UNIVERSITY OF NIGERIA

National Open University of Nigeria

Headquarters
14/16 Ahmadu Bello Way
Victoria Island
Lagos

Abuja Office
5, Dar es Salaam Street
Off Aminu Kano Crescent
Wuse II, Abuja

e-mail: centralinfo@noun.edu.ng
URL: www.noun.edu.ng

Published By:
National Open University of Nigeria

First Printed 2012

ISBN: 978-058-431-5

All Rights Reserved

vii
PHY312 MATHEMATICAL METHODS OF PHYSICS I

viii
PHY312 MATHEMATICAL METHODS OF PHYSICS I

CONTENTS PAGE

Module 1 Partial Differential Equations with Applications in

Physics…………………………….…………….……. 1

Unit 1 Partial Differential Equations……………………….… 1

Unit 2 Fourier Series…………………………………………..18

Module 2 Application of Fourier to PDEs (Legendre

Polynomials and Bessel Functions)……………….... 33

Unit 1 Legendre Polynomials…………….………………… 33

Unit 2 Bessel functions……………………………..………. 46

Module 3 Application of Fourier to PDEs (Hermite

Polynomials and Laguerre Polynomials)………..… 64

Unit 1 Hermite Polynomials…………………..…………….. 64

Unit 2 Laguerre Polynomials…………………………………. 80

ix
PHY312 MATHEMATICAL METHODS OF PHYSICS I

MODULE 1 PARTIAL DIFFERENTIAL EQUATIONS

WITH APPLICATIONS IN PHYSICS

Unit 1 Partial Differential Equations

Unit 2 Fourier Series

UNIT 1 PARTIAL DIFFERENTIAL EQUATIONS

CONTENTS

1.0 Introduction
2.0 Objectives
3.0 Main Content
3.1 Definition
3.2 Linear Second-Order Partial Differential Equations
3.2.1 Laplace‘s Equation
3.2.2 Wave Equation
3.2.3 Heat Conduction Equation
3.2.4 Poisson’s Equation
3.3 Method of Separation of Variables
3.3.1 Application to Wave Equation
3.3.2 Application to Heat Conduction Equation
3.4 Laplace Transform Solutions of Boundary-Value
Problems
4.0 Conclusion
5.0 Summary
6.0 Tutor -Marked Assignment
7.0 References/Further Reading

1.0 INTRODUCTION

In this unit, we shall study some elementary methods of solving partial

differential equations which occur frequently in physics and in
engineering. In general, the solution of the partial differential equation
presents a much more difficult problem than the solution of ordinary
differential equations.

We are therefore going to limit ourselves to a few solvable partial

differential equations that are of physical interest.

1
PHY312 MATHEMATICAL METHODS OF PHYSICS I

2.0 OBJECTIVES

At the end of this unit, you should be able to:

• define linear second-order partial differential equation in more

than one independent variable
• use the technique of separation of variables in solving important
second order linear partial differential equations in physics
• solve the exercises at the end of this unit.

3.0 MAIN CONTENT

3.1 Definition

An equation involving one or more partial derivatives of (unknown)

functions of two or more independent variables is called a partial
differential equation. The order of a PDE is the highest order partial
derivative or derivatives which appear in the equation. For example,

∂U ∂ 3U  ∂ 2U ∂ 2U 
U +  2  = ez
2 
(1)
∂z ∂y 3
 ∂y ∂z 
is a third order PDE since the highest order term is given by

∂ 3U
∂y 3

A PDE is said to be linear if it is of the first degree, i.e. not having

exponent greater than 1 in the dependent variable or its partial
derivatives and does not contain product of such terms in the equation.
Partial derivatives with respect to an independent variable are written for
brevity as a subscript; thus

∂ 2U ∂ 2U
U tt = and U =
∂t 2 ∂x∂y
xy

The PDE

1
U tt = U xx + U yy + U zz (2)
c2

(Where c is a constant) is linear and is of the second order while eq. (1)
is an example of a nonlinear PDE.

2
PHY312 MATHEMATICAL METHODS OF PHYSICS I

Example 1: Important linear partial differential equations of

second order

(1)
∂ 2u 2 ∂ u
2
= c One − dim ensional wave equation
∂t 2 ∂x 2
(2)
∂u ∂ 2u
= c2 2 One − dim ensional heat equation
∂t ∂x
(3)
∂ 2u ∂ 2u
+ =0 Two − dim ensional Laplace equation
∂x 2 ∂y 2
(4)
∂ 2u ∂ 2u
+ = f ( x, y ) Two − dim ensional poisson equation
∂x 2 ∂y 2
(5)
∂ 2u ∂ 2u ∂ 2u
+ + =0 Three − dim ensional Laplace equation
∂x 2 ∂y 2 ∂z 2

3.2 Linear Second-Order Partial Differential Equations

Many important PDEs occurring in science and engineering are second

order linear PDEs. A general form of a second order linear PDE in two
independent variables x and y can be expressed as

∂ 2u ∂ 2u ∂ 2u ∂u ∂u
A + B + C +D +E + Fu = G (3)
∂x 2
∂x∂y ∂y 2
∂x ∂y

where A, B, C……,G may be dependent on variables x and y. If G = 0 ,

then eq. (3) is called homogeneous; otherwise it is said to be a non-
homogeneous.

The homogeneous form of Eq. (3) resembles the equation of a general

conic:

ax 2 + bxy + cy 2 + dx + ey + f = 0

We thus say that eq. (3) is of

elliptic   B 2 − 4 AC < 0
 
hyperbolic type when  B 2 − 4 AC > 0
parabolic   B 2 − 4 AC = 0


3
PHY312 MATHEMATICAL METHODS OF PHYSICS I

For example, according to this classification the two-dimensional

Laplace equation

∂ 2u ∂ 2u
+ = 0
∂x 2 ∂y 2

is of elliptic type ( A = C = 1, B = D = E = G = 0), and the equation

∂ 2u ∂ 2u
−α 2
= 0 ( α is a real constant)
∂x 2 ∂y 2

is of hyperbolic type. Similarly, the equation

∂ 2u ∂ u
− α = 0 ( α is a real constant)
∂ x 2 ∂ y

is of parabola type.

Some important linear second-order partial differential equations that

are of physical interest are listed below.

Example 2

Eliminate A and P from the function Z = Ae pt sin px

∂Z ∂2Z
Solution Let = pAe pt sin px and = p 2 Ae pt sin px
∂t ∂t 2

∂Z ∂2Z
also = pAe pt cos px and = − p 2 Ae pt sin px
∂x ∂x 2
∂2Z ∂2Z
+ =0
∂t 2 ∂x 2
i.e. p 2 Ae pt sin px − p 2 Ae pt sin px = 0

Example 3

Solve the equation

∂ 2u ∂ 2u ∂ 2u
− 7 + 6 =0
∂x 2 ∂x∂y ∂y 2
Solution: Let u ( x, y ) = f ( y + m1 x) + g ( y + m2 x)

So that m 2 − 7m + 6 = 0

This implies that m = 1 or 6

Hence u ( x, y ) = H ( y + x ) + G ( y + 6 x )

4
PHY312 MATHEMATICAL METHODS OF PHYSICS I

3.2.1 Laplace‘s Equation

∇ 2u = 0 (4)
 ∂ 2
∂ 2
∂ 
2
Where ∇ 2 is the Laplacian operator  ∇ 2 = 2 + 2 + 2  . The
 ∂x ∂y ∂y 
function u may be the electrostatic potential in a charge-free region or
gravitational potential in a region containing no matter.

3.2.2 Wave Equation

1 ∂ 2u
∇ 2
u = (5)
v 2 ∂ t 2
Where u represents the displacement associated with the wave and v, the
velocity of the wave.

3.2.3 Heat Conduction Equation

∂u
= α∇ 2 u (6)
∂t
Where u is the temperature in a solid at time t. The constant α is called
the diffusivity and is related to the thermal conductivity, the specific
heat capacity, and the mass density of the object.

3.2.4 Poisson’s Equation

∇ 2 u = ρ ( x, y , z ) (7)

Where the function ρ ( x, y, z ) is called the source density. For example, if

u represents the electrostatic potential in a region containing charges,
then ρ is proportional to the electric charge density.

Example 4

Laplace’s equation arises in almost all branches of analysis. A simple

example can be found from the motion of an incompressible fluid. Its
velocity v ( x, y, z , t ) and the fluid density ρ ( x, y, z , t ) must satisfy the
equation of continuity:

∂ρ
+ ∇.( ρv ) = 0
∂t
If ρ is constant we then have

∇•v = 0

5
PHY312 MATHEMATICAL METHODS OF PHYSICS I

If furthermore, the motion is irrotational, the velocity vector can be

expressed as the gradient of a scalar function V:

v = −∇V
and the continuity becomes Laplace’s equation:

∇.v = ∇.(−∇V ) = 0, or ∇ 2V = 0
The scalar function V is called the velocity potential

3.3 Method of Separation of Variables

The technique of separation of variables is widely used for solving many

of the important second order linear PDEs.
The basic approach of this method in attempting to solve a differential
equation (say, two independent variables x and y) is to write the
dependent variable u ( x, y ) as a product of functions of the separate
variables u ( x, t ) = X ( x) T (t ) . In many cases the partial differential
equation reduces to ordinary equations for X and T.

3.3.1 Application to Wave Equation

Let us consider the vibration of an elastic string governed by the one-

dimensional wave equation
∂ 2u 2 ∂ u
2
= c (8)
∂t 2 ∂x 2
where u(x, y) is the deflection of the string. Since the string is fixed at
the ends x = 0 and x = l , we have the two boundary conditions

u (0, t ) = 0, u (l , t ) = 0 for all t (9)

The form of the motion of the string will depend on the initial deflection
(deflection at t = 0) and on the initial velocity (velocity at t = 0).
Denoting the initial deflection by f(x) and the initial velocity by g(x), the
two initial conditions are

∂u
u ( x, 0) = f ( x) = g ( x) (10)
∂t t =0

This method expresses the solution of u ( x, t ) as the product of two

functions with their variables separated, i.e.

U ( x, t ) = X ( x) T (t ) (11)

where X and T are functions of x and t respectively.

6
PHY312 MATHEMATICAL METHODS OF PHYSICS I

Substituting eq. (11) in eq. (8), we obtain

X T ′′ = c 2 X ′′ T
or
X ′′( x) 1 T ′′(t )
= (12)
X ( x) c 2 T (t )

In other words

X ′′ 1 T ′′
= 2 =λ (13)
X c T

The original PDE is then separated into two ODEs, viz.

X ′′( x) − λX ( x ) = 0 (14)
and
T ′′(t ) − λc 2 T (t ) = 0 (15)

The boundary conditions given by eq. (9) imply

X (0) T (t ) = 0
and
X (l ) T (t ) = 0

Since T(t) is not identically zero, the following conditions are satisfied

X ( 0) = 0 and X (l ) = 0 (16)

Thus eq. (14) is to be solved subject to conditions given by eq. (16).

There are 3 cases to be considered.

Case 1 λ >0

The solution to eq. (14) yields

λx λx
X ( x ) = Ae − + Be (17)

To satisfy the boundary condition given by eq. (16), we must have

λl λl
Ae − + Be =0

Since the determinant formed by the coefficients of A and B is non-zero,

the only solution is A = B = 0. This yields the trivial solution X(x) = 0.

7
PHY312 MATHEMATICAL METHODS OF PHYSICS I

Case 2 λ =0

The solution to eq. (14) yields

X ( x) = A + Bx

To satisfy the boundary condition given by eq. (16), we must have

A=0
and
A + Bl = 0
implying
A = 0, B=0

Again for this case, a trivial solution is obtained

Case 3 λ <0

Let λ = − k 2 . The solution to eq. (14) yields

X ( x ) = A cos kx + B sin kx (18)

To satisfy the boundary condition given by eq. (16), we must have

A=0
and
B sin kl = 0
To obtain a solution where B ≠ 0 , we must have
kl = nπ n = 1, 2,...
Thus
 nπ 
2

λ =− k2 =−  (19)
 l 
(n=0 corresponds to the trivial solution). The specific values of λ are
known as the eigenvalues of eq. (14) and the corresponding solutions,
 nπ 
viz, sin  x  are called the eigenfunctions. Since there are many
 l 
possible solutions, each is subscripted by n. Thus

 nπ 
X n ( x) = Bn sin  x n =1, 2, 3, ... (20)
 l 

The solution to Eq. (15) with λ given by Eq. (19) is

 nπ   nπ 
Tn (t ) = E n cos ct  + Fn sin  ct  n =1, 2, 3, ... (21)
 l   l 
8
PHY312 MATHEMATICAL METHODS OF PHYSICS I

Where En and Fn are arbitrary constants. There are thus many solutions
for eq. (8) which is given by

U n ( x, t ) = X n ( x ) Tn (t )
  nπ   nπ  nπ
a n cos l ct  + bn sin  l ct  sin l x (22)
    

Where an = Bn En and bn = Bn Fn . Since eq. (8) is linear and homogeneous,

the general solution is obtained as the linear superposition of all the
solutions given by eq. (22), i.e.

∞
 nπc nπc  nπ
U ( x, t ) = ∑  a n cos t + bn sin t  sin x (23)
n =1  l l  l

Differentiating with respect to t, we have

∞
nπc  nπc nπc  nπ
U t ( x, t ) = ∑  − a n sin t + bn cos t  sin x (24)
n =1 l  l l  l

The coefficients an and bn are obtained by applying the initial

conditions in eq. (10). Thus,

∞
nπ
U ( x,0) = f ( x) = ∑ a n sin x (25)
n =1 l
∞
 nπ  nπ
U ( x,0) = g ( x) = ∑ bn  c  sin x (26)
n =1  l  l

In order to determine a n and bn we use the orthogonality properties of

nπ
sin x in the range 0 ≤ x ≤ l , i.e.
l
mπ nπ
l
l
∫ sin x sin x dx = δ m n (27)
0
l l 2

Where δ m n is the Kronecker delta function having the property

0 n≠m
δ m n =  (28)
1 n =m

mπ
Multiply eq. ((25) by sin x and integrating between the limits x = 0
l
and x = l, we get

9
PHY312 MATHEMATICAL METHODS OF PHYSICS I

mπ ∞
mπ nπ
l l

∫ f ( x) sin x dx = ∑ ∫ a n sin x sin x dx

0
l n =1 0 l l
l
= am (29)
2
mπ
l
2
i.e. am = ∫
l 0
f ( x) sin
l
x dx

mπ
Similarly multiplying eq. (26) by sin x and integrating between the
l
limits x = 0 and x = l , we get
mπ ∞
 nπ  nπ mπ
l l

∫0 g ( x ) sin
l
x dx = ∑ ∫
n =1 0
bn  c  sin
 l  l
x sin
l
x dx

 mπ  l
= bm  c (30)
 l 2
mπ
l
2
i.e. bm =
mπc 0 ∫ g ( x) sin
l
x dx

With a m and bm obtained for m =1, ... ∞, eq. (23) is the solution to PDE
given by eq. (8) subject to the initial conditions and the boundary
conditions.

3.3.2 Application to Heat Conduction Equation

The one-dimensional heat flow in a rod bounded by the planes x = 0 and

x = a is of practical interest. The solution applies to the case where the y
and z dimensions extend to infinity. The temperature distribution is
determined by solving the one-dimensional heat conduction equation

∂ 2θ 1 ∂θ
= (31)
∂x 2 v ∂t

Where θ represents the temperature and

k
v= (32)
Cρ

k, C and ρ are the thermal conductivity, specific heat and density of the
material respectively. We shall treat the case where the boundary
conditions are given by

θ ( x = 0, t ) = 0 (33)

10
PHY312 MATHEMATICAL METHODS OF PHYSICS I

θ ( x = a, t ) = 0 (34)

The initial temperature distribution is given by

θ ( x , t = 0) = f ( x ) (35)

Solution: Using the method of separation of variables, the x-dependence

and t-dependence are separated out as expressed by

θ ( x, t ) = X ( x ) T (t ) (36)

Substituting eq. (36) into eq. (31) yields

1 d 2 X 1 1 dT
= =α (37)
X dx 2 v T dt

We shall now consider three cases corresponding to different values of

the constant α .
Case 1 λ =0

The separated ODE for X ( x) becomes

d2X
=0 (38)
dx 2
i.e. X ( x) = Ax + B

The boundary conditions expressed by eqs. (33) and (34) are

respectively

X ( x = 0) = 0 and X ( x = a) = 0 (39)

Since T(t) should not be identically zero. Thus for eq. (38) to satisfy the
boundary conditions given by eq. (39), we must have A = 0, B = 0. This
gives the steady-state solution where temperature in the rod is
everywhere zero.

Case 2 λ >0

Let α = k 2 . The ODE for X becomes

d2X
2
= k2X (40)
dx
Therefore X ( x ) = Ae kx + Be − kx

Applying the boundary conditions given in eq. (39), we get

11
PHY312 MATHEMATICAL METHODS OF PHYSICS I

0=A+B
0 = Ae ka + Be − ka

Again we have A = B = 0

Case 3 λ <0

Let α = −λ2 . The ODE for X ( x) becomes

d2X
= − λ2 X (41)
dx 2

Thus X ( x) = A cos λx + B sin λx

The boundary conditions require

A=0
B sin λa = 0 (42)
i.e. λ a = nπ n =1, 2, ... (43)

Since there are multiple solutions, each λ is designated by a subscript n

as λn . The solution of the ODE for T(t) is readily obtained as

T (t ) = Ce − λnvt
2
(44)

Thus the general solution which is a superposition of all admissible

solution is given by

∞
nπ
θ ( x, t ) = ∑ Dn e −λ vt sin
2
n
x (45)
n =1 a
∞
 n 2 π 2  nπ
= ∑ Dn exp − 2 vt  sin x (46)
n =1  a  a

To complete the solution Dn must be determined from the remaining

initial condition

∞
nπ
i.e. f ( x) = ∑ Dn sin x (47)
n =1 a

12
PHY312 MATHEMATICAL METHODS OF PHYSICS I

mπ
In order to determine Dn , we multiply eq. (47) by sin x and integrate
a
the limits x=0 and x = a to obtain

mπ ∞
nπ mπ
a a
a
∫ f ( x) sin x dx = ∑ Dn ∫ a n sin x sin x dx = Dm
0
a n =1 0
a a 2

Thus

mπ
a
2
Dm = ∫ f ( x) sin x dx (48)
a0 a

For the specific case where f ( x) = θ 0 (constant), the solution is given by

4θ 0 ∞
1 (2n + 1)πx  v(2n + 1) 2 π 2 
θ ( x, t ) = ∑
π n =0 2 n + 1
sin
a
exp −
a2

 

0 < x < a (49)

From eq. (49), it can be deduced that a rectangular pulse of height θ 0 for
0 < x < a has the Fourier series expansion given by

4θ 0 ∞
1 ( 2n + 1)πx
π
∑ 2n + 1 sin
n =0 a

Also if f ( x) = γx, then

2 aγ (−1) n−1  nπ 
∞
 vn 2π 2 
θ ( x, t ) = ∑
π n=1 n
sin  x  exp −
 a  a2 
t (50)


If the end boundaries are maintain at different temperature i.e.

θ ( x = 0, t ) = θ1
θ ( x = a, t ) = θ 2 (51)

Then case 1 of the solution where α = 0, would yield the steady-state

solution given by θ1 +
x
(θ 2 − θ1 ) . The general solution is given by
a
θ ( x, t ) = φ ( x, t ) + θ1 + (θ 2 − θ1 )
x
(52)
a

Where φ ( x, t ) is the transient solution.

13
PHY312 MATHEMATICAL METHODS OF PHYSICS I

The boundary conditions for φ ( x, t ) are obtained as follows:

at x = 0: θ ( x = 0, t ) = θ1 = φ ( x = 0, t ) + θ1 ⇒ φ ( x = 0, t ) = 0
at x = a: θ ( x = a, t ) = θ 2 = φ ( x = a, t ) + θ 2 ⇒ φ ( x = a, t ) = 0
φ ( x, t ) is obtained under case 3.

SELF-ASSESSMENT EXERCISE 1

1. State the nature of each of the following equations (that is,

whether elliptic, parabolic or hyperbolic)

∂2 y ∂2 y
(a) + α =0
∂t 2 ∂x 2
∂ 2u ∂ 2u ∂u
(b) x 2 + y 2 + 3y 2
∂x ∂y ∂x

2(a) Show that y ( x, t ) = F (2 x + 5t ) + G (2 x − 5t ) is a general solution of

∂2 y ∂2 y
4 = 25
∂t 2 ∂x 2
(b) Find a particular solution satisfying the conditions
y (0, t ) = y (π , t ) = 0, y ( x,0) = sin 2 x, y ′( x,0) = 0.

3. Solve the following PDEs

∂ 2u
(a) = 8 xy 2 + 1
∂x 2

∂ 2 u ∂u
(b) − = 6 xe x
∂xy ∂y

3.4 Laplace Transform Solutions of Boundary-Value

Problems

Laplace and Fourier transforms are useful in solving a variety of partial

differential equations; the choice of the appropriate transforms depends
on the type of boundary conditions imposed on the problem. Laplace
transforms can be used in solving boundary-value problems of partial
differential equation.

14
PHY312 MATHEMATICAL METHODS OF PHYSICS I

Example 5

Solve the problem

∂u ∂ 2u
=2 2 (53)
∂t ∂x
u (0, t ) = u (3, t ) = 0, u ( x,0) = 10 sin 2πx − 6 sin 4πx (54)

Solution: Taking the Laplace transform L of Eq. (53) with respect to t

gives
 ∂u   ∂ 2u 
L   = 2L  2 
 ∂t   ∂x 
Now
 ∂u 
L   = pL(u ) − u ( x,0)
 ∂t 
and
 ∂ 2u  ∂ 2 ∞ ∂2
L  2  = 2 ∫ e − pt u ( x, t )dt = 2 L[u ]
 ∂x  ∂x 0 ∂x
∞
Here ∂ 2 ∂x 2 and ∫0
...dt are interchangeable because x and t are
independent.

For convenience, let

∞
U = U ( x, p ) = L[u ( x, t )] = ∫ e − pt u ( x, t )dt
0

We then have
∂ 2U
pU − u ( x,0) = 2 L
∂x 2
from which we obtain, using the given conditions (54),
∂ 2U 1
− pU = 3 sin 4πx − 5 sin 2πx. (55)
∂x 2 2
Then taking the Laplace transform of the given conditions
u (0, t ) = u (3, t ) = 0, we have

L[u (0, t )] = 0, L[u (3, t )] = 0

Or
U (0, p ) = 0, U (3, p ) = 0.

These are the boundary conditions on U ( x, p ) . Solving eq. (55) subject to

these conditions we find

5 sin 2πx 3 sin 4πx

U ( x, p ) = −
p + 16π 2 p + 64π 2

15
PHY312 MATHEMATICAL METHODS OF PHYSICS I

The solution to eq. (55) can now be obtained by taking the inverse
Laplace transform
u ( x, t ) = L−1 [U ( x, p)] = 5e16π t sin 2πx − 36e 64π t sin 4πx.
2 2

SELF-ASSESSMENT EXERCISE 2

1. Differentiate between ordinary differential equation and partial

differential equation.

2. Derive the PDE that give rise to the function

Z = a( x + y ) + b( x − y ) + abt + c = 0
3. Use the method of separation of variable to find the solution of
the boundary value problem
∂2 y ∂2 y
=
∂x 2 ∂t 2
y (0, t ) = 0 t>0
y (1, t ) = 0 t>0
y ( x,0) = sin 2 x
y ′( x,0) = 0 0≤ x<∞

4.0 CONCLUSION

In this unit, we have studied the notion of a solution of partial

differential equation. Also some elementary methods of solving linear
partial differential equations which occur frequently in physics and
engineering were dealt with.

5.0 SUMMARY
Here in this unit you have learnt about second order partial differential
equation. The classical method of separation of variables was
extensively studied along with the Laplace transform solutions of
boundary-value problems.

6.0 TUTOR- MARKED ASSIGNMENT

1. Form the PDEs whose general solutions are as follow:

z = Ae − p t cos px
2
(a)
 y
(b) z = f  
x
2. Solve the equation
∂ 2u ∂ 2u
2 =3 2 =0
∂x∂y ∂y

16
PHY312 MATHEMATICAL METHODS OF PHYSICS I

3. Find the solution of the differential equation

∂2 y ∂2 y
α2 =
∂x 2 ∂t 2
y (0, t ) = 0 0<t <∞
Where
y (0, t ) = 0 0≤t <∞
y ( x,0) = f ( x) 0≤ x≤L
y ( x,0) = g ( x) 0≤ x<L

4. Solve by Laplace transforms the boundary-value problem

∂ 2 u 1 ∂u
= for x > 0, t > 0
∂x 2 k ∂t
given that u = u0 (a constant) on x = 0 for t > 0 , and u = 0 for
x > 0, t = 0

7.0 REFERENCES/ FURTHER READING

Erwin, K. (1991). Advanced Engineering Mathematics. John Wiley &

Sons, Inc.

Pinsky, M.A. (1991).Partial Differential Equations and Boundary-Value

Problems with Applications. New York: McGraw-Hill.

17
PHY312 MATHEMATICAL METHODS OF PHYSICS I

UNIT 2 FOURIER SERIES

CONTENTS

1.0 Introduction
2.0 Objectives
3.0 Main Content
3.1 Periodic Functions
3.2 Even and Odd Functions
3.3 Fourier Theorem
3.4 Evaluation of Fourier Coefficients
3.5 Application of Fourier Series in Forced Vibrations
3.6 Half-Range Expansions
3.7 Fourier Integral
3.8 Fourier Integrals of Even and Odd Functions
4.0 Conclusion
5.0 Summary
6.0 Tutor-Marked Assignment
7.0 References/Further Reading

1.0 INTRODUCTION

In this unit, we shall discuss basic concepts, facts and techniques in

connection with Fourier series. Illustrative examples and some important
applications of Fourier series to Partial differential equations will be
studied.
We will also study the concept of periodic functions, even and odd
functions and the conditions for Fourier expansion.

2.0 OBJECTIVES

At the end of this unit, you should able to:

• identify whether a given function is even, odd or periodic

• evaluate the Fourier coefficients
• derive and apply Fourier series in forced vibration problems
• use Fourier Integral for treating various problems involving
periodic function
• apply half range expansion to solutions of some problems.

18
PHY312 MATHEMATICAL METHODS OF PHYSICS I

3.0 MAIN CONTENT

3.1 Periodic Functions

A function f ( x) is said to be periodic if it defined for all real x and if

there is some positive number T such that

f ( x + T ) = f ( x) (1)

This number T is then called a period of f ( x) .

Periodic functions occur very frequently in many application of

mathematics to various branches of science. Many phenomena in nature
such as propagation of water waves, light waves, electromagnetic waves,
etc are periodic and we need periodic functions to describe them.
Familiar examples of periodic functions are the sine and cosine
functions.

Example 1

Find the period of Tan x.

Solution: Suppose T is its period

f ( x + T ) = Tan ( x + T ) = Tanx
so that

Tan ( x + T ) − Tanx = 0
using trigonometric identity, we have

Tan T (1 − Tan 2 x)
=0
1 − TanxTanT

This implies that

Tan T = 0 If and only if 1 − Tan 2 x ≠ 0
T = Tan −1 0
Hence T =π

3.2 Even and Odd Functions

A function f ( x) defined on interval [a, b] is said to be a even function if

f (− x ) = f ( x ) (2)

19
PHY312 MATHEMATICAL METHODS OF PHYSICS I

It is odd otherwise, that is

f (− x) = − f ( x) (3)

Example 2

Let f ( x) = sin x
Then f (− x) = − f ( x) i.e. sin( − x) − = − sin x

Thus it is obvious that sine function is always an odd function while

cosine function is an even function.

3.3 Fourier Theorem

According to the Fourier theorem, any finite, single valued periodic

function f (x) which is either continuous or possess only a finite number
of discontinuities (of slope or magnitude), can be represented as the sum
of the harmonic terms as
1
f ( x) = a 0 + a1 cos x + a 2 cos 2 x + . . . + a n cos nx
2
+ b1 sin x + b2 sin 2 x + . . . + bn sin nx

∞
a 0 + ∑ (a n cos nx + bn sin nx )
1
= (4)
2 n =1

3.4 Evaluation of Fourier Coefficients

Let us assume that f ( x) is a periodic function of period 2π which can be

represented by a trigonometric series
∞
f ( x ) = a0 + ∑ (a n cos nx + bn sin nx) (5)
n =1

Given such a function f ( x) we want to determine the coefficients of

a n and bn in the corresponding series in eq. (5).

We first determine a0 . Integrating on both sides of eq. (4) from − π to

π , we have
π  ∞
π 
∫π f ( x)dx = ∫ a 0 + ∑ ( a n cos nx + bn sin nx)  dx
− −π
 n =1 

If term-by-term integration of the series is allowed, then we obtain

π π ∞ π π
∫π f ( x)dx = a0 ∫ dx + ∑  an ∫ cosnxdx + bn ∫ sinnxdx 
n =1  
− −π − π −π

20
PHY312 MATHEMATICAL METHODS OF PHYSICS I

The first term on the right equals 2πa0 . All other integrals on the right
are zero, as can be readily seen by performing the integration. Hence our
first result is

1 π
a0 =
2π ∫ π f ( x)dx
−
(6)

We now determine a1 , a2 , . . . by a similar procedure. We multiply

Eq. (5) by cos mx , where m is any fixed positive integer, and then
integrate from − π to π ,

π  ∞
π 
∫π f ( x ) cos mxdx = ∫ a0 + ∑ (an cos nx + bn sin nx) cos mxdx (7)
− −π
 n =1 

Integrating term-by-term, we see that the right-hand side becomes

π π ∞ π
a 0 ∫ cos mxdx + ∑ a n ∫ cos nx cos mxdx + bn ∫ sin nx cos mxdx 
−π  −π
n =1 
−π 

The first integration is zero. By applying trigonometric identity, we

obtain

π 1 π 1 π
∫ πcos nx cos mxdx = 2 ∫ π cos(n + m) xdx + 2 ∫ π cos(n − m) xdx
− − −

π 1 π 1 π
∫ πsin nx cos mxdx = 2 ∫ π sin(n + m) xdx + 2 ∫ π sin(n − m) xdx.
− − −

Integration shows that the four terms on the right are zero, except for the
last term in the first line which equals π when n = m. since in eq. (7)
this term is multiplied by a m , the right-hand side in eq. (7) is equal
to a mπ , and our second result is

1 π

π ∫−π
am = f ( x) cos mxdx m = 1, 2 . . . (8)
We finally determine b1, b2, ...... in eq.(5) by sin mx , where m is any fixed
positive integer, and the integrate from − π to π , we have

∞
π π  
∫ π f ( x) sin mxdx = ∫ π a + ∑ (a
− −
0
n =1
n cos nx + bn sin nx) sin mxdx

(9)

Integrating term-by-term, we see that the right-hand side becomes

21
PHY312 MATHEMATICAL METHODS OF PHYSICS I

π ∞π π
a0 ∫ sin mxdx + ∑ an ∫ cos nx sin mxdx + bn ∫ sin nx sin mxdx
−π  −π
n =1 
−π 

The first integral is zero. The next integral is of the type considered
before, and we know that it is zero for all n = 1, 2,…. For the integral we
obtain
π 1 π 1 π
∫ πsin nx sinmxdx = 2 ∫ π cos(n − m) xdx − 2 ∫ π cos(n + m) xdx
− − −

The last term is zero. The first term on the right is zero when n ≠ m and
is π when n = m . Since in eq. (9) this term is multiplied by bm , the right-
hand side in eq. (6) is equal to bmπ , and our last result is

1 π

π ∫π
bm = f ( x) sin mxdx m = 1, 2 . . .
−

Writing n in place of m, we altogether have the so-called Euler

formulas

1 π

π ∫π
(a) a0 = f ( x)dx
−

1 π
(b) an =
π ∫ π f ( x) cos nxdx
−
n = 1, 2 . . . (10)
1 π

π ∫π
(c) bn = f ( x) sin nxdx
−

Example 3 Square wave

Find the Fourier coefficients of the periodic function

− k when −π < x < 0

f ( x) =  and f ( x + 2π ) = f ( x)
k when 0< x <π

Functions of this type may occur as external forces acting on mechanical

systems, electromotive forces in electric circuits, etc

Solution: From eq. (10a) we obtain a 0 = 0 . This can also be seen without
integration since the area under curve of f(x) between − π and π is zero.
From eq. (10b)

1 π 1 0 π
(−k ) cos nxdx + ∫ k cos nxdx 
π ∫π ∫
an = f ( x) cos nxdx =
− π  −π 0 

22
PHY312 MATHEMATICAL METHODS OF PHYSICS I

π
1  sin nx sin nx 
0

= − k +k =0
π  n −π n 0 

Because sin nx = 0 at − π , 0 and π for all n = 1, 2….Similarly, from Eq.

(10c) we obtain
1 π 1 0 π
(−k ) sin nxdx + ∫ k sin nxdx 
π ∫π ∫
bn = f ( x) sin nxdx =
− 
π −π 0 

π
1  cos nx cos nx 
0

= − k +k =0
π  n −π n 0 
Since cos(−α ) = cos α and cos 0 = 1 , this yields
k
[cos 0 − cos(−nπ ) − cos nπ + cos 0] = 2k (1 − cos nπ )
bn =
nπ nπ
Now, cos π = −1, cos 2π = 1, cos 3π = −1 etc, in general
− 1 for odd n,
cos nπ =  and thus
1 for even n,
2 for odd n,
1 − cos nπ = 
0 for even n,

Hence the Fourier coefficients bn of our function are

4k 4k 4k
b1 = , b2 = 0, b3 = , b4 = 0, b5 =
π 3π 5π
and since the a n are zero, the corresponding Fourier series is
4k  1 1 
 sin x + sin 3x + sin 5 x + ........ (11)
π  3 5 
The partial sums are
4k 4k  1 
S1 = sin x, S2 =  sin x + sin 3 x , etc,
π π  3 

Furthermore, assuming that f (x) is the sum of the series and

setting x = π 2 , we have

π  4k  1 1 
f =k = 1 − + − +.....
2 π  3 5 
1 1 1 π
or 1 − + − + −......... =
3 5 7 4

23
PHY312 MATHEMATICAL METHODS OF PHYSICS I

SELF-ASSESSMENT EXERCISE 1

1. Define the periodic function. Give five examples.

2. Find the smallest positive period T of the following functions.

a. sin x

3. Are the following functions odd, even, or neither odd nor even?
a. e x
b. x sin x
4. Find the Fourier series of the following functions which are
assumed to have the

a. period 2π
b. f ( x) = x 2 4 −π < x < π
c. f ( x ) = sin x −π < x < π

3.5 Application of Fourier Series in Forced Vibrations

We now consider an important application of Fourier series in solving a

differential equation of the type

d 2x dx
m 2
+ Γ + kx(t ) = F (t ) (12)
dt dt

For example, the above equation would represent the forced vibrations
of a damped oscillator with Γ representing the damping constant, F(t)
the external force and m and k representing the mass of the particle and
the force constant respectively. We write eq. (12) in the form

d 2x dx
2
+ 2K + ω 02 x(t ) = G (t ) (13)
dt dt
Γ k F (t )
Where K = , ω 02 = and G (t ) = . The solution of the
2m m m
homogeneous part of eq. (13) can be readily obtained and is given by

x(t ) = A1e = Kt cos [ (ω 2

0 )
− K 2 t +θ ] for ω 02 > K 2 (14)
x(t ) = ( A2 t + B )e − Kt
for ω >K
2
0
2
(15)

In order to obtain the solution of the inhomogeneous part of eq. (13), we

first assume F(t) to be a sine or cosine function; for definiteness we
assume

24
PHY312 MATHEMATICAL METHODS OF PHYSICS I

G (t ) = b sin ωt (16)

The particular solution of eq. (13) can be written in the form

x(t ) = C sin ωt + D cos ωt (17)

The values of C and D can readily be obtained by substituting eq. (17) in

eq. (13),
and comparing coefficients of sin ωt and cos ωt we obtain

2ωK
D = − b
(ω 2
0 −ω 2 ) 2
+ 4ω 2 K 2

(18)
ω 02 − ω 2
C = − b
(ω 2
0 −ω 2 ) 2
+ 4ω 2 K 2

Now, if G(t) is not a sine or cosine function, a general solution of eq. (13)
is difficult to obtain. However, if we make a Fourier expansion of G(t)
then the general solution of eq.

(13) can easily be written down. As a specific example, we assume

G (t ) = αt (19)

The Fourier expansion of G (t ) can readily be obtained as

∞
G (t ) = ∑ bn sin nωt (20)
n =1

Proceeding in a manner similar to that described above we obtained the

following solution for the inhomogeneous part of eq. (13)
∞
x(t ) = ∑ [C n sin nωt + Dn cos nωt ] (21)
n =1

Where
2nωK
Dn = − bn
(ω 2
0 − n ω2
2
) 2
+ 4 n 2ω 2 K 2
(22)
(ω02 − n 2ω 2 )
Cn = − bn
(ω 2
0 − n 2ω 2 )2
+ 4 n 2ω 2 K 2
thus, if G(t) is a periodic function with period T then eq. (21) will be
valid for all values
of t.

3.6 Half-Range Expansions

In various physical and engineering problems there is a practical need

for applying Fourier series to functions f (t ) which are defined merely

25
PHY312 MATHEMATICAL METHODS OF PHYSICS I

on some finite interval. The function f (t ) is defined on an interval

0< t < l and on this interval we want represent f (t ) by a Fourier series.

A half-range Fourier series for a function f (x) is a series consisting of

the sine and cosine terms only.

Such functions are defined on an interval (0, l ) and we then obtain a

Fourier cosine series which represents an even periodic function f1 (t ) of
period T = 2l so that
∞
nπ
f (t ) = a 0 + ∑ a n cos t 0< t < l
n =1 l
(23)
and the coefficients are
1 l 2 l nπ
l ∫0 ∫
a0 = f (t )tdt , an = f (t ) cos tdt n = 1, 2 . . .
l 0 l
(24)

Then we obtain a Fourier sine series which represents an odd periodic

function f 2 (t ) of period T = 2l so that
∞
nπ
f (t ) = ∑ bn sin t 0< t < l
n =1 l
(25)
and the coefficients are

2 l nπ
bn =
l ∫0
f (t ) sin
l
tdt n = 1, 2 . . .
(26)

The series in eqs.(23) and (25) with the coefficients in eqs.(24) and (26)
are called half-range expansions of the given function f (t )

Example 4

Find the half-range expansions of the function

 2k l
 l t when 0<t <
2
f (t ) = 
 2k (l − t ) when l
<t <l
 l 2

26
PHY312 MATHEMATICAL METHODS OF PHYSICS I

Solution: From eq. (24) we obtain

1  2k l / 2 2k l  k
a 0 =  ∫ tdt + ∫ (l − t ) dt  =
l l 0 l l/2  2
2  2k l / 2 nπ 2k l nπ 
a n =  ∫ t cos tdt + ∫ (l − t ) cos tdt 
l l 0 l l l/2 l 

Now by integration by part

l/2 nπ lt nπ l / 2 1 l / 2 nπ
∫
0
t cos
l
tdt =
nπ
sin
l
t0 −
nπ ∫0
sin
l
tdt

l2 nπ l2  nπ 
= sin + 2 2  cos − 1
2nπ 2 nπ  2 
Similarly,
l nπ l2 nπ l2  nπ 
∫ (l − t ) cos tdt = − sin − 2 2  cos nπ − cos 
l/2 l 2nπ 2 nπ  2 

By inserting these two results we obtain

4k  nπ 
un = 2 2 
2 cos − cos nπ − 1
nπ  2 
Thus,
a 2 = −16k 2 2 π 2 , a 6 = −16k 6 2 π 2 , a10 = −16k 10 2 π 2 , ...

And a n = 0 when n ≠ 2, 6, 10, 14, ... Hence the first half-range expansion
of f (t ) is
k 16k  1 2π 1 6π 
f (t ) = − 2  2 cos t + 2 cos t + ...
2 π 2 l 6 l 

This series represents the even periodic expansion of the function f (t ) .

Similarly from eq. (26)

8k nπ
bn = sin
nπ 2 2
2
and the other half-range expansion of f (t ) is

8k  1 π 1 3π 1 5π 
f (t ) = 2  2
sin t − 2 sin t + 2 sin t − +...
π 1 l 3 l 5 l 

This series represents the odd periodic extension of f (t ) .

27
PHY312 MATHEMATICAL METHODS OF PHYSICS I

Example 5

Find a Fourier sine series for

0 x≤2
f ( x) =  on (0, 3).
2 x>2

Solution: Since the function is odd, then a 0 = 0

2 l nπ
Then bn = ∫ f ( x) sin xdx
l 0 l
2 3 nπ
= ∫ f ( x) sin xdx
3 0 3
2 2 nπ 2 3 nπ
= ∫ 0.sin xdx + ∫ 2 sin xdx
3 0 3 3 2 3
Now by integration, we have
4  2nπ 
bn =  cos − cos nπ 
nπ  3 
The series thus becomes
∞
4  2 nπ  nπx
f ( x) = ∑  cos − ( −1) n  sin
n =1 nπ  3  3

So that
41 πx 3 2πx 2 3πx 
f ( x) =  sin − sin + sin − +...
π 2 3 4 3 3 3 

Example 6

Find the Fourier cosine series for

f ( x) = e x on (0, π )

Solution: Since f (x) is an odd function, then

b0 =
1 l x
∫
l 0
1
e dx = e π − 1
π
( )
Also
nπx
bn =
2
π∫ 0
π
e x cos
π
dx =
2 1  π

π 1 + n2 
(
 e cos nπ − 1 )

Thus the series becomes

(e − 1) + [(−1) ]
∞
2 2 1
ex =
π
π
∑
π 1+ n n =1
2
n
e π − 1 cos nx

28
PHY312 MATHEMATICAL METHODS OF PHYSICS I

SELF-ASSESSMENT EXERCISE 2

1. Find the Fourier sine series for

f ( x) = e x on (0, π )

2. Find the Fourier series for

f ( x) = x on 0 < x < 2
consisting of (a) sine series only (b) cosine series only

3.7 Fourier Integral

Fourier series are powerful tools in treating various problems involving

periodic functions. When the fundamental period is made infinite, the
limiting form of the Fourier series becomes an integral which is called
Fourier Integral.

3.7.1 Definition

Let f (x) be defined and single valued in the interval [− L, L ] . If

f (x) satisfies the following conditions:

(i) f (x) is periodic and of period 2L

(ii) f (x) and f ′(x) are piecewise continuous
∞
(iii) ∫−∞
f ( x) dx is convergent, then f (x) can be expressed as
∞
f ( x) = ∫ ( A(α ) cos αx + B (α ) sin αx )dx
0

(27)
1 ∞
A(α ) = ∫ f ( x) cos αxdx
π −∞

(28)
1 ∞
B (α ) = f ( x) sin αxdx
π∫ −∞

(29)

3.8 Fourier Integrals of Even and Odd Functions

It is of practical interest to note that if a function is even or odd and can

be represented by a Fourier integral, and then this representation will be
simpler than in the case of an arbitrary function. This follows
immediately from our previous formulas, as we shall now see.

If f (x) is an even function, then B (α ) = 0

2 ∞
A(α ) = ∫ f ( x) cos αxdx (30)
π 0

29
PHY312 MATHEMATICAL METHODS OF PHYSICS I

and eq. (27) reduces to the simpler form

∞
f ( x) = ∫ A(α ) cosαxdx ( f even) (31)
0

Similarly, if f (x) is odd, then A(α ) = 0 in eq. (28), also

2 ∞
B(α ) = f ( x ) sin αxdx
π∫ 0

(32)
and
∞
f ( x ) = ∫ B (α ) sin αxdx ( f odd ) (33)
0

These simplifications are quite similar to those in the case of a Fourier

series discussed.

Example 7

Find the Fourier Integral of f ( x) = x 2 −π ≤ x ≤ π

Solution:
1 ∞
A(α ) = f ( x) cos αxdx
π∫ −∞

1 ∞
x 2 cos αxdx
π∫
=
−∞

Using integration by parts, we obtain

π
2 x 1 
A(α ) =  cos αx − 2 sin αx  = 0
πα α α  −π
Also
1 ∞
B (α ) = ∫ f ( x) sin αxdx
π −∞

1 ∞
x 2 sin αxdx
π∫
=
−∞

So that
π
1  x2 2  2π
B (α ) = −  cos αx − 3 cos αx  = (−1)α
π α α  −π α

30
PHY312 MATHEMATICAL METHODS OF PHYSICS I

From eq. (27)

∞
f ( x) = ∫ ( A(α ) cosαx + B (α ) sin αx )dx
0
and
∞ 2π  2π
f ( x ) = ∫  0 • cos αx + (−1)α sin αx dx = 2 (−1)α
0
 α  α
Hence
2π ∞ 2π
f ( x) = x 2 = (−1)α ∫ sin αxdx = (−1)α
α 0 α 2

4.0 CONCLUSION

In this unit, you have studied the concept of periodic functions,

representations of functions by Fourier series, involving sine and cosine
function are given special attention. We also use the series expansion in
the determination of Fourier coefficients and the half-range expansions.

5.0 SUMMARY

In this unit, you have studied:

• Even and odd functions

• Fourier Integral representations and Fourier series expansion.
• Application of Fourier Integral technique in the simplification of
even and odd functions.

6.0 TUTOR- MARKED ASSIGNMENT

1. Find the smallest positive period T of the following functions

a. (i) sin 2πx
2πnx
b. (ii) cos
k

2. Find the Fourier series for

0 −5 < x < 0
f ( x) =  where f (x) has period 10
3 0< x<5

3. Find the Fourier series for

f ( x) = x 2 for 0 < x < 2π

4. Find the Fourier series of function

f ( x ) = x + π when − π < x < π and f ( x + 2π ) = f ( x)

31
PHY312 MATHEMATICAL METHODS OF PHYSICS I

5. Expand the function

T T
f (t ) = t 2 −
< x < in a Fourier series to show that
2 2
T π
2 2
 1 1 
f (t ) = t 2 = 2 
− 4 cos ωt − cos 2ωt + cos ωt − ...
4π  3  4 9 
take ω = 2π T

6. Represent the following functions f (t ) by a Fourier cosine series

π
(a) f (t ) = sin t (0 < t < l )
l
(b) f (t ) = e t (0 < t < l )

7. Find the Fourier integral representation of the function

1 when x < 1,
f ( x) = 
0 when x > 1.

7.0 REFERENCES/FURTHER READING

Puri, S.P. (2004). Textbook of Vibrations and Waves. Macmillan India

Ltd.

Ghatak, A.K.; Goyal, I.C. & Chua, S.J. (1995). Mathematical Physics.
Macmillan India Ltd.

Carslaw, H.S. (1950). Introduction to the Fourier Series and Integral.

New York:

Dover Publications.

32
PHY312 MATHEMATICAL METHODS OF PHYSICS I

MODULE 2 APPLICATION OF FOURIER TO PDES

(LEGENDRE POLYNOMIALS AND
BESSEL FUNCTIONS)

Unit 1 Legendre Polynomials

Unit 2 Bessel Functions

UNIT 1 LEGENDRE POLYNOMIALS

CONTENTS

1.0 Introduction
2.0 Objectives
3.0 Main Content
3.1 Legendre Equation
3.2 The Polynomial Solution of the Legendre’s Equation
3.3 The Generating Function
3.4 Rodrigue’s Formula
3.5 Orthogonality of the Legendre Polynomials
3.6 The Angular Momentum Problem in Quantum Mechanics
3.7 Important Integrals Involving Legendre Functions
4.0 Conclusion
5.0 Summary
6.0 Tutor- Marked Assignment
7.0 References/Further Reading

1.0 INTRODUCTION

In this unit, you will be introduced to the polynomial solutions of the

Legendre equation, the generating function as well as the orthogonality
of Legendre polynomials. Also we shall consider some important
integrals involving Legendre functions which are of considerable use in
many areas of physics.

2.0 OBJECTIVES

At the end of this unit, you should be able:

• derive the polynomial solution of the Legendre equation

• use the generating functions to derive some important identities
• determine the orthogonality of the Legendre polynomials.

33
PHY312 MATHEMATICAL METHODS OF PHYSICS I

3.0 MAIN CONTENT

3.1 Legendre Equation

The equation

(1 − x 2 y ′′( x) − 2 xy ′( x) + n( n + 1) y ( x) = 0 (1)
where n is a constant is known as the Legendre’s differential equation.
In this unit we will discuss the solutions of the above equation in the
domain − 1 < x < 1 . We will show that when
n = 0, 1, 2, 3, . ..
one of the solutions of eq. (1) becomes a polynomial. These polynomial
solutions are known as the Legendre polynomials, which appear in
many diverse areas of physics and engineering.

3.2 The Polynomial Solution of the Legendre’s Equation

If we compare eq. (1) with homogeneous, linear differential equations of

the type

y ′′( x) + U ( x) y ′( x) + V ( x) y ( x) = 0 (2)

we find that the coefficients

2x n(n + 1)
U ( x) = − and V ( x) = (3)
1− x 2
1− x2

are analytical at the origin. Thus the point x = 0 is an ordinary point and
a series solution of eq. (1) using Frobenius method should be possible.
Such that

y ( x) = C 0 S n ( x) + C1Tn ( x)
where
n(n + 1) 2 n(n − 2)(n + 1)(n + 3) 4
S n ( x) = 1 − x + x −... (4a)
2! 4!
And
(n − 1)(n + 2) 3 (n − 1)(n − 3)(n + 2)(n + 4) 5
Tn ( x) = x − x + x − . . . (4b)
3! 5!

If n ≠ 0. 1, 2, . . . both eqs. (4a) and (4b) are infinite series and converge
only if x < 1.

It may be readily seen that when

n = 0. 2, 4, . . .

34
PHY312 MATHEMATICAL METHODS OF PHYSICS I

The even series becomes a polynomial and the odd series remains an
infinite series. Similarly for

n = 1, 3, 5, . . .
the odd series becomes a polynomial and the even series remains an
infinite series.
Thus when
n = 0, 1, 2, 3, . . .
one of the solutions becomes a polynomial. The Legendre polynomial,
or the Legendre function of the first kind is denoted by Pn (x) and is
defined in terms of the terminating series as below:

 S n ( x)
 S (1) for n = 0, 2, 4, 6, . ..
 n
Pn ( x) =  (5)
Tn ( x) for n =1, 3, 5, 7, . ..
 Tn (1)
Thus,
P0 ( x) = 1, P1 ( x) = x, P2 ( x) =
2
(
1 2
3x − 1 ,)
P3 ( x) =
1 3
2
( ) 1
(
5 x − 3 x , P4 ( x) = 35 x 4 − 30 x 2 + 3 ,
8
) (6)
1
( )
P5 ( x) = 63 x 5 − 70 x 3 + 15 x ,. . .
8

Obviously, Pn (1) = 1 (7)

Higher order Legendre polynomials can easily be obtained by using the

recurrence relation

nPn ( x) = ( 2n − 1) x Pn−1 ( x) − ( n − 1) Pn− 2 ( x)

Since for even values of n the polynomials Pn (x) contain only even
powers of x and for odd values of n the polynomials contain only odd
powers of x, we readily have

Pn (− x) = (−1) n Pn ( x) and obviously (8)

Pn (−1) = (−1) n
(9)

3.3 The Generating Function

The generating function for the Legendre polynomials is given by

∞
G ( x, t ) = (1 − 2 xt + t 2 ) −1 2 = ∑ Pn ( x)t n ; − 1 ≤ x ≤ 1, t < 1 (10)
n =0

35
PHY312 MATHEMATICAL METHODS OF PHYSICS I

Let us assume that

∞
G ( x, t ) = (1 − 2 xt + t 2 ) −1 2 = ∑ K n ( x)t n (11)
n =0

Where K n (x) is a polynomial of degree n. Putting x = 1 in eq. (11), we

obtain
∞

∑K
n =0
n ( x)t n = (1 − 2t + t 2 ) −1 2

= (1 − t ) −1
= 1+ t + t 2 + t 3 + . . . + t n + . . .

Equating the coefficients of t n from both sides, we have

K n (1) = 1 (12)

Now, if we can show that K n (x) satisfies eq. (1), then K n (x) will be
identical to Pn (x) . Differentiating G(x, t) with respect to x and t, we
obtain
∂G
(1 − 2 xt + t 2 ) = ( x − t )G ( x, t ) (13)
∂t
and
∂G ∂G
t = (x − t) (14)
∂t ∂x
Using eqs. (11), (13) and (14), we have
∞ ∞
(1 − 2t + t 2 )∑ nK n ( x)t n−1 = ( x − t )∑ K n ( x)t n (15)
n =0 n =0

and
∞ ∞
t ∑ nK n ( x)t n−1 = ( x − t )∑ K n′ ( x)t n (16)
n =0 n =0

Equating the coefficient of t n −1 on both sides of eqs. (15) and (16), we

get
nK n ( x) − ( 2n − 1) xK n −1 ( x ) + ( n − 1) K n − 2 ( x) = 0 (17)
and
xK n′ −1 ( x) − K n′ − 2 ( x) = ( n − 1) K n−1 ( x) (18)

Replacing n by n+1 in Eq. (18), we obtain

xK n′ ( x) − K n′ −1 ( x) = nK n ( x) (19)

We next differentiate Eq. (17) with respect to x and eliminate K n′ −2 with

help of Eq. (18) to obtain

36
PHY312 MATHEMATICAL METHODS OF PHYSICS I

K n′ ( x ) − xK n′ −1 ( x) − nK n−1 ( x) = 0 (20)

If we multiply eq. (19) by x and subtract it from eq. (20), we would get

(1 − x 2 ) K n′ − n( K n−1 − xK n ) = 0 (21)

Differentiating the above equation with respect to x, we have

(1 − x 2 ) K n′′ − 2 xK n′ − n( K n′ −1 − xK n′ − K n ) = 0 (22)

Using eqs. (19) and (22), we obtain

(1 − x 2 ) K n′′( x) − 2 xK n′ ( x) − n(n + 1) K n ( x) = 0 (23)

which shows that K n (x) is a solution of Legendre equation. In view of

eqs. (7) and (12) and the fact that K n (x) is a polynomial in x of degree n,
it follows that K n (x) is nothing but Pn (x) . eq. (17) gives the recurrence
relation for Pn (x)
nPn ( x) = ( 2n − 1) x Pn−1 ( x) − ( n − 1) Pn− 2 ( x) (24)

3.4 Rodrigues’ Formula

Let
φ ( x ) = ( x 2 − 1) n (25)

Differentiating eq. (25), we get

dφ
= 2nx( x 2 − 1) n−1
dx
or
d 2φ dφ
(1 − x 2 ) 2
+ 2 x(n − 1) + 2nφ = 0
dx dx

Differentiating the above equation n times with respect to x, we would

get

d 2φn dφ
(1 − x 2 ) 2
+ 2 x n + n( n + 1)φn = 0 (26)
dx dx
where
φn =
d nφ d n
dx n
= n x2 −1
dx
n
[( )] (27)

37
PHY312 MATHEMATICAL METHODS OF PHYSICS I

This shows that φn (x) is a solution of the Legendre’s equation. Further, it

is obvious from eq. (27) that φn ( x) is a polynomial of degree n in x.
Hence φn ( x) should be a constant multiple of Pn ( x) , i.e.

[(
d n x2 −1 )]
n

= CPn ( x) (28)
dx n
dn
dx n
x[(
2
− 1
n
=)]dn
dx n
[
(x + 1)n (x − 1)n ]
= n!( x − 1) n + n ( x + 1)n(x − 1)
n! n −1

1!
n(n − 1) n!
+ (x + 1)2 n(n − 1)( x − 1) n−2 + . . . + ( x + 1) n n! (29)
2! 2!

It may be seen that all terms on the right hand side of eq. (29) contain a
factor (x-1) except for the last term. Hence

dn 2
dx n
(
x −1
n
) = 2 n n! (30)
x =1

Using Eqs. (7), (28) and (29), we obtain

C = 2 n n! (31)

Therefore
Pn ( x ) =
1 dn 2
n
2 n! dx n
x −1
n
( ) (32)

This is known as the Rodrigues formula for the Legendre polynomials.

For example
P2 ( x ) =
1 d2 2
2
2 2! dx 2
x −1
2
( )
1
(
= 3x 2 − 1
2
)

Which is consistent with eq. (6)

3.5 Orthogonality of the Legendre Polynomials

Since the Legendre’s differential equation is of the Sturm-Liouville form

in the interval − 1 ≤ x ≤ 1, with Pn (x) satisfying the appropriate boundary
conditions at x = ±1 . The Legendre polynomials form an orthogonal set
of functions in the interval − 1 ≤ x ≤ 1, i.e

38
PHY312 MATHEMATICAL METHODS OF PHYSICS I

1
∫−1
p n ( x) p m ( x)dx = 0 m≠n (33)

The Orthogonality of the Legendre polynomials can be proved as

follows: Pn ( x) satisfies eq. (1) which can be written in the Sturm-
Liouville form as

d  2

dx 
(
dP ( x) 
)
x − 1 n  + n( n + 1) Pn ( x) = 0
dx 
(34)

Similarly
d  2

dx 
(
dP ( x) 
)
x − 1 m  + m( m + 1) Pm ( x) = 0
dx 
(35)

Multiply eq. (34) by Pm ( x ) and eq. (35) by Pn ( x) and subtracting eq. (35)
from eq. (34), we get

d
dx
[
(1 − x 2 )(Pn′( x) Pm ( x) − Pm′ ( x) Pn ( x) ) ]
= (m − n)(n + m + 1) Pn ( x) Pm ( x)

Integrating the above equation from x = -1 to x = 1, we get

(1 − x 2 )[(Pn′ ( x) Pm ( x) − Pm′ ( x) Pn ( x) )] −1
+1

1
= (m − n)(n + m + 1) ∫ Pn ( x) Pm ( x)dx
−1

Because of the factor (1 − x 2 ) the left hand side of the above equation
vanishes; hence

1
∫−1
Pn ( x) Pm ( x )dx for m ≠ n

To determine the value of the integral

1
∫−1
Pm2 ( x )dx

we square both sides of eq. (10) and obtain

∞ ∞
(1 − 2 xt + t 2 ) −1 = ∑∑ Pm ( x) Pn ( x)t m + n (36)
m =0 n =0

Integrating both sides of the above equation with respect to x from -1 to

+1 and using eq. (33), we get

39
PHY312 MATHEMATICAL METHODS OF PHYSICS I

∞
1 1 1+ t
∑t ∫
1 1
2n
Pn2 ( x)dx = ∫ dx = ln
n =0
−1 −1 1 − 2 xt + t 2
t 1− t
 1 1 1 2n 
= 21 + t 2 + t 4 + . . . + t + ... 
 3 5 2 n + 1 

Equating the coefficients of t 2 n on both sides of the above equation, we

have

1 2
∫−1
Pn2 ( x)dx =
2n + 1
n = 0, 1, 2, 3, .... (37)

Thus we may write

1 2
∫ Pn ( x) Pm ( x)dx = δ nm
−1 2n + 1
where
0 if n ≠ m
δ nm = 
1 if n = m

Example

We consider the function cos πx 2 and expand it in a series (in the

domain − 1 < x < 1 ) up to the second power of x:

πx 2
cos = ∑ C n Pn ( x)
2 n=0

Now
2 n + 1 +1 πx
Cn =
2 ∫−1
cos Pn ( x) dx
2

Substituting for Pn ( x) from eq. (6) and carrying out brute force
integration, we readily get

2 10  12 
C0 = ; C1 = 0; C 2 = 1 − 
π π  π2 
Thus
πx 2 10  12  3 x 2 − 1 
cos = + 1 −  
2 π π  π 2  2 

40
PHY312 MATHEMATICAL METHODS OF PHYSICS I

3.6 The Angular Momentum Problem in Quantum

Mechanics

In electrostatics the potential Φ satisfies the Laplace equation

∇ 2Φ = 0 (38)

We wish to solve the above equation for a perfectly conducting sphere

(of radius a), place in an electric field which is in the absence of the
sphere was of uniform magnitude E0 along z-direction. We assume the
origin of our coordinate system to be at the centre of the sphere. Because
the sphere is a perfect conductor, the potential on its surface will be
constant which, without any loss of generality, may be assumed to be
zero. Thus, eq. (35) is said to be solved subject to the boundary
condition

Φ (r = a) = 0 (39)

At a large distance from the sphere the field should remain unchanged
and thus
E (r → ∞) = E 0 zˆ
Since
E = −∇Φ
we have
Φ ( r → ∞ ) = − E0 z + C
= − E 0 r cos θ + C (40)

Where C is a constant. Obviously, we should use the spherical system of

coordinates so that

1 ∂  2 ∂Φ  1 ∂  ∂Φ 
∇ 2Φ = r + 2  sin θ 
r ∂r  ∂r  r sin θ ∂θ 
2
∂θ 
1 ∂ 2Φ
+ 2 2 =0 (41)
r sin θ ∂φ 2

From the symmetry of the problem it is obvious that Φ would be

independent of the azimuthal coordinate φ so that eq. (41) simplifies to

1 ∂  2 ∂Φ  1 ∂  ∂Φ 
r + 2  sin θ =0 (42)
r 2
∂r  ∂r  r sin θ ∂θ  ∂θ 

Separation of variables

Φ = R ( r )Θ(θ )
will yield

41
PHY312 MATHEMATICAL METHODS OF PHYSICS I

1 d  2 dR  1 d  dΘ 
r =  sin θ  = a cons tan t (= λ ) (43)
R dr  dr  Θ sin θ dθ  dθ 

Changing the independent variable from θ to µ by the relation

µ = cosθ
In the angular equation, we get
d 2Θ dΘ
(1 − µ 2 ) − 2µ + λΘ = 0 (44)
dµ 2
dµ

In order that the solution of eq. (44) does not diverge

at µ = ±1(θ = 0 and π ) , we must have
λ = l (l + 1); l = 0, 1, 2, . . .
and then
2l + 1
Θ(θ ) = Pl (cosθ ) (45)
2

Thus the radial equation can be written as

1 d  2 dR 
r  = l (l + 1)
R dr  dr 
or
d 2R dR
r2 2
+ 2r − l (l + 1) R = 0 (46)
dr dr

The above equation is the Cauchy’s differential equation and its solution
can readily be written as

Bl
R = Al r l +
r l +1

Hence the complete solution of eq. (42) is given by

∞ ∞
Bl
Φ ( r ,θ ) = ∑ Al r l Pl (cosθ ) + ∑ P (cosθ )
l +1 l
l =0 l =0 r

[
= Ao Po (cosθ ) + A1rP1 (cosθ ) + A2 r 2 P2 (cosθ ) + . . . ]
Bo B
+ Po (cos θ ) + 21 P1 (cos θ ) + . . .
r r

Applying the boundary condition given by eq. (40), we get

A0 = C , A1 = − E o , A2 = A3 = . . . = 0
Thus
 B   B 
Φ ( r ,θ ) =  C + 0  P0 (cosθ ) +  − E 0 r + 21  P1 (cosθ )
 r   r 

42
PHY312 MATHEMATICAL METHODS OF PHYSICS I

B2
+ P2 (cosθ ) + . . .
r3
Applying the condition at r = a [see eq. (39)], we get

 B   B 
 C + 0  +  − E 0 a + 21  P1 (cosθ )
 a   a 
B2 B
+ 3 P2 (cosθ ) + 42 P3 (cosθ ) + . . . = 0
a a

Since the above equation has to be satisfied for all values of θ and
since Pn (cosθ ) form a set of orthogonal functions, the coefficients
of Pn (cosθ ) should be zero giving

B0 = − aC , B1 = E 0 a 3
B2 = B3 = B4 . . . = 0
Thus
 a  a3 
Φ (r ,θ ) = C 1 +  − E0 1 + 3 r cosθ (47)
 r  r 

The 1 r potential would correspond to a charged sphere and, therefore,

for an uncharged sphere we must have C = 0 giving

 a3 
Φ (r ,θ ) = − E0 r cosθ 1 + 3  (48)
 r 

This is the required solution to the problem. One can easily determine
the components of the electric field as:

∂Φ  a3 
Er = − = E0 cosθ 1 + 2 3 
∂r  r 
1 ∂Φ  a3 
Eθ = − = E0 sin θ 1 − 3 
r ∂θ  r 
1 ∂Φ
Eφ = − =0
r sin θ ∂φ

3.7 Important Integrals Involving Legendre Functions

We give below some important integrals involving Legendre functions

which are of considerable use in many areas of physics.

[x + ( x ]
∞ n
1
− 1)1 2 cosθ dθ
π∫
Pn ( x) = 2
(49)
0

43
PHY312 MATHEMATICAL METHODS OF PHYSICS I

π n

∫0 (cosφ + i sin φ cosθ ) dθ

1
Pn (cos φ ) = (50)
π
1 23 2
∫−1 (1 − x) −1 2 Pn ( x)dx = (51)
2n + 1
1 1
[ 2
]
(n + m)!
∫−1 (1 − x 2 ) Pn ( x) dx = m(n − m)!
m
(52)
1
[ ]
2 1 (n + m)!
∫−1 Pn ( x) dx =  1  (n − m)!
m
(53)
n + 
 2

SELF-ASSESSMENT EXERCISE

1. Show that (n + 1) Pn ( x ) = Pn′+1 ( x) − Pn′−1 ( x)

2. Using the Rodrigue’s formula show that
1
Pn′ ( x) = n(n + 1)
2

4.0 CONCLUSION

The concept of generating function for the Legendre polynomials allows

us to readily derive some important identities.

We have also established in this unit, relationship between

Orthogonality of the Legendre polynomials and the generating function.

5.0 SUMMARY

This unit deals with Legendre functions and its applications to physical
problems especially in quantum mechanics.

6.0 TUTOR-MARKED ASSIGNMENT

1. Show that
(1 − x 2 ) Pn′( x) = nPn −1 ( x) − nxPn ( x)
n(n + 1)
=− [Pn+1 ( x) − Pn−1 ( x)]
2n + 1

2. Determine the coefficients

C 0 , C1 , C 2 , C3 , in the expansion
 nx  3
sin   = ∑ C n Pn ( x) −1 < x < 1
 2  n =0

44
PHY312 MATHEMATICAL METHODS OF PHYSICS I

3. Consider the function

0 −1 ≤ x < 0
f ( x) = 
1 0 < x ≤1
Show that
1 1 ∞
f ( x) = − ∑ [Pn+1 (0) − Pn−1 (0) −]Pn ( x) −1 < x < 1
2 2 n =1

4. Show that the generating function

∞
1
= ∑ Pn ( x)u n
1 − 2 xu + u 2 n =0

Hint: Start from the binomial expansion of 1 1 − v , set v = 2 xu − u 2 ,

multiply the powers of 2 xu − u 2 out, collect all the terms involving u n ,
and verify that the sum of these terms is Pn ( x)u n .

7.0 REFERENCES/FURTHER READING

Ghatak, A.K.; Goyal, I.C. & Chua, S.J. (1995). Mathematical Physics.
Macmillan India Ltd.

Erwin, K. (1991). Advanced Engineering Mathematics. John Wiley &

Sons, Inc.

45
PHY312 MATHEMATICAL METHODS OF PHYSICS I

UNIT 2 BESSEL FUNCTIONS

CONTENTS

1.0 Introduction
2.0 Objectives
3.0 Main Content
3.1 Bessel Differential Equation
3.2 Series Solution and Bessel Function of the First Kind
3.3 Recurrence Relations
3.4 The Generating Function
3.4.1 Derivation of Recurrence Relation from the
Generating Function
3.5 Some Useful Integrals
3.6 Spherical Bessel functions
3.7 Bessel Function of the Second Kind
3.8 Modified Bessel functions
4.0 Conclusion
5.0 Summary
6.0 Tutor-Marked Assignment
7.0 References/Further Reading

1.0 INTRODUCTION

In this unit we shall consider the series solution as well as Bessel

functions of the first and second kinds of order n.
We will also be introduced to some integrals which are useful in
obtaining solutions of some problems.

2.0 OBJECTIVES

At the end of this unit, you should be able to:

• derive the solution of Bessel function of the first kind

• prove a relationship between the recurrence relation and the
generating functions
• derive the solution of Bessel function of the second kind.

46
PHY312 MATHEMATICAL METHODS OF PHYSICS I

3.0 MAIN CONTENT

3.1 Bessel Differential Equation

The equation

d2y dy
x2 2
+ x ( x 2 − n 2 ) y ( x) = 0 (1)
dx dx

Where n is a constant known as Bessel’s differential equation.

Since n 2 appears in eq. (1), we will assume, without any loss of
generality, that n is either zero or a positive number. The two linearly
independent solutions of eq. (1) are

J n (x ) and J −n (x)
Where J n ( x) is defined by the infinite series
∞ n+ 2 r
1  x
J n ( x) = ∑ (−1) r   (2)
r =0 r!Γ(n + r + 1)  2 
or
xn  x2 x4 
J n ( x) =  2(2n + 2) 2.4(2n + 2)(2n + 4) − .. .
1 − + (3)
2 Γ(n + 1) 
n

where Γ(n + r + 1) represents the gamma function.

3.2 Series Solution and Bessel Function of the First Kind

If we use eq. (1) with the homogeneous, linear differential equation of

the type

y ′′( x) + U ( x) y ′( x ) + V ( x) y ( x) = R ( x) (4)
we find the coefficients
1 n2
U ( x) = and V ( x) = 1 −
x x2
are singular at x = 0. However, x = 0 is a regular singular point of the
differential equation and a series solution of eq. (1) in ascending powers
of x. Indeed, one of the solutions of eq. (1) is given by

 x2 x4 
J n ( x) = C0 x n 1 − + − .. . (5)
 2(2n + 2) 2.4(2n + 2)(2n + 4) 

and where C0 is an arbitrary constant. This solution is analytic at x = 0

for n ≥ 0 and converges for all finite values of x. If we choose

C 0 = 2 n Γ(n + 1) (6)

47
PHY312 MATHEMATICAL METHODS OF PHYSICS I

then the eq. (5) is denoted by J n(x) and is known as the Bessel function
of the first kind of order n.

∞ n+ 2 r
1  x
J n ( x) = ∑ (−1) r  
r =0 r!Γ(n + r + 1)  2 
n n+ 2 n+4
1  x 1 x 1 x
=   −   +   − ... (7)
Γ(n + 1)  2  1!Γ(n + 2)  2  2!Γ(n + 3)  2 

In particular
J 0 ( x) = 1 −
(x 2)2 + (x 2)4 − (x 2)6 + ... (8)
(1!) 2 (2!) 2 (3!) 2

J 1 2 ( x) =
(x 2)1 2 − (x 2)5 2 + (x 2)9 2 +. . .
Γ(3 2 ) 1!Γ(5 2) 2!Γ(7 2 )
2  x3 x5 
= x − + −. . .
πx  3! 5! 
2
= sin x (9)
πx

It follows immediately from eqs. (7) and (8) that

J n ( 0) = 0 for n > 0
and
J n ( 0) = 1
If n ≠ 0, 1, 2, 3, . . . then

J −n ( x) = ∑
∞
(x 2)− n+ 2r (10)
r =0 r!Γ(− n + r + 1)

Example 1

In this example we will determine the value of J −1 2 ( x) from eq. (10).

Thus
J −1 2 ( x) =
( x 2)
−1 2
−
( x 2)
32
+
( x 2)
72
+. . .
Γ(1 2) 1!Γ(3 / 2) 2!Γ(5 2 )
2  x2 x4 
= 1− + −. . .
πx  2! 4! 
2
= cos x
πx

Which is linearly independent of J 1 2 ( x ) [see eq. (9)] and it can be

verified that J −1 2 ( x) does in fact satisfy eq. (1) for n = ½. Thus

48
PHY312 MATHEMATICAL METHODS OF PHYSICS I

2
J 1 / 2 ( x) = sin x (11)
πx
and
2
J −1 / 2 ( x) = cos x (12)
πx

Using the above two equations and the recurrence relation [see Eq. (21)]
2n
J n +1 ( x) = J n ( x) − J n −1 ( x) (13)
x

We can readily obtain closed form expression for J ±3 / 2 ( x) ,

J ± 5 / 2 ( x) , J ±7 / 2 ( x ) , …
2  sin x 
J 3 / 2 ( x) =  − cos x  (14)
πx  x 
2 1 
J −3 / 2 ( x ) =  − cos x − sin x  (15)
πx  x 
2  (3 − x 2 ) 3 
J 5 / 2 ( x) =  sin x − cos x  (16)
πx  x 2
x 
2  (3 − x 2 ) 3 
J −5 / 2 ( x ) =  cos x − sin x  (17)
πx  x 2
x 
etc.

Next, we will examine eq.(10) when n is a positive integer. To be

specific we assume n = 4; then the first, second, third and fourth terms in
the series given by eq. (10) will contain the terms

1 1 1 1
, , , and
Γ(−3) Γ(−2) Γ(−1) Γ(0)
respectively and all these terms are zero. In general the first n terms of
the series would vanish giving

J −n ( x) = ∑
∞
(x 2)− n+ 2r (18)
r =n r!Γ(− n + r + 1)

If we put r = k+n, we would obtain

∞
J −n ( x) = ∑ (−1) k + n
(x 2)n+2 k
k =0 (k + n)!Γ(k + 1)
∞
= ∑ ( −1) k (x 2)n+ 2k
k =0 k!Γ( k + n + 1)
= (−1) J n ( x)
n
(19)

49
PHY312 MATHEMATICAL METHODS OF PHYSICS I

Thus for n = 0, 1, 2, 3, . . ., J −n ( x) does not represent the second

independent solution of eq. (1). The second independent solution will be
discussed later.

3.3 Recurrence Relations

The following are some very useful relations involving J n ( x) :

xJ n′ ( x) = nJ n ( x) − xJ n +1 ( x) (20a)
= xJ n−1 ( x) − nJ n ( x) (20b)
Thus
2n
J n−1 ( x) + J n+1 ( x) = J n ( x) (21)
x
Also
d n
dx
[ ]
x J n ( x) = x n J n−1 ( x) (22)

In order to prove eq. (20a) w.r.t x to obtain

n + 2 r −1
∞
( n + 2r )  x  1
xJ n′ ( x) = ∑ (−1) r   x (23)
r =0 r!Γ(n + r + 1)  2  2
or
∞ n+ 2 r
1 x
xJ n′ ( x) = n∑ (−1) r  
r =0 r!Γ(n + r + 1)  2 
∞ n + 2 r −1
1 x
+ x ∑ (−1) r
 
r =0 (r − 1)!Γ(n + r + 1)  2 
∞ n + 2 r −1
1 x
= nJ n ( x) − x ∑ (−1) r
  (24)
r =0 r!Γ(n + r + 1)  2 
or
xJ n′ ( x) = nJ n ( x ) − xJ n+1 ( x) (25)

Which proves eq. (20a). eq. (23) can also be written as

∞
2(n + r )( x / 2) n + 2 r −1 x
xJ n′ ( x) = ∑ (−1) r
r =0 r!Γ(n + r + 1) 2
∞
( x / 2) n + 2 r
− n∑ (−1) r
r =0 r!Γ(n + r + 1)
∞ n −1+ 2 r
( x / 2)
= x ∑ (−1) r − nJ n ( x)
r =0 r!Γ(n + r )
or
xJ n′ ( x) = xJ n−1 ( x ) − nJ n ( x) (26)

50
PHY312 MATHEMATICAL METHODS OF PHYSICS I

Which proves eq. (20b). From eq. (25) we readily obtain

d −n
dx
[ ]
x J n ( x) = x − n J n+1 ( x) (27)

Further, adding eqs. (25) and (26) we get

J n−1 ( x ) − J n+1 ( x ) = 2 J n′ ( x) (28)

Using eq. (21) we may write

2
J 2 ( x) = J 1 ( x) − J 0 ( x) (29)
x
4
J 3 ( x ) = J 2 ( x) − J 1 ( x)
x
 8  4
=  2 − 1 J 1 ( x ) − J 0 ( x ) (30)
x  x
6
J 4 ( x) = J 3 ( x ) − J 2 ( x)
x
 48 8   24 
=  3 −  J 1 ( x) −  2 + 1 J 0 ( x ) (31)
x x x 
etc.

The proof of eq. (22) is simple

d n
dx
[ ]
x J n ( x) = x n J n′ ( x) + nx n−1 J n ( x)

 n 
= x n  J n−1 ( x) − J n ( x)  + nx n−1 J n ( x)
 x 
= x J n−1 ( x)
n
[Using eq. (20b)] (32)

Now using eq. (20a)

J 0′ ( x) = − J 1 ( x) (33)

Therefore
∫ J ( x)dx = − J
1 0 ( x) + Constant (34)
or
∞
∫
0
J 1 ( x )dx = 1 [Because J 0 (0) = 1 ] (35)

Equation (32) gives us

∫x J n−1 ( x) dx = x n J n ( x)
n
(36)

51
PHY312 MATHEMATICAL METHODS OF PHYSICS I

Example 2

In this example we will evaluate the integral

∫x
4
J 1 ( x)dx
in terms of J 0 ( x) and J 1 ( x) . Since
d p
dx
[ ]
x J p ( x) = x p J p −1 ( x) [see eq. (22)]
we have
∫x J p −1 ( x )dx = x p J p ( x)
p

Thus
∫x
4
[
J 1 ( x) dx = ∫ x 2 x 2 J 1 ( x) dx ]
= x2 [x J2
2 ( x)]− ∫ 2 x J
3
2 ( x) dx
= x 4 J 2 ( x) − 2 x 3 J 3 ( x)
4 
= x 4 J 2 ( x) − 2 x 3  J 2 ( x) − J 1 ( x)
x 
 2 
= ( x 4 − 8 x 2 ) J 1 ( x) − J 0 ( x)  + 2 x 3 J 1 ( x)
x 
( )
= 4 x 16 x J 1 ( x) − ( x 4 − 8 x 2 ) J 0 ( x)
3

plus, of course, a constant of integration.

3.4 The Generating Function

Bessel functions are often defined through the generating function G(z,t)
which is given by the following equation

 z  1 
G ( z, t ) = exp   t −  (37)
 2  t 

For every finite value of z, the function G(z,t) is a regular function of t

for all (real or complex) values of t except at point t = 0. Thus it can be
expanded in a Laurent series

 z  1  +∞
exp   t −  = ∑ t n J n ( z ) (38)
 2  t  n =−∞

In the above equation, the coefficient of t n is defined as J n ( z ) ; we will

presently show that this definition is consistent with series given by eq.
(3). Now, for any finite value of z and for 0 < t < ∞ we may write

52
PHY312 MATHEMATICAL METHODS OF PHYSICS I

n
 zt  ∞ 1  zt 
exp   = ∑  
 2  n =0 n!  2 
2 3
z 1  z  t2  z  t3
= 1+ +  +  +. . . (39)
2 1!  2  2!  2  3!
and
 z  ∞ (−1)  z 
n n

exp −  = ∑  
 2t  n=0 n!  2t 
2 3
z1 z 1 z 1
= 1− +  −  +. . . (40)
2 t  2  2!t 2  2  3!t 3
Thus the generating function can be expressed as a series of the form

 z  1  +∞
G ( z , t ) = exp   t −  = ∑ An ( z )t n (41)
 2  t  n= −∞
or
+∞  z 1  z  2 t 2  z 3 t 3 
∑ An ( z )t = 1 +
n
+  +  + . . .
n = −∞  2 1!  2  2!  2  3! 
 z 1  z 2 1  z 3 1 
× 1 − +  2
−  3
+ . . . (42)
 2 1!t  2  2!t  2  3!t 

On the other hand, the coefficient of t 0 will be given by

2 4 6
z 1 z 1 z 1
A0 ( z ) = 1 −   2
+  2
−  2
+.. . (43)
 2  (1!)  2  (2!)  2  (3!)

Comparing the above equation with eq. (8), we find

A0 ( z ) = J 0 ( z )

Similarly, the coefficient of t n on the right hand side of eq. (42) will be
given by
n n+2 n+ 4
z 1 z 1 z 1
An ( z ) =   +  +  −. . .
 2  n!  2  (n + 1)!1!  2  (n + 2)!

which when compared with eq. (7) gives us

An ( z ) = J n ( z )
Proving
 z  1   +∞
exp   t −   = ∑ t n J n ( z )
 2  t   n= −∞
In the above equation, if we replace t by -1/y, we obtain
z  1  + ∞ +∞
exp   y −  = ∑ (−1) n y − n J n ( z ) = ∑ y n J n ( z )
2  y  n= −∞ n = −∞

53
PHY312 MATHEMATICAL METHODS OF PHYSICS I

Thus
J n ( z ) = (−1) n J −n ( z )

3.4.1 Derivation of the Recurrence Relations from the

Generating Function

Differentiating eq. (38) w.r.t z, we obtain

1  1  z  1  +∞ n
 t −  exp   t −  = ∑ t J n′ ( z ) (44)
2 t   2  t  n=−∞
Thus
+∞ +∞ +∞

∑t
n = −∞
n +1
J n ( z ) − ∑ t n−1 J n ( z ) =
n = −∞
∑t
n = −∞
n
2 J n′ ( z )

Comparing the coefficients of t n , we obtain

J n−1 ( z ) − J n+1 ( z ) = 2 J n′ ( z )

Similarly, if we differentiate eq. (38) w.r.t t we will obtain

z 1  +∞ n +∞
1 + 2  ∑ t J n ( z ) = ∑ nt J n ( z )
n −1

2  t n= −∞ n = −∞

Comparing the coefficients of t n −1 , we get

z[J n−1 ( z ) − J n+1 ( z )] = 2nJ n ( z )

3.5 Some Useful Integrals

cos[x sin θ − nθ ]dθ

1 π

π∫
Using J n ( z ) =
0

2 π /2
cos( x sin θ )dθ
π∫
J 0 ( z) = (45)
0

Thus
∞ 2  ∞ −αx e ix sin θ + e −ix sin θ 
π /2
∫ e −αx J 0 ( x)dx = dx  dθ
π ∫0  ∫0
e
0 2 
1 π / 2  1 1 
= ∫  + dθ
π 0
α − i sin θ α + i sin θ 
2α π / 2 dθ
= ∫
π 0 α + sin 2 θ
2
(46)
or
∞ 1
∫ e −αx J 0 ( x)dx = (47)
0
1+α 2

54
PHY312 MATHEMATICAL METHODS OF PHYSICS I

where in evaluating the integral on the right hand side of eq. (46), we
have used the substitution y = α cot θ . By making α → 0 , we get
∞
∫0
J 0 ( x)dx = 1 (48)

From eq. (28), we have

2 J n′ ( x) = J n−1 ( x) − J n+1 ( x)
Thus
∞ ∞ ∞
2 ∫ J n′ ( x) = ∫ J n−1 ( x) − ∫ J n+1 ( x)
0 0 0

But
∞ ∞
∫ J n′ ( x) = J n ( x) 0
0
= 0 for n > 0
Thus
∞ ∞
∫0
J n+1 ( x) = ∫ J n−1 ( x)
0
n>0 (49)
Since
∞
∫0
J 1 ( x) = 1 [see eq. (35)]
and
∞
∫0
J 0 ( x) = 1 [see eq. (48)]

Using eq. (49), we have

∞
∫0
J n ( x) = 1 n = 0, 1, 2, 3, ... (50)

Replacing α by α + iβ in eq. (47), we get

∞ 1
∫ e −(α +iβ ) x J 0 ( x )dx = (51)
0
(α + iβ ) 2 + 1
which in the limit of α → 0 becomes
∞ 1
∫ e −iβx J 0 ( x)dx = (52)
0
1− β 2

For β < 1, the right hand side is real and we have

∞ 1
∫ J 0 ( x) cos βxdx = (53)
0
1− β 2
and
∞
∫ J 0 ( x) sin β xdx = 0
0

Similarly, β > 1, the right hand side of Eq. (52) is imaginary and we have
∞
∫ J 0 ( x) cos βxdx = 0
0

55
PHY312 MATHEMATICAL METHODS OF PHYSICS I

∞ 1
∫ J 0 ( x) sin βxdx = (54)
0
1− β 2

3.6 Spherical Bessel Functions

1
We start with the Bessel equation eq. (1)] with n = l + , i.e.
2
dy   1  
2
d2y
x 2
+ x  x −  l +   y ( x) = 0 (55)
dx 2 dx   2  
where
l = 0, 1, 2, . . .

The solutions of eq. (55) are

J 1 ( x) and J 1 ( x)
l+ −l−
2 2

If we make the transformation

1
f ( x) = y ( x) (56)
x
we would readily obtain
1 d  2 df   l (l + 1) 
x  + 1− f ( x) = 0 (57)
x 2 dx  dx   x 2 

The above equation represents the spherical Bessel equation. From eqs.
(55) and (56) it readily follows that the two independent solutions of
eq.(57) are

1 1
J 1 ( x ) and J 1 ( x)
x 2
l+ x −l− 2

The spherical Bessel functions are defined through the equations

π
jl ( x) = J 1 ( x) (58)
2x l+
2
and
π
nl ( x) = (−1) l J 1 ( x) (59)
2x −l −
2

and represent the two independent solutions of eq. (57). Now, if we

define the function
u(x) = xf(x)

then eq. (57) takes the form

d 2 u  l (l + 1) 
+ 1− u ( x) = 0 (60)
dx 2  x 2 

56
PHY312 MATHEMATICAL METHODS OF PHYSICS I

The above equation also appears at many places and the general solution
is given by
u ( x) = c1[xJ l ( x)] + c2 [xnl ( x)] (61)
which also be written in the form
   
u ( x) = A1  x J 1 ( x) + A2  x J 1 ( x) (62)
l+ −l −
 2   2 
For l = 0 , the solutions of eq. (60) are
sin x and cos x

Thus, for l = 0 the two independent solutions of eq.(57) are

sin x cos x
and
x x

Indeed if we use the definitions of jl (x) and nl (x) given eqs. (58) and (59)
respectively, we would readily obtain
sin x
j0 ( x) = (63)
x
cos x
n0 ( x) = (64)
x
sin x cos x
j1 ( x) = 2 − (65)
x x
cos x sin x
n1 ( x ) = − etc (66)
x2 x

Further, if we multiply the recurrence relation [Eq. (21)]

2n
J n +1 ( x) = J n ( x) − J n −1 ( x)
x
π l
by and assume n = l − , we would get
2x 2
(2l − 1)
jl ( x) = nl −1 ( x) − nl − 2 ( x ) (67)
x
using which we can readily obtain analytic expression for j2 ( x), j3 ( x), . ..
etc.
Similarly,
(2l − 1)
nl ( x) = nl −1 ( x) − nl − 2 ( x) (68)
x

3.7 Bessel Functions of the Second Kind: Yn

The Bessel functions of the second kind, denoted by Yn (x) , are solutions
of the Bessel differential equation. They have a singularity at the origin
(x = 0). Yn (x) is sometimes also called the Neumann function. For non-
integer n, it is related to J n (x ) by:

57
PHY312 MATHEMATICAL METHODS OF PHYSICS I

J n ( x ) cos µπ − J − n ( x)
Yn ( x) =
sin µπ
(69)
or
1 ∂ ∂ 
Yn ( x) =  J µ ( x ) − ( −) n J − µ ( x) (70)
n  ∂µ ∂µ  µ =n

We need to show now that Yn (x ) defined by eq.(70) satisfies Eq.(1)

where n is either zero or an integer. We know that

1  µ2 
J µ′′ ( x) + J µ′ ( x) + 1 − 2  J µ ( x) = 0 (71)
x  x 
for any value of µ . Differentiating the above equation with respect to µ ,
we get
d 2 ∂J µ ( x) 1 d ∂J µ ( x)  µ 2  ∂J µ ( x) 2 µ
+ + 1 − 2  = 2 J µ ( x) (72)
dx 2 ∂µ x dx ∂µ  x  ∂µ x

Similarly

d 2 ∂J − µ ( x ) 1 d ∂J − µ ( x)  µ 2  ∂J − µ ( x) 2 µ
+ + 1 − 2  = 2 J − µ ( x) (73)
dx 2 ∂µ x dx ∂µ  x  ∂µ x
From eqs. (72) and (73), it is easy to show that
d2 1 d  µ2 
S µ ( x) + S µ ( x) + 1 − 2  S µ ( x)
dx 2 x dx  x 
2µ
[
= 2 J µ ( x) − (−1) n J − µ ( x)
x
] (74)
where
∂ ∂
S µ ( x) = J µ ( x) − (−1) n J −µ ( x) (75)
∂µ ∂µ

Thus Yn (x ) is the second solution of Bessel’s equation for all real values
of n and is known as the Bessel function of the second kind of order n.
The general solution of eq.(1) can, therefore, be written as

y = C1 J n ( x) + C 2Yn ( x) (76)
where C1 and C 2 are arbitrary constants.

The expression for Yn (x ) for n = 0, 1, 2, . . . can be obtained by using eqs.

(2) and (70) and is given below

58
PHY312 MATHEMATICAL METHODS OF PHYSICS I

− n n −1 r
1  x (n − r − 1)!  x 2 
Yn ( x) = (ln( x / 2) + γ )J n ( x) −  
2
π π 2
∑ r!
 
r =0  4 
−n ∞
1  x ( x 2 / 4) r
−   ∑ (−1) r [ϕ (r ) + ϕ (r + n)] (77)
π 2 r =0 r!(n + r )!
m
Where ϕ (r ) = ∑ s −1 ; ϕ (0) = 0
s =1

and γ = Lim[ϕ (n) − ln n]

n →∞

Example 3

In this example we will solve the radial part of the Schrodinger equation
1 d  2 dR   2 µE l (l + 1) 
2 r + − 2  R( x) = 0; l = 0, 1, 2, (78)
r dr  dr   h r 
in the region 0 < r < a subject to the following boundary conditions that
R (a) = 0 (79)

and R (r ) is finite in the region 0 < r < a . Equation (78) can be

conveniently written in the form

1 d  2 dR   l (l + 1) 
ρ  + 1 −  R( ρ ) = 0
ρ 2 dρ  dρ   ρ 2 
Where
ρ = kr; k = (2 µE / h 2 )1 / 2

Thus the general solution of the above equation is given by

R ( ρ ) = Aj l ( ρ ) + Bnl ( ρ ) (80)

But nl ( ρ ) diverges at ρ = 0 , therefore, we must choose B=0. The

boundary condition R(a)=0 leads to the transcendental equation
jl (ka) = 0 (81)

Thus, for l = 0 , we have

ka = nπ ; n = 1, 2,. . . . (82)

Which will give allowed values of k. Similarly, for l = 1 , we get

tan ka = ka (83)

3.8 Modified Bessel Functions

If we replace x by ix in eq. (1), we obtain

x 2 y ′′ + xy ′ − ( x 2 + n 2 ) y = 0 (84)

59
PHY312 MATHEMATICAL METHODS OF PHYSICS I

The two solutions of the above equation will obviously be

J n (ix ) and Yn (ix)

As these functions are real for all values of n, let us define a real
function as

I n ( x) = i − n J n (ix) (85)

or
∞
( x / 4) n + 2 r
I n ( x) = ∑ (86)
r = 0 r!( n + r + 1)!

This function will be the solution of eq. (84) and is known as the
Modified Bessel function of the first kind. For very large values of x
ex
I n ( x) ~ (87)
2πx

The other solution known as the Modified Bessel function of the second
kind is defined as

π I −n ( x) − I n ( x)
K n ( x) = (88)
2 sin nπ

For non-integer values of n, I n and I −n are linearly independent and as

such K n (x) is a linear combination of these functions [compare with eq.
(69) which gives the definition of Yn (x)] . When n is an integer, it can be
shown [see eq. (86)] that

I −n = I n (89)
and therefore K n (x) becomes indeterminate for n = 0 or an integer. As
in the case of Yn (x ) for n = 0 or an integer, we define K n (x) as
 π I − µ ( x) − I µ ( x) 
K n ( x) = Lim   (90)
µ →n 2
 sin µπ 
or
(−1) n  ∂I − µ ( x) ∂I µ ( x) 
K n ( x) =  −  (91)
2  ∂µ ∂µ  µ = n
For x very large
π −x
K n ( x) ~ e (92)
2x
From eq. (88) it follows that
K − n ( x) = K n ( x) (93)

60
PHY312 MATHEMATICAL METHODS OF PHYSICS I

Which is true for all values of n. recurrence relations for I n can be

derived from those of J n (x ) and Eq. (85). They are

xI n′ ( x ) = xI n −1 ( x) − nI n ( x)
(94) xI n′ ( x ) = nI n ( x) + xI n +1 ( x)
95) I n −1 ( x) + I n +1 ( x) = 2 I n′ ( x) (96)
and similarly
xK n′ = (nK n − xK n −1 ) (97)
xK n′ = nK n + xK n +1 (98)
K n −1 + K n +1 = −2 K n′ (99)

Example 4

In this example we will consider the solutions of the equation

r2
d 2R
dr 2
+r
dR
dr
[ ]
+ (k 02 n 2 (r ) − β 2 )r 2 − l 2 R (r ) = 0 l = 0, 1, . . . (100)
n(r ) = n1 0<r<a
Where (101)
= n2 r>a
and n2 < n1; k 0 (ω / c) represents the free space wave number. The
quantity β represents the propagation constant and for guided modes
β 2 takes discrete values in the domain
k 02 n 22 < β 2 < k 02 n12 (102)

Thus, in the regions 0 < r < a and r > a , eq. (100) can be written in the
form

d 2R dR  2 r 2 
r 2
2
+r + U 2 − l 2  R(r ) = 0 0 < r < a. (103)
dr dr  a 
and
d 2R dR  2 r 2 
r2 2
+ r + W 2 + l 2  R (r ) = 0 r > a. (104)
dr dr  a 
where
[
U 2 = a 2 k 02 n12 − β 2 ] (105)
and
[
W 2 = a 2 β 2 − k 02 n22 ] (106)
so that
V 2 = U 2 + W 2 = a 2 k 02 (n12 − n 22 ) (107)
is a constant. The solutions of Eq. (103) are
 r  r
J l U  and Yl  U  (108)
 a  a

61
PHY312 MATHEMATICAL METHODS OF PHYSICS I

and the latter solution has to be rejected as it diverges at r = 0. Similarly,

the solutions of eq. (104) are

 r  r
K l W  and I l W 
 a  a
and the second solution has to be rejected because it diverges as
r → ∞ . Thus
 A  r
 J (U ) J l U  0<r<a
  a
and R(r ) =  l (109)
 A  r
K l W  r>a
 K l (W )  a
where the constants have been so chosen and R(r) is continuous at r = a.
Continuity of dR/dr at r = a gives us

J l′ (U ) K ′ (U )
U = WU l (110)
J l (U ) K l (U )

which is the fundamental equation determining the eigenvalues β / k 0 .

SELF-ASSESSMENT EXERCISE

1. Using
J 0 (2) = 0.22389, J 1 (2) = 0.57672, calculate J 2 (2), J 3 (2), and J 4 (2).

Hint: Use Eq. (21)

2. Show that
a 1 2 2  J n−1 (a ) J n +1 (a ) 
∫0
J n2 ( x)xdx =
2
a J n (a ) 1 −
 J n2 (a )



4.0 CONCLUSION

In this unit, we have considered Bessel function and spherical Bessel

function.

We have also established in this unit, relationship between the

recurrence relation and the generating function.

5.0 SUMMARY

This unit is on Bessel functions. It has a lot of application that arises in

numerous diverse areas of applied mathematics. This unit will be of
significant importance in the subsequent course in quantum mechanics.

62
PHY312 MATHEMATICAL METHODS OF PHYSICS I

6.0 TUTOR- MARKED ASSIGNMENT

1. Using

J 1 (2) = 0.57672, J 2 (2) = 0.35283 calculate J 3 (2), J 4 (2), and J 5 (2).

Hint: Use Eq. (21)

2. Using the integral

1 Γ(n + r + 1)Γ(m + 1)
∫ (1 − x ) x 2 n + 2 r +1 dx =
2 m
; m > −1, n > −1
0 2Γ(m + n + r + 2)
Prove that
m +1
2 x 1
J n + m +1 ( x) =   ∫ (1 − y ) y n +1 J n ( xy )dy
2 m

Γ(m + 1)  2  0

3. Hint: Use the expansion given by eq. (2) and integrate term by
term.

1
In problem 2 assume m = n = − , and use eq. (12) to deduce
2
2 1 cos xy
π∫
J 0 ( x) = dy
0
1− y2

4. Show that the solution of the differential equation

y ′′( x) + (ae x − b) y ( x) = 0
is given by y ( x) = AJ µ (ξ ) + BJ µ (ξ ); ξ = 2 ae x / 2 ; µ = 2 b

7.0 REFERENCES/FURTHER READING

Erwin, Kreyszig (1991). Advanced Engineering Mathematics. John

Wiley & Sons, Inc.

Arfken, G. (1990). Mathematical Methods for Physicists. New York:

Academic Press

63
PHY312 MATHEMATICAL METHODS OF PHYSICS I

MODULE 3 APPLICATION OF FOURIER TO PDES

(HERMITE POLYNOMIALS AND
LAGUERRE POLYNOMIALS)

Unit 1 Hermite Polynomials

Unit 2 Laguerre Polynomials

UNIT 1 HERMITE POLYNOMIALS

CONTENTS

1.0 Introduction
2.0 Objectives
3.0 Main Content
3.1 Hermite Differential Equation
3.2 The Generating Function
3.3 Rodrigues Formula
3.4 Orthogonality of Hermite Polynomials
3.5 The Integral Representation of the Hermite Polynomials
3.6 Fourier Transform of Hermite-Gauss Functions
3.7 Some Important Formulae Involving Hermite Polynomials
4.0 Conclusion
5.0 Summary
6.0 Tutor-Marked Assignment
7.0 References/Further Reading

1.0 INTRODUCTION

In this unit, we shall consider certain boundary value problems whose

solutions form orthogonal set of functions. It can also be seen in this unit
how the generating function can readily be used to derive the Rodrigues’
formula.

2.0 OBJECTIVES

At the end of this unit, you should be able to:

• define Hermite polynomials as the polynomial solutions of the

Hermite differential equation
• prove the Orthogonality of Hermite polynomials
• derive the Rodrigues’ formula which can be used to obtain
explicit expressions for Hermite polynomials
• solve the exercises at the end of this unit.

64
PHY312 MATHEMATICAL METHODS OF PHYSICS I

3.0 MAIN CONTENT

3.1 Hermite Differential Equation

The equation

y ′′( x) − 2 xy ′( x) + (λ − 1) y ( x) = 0 (1)
where λ is a constant is known as the Hermite differential equation.
When λ is an odd integer, i.e. when
λ = 2n + 1; n = 0, 1, 2, . . . (2)

One of the solutions of eq. (1) becomes a polynomial. These polynomial

solutions are called Hermite polynomials. Hermite polynomials appear
in many diverse areas, the most important being the harmonic oscillator
problem in quantum mechanics.

Using Frobenius method to solve eq.(1), and following the various steps,
we have

Step1: We substitute the power series

∞
y ( x) = ∑ C r x p + r (3)
r =0

in eq. (1) and obtain the identity

C 0 p ( p − 1) + C1 ( p + 1) px +
∞

∑ [C
r =2
r ( p + r )( p + r − 1) −C r − 2 ( 2 p + 2r − 3 − λ )] x r = 0

Step 2: Equating to zero the coefficients of various powers of x, we

obtain

(i) p=0 or p =1 (4a)

(ii) p ( p + 1)C1 = 0 (4b)
2 p + 2r − 3 − λ
(iii) Cr = C r −2 for r ≥ 2 (4c)
( p + r )( p + r − 1)

When p = 0, C1 becomes indeterminate; hence p = 0 will yield both the

linearly independent solutions of eq. (1). Thus, we get

2r − 3 − λ
Cr = C r −2 for r ≥ 2 (5)
r (r − 1)
which gives
1− λ
C2 = C0 .
2!
3−λ
C3 = C1 , . .
3!

65
PHY312 MATHEMATICAL METHODS OF PHYSICS I

(1 − λ )(5 − λ )
C4 = C0
4!
(3 − λ )(7 − λ )
C5 = C1 , . . . etc
5!

Because C 2 , C 4 ,. . . . are related to C 0 and C 3 , C 5 , . . . are related to

C1 , we can split the solution into even and odd series. Thus, we may
write

y ( x) = (C 0 + C 2 x 2 + C 4 x 4 + . . .) + (C1 x + C 3 x 3 + + . . .)
 1 − λ 2 (1 − λ )(1 − λ ) 4 
= C 0 1 + x + x + . . .
 2! 4! 
 (3 − λ ) (3 − λ )( 7 − λ ) 
+ C1  x + x3 + x5 + . . . (6)
 3 ! 5! 

It may be readily seen that when

λ = 1, 5, 9, . . .
the even series becomes a polynomial and the odd series remains an
infinite series. Similarly, for
λ = 3, 7, 11, . . .
the odd series becomes a polynomial and the even series remains an
infinite series. Thus, when

λ = 2n + 1; n = 0, 1, 2, . . .

One of the solutions becomes a polynomial. If the multiplication

constant C 0 or C1 is chosen that the coefficient of the highest power of x
in the polynomial becomes 2 n , then these polynomials are known as
Hermite polynomials of order n and are denoted by H n (x) . For example,
for λ = 9 (n = 4) , the polynomial solution
 4 
y ( x) = C 0 1 − 4 x + x 4 
 3 
If we choose
C 0 = 12
the coefficient of x 4 becomes 2 4 and, therefore
H 4 ( x) = 16 x 4 − 48 x 2 + 12

Similarly,

for λ = 7 (n = 3) , the polynomial solution is given by

 2 
y ( x) = C1  x − x 3 
 3 

66
PHY312 MATHEMATICAL METHODS OF PHYSICS I

Choosing
C1 = −12
we get
H 3 ( x ) = 8 x 3 − 12 x

In general

n!(2 x) n − 2 r
N
H n ( x) = ∑ (7)
r = 0 r!( n − 2 r )!

where
n
 2 if n is even
N =
n −1 if n is odd
 2

Using eq. (7) one can obtain Hermite polynomials of various orders, the
first few are given below:

H 0 ( x ) = 1; H 1 ( x) = 2 x; H 2 ( x ) = 4 x 2 − 2; 

H 3 ( x) = 8 x 3 − 12 x; H 4 ( x) = 16 x 4 − 48 x 2 + 12  (8)

Higher order Hermite polynomials can easily be obtained either by using

eq. (7) or by using the recurrence relation (see eq. 20)

3.2 The Generating Function

The generating function for Hermite polynomials is given by

∞
1
G ( x, t ) = e − t + 2 xt
=∑
2
H n ( x)t n (9)
n =0 n!

Expanding e −t and e 2 xt in power series, we have

2

1 4 1 6
e −t = 1 − t 2 + t − t +. . .
2

2! 3!
(2 x ) 2 2 ( 2 x ) 3 3
e 2 xt = 1 + (2 x )t + t + t +. .
2! 3!

Multiplying the above two series, we shall obtain a power series in t

with

67
PHY312 MATHEMATICAL METHODS OF PHYSICS I

1
Coefficient of t 0 = 1 = H 0 ( x)
0!
1
“ “ t=2 = H 1 ( x)
1!
1
“ “ t 2 = 2x 2 − 1 = H 2 ( x) etc
2!

It is also evident that the coefficient of t 2 in the multiplication of the two

series will be a polynomial of degree n and will contain odd powers
when n is odd and even powers when n is even. In this polynomial, the
coefficient of x n can easily be seen to be (2 n / n!) . We then assume that

∞
1
G ( x, t ) = e − t + 2 xt
=∑
2
K n ( x)t n (10)
n=0 n!

Where K n (x) is a polynomial of degree n. Differentiating eq. (10) with

respect to t, we get
∞ ∞
n 1
( 2 x − 2t )e −t + 2 xt
=∑ K n ( x)t n −1 = ∑ K n ( x)t n −1
2

n = 0 n! n = 0 ( n − 1)!

or
∞ ∞
1 1
2( x − t ) ∑ K n ( x)t n = ∑ K n +1 ( x)t n (11)
n = 0 n! n = 0 n!

Comparing the coefficients of t n on both sides of eq. (11), we obtain

2 xK n ( x) − 2nK n −1 ( x) = K n +1 ( x ) (12)

We next differentiate eq.(10) with respect to x to obtain

∞ ∞
1 1
2t ∑ K n ( x)t n = ∑ K n′ ( x)t n (13)
n = 0 n! n = 0 n!

Comparing the coefficients of t n on both sides of eq. (11), we get

K n′ ( x) = 2nK n −1 ( x) (14)

If we replace n by (n+1) in eq.(14), we would get

K n′ +1 ( x) = 2(n + 1) K n ( x) (15)

Differentiating eqs.(14) and (12) with respect to x, we obtain

respectively

K n′′ ( x) = 2nK n′ −1 ( x) (16)

and

68
PHY312 MATHEMATICAL METHODS OF PHYSICS I

2 xK n′ ( x) + 2 K n ( x) − 2nK n′ −1 ( x) = K n′ +1 ( x) (17)

Subtracting eqs. (17) and (16) and using (15), we get

K n′′ ( x ) − 2 xK n′ ( x) + 2nK n ( x) = 0 (18)

which shows that K n (x) is a solution of the Hermite equation (1) with
λ = 2n + 1 , i.e. of the equation
y ′′( x) − 2 xy ′( x) + 2ny ( x) = 0 (19)

Since, as discussed before, K n (x) is also a polynomial of degree n (with

coefficient of x n equal to 2 n ), K n (x) is, therefore, nothing but H n (x) .
Equations (12) and (14), thus, give recurrence relations for H n (x)
2 xH n ( x) = 2nH n −1 ( x) + H n +1 ( x) (20)
and
H n′ ( x) = 2nH n −1 ( x) (21)
3.3 Rodrigues Formula

In the preceding section we have shown that

∞
1
G ( x, t ) = e − t + 2 xt
=∑
2
H n ( x)t n (22)
n = 0 n!

One can rewrite the generating function G(x, t) in the form

G ( x, t ) = e x e − ( t + x )
2 2

It may be easily seen that

∂ nG n ∂
n
= x2
− e −(t − x )
2
e ( 1) (23)
∂t n
∂x n

From eq. (22) it follows that

∂ nG
= H n ( x) (24)
∂t n t =0

Using eqs (23) and (24), we obtain

d n −x2
H n ( x) = (−1) n e x
2
e (25)
dx n
which is known as Rodrigues formula for Hermite polynomials. For
example,
d 2 −x2 2 d
H 2 ( x) = e x e = ex (−2 xe − x )
2 2

2
dx dx
[
− x2
= e − 2e + 4 x e
x2 2 − x2
]
= 4x 2 − 2

69
PHY312 MATHEMATICAL METHODS OF PHYSICS I

Which is consistent with eq. (8). Similarly, we can determine other

Hermite polynomials by elementary differentiation of eq. (25).

3.4 Orthogonality of Hermite Polynomials

The Hermite polynomials satisfy eq.(1) for λ = 2n + 1 . Thus, we have

d 2Hn dH n
2
− 2x + 2nH n ( x) = 0 (26)
dx dx

In order to derive the Orthogonality condition we transform eq. (26) to

the Sturm-Liouville form by multiplying it by

[
exp − ∫ 2 xdx = e − x ] 2
(27)
to obtain
d  − x 2 dH n 

dx 
e
dx  = −2n e − x H n ( x )
2
[ ] (28)

Similarly
d  − x 2 dH m 

dx 
e
dx  = −2m e − x H m ( x)
2
[ ] (29)

We multiply eq.(28) by H m (x) and eq.(29) by H n (x) , subtract them and

integrate the resulting equation with respect to x from − ∞ to ∞ to obtain

+∞ d  2 dH n  d  2 dH m  
∫  H m ( x ) e − x  − H n ( x ) e − x dx
−∞
 dx  dx  dx  dx  
+∞
= 2(m − n) ∫ e − x H m ( x) H n ( x)dx
2

−∞

Now
+∞ d   − x 2 dH n   2 dH m  
LHS= ∫  H m ( x) e  − H n ( x ) e − x dx
−∞ dx   dx   dx  
+∞
 2 dH 2 dH
m 
=  H m ( x )e − x n
− H n ( x )e − x
 dx dx  −∞
=0
Thus
+∞
∫ e − x H m ( x) H n ( x)dx = 0; m≠n
2
(30)
−∞

which shows that the Hermite polynomials are Orthogonal with respect
to the weight function e − x . Thus if we define the functions
2

φ n ( x) = N n e − x / 2 H n ( x); n = 0, 1, 2, . . .
2
(31)
then eq. (30) assumes the form
+∞
∫ φ m ( x)φ n ( x)dx = 0; m≠n (32)
−∞

70
PHY312 MATHEMATICAL METHODS OF PHYSICS I

3.5 The Integral Representation of the Hermite Polynomials

The integral representation of the Hermite polynomial is given by

2 n ( −i ) n +∞
H n ( x) = ∫ t n e −t + 2 ixt
2 2
ex dt (33)
π −∞

In order to prove the above relation we start with the relation

1 +∞
e−x = ∫ e −t + 2 ixt
2 2 2
ex dt
π −∞

which can easily be obtained from the well known formula

+∞
−αt + βt π β 2 
∫−∞ =
2
e dt exp  
α  4α 
by assuming α = 1 and β = 2ix . Now according to the Rodrigues formula
d n −x
H n ( x) = (−1) n e x
2 2
e
dx n
1 d n + ∞ −t + 2ixt
π dx ∫−∞
= (−1) n e x
2 2

n
e dt

1 x +∞
= (−1) n e ∫ (2i ) n t n e −t + 2ixt dt
2 2

π −∞

from which eq. (50) readily follows.

3.6 Fourier Transform of Hermite-Gauss Functions

In this section we will show that

e−x
2
/2
H n ( x) =
1
i 2π
n ∫
−∞
+∞
[e k2 / 2
H n (k ) e ikx dk] (34)

Implying that the Fourier transform of the Hermite-Gauss function is a

Hermite-Guass function. In order to prove eq. (34) we start with the
generating function

∞
1
G ( x, t ) = e 2 kt −t = ∑
2
H n ( k )t n
n = 0 ,1, . . . n!

 1 
We multiply the above by  ikx − k 2  and integrate over k to obtain
 2 
∞
+∞  1 2  t n +∞
e −t ∫−∞  2
− + + = ∑ ∫ H n ( x)e − k / 2 e ikx dk
2 2
e k ( 2t ix ) k  dk (35)
−∞
n = 0 n!

71
PHY312 MATHEMATICAL METHODS OF PHYSICS I

Now
 (2t + ix ) 2 
LHS = e −t 2π exp
2


 2 
= 2π e t + 2ixt
e−x
2 2
/2

H n ( x)
= 2π e − x ∑
2
/2
(it ) n
n n!

Comparing coefficients of t n on both sides of eq. (35), we get eq. (34).

3.7 Some Important Formulae Involving Hermite

Polynomials
n
n!
H n ( x + y ) = 2 −n / 2 ∑ H n − p ( x 2 ) H p ( y 2) (36)
p p!(n − p)!
nπ 
n/2
 2n  
H n ( x) → ( 2 n + 1) x −
2
2   e x / 2 cos  (37)
n →∞
  e  2 
(2s)! s
H 2 s − 2 n ( x) 
x 2s = 2s ∑ 
2 n =0,1, ... n!(2s − 2n)! 
 s = 0, 1, 2, . . . (38)
(2s + 1)! s H 2 s +1− 2 n ( x) 
x 2 s +1
=
2 2 s +1 ∑
n = 0 , 1, . . . n!(2 s + 1 − 2n )!


SELF-ASSESSMENT EXERCISE

Using the generating function for H n (x) , show that

∞
1 1
(a) cosh 2 x = ∑ H 2 n ( x)
e n = 0 , 1, . . . ( 2 n)!
∞
1 1
(b) sinh 2 x = ∑ H 2 n+1 ( x)
e n = 0, 1, . . . ( 2n + 1)!
∞
1
(c) e cos 2 x = ∑
n = 0 , 1, . . .
(−1) n
(2n)!
H 2 n ( x)
∞
1
(d) e sin 2 x = ∑
n = 0 , 1, . . .
( −1) n
( 2n + 1)!
H 2 n +1 ( x)

Hint: To obtain (a) and (b) substitute t = 1 and t = -1in eq. (9) add and
subtract the resulting equations. Similarly for (c) and (d), substitute t = I
and equate real and imaginary parts.

Prove that
+∞ (2n)! 2
∫ e − x H 2 n (α , x)dx = n (α − 1) n
2

−∞ n!

72
PHY312 MATHEMATICAL METHODS OF PHYSICS I

Hint: Replace x by αy in eq.(9), multiply the resulting equation

by e − y and integrate with respect to y.
2

4.0 CONCLUSION

Here, in this unit, we have dealt with the Hermite polynomials which are
Orthogonal with respect to the weight function e − x . We have also
2

established that the Fourier transform of the Hermite-Gauss function is a

Hermite-Guass function.

5.0 SUMMARY

This unit was on the Hermite polynomials. It has a lot of application in

linear harmonic oscillator problem in quantum mechanics. The unit will
be of immense importance in the subsequent course in classical
mechanics.

6.0 TUTOR- MARKED ASSIGNMENT

1. If two operators are defined as

1  d 
a= x + 
2 dx 
1  d 
a= x− 
2 dx 
Show that
aφ n ( x) = nφ n −1 ( x)
a φ n ( x) = nφ n −1 ( x)

2. Prove that
+∞ 1  2
H n  x + x0 e − x / 2 dx = n x 0n
∫
−∞
 2 
  1  
2

Hint: Multiply Eq. (9) by −  x + x0   and integrate over x.

  2  

7.0 REFERENCES/FURTHER READING

Erwin, K. (1991). Advanced Engineering Mathematics. John Wiley &

Sons, Inc.

Arfken, G. (1990). Mathematical Methods for Physicists. New York:

Academic Press.

73
PHY312 MATHEMATICAL METHODS OF PHYSICS I

UNIT 2 LAGUERRE POLYNOMIALS

CONTENTS

1.0 Introduction
2.0 Objectives
3.0 Main Content
3.1 Laguerre Differential Equation
3.2 The Generating Function
3.3 Rodrigues Formula
3.4 Orthogonality of Hermite Polynomials
3.5 The Integral Representation of the Laguerre Polynomials
3.6 Some Important Results Involving Laguerre Polynomials
3.7 The Second Solution of the Laguerre Differential Equation
3.8 Associated Laguerre Polynomials
4.0 Conclusion
5.0 Summary
6.0 Tutor-Marked Assignment
7.0 References/Further Reading

1.0 INTRODUCTION

In the previous unit, you came across solutions of orthogonal set of

functions. This unit which is the last one in this book will examine
critically how a Laguerre differential equation can be transformed to
Sturm-Liouville form.

It shows that Laguerre polynomials and the associated functions arise in

many branches of physics, e.g. in the hydrogen atom problem in
quantum mechanics, in optical fibers characterised by parabolic
variation of refractive index, etc.

We also show that Laguerre polynomials are orthogonal in the interval

0 ≤ x ≤ ∞ with respect to the weight function e − x .

2.0 OBJECTIVES

At the end of this unit, you should be able to:

• use Frobenius method to obtain the polynomial solution of the

Laguerre differential equations
• determine the Orthogonality of the Laguerre polynomials
• derive the Rodrigues formula
• derive the second solution of the Laguerre differential equation.

74
PHY312 MATHEMATICAL METHODS OF PHYSICS I

3.0 MAIN CONTENT

3.1 Laguerre Differential Equation

The equation
xy ′′( x) − (1 − x) y ′( x) + ny ( x ) = 0
(1)
where n is a constant known as the Laguerre differential equation.
When n = 0, 1, 2, . . . (2)

One of the solutions of eq. (1) becomes a polynomial. These polynomial

solutions are known as the Laguerre polynomials.

Using Frobenius method to solve eq.(1), and following the various steps,
we have

Step We substitute the power series

∞
y ( x) = ∑ C r x p + r , C0 ≠ 0
r =0

Eq. (1) and obtain the identity

∞ ∞

∑C
r =0
r ( p + r ) 2 x p + r −1 − ∑ C r ( p + r − n) x p + r = 0
r =0

or
[ ]
∞
C 0 p 2 x p −1 − ∑ C r ( p + r ) 2 − C r −1 ( p + r − n − 1) x p + r −1 = 0 (3)
r =1

Step 2 Equating to zero the coefficients of various powers of x in the

identity (3), we obtain

(i) p2 = 0 INDICIAL EQUATION (4)

p + r − n −1
(ii) Cr = C r −1 r ≥ 1 RECURRENCE RELATION (5)
( p + r)2
Substituting p = 0 in eq. (5), we get
r − n −1
Cr = C r −1 r ≥1
r2
which gives
n n(n − 1)
C1 = − C0 C2 = C0
(1!) 2 (2!) 2
n(n − 1)(n − 2)
C3 = C0 etc
(3!) 2
n! ( −1) n
C n = ( −1) n =
( n!) 2 n!
and
C n+1 = C n + 2 = 0 . . . . = 0

75
PHY312 MATHEMATICAL METHODS OF PHYSICS I

Therefore one of the solutions of eq. (1) can be written as

 n n(n − 1) 2 n x 
n
y ( x) = C 0 1 − 2
x + x − . . . + ( − 1)  (6)
 (1!) (2!) 2 n! 
which is a polynomial of degree n. If the multiplication constant C 0 is
chosen to be unity so that the constant term becomes unity, the
polynomial solution given by eq. (6) is known as Laguerre Polynomial
of degree n and denoted by Ln (x) . Thus
n n( n − 1) 2 n x
n
Ln ( x) = 1 − x + x − . . . + ( −1)
(1!) 2 ( 2!) 2 n!
or
n
n!
Ln ( x ) = ∑ (−1) n xr (7)
r =0 (n − r )!(r!) 2
with
Ln (0) = 1 (8)

The first four Laguerre polynomials can be written as:

L0 ( x) = 1,
L1 ( x) = 1 − x,
1 2
L2 ( x) = 1 − 2 x + x , (9)
2
3 1
L3 ( x) = 1 − 3 x + x 2 − x 3 , . . .
2 6

Higher order polynomials can easily be obtained either by using eq.(7)

or by using the recurrence relation [see eq. (20)].

3.2 The Generating Function

The generating function for Laguerre polynomials is given by

1  xt  ∞
G ( x, t ) = exp −  = ∑ Ln ( x)t ; t <1
n
(10)
1− t  1 − t  n =0

We expand the left hand side of eq. (10) to obtain

 xt 
(1 − t ) −1 exp − 
 1− t 
x 2 t 2 (1 − t ) 3
= (1 − t ) −1 − xt (1 − t ) − 2 + − ..
2!
x 2t 2
= (1 + t + t 2 + . . .) − xt (1 + 2t + 3t 2 + . . .) + (1 + 3t + 6t 2 + . . .) − . . . (11)
2!

76
PHY312 MATHEMATICAL METHODS OF PHYSICS I

The right hand side of eq.(11) can be written as a power series in t with
Coefficient of t 0 = 1 = L0 ( x)
“ “ t =1− x = L1 ( x)
“ “ t = 1 − 2x + x / 2
2 2
= L2 ( x)

etc. It is also evident that the coefficient of t 2 on the right hand side of
eq.(11) will be a polynomial of degree n and that the constant term in
this polynomial will be unity. We can then assume that

1  xt  ∞
G ( x, t ) = exp−  = ∑ K n ( x)t
n
(12)
1− t  (1 − t )  n =0

where K n (x) is a polynomial of degree n. Differentiating eq. (12) with

respect to t, we get

(1 − x − t )  xt  ∞
 = ∑ nK n ( x)t
n −1
exp  −
(1 − t ) 3  (1 − t )  n = 0

or
∞ ∞
(1 − x − t )∑ K n ( x)t n = (1 − 2t + t 2 )∑ nK n ( x)t n−1
n =0 n =1

Comparing the coefficients of t n on both sides of the above equation,

we get
( n + 1) K n+1 ( x) − ( 2n + 1 − x) K n ( x ) + nK n−1 ( x) = 0; n ≥1 (13)

We next differentiate eq.(12) with respect to x to obtain

∞ ∞
− t ∑ K n ( x)t n = (1 − t )∑ K n′ ( x)t n (14)
n =0 n =0

Comparing the coefficients of t n on both sides of the above equation,

we get
K n′ ( x ) − K n′ −1 ( x) = − K n−1 ( x) (15)

If we replace n by (n+1) in the above equation, we would get

K n′ +1 ( x) = K n′ ( x) − K n ( x) (16)

Differentiating eq. (13) with respect to x, we obtain

(n + 1) K n′ +1 ( x) − (2n + 1 − x ) K n′ ( x ) + K n ( x) + nK n′ −1 ( x ) = 0 (17)

Substituting K n′ −1 ( x) and K n′ +1 ( x) from eqs. (15) and (16) respectively in

eq. (17), we get
xK n′ ( x) = nK n ( x) − nK n−1 ( x) (18)

77
PHY312 MATHEMATICAL METHODS OF PHYSICS I

Differentiating the above equation with respect to x and using eq. (15),
we have
xK n′′ ( x) + K n′ ( x ) = −nK n −1 ( x) (19)

Subtracting eq. (18) from eq. (19), we get

xK n′′( x) + (1 − x) K n′ ( x) + nK n ( x) = 0

Showing that K n (x) is a solution of the Laguerre equation, i.e. of the

equation
xy ′′( x) − (1 − x) y ′( x ) + ny ( x) = 0

Hence K n (x) nothing but Ln (x) . Equations (13) and (18) give the
following recurrence relations respectively:
( n + 1) Ln+1 ( x) = ( 2n + 1 − 1) Ln ( x) − nLn−1 ( x) (20)
xLn′ ( x) = nLn ( x) − nLn−1 ( x) (21)
We also have
Ln ( x ) = Ln′ ( x) − Ln′ +1 ( x) (22)

3.3 Rodrigues Formula

In the preceding section we have shown that

1  xt  ∞
G ( x, t ) = exp −  = ∑ Ln ( x)t
n

1− t  1 − t  n =0

We can write the above equation as

∞
1  x(1 − t − 1) 
∑ L ( x)t
n =0
n
n
=
1− t
exp −
 1 − t 
or
∞
 1  x 
∑ L ( x)t n
= ex  exp −  (23)
1 − t  1 − t 
n
n =0

Differentiating eq. (23) n times with respect to t and then putting t = 0,

we will have
 ∂n  1  x  
n! Ln ( x) = e x  n  exp −  
 ∂t 1 − t  1 − t    t =0
 ∂ n  ∞ (−1) r x r 
= e x  n ∑ r +1 
 ∂t  r =0 (1 − t ) r! t =0

78
PHY312 MATHEMATICAL METHODS OF PHYSICS I

∞ (r + 1)(r + 2) . . .(r + n) 
= e x ∑ (−1) r 
 r =0 (1 − t ) r + n+1 r!  t =0
∞
(n + r )! r
= e x ∑ (−1) r x
r =0 (r!) 2
or
ex ∞ ( n + r )! 2
Ln ( x) = ∑
n! r =0
( −1) r
( r!) 2
x (24)

d n n −x d2  n ∞ r x 
r
( x e ) =  ∑
x ( − 1) 
dx n dx 2  r =0 r! 
∞
( n + r )( n + r − 1) . . .(r + 1) r
= ∑ ( −1) r x
r =0 r!
∞
(n + r )! r
= ∑ (−1) r x
r =0 (r!) 2
Thus
e x d n n −x
Ln ( x ) = (x e ) (25)
n! dx n

This is known as Rodrigues formula for the Laguerre polynomials. For

example, putting n = 2 in the Rodrigues’ formula, we have

e x d 2 2 −x
L2 ( x ) = (x e )
2! dx 2

ex d
= (2 x 2 e − x − x 2 e − x )
2! dx
ex d
= (2e − x − 4 xe − x + x 2 e − x )
2! dx
x2
= 1 − 2x +
2
Which is consistent with eq. (9). Similarly, we can determine other
Laguerre polynomials by elementary differentiation of the result
expressed by eq. (25).

3.4 Orthogonality of Hermite Polynomials

As Laguerre differential equation is not of the form of Sturm-Liouville

differential equation, its solutions Ln (x) , therefore, do not by themselves
form an Orthogonal set. However, in order to transform Laguerre
differential equation to the Sturm-Liouville form, we may write eq. (1)
as
(1 − x) n
y ′′( x) − y ′( x) + y ( x) = 0
x x

79
PHY312 MATHEMATICAL METHODS OF PHYSICS I

Multiplying the above equation by

 1− x 
p ( x) = exp  ∫ = xe − x
2

 (26)
 x 

We obtain
d  dy 
 p ( x)  + ne − x y ( x) = 0 (27)
dx  dx 

Thus for Laguerre polynomials, the Sturm-Liouville form is given by

d  dL ( x ) 
 p ( x) n  = − ne − x Ln ( x) (28)
dx  dx 

Similarly
d  dL ( x) 
 p ( x) m  = −me − x Lm ( x) (29)
dx  dx 

Multiply eq.(28) by Lm (x) and eq.(29) by Ln (x) and subtracting the

resulting equations, we obtain
d  dL ( x)  d  dL ( x ) 
Lm ( x)  p ( x ) n  − Lm ( x )  p ( x) m 
dx  dx  dx  dx 
= (m − n) Lm ( x) Ln ( x) (30)

The left hand side of eq.(30) is simply

d  dL ( x) dL ( x) 
 Lm ( x) p ( x) n − Lm ( x) p ( x) m  (31)
dx  dx dx 

Integrating eq.(30) and using eq.(31), we get

∞
∞   dL ( x) dL ( x) 
(m − n) ∫ e − x Lm ( x ) Ln ( x)dx =  p ( x) Lm ( x) n − Lm ( x) m 
0
  dx dx  0

Since p(x) = 0 at x = 0 and at x = ∞ , the right hand side vanishes and we

readily obtain
∞
∫ e − x Lm ( x) Ln ( x)dx = 0 for m ≠ n (32)
0

The above equation shows that the Laguerre polynomials are Orthogonal
in the interval 0 ≤ x ≤ ∞ with respect to the weight function e − x . We now
define the functions
φ n ( x) = N n Ln ( x)e − x / 2
2
(33)

The constant N n is chosen so that the functions φn (x) are normalised, i.e.
+∞
∫ φn2 ( x)dx = 1 for m=n (32)
0

80
PHY312 MATHEMATICAL METHODS OF PHYSICS I

3.5 The Integral Representation of the Laguerre Polynomials

The integral representation of the Laguerre polynomial is given by

Ln ( x ) =
e x + ∞ −t n
n! ∫0
[
e t dt J 0 2( xt )1 / 2 dt ] (33)

In order to prove the above relation we start with the relation

∫0
+∞
[ ]
e −t t n dt J 0 2( xt )1 / 2 dt
+∞ ∞
( −1) r (tx) r
= ∫ e −t t n ∑ dt
0
r =0 ( r!) 2
∞
(−1) r ( x) r + ∞ −t n + r
=∑
(r!) 2 ∫0
e t dt
r =0
∞
( −1) r x r Γ( n + r + 1)
=∑
r =0 ( r!) 2
∞
(−1) r (n + r )! r
=∑ x (34)
r =0 (r!) 2

Using eqs. (24) and (33), we get

∫0
+∞
[ ]
e −t t n dt J 0 2( xt )1 / 2 dt = e − x n! Ln ( x) (35)
from which eq. (33) readily follows.

3.6 Some Important Results Involving Laguerre Polynomials

We give some important results involving Laguerre polynomials which

can be readily derived:

x
∫0
Ln ( x)dx = Ln ( x ) − Ln+1 ( x) [Use Eq. (22)] (36)

[ ]
∞
y n Ln ( x)
∑
n =0 n!
= e y J 0 2( xy )1 / 2 (37)

+∞
m −x 0 if m < n
∫0 x e Ln ( x)dx = (−1) n n! if m = n (38)
N
( N + 1)
∑ Ln ( x) Ln ( y ) = [LN ( x) LN +1 ( y ) − LN +1 ( x) LN ( y)] (39)
n =0 x− y 0
from which eq. (50) readily follows.

3.7 The Second Solution of the Laguerre Differential

Equation

Since the indicial equation [eq. (4)] has two equal roots, the two
independent solutions of eq. (1)

81
PHY312 MATHEMATICAL METHODS OF PHYSICS I

 ∂y 
( y ) p =0 and  
 ∂p  p =0
Now
 p−n ( p − n)( p − n + 1) 2
y ( x, p ) = x p 1 + x+ x +
 ( p + 1) ( p + 1) 2 ( p + 2) 2
2

( p − n)( p − n + 1)( p − n + 2) 3 
x + . . . (40)
( p + 1) ( p + 2) ( p + 3)
2 2 3

Thus,
y1 ( x) = y ( x, p = 0)
n(n − 1) 2 n( n − 1)( n − 2) 3
= 1 − nx + x − x .. (41)
(2!) 2 (3!) 2
and
∂y  p  p−n ( p − n)( p − n + 1) 2 
y 2 ( x) = =  x ln x 1 + x+ x + . . .
∂p   ( p + 1)
2
( p + 1) ( p + 2)
2 2


 1 2  p−n  1 2 2 2 
+ x p  −  x +  − − − 
 p − n p + 1  ( p + 1)  p − n p − n +1 p +1 p + 2 
2

 ( p − n)( p − n + 1) 2 
× x + . . .
 ( p + 1) ( p + 2)
2 2
 p = 0
 3n 2 − n − 1 2 
= y1 ( x) ln x + (2n + 1) x − 2
x + . . . (42)
 (2!) 

For example, for n = 0

y1 ( x) = 1 = L0 ( x)
and

x2 2! x 3 3! x 4
y 2 ( x) = ln x + x + + + +. . . (43)
( 2!) 2 (3!) 2 ( 4!) 2
Similarly, for n = 1
y1 ( x) = 1 − x = L1 ( x)
and

x2 x3
y 2 ( x) = (1 − x) ln x + 3 x − + −. . . (44)
(2!) 2 (3!) 2

3.8 Associated Laguerre Polynomials

Replace n by (n + k ) in eq. (1), it is obvious that Ln +k (x) will be a

solution of the following differential equation.

82
PHY312 MATHEMATICAL METHODS OF PHYSICS I

xy ′′ − (1 − x) y ′ + (n + k ) y = 0 (45)

Differentiating the above equation k times, it can easily be shown that

dk
y = k [Ln+ k (x)] (46)
dx
or a constant multiple of it is a solution of the differential equation
xy ′′ − ( k + 1 − x) y ′ + ny = 0 (47)

Where n and k are positive integers or zero. The above equation is

known as the Associated Laguerre Equation. Its polynomial solutions
[see eq.(45)] are denoted by Lkn (x) and are defined by

dk
Ln ( x ) = (−1) n [Ln+k ( x) ] (48)
dx k

This is known as the Associated Laguerre Polynomials. It is obvious

from eq. (48) that
Lkn (x) is polynomial of degree n in x and that
L0n ( x) = Ln ( x) (49)

Using eqs. (7) and (48), it follows that

n
(n + k )!
Lkn ( x) = ∑ ( −1) r xr (50)
r =0 ( n − r )!( r + k )!r!

We will define Lkn (x) for non-integer values of k, we may, therefore,

write the above equation as
n
Γ(n + k + 1)
Lkn ( x) = ∑ (−1) r xr (51)
r =0 (n − r )!Γ(r + k ) r!

Using the above equation, the first three polynomials can easily be
written as:
Lk0 ( x ) = 1
L1k ( x) = k + 1 − x (52)
1 1
Lk2 ( x ) = (k + 2)(k + 1) − (k + 2) x + x 2
2 2
Differentiating the Laguerre generating function [eq. (10)] k times with
respect to x, one can easily obtain the generating function for the
associated Laguerre polynomials. Thus
1  xt  ∞ k
g ( x, t ) ≡ − ∑ Ln ( x)t
n
k +1
exp (53)
(1 − t )  1 − t  n =0

Furthermore, from eq.(51)

83
PHY312 MATHEMATICAL METHODS OF PHYSICS I

Γ(n + k + 1)
Lkn (0) = (54)
n!Γ(k + 1)

SELF-ASSESSMENT EXERCISE

1. Show that
4
(−1) r Lkr ( x)
x 4 = Γ(5 + k ) 4!∑
r = 0 Γ ( r + k + 1)( 4 − r )!

2. Hint: Use eq. (51 Show that

Ln (0) = 1
Ln′ (0) = − n
1
Ln′′(0) = n(n − 1)
2
Hint: Use Eq. (7).

4.0 CONCLUSION

In this unit, we have established the relationship between Laguerre and

associated Laguerre polynomials. The generating function and some
important results involving Laguerre polynomials were also dealt with.

5.0 SUMMARY

This unit deals with Laguerre functions and its applications to physical
problems especially in Quantum mechanics.

6.0 TUTOR-MARKED ASSIGNMENT

1. Show that
n
Lγn+ k +1 ( x + y ) = ∑ Lγr ( x ) Lkn − r ( y ), n = 0, 1, 2, . . .
r =0

Hint: Use the generating function.

2. Show that
(−1) n H 2 n +1 ( x1 / 2 )
L1n/ 2 ( x) =
2 2 n+1 n! x1 / 2
(−1) n
L−n1 / 2 ( x ) = 2 n H 2 n ( x1 / 2 )
2 n!

Hint: Use the integral representation of Lkn (x) and H n (x) .

3. Using eq. (53), prove the identity

84
PHY312 MATHEMATICAL METHODS OF PHYSICS I

∂g
(1 − t ) + [x − (1 − t )(1 + k )]g = 0
∂t
and then derive the recurrence relation [eq. (56)].

4. Using eq.(53), prove the identity

∂g
(1 − t ) + tg ( x, t ) = 0
∂x
and hence derive the following relation
dLkn ( x ) dLkn−1 ( x)
− + dLkn −1 ( x) = 0
dx dx
n = 1, 2, . . .
5. Show that
x
∫0
Ln (t )dt = Ln ( x) − Ln+1 ( x)
Hint: Use the relation derived in problem 4.

7.0 REFERENCES/FURTHER READING

Erwin, K. (1991). Advanced Engineering Mathematics. John Wiley &

Sons, Inc.

Arfken, G. (1990). Mathematical Methods for Physicists. New York:

Academic Press.

85
PHY312 MATHEMATICAL METHODS OF PHYSICS I

MODULE 1 PARTIAL DIFFERENTIAL EQUATIONS

WITH APPLICATIONS IN PHYSICS

Unit 1 Partial Differential Equations

Unit 2 Fourier Series

UNIT 1 PARTIAL DIFFERENTIAL EQUATIONS

CONTENTS

8.0 Introduction
9.0 Objectives
10.0 Main Content
10.1 Definition
10.2 Linear Second-Order Partial Differential Equations
3.2.1 Laplace‘s Equation
3.2.2 Wave Equation
3.2.3 Heat Conduction Equation
3.2.4 Poisson’s Equation
10.3 Method of Separation of Variables
10.3.1 Application to Wave Equation
3.3.2 Application to Heat Conduction Equation
10.4 Laplace Transform Solutions of Boundary-Value
Problems
11.0 Conclusion
12.0 Summary
13.0 Tutor -Marked Assignment
14.0 References/Further Reading

7.0 INTRODUCTION

In this unit, we shall study some elementary methods of solving partial

differential equations which occur frequently in physics and in
engineering. In general, the solution of the partial differential equation
presents a much more difficult problem than the solution of ordinary
differential equations.

We are therefore going to limit ourselves to a few solvable partial

differential equations that are of physical interest.

1
PHY312 MATHEMATICAL METHODS OF PHYSICS I

8.0 OBJECTIVES

At the end of this unit, you should be able to:

• define linear second-order partial differential equation in more

than one independent variable
• use the technique of separation of variables in solving important
second order linear partial differential equations in physics
• solve the exercises at the end of this unit.

9.0 MAIN CONTENT

3.9 Definition

An equation involving one or more partial derivatives of (unknown)

functions of two or more independent variables is called a partial
differential equation. The order of a PDE is the highest order partial
derivative or derivatives which appear in the equation. For example,

∂U ∂ 3U  ∂ 2U ∂ 2U 
U +  2  = ez
2 
(1)
∂z ∂y 3
 ∂y ∂z 
is a third order PDE since the highest order term is given by

∂ 3U
∂y 3

A PDE is said to be linear if it is of the first degree, i.e. not having

exponent greater than 1 in the dependent variable or its partial
derivatives and does not contain product of such terms in the equation.
Partial derivatives with respect to an independent variable are written for
brevity as a subscript; thus

∂ 2U ∂ 2U
U tt = and U =
∂t 2 ∂x∂y
xy

The PDE

1
U tt = U xx + U yy + U zz (2)
c2

(Where c is a constant) is linear and is of the second order while eq. (1)
is an example of a nonlinear PDE.

2
PHY312 MATHEMATICAL METHODS OF PHYSICS I

Example 1: Important linear partial differential equations of

second order

(1)
∂ 2u 2 ∂ u
2
= c One − dim ensional wave equation
∂t 2 ∂x 2
(2)
∂u ∂ 2u
= c2 2 One − dim ensional heat equation
∂t ∂x
(3)
∂ 2u ∂ 2u
+ =0 Two − dim ensional Laplace equation
∂x 2 ∂y 2
(4)
∂ 2u ∂ 2u
+ = f ( x, y ) Two − dim ensional poisson equation
∂x 2 ∂y 2
(5)
∂ 2u ∂ 2u ∂ 2u
+ + =0 Three − dim ensional Laplace equation
∂x 2 ∂y 2 ∂z 2

3.10 Linear Second-Order Partial Differential Equations

Many important PDEs occurring in science and engineering are second

order linear PDEs. A general form of a second order linear PDE in two
independent variables x and y can be expressed as

∂ 2u ∂ 2u ∂ 2u ∂u ∂u
A + B + C +D +E + Fu = G (3)
∂x 2
∂x∂y ∂y 2
∂x ∂y

where A, B, C……,G may be dependent on variables x and y. If G = 0 ,

then eq. (3) is called homogeneous; otherwise it is said to be a non-
homogeneous.

The homogeneous form of Eq. (3) resembles the equation of a general

conic:

ax 2 + bxy + cy 2 + dx + ey + f = 0

We thus say that eq. (3) is of

elliptic   B 2 − 4 AC < 0
 
hyperbolic type when  B 2 − 4 AC > 0
parabolic   B 2 − 4 AC = 0


3
PHY312 MATHEMATICAL METHODS OF PHYSICS I

For example, according to this classification the two-dimensional

Laplace equation

∂ 2u ∂ 2u
+ = 0
∂x 2 ∂y 2

is of elliptic type ( A = C = 1, B = D = E = G = 0), and the equation

∂ 2u ∂ 2u
−α 2
= 0 ( α is a real constant)
∂x 2 ∂y 2

is of hyperbolic type. Similarly, the equation

∂ 2u ∂ u
− α = 0 ( α is a real constant)
∂ x 2 ∂ y

is of parabola type.

Some important linear second-order partial differential equations that

are of physical interest are listed below.

Example 2

Eliminate A and P from the function Z = Ae pt sin px

∂Z ∂2Z
Solution Let = pAe pt sin px and = p 2 Ae pt sin px
∂t ∂t 2

∂Z ∂2Z
also = pAe pt cos px and = − p 2 Ae pt sin px
∂x ∂x 2
∂2Z ∂2Z
+ =0
∂t 2 ∂x 2
i.e. p 2 Ae pt sin px − p 2 Ae pt sin px = 0

Example 3

Solve the equation

∂ 2u ∂ 2u ∂ 2u
− 7 + 6 =0
∂x 2 ∂x∂y ∂y 2
Solution: Let u ( x, y ) = f ( y + m1 x) + g ( y + m2 x)

So that m 2 − 7m + 6 = 0

This implies that m = 1 or 6

Hence u ( x, y ) = H ( y + x ) + G ( y + 6 x )

4
PHY312 MATHEMATICAL METHODS OF PHYSICS I

3.2.1 Laplace‘s Equation

∇ 2u = 0 (4)
 ∂ 2
∂ 2
∂ 
2
Where ∇ 2 is the Laplacian operator  ∇ 2 = 2 + 2 + 2  . The
 ∂x ∂y ∂y 
function u may be the electrostatic potential in a charge-free region or
gravitational potential in a region containing no matter.

3.2.2 Wave Equation

1 ∂ 2u
∇ 2
u = (5)
v 2 ∂ t 2
Where u represents the displacement associated with the wave and v, the
velocity of the wave.

3.4.3 Heat Conduction Equation

∂u
= α∇ 2 u (6)
∂t
Where u is the temperature in a solid at time t. The constant α is called
the diffusivity and is related to the thermal conductivity, the specific
heat capacity, and the mass density of the object.

3.4.4 Poisson’s Equation

∇ 2 u = ρ ( x, y , z ) (7)

Where the function ρ ( x, y, z ) is called the source density. For example, if

u represents the electrostatic potential in a region containing charges,
then ρ is proportional to the electric charge density.

Example 4

Laplace’s equation arises in almost all branches of analysis. A simple

example can be found from the motion of an incompressible fluid. Its
velocity v ( x, y, z , t ) and the fluid density ρ ( x, y, z , t ) must satisfy the
equation of continuity:

∂ρ
+ ∇.( ρv ) = 0
∂t
If ρ is constant we then have

∇•v = 0

5
PHY312 MATHEMATICAL METHODS OF PHYSICS I

If furthermore, the motion is irrotational, the velocity vector can be

expressed as the gradient of a scalar function V:

v = −∇V
and the continuity becomes Laplace’s equation:

∇.v = ∇.(−∇V ) = 0, or ∇ 2V = 0
The scalar function V is called the velocity potential

3.5 Method of Separation of Variables

The technique of separation of variables is widely used for solving many

of the important second order linear PDEs.
The basic approach of this method in attempting to solve a differential
equation (say, two independent variables x and y) is to write the
dependent variable u ( x, y ) as a product of functions of the separate
variables u ( x, t ) = X ( x) T (t ) . In many cases the partial differential
equation reduces to ordinary equations for X and T.

3.3.1 Application to Wave Equation

Let us consider the vibration of an elastic string governed by the one-

dimensional wave equation
∂ 2u 2 ∂ u
2
= c (8)
∂t 2 ∂x 2
where u(x, y) is the deflection of the string. Since the string is fixed at
the ends x = 0 and x = l , we have the two boundary conditions

u (0, t ) = 0, u (l , t ) = 0 for all t (9)

The form of the motion of the string will depend on the initial deflection
(deflection at t = 0) and on the initial velocity (velocity at t = 0).
Denoting the initial deflection by f(x) and the initial velocity by g(x), the
two initial conditions are

∂u
u ( x, 0) = f ( x) = g ( x) (10)
∂t t =0

This method expresses the solution of u ( x, t ) as the product of two

functions with their variables separated, i.e.

U ( x, t ) = X ( x) T (t ) (11)

where X and T are functions of x and t respectively.

6
PHY312 MATHEMATICAL METHODS OF PHYSICS I

Substituting eq. (11) in eq. (8), we obtain

X T ′′ = c 2 X ′′ T
or
X ′′( x) 1 T ′′(t )
= (12)
X ( x) c 2 T (t )

In other words

X ′′ 1 T ′′
= 2 =λ (13)
X c T

The original PDE is then separated into two ODEs, viz.

X ′′( x) − λX ( x ) = 0 (14)
and
T ′′(t ) − λc 2 T (t ) = 0 (15)

The boundary conditions given by eq. (9) imply

X (0) T (t ) = 0
and
X (l ) T (t ) = 0

Since T(t) is not identically zero, the following conditions are satisfied

X ( 0) = 0 and X (l ) = 0 (16)

Thus eq. (14) is to be solved subject to conditions given by eq. (16).

There are 3 cases to be considered.

Case 1 λ >0

The solution to eq. (14) yields

λx λx
X ( x ) = Ae − + Be (17)

To satisfy the boundary condition given by eq. (16), we must have

λl λl
Ae − + Be =0

Since the determinant formed by the coefficients of A and B is non-zero,

the only solution is A = B = 0. This yields the trivial solution X(x) = 0.

7
PHY312 MATHEMATICAL METHODS OF PHYSICS I

Case 2 λ =0

The solution to eq. (14) yields

X ( x) = A + Bx

To satisfy the boundary condition given by eq. (16), we must have

A=0
and
A + Bl = 0
implying
A = 0, B=0

Again for this case, a trivial solution is obtained

Case 3 λ <0

Let λ = − k 2 . The solution to eq. (14) yields

X ( x ) = A cos kx + B sin kx (18)

To satisfy the boundary condition given by eq. (16), we must have

A=0
and
B sin kl = 0
To obtain a solution where B ≠ 0 , we must have
kl = nπ n = 1, 2,...
Thus
 nπ 
2

λ =− k2 =−  (19)
 l 
(n=0 corresponds to the trivial solution). The specific values of λ are
known as the eigenvalues of eq. (14) and the corresponding solutions,
 nπ 
viz, sin  x  are called the eigenfunctions. Since there are many
 l 
possible solutions, each is subscripted by n. Thus

 nπ 
X n ( x) = Bn sin  x n =1, 2, 3, ... (20)
 l 

The solution to Eq. (15) with λ given by Eq. (19) is

 nπ   nπ 
Tn (t ) = E n cos ct  + Fn sin  ct  n =1, 2, 3, ... (21)
 l   l 
8
PHY312 MATHEMATICAL METHODS OF PHYSICS I

Where En and Fn are arbitrary constants. There are thus many solutions
for eq. (8) which is given by

U n ( x, t ) = X n ( x ) Tn (t )
  nπ   nπ  nπ
a n cos l ct  + bn sin  l ct  sin l x (22)
    

Where an = Bn En and bn = Bn Fn . Since eq. (8) is linear and homogeneous,

the general solution is obtained as the linear superposition of all the
solutions given by eq. (22), i.e.

∞
 nπc nπc  nπ
U ( x, t ) = ∑  a n cos t + bn sin t  sin x (23)
n =1  l l  l

Differentiating with respect to t, we have

∞
nπc  nπc nπc  nπ
U t ( x, t ) = ∑  − a n sin t + bn cos t  sin x (24)
n =1 l  l l  l

The coefficients an and bn are obtained by applying the initial

conditions in eq. (10). Thus,

∞
nπ
U ( x,0) = f ( x) = ∑ a n sin x (25)
n =1 l
∞
 nπ  nπ
U ( x,0) = g ( x) = ∑ bn  c  sin x (26)
n =1  l  l

In order to determine a n and bn we use the orthogonality properties of

nπ
sin x in the range 0 ≤ x ≤ l , i.e.
l
mπ nπ
l
l
∫ sin x sin x dx = δ m n (27)
0
l l 2

Where δ m n is the Kronecker delta function having the property

0 n≠m
δ m n =  (28)
1 n =m

mπ
Multiply eq. ((25) by sin x and integrating between the limits x = 0
l
and x = l, we get

9
PHY312 MATHEMATICAL METHODS OF PHYSICS I

mπ ∞
mπ nπ
l l

∫ f ( x) sin x dx = ∑ ∫ a n sin x sin x dx

0
l n =1 0 l l
l
= am (29)
2
mπ
l
2
i.e. am = ∫
l 0
f ( x) sin
l
x dx

mπ
Similarly multiplying eq. (26) by sin x and integrating between the
l
limits x = 0 and x = l , we get
mπ ∞
 nπ  nπ mπ
l l

∫0 g ( x ) sin
l
x dx = ∑ ∫
n =1 0
bn  c  sin
 l  l
x sin
l
x dx

 mπ  l
= bm  c (30)
 l 2
mπ
l
2
i.e. bm =
mπc 0∫ g ( x) sin
l
x dx

With a m and bm obtained for m =1, ... ∞, eq. (23) is the solution to PDE
given by eq. (8) subject to the initial conditions and the boundary
conditions.

3.3.2 Application to Heat Conduction Equation

The one-dimensional heat flow in a rod bounded by the planes x = 0 and

x = a is of practical interest. The solution applies to the case where the y
and z dimensions extend to infinity. The temperature distribution is
determined by solving the one-dimensional heat conduction equation

∂ 2θ 1 ∂θ
= (31)
∂x 2 v ∂t

Where θ represents the temperature and

k
v= (32)
Cρ

k, C and ρ are the thermal conductivity, specific heat and density of the
material respectively. We shall treat the case where the boundary
conditions are given by

θ ( x = 0, t ) = 0 (33)

10
PHY312 MATHEMATICAL METHODS OF PHYSICS I

θ ( x = a, t ) = 0 (34)

The initial temperature distribution is given by

θ ( x , t = 0) = f ( x ) (35)

Solution: Using the method of separation of variables, the x-dependence

and t-dependence are separated out as expressed by

θ ( x, t ) = X ( x ) T (t ) (36)

Substituting eq. (36) into eq. (31) yields

1 d 2 X 1 1 dT
= =α (37)
X dx 2 v T dt

We shall now consider three cases corresponding to different values of

the constant α .
Case 1 λ =0

The separated ODE for X (x) becomes

d2X
=0 (38)
dx 2
i.e. X ( x) = Ax + B

The boundary conditions expressed by eqs. (33) and (34) are

respectively

X ( x = 0) = 0 and X ( x = a) = 0 (39)

Since T(t) should not be identically zero. Thus for eq. (38) to satisfy the
boundary conditions given by eq. (39), we must have A = 0, B = 0. This
gives the steady-state solution where temperature in the rod is
everywhere zero.

Case 2 λ >0

Let α = k 2 . The ODE for X becomes

d2X
2
= k2X (40)
dx
Therefore X ( x ) = Ae kx + Be − kx

Applying the boundary conditions given in eq. (39), we get

11
PHY312 MATHEMATICAL METHODS OF PHYSICS I

0=A+B
0 = Ae ka + Be − ka

Again we have A = B = 0

Case 3 λ <0

Let α = −λ2 . The ODE for X (x) becomes

d2X
= − λ2 X (41)
dx 2

Thus X ( x) = A cos λx + B sin λx

The boundary conditions require

A=0
B sin λa = 0 (42)
i.e. λ a = nπ n =1, 2, ... (43)

Since there are multiple solutions, each λ is designated by a subscript n

as λn . The solution of the ODE for T(t) is readily obtained as

T (t ) = Ce − λnvt
2
(44)

Thus the general solution which is a superposition of all admissible

solution is given by

∞
nπ
θ ( x, t ) = ∑ Dn e −λ vt sin
2
n
x (45)
n =1 a
∞
 n 2 π 2  nπ
= ∑ Dn exp − 2 vt  sin x (46)
n =1  a  a

To complete the solution Dn must be determined from the remaining

initial condition

∞
nπ
i.e. f ( x) = ∑ Dn sin x (47)
n =1 a

12
PHY312 MATHEMATICAL METHODS OF PHYSICS I

mπ
In order to determine Dn , we multiply eq. (47) by sin x and integrate
a
the limits x=0 and x = a to obtain

mπ ∞
nπ mπ
a a
a
∫ f ( x) sin x dx = ∑ Dn ∫ a n sin x sin x dx = Dm
0
a n =1 0
a a 2

Thus

mπ
a
2
Dm = ∫ f ( x) sin x dx (48)
a0 a

For the specific case where f ( x) = θ 0 (constant), the solution is given by

4θ 0 ∞
1 (2n + 1)πx  v(2n + 1) 2 π 2 
θ ( x, t ) = ∑
π n =0 2 n + 1
sin
a
exp −
a2

 

0 < x < a (49)

From eq. (49), it can be deduced that a rectangular pulse of height θ 0 for
0 < x < a has the Fourier series expansion given by

4θ 0 ∞
1 ( 2n + 1)πx
π
∑ 2n + 1 sin
n =0 a

Also if f ( x) = γx, then

2 aγ (−1) n−1  nπ 
∞
 vn 2π 2 
θ ( x, t ) = ∑
π n=1 n
sin  x  exp −
 a  a2 
t (50)


If the end boundaries are maintain at different temperature i.e.

θ ( x = 0, t ) = θ1
θ ( x = a, t ) = θ 2 (51)

Then case 1 of the solution where α = 0, would yield the steady-state

solution given by θ1 +
x
(θ 2 − θ1 ) . The general solution is given by
a
θ ( x, t ) = φ ( x, t ) + θ1 + (θ 2 − θ1 )
x
(52)
a

Where φ ( x, t ) is the transient solution.

13
PHY312 MATHEMATICAL METHODS OF PHYSICS I

The boundary conditions for φ ( x, t ) are obtained as follows:

at x = 0: θ ( x = 0, t ) = θ1 = φ ( x = 0, t ) + θ1 ⇒ φ ( x = 0, t ) = 0
at x = a: θ ( x = a, t ) = θ 2 = φ ( x = a, t ) + θ 2 ⇒ φ ( x = a, t ) = 0
φ ( x, t ) is obtained under case 3.

SELF-ASSESSMENT EXERCISE 1

8. State the nature of each of the following equations (that is,

whether elliptic, parabolic or hyperbolic)

∂2 y ∂2 y
(a) + α =0
∂t 2 ∂x 2
∂ 2u ∂ 2u ∂u
(b) x 2 + y 2 + 3y 2
∂x ∂y ∂x

2(a) Show that y ( x, t ) = F (2 x + 5t ) + G (2 x − 5t ) is a general solution of

∂2 y ∂2 y
4 = 25
∂t 2 ∂x 2
(b) Find a particular solution satisfying the conditions
y (0, t ) = y (π , t ) = 0, y ( x,0) = sin 2 x, y ′( x,0) = 0.

3. Solve the following PDEs

∂ 2u
(a) = 8 xy 2 + 1
∂x 2

∂ 2 u ∂u
(b) − = 6 xe x
∂xy ∂y

3.6 Laplace Transform Solutions of Boundary-Value

Problems

Laplace and Fourier transforms are useful in solving a variety of partial

differential equations; the choice of the appropriate transforms depends
on the type of boundary conditions imposed on the problem. Laplace
transforms can be used in solving boundary-value problems of partial
differential equation.

14
PHY312 MATHEMATICAL METHODS OF PHYSICS I

Example 5

Solve the problem

∂u ∂ 2u
=2 2 (53)
∂t ∂x
u (0, t ) = u (3, t ) = 0, u ( x,0) = 10 sin 2πx − 6 sin 4πx (54)

Solution: Taking the Laplace transform L of Eq. (53) with respect to t

gives
 ∂u   ∂ 2u 
L   = 2L  2 
 ∂t   ∂x 
Now
 ∂u 
L   = pL(u ) − u ( x,0)
 ∂t 
and
 ∂ 2u  ∂ 2 ∞ ∂2
L  2  = 2 ∫ e − pt u ( x, t )dt = 2 L[u ]
 ∂x  ∂x 0 ∂x
∞
Here ∂ 2 ∂x 2 and ∫0
...dt are interchangeable because x and t are
independent.

For convenience, let

∞
U = U ( x, p ) = L[u ( x, t )] = ∫ e − pt u ( x, t )dt
0

We then have
∂ 2U
pU − u ( x,0) = 2 L
∂x 2
from which we obtain, using the given conditions (54),
∂ 2U 1
− pU = 3 sin 4πx − 5 sin 2πx. (55)
∂x 2 2
Then taking the Laplace transform of the given conditions
u (0, t ) = u (3, t ) = 0, we have

L[u (0, t )] = 0, L[u (3, t )] = 0

Or
U (0, p ) = 0, U (3, p ) = 0.

These are the boundary conditions on U ( x, p ) . Solving eq. (55) subject to

these conditions we find

5 sin 2πx 3 sin 4πx

U ( x, p ) = −
p + 16π 2 p + 64π 2

15
PHY312 MATHEMATICAL METHODS OF PHYSICS I

The solution to eq. (55) can now be obtained by taking the inverse
Laplace transform
u ( x, t ) = L−1 [U ( x, p )] = 5e16π t sin 2πx − 36e 64π t sin 4πx.
2 2

SELF-ASSESSMENT EXERCISE 2

4. Differentiate between ordinary differential equation and partial

differential equation.

5. Derive the PDE that give rise to the function

Z = a ( x + y ) + b( x − y ) + abt + c = 0
6. Use the method of separation of variable to find the solution of
the boundary value problem
∂2 y ∂2 y
=
∂x 2 ∂t 2
y (0, t ) = 0 t>0
y (1, t ) = 0 t>0
y ( x,0) = sin 2 x
y ′( x,0) = 0 0≤ x<∞

7.0 CONCLUSION

In this unit, we have studied the notion of a solution of partial

differential equation. Also some elementary methods of solving linear
partial differential equations which occur frequently in physics and
engineering were dealt with.

8.0 SUMMARY
Here in this unit you have learnt about second order partial differential
equation. The classical method of separation of variables was
extensively studied along with the Laplace transform solutions of
boundary-value problems.

9.0 TUTOR- MARKED ASSIGNMENT

1. Form the PDEs whose general solutions are as follow:

z = Ae − p t cos px
2
(a)
 y
(b) z = f  
x
2. Solve the equation
∂ 2u ∂ 2u
2 =3 2 =0
∂x∂y ∂y

16
PHY312 MATHEMATICAL METHODS OF PHYSICS I

3. Find the solution of the differential equation

∂2 y ∂2 y
α2 =
∂x 2 ∂t 2
y (0, t ) = 0 0<t <∞
Where
y (0, t ) = 0 0≤t <∞
y ( x,0) = f ( x) 0≤ x≤L
y ( x,0) = g ( x) 0≤ x<L

4. Solve by Laplace transforms the boundary-value problem

∂ 2 u 1 ∂u
= for x > 0, t > 0
∂x 2 k ∂t
given that u = u0 (a constant) on x = 0 for t > 0 , and u = 0 for
x > 0, t = 0

7.0 REFERENCES/ FURTHER READING

Erwin, K. (1991). Advanced Engineering Mathematics. John Wiley &

Sons, Inc.

Pinsky, M.A. (1991).Partial Differential Equations and Boundary-Value

Problems with Applications. New York: McGraw-Hill.

17
PHY312 MATHEMATICAL METHODS OF PHYSICS I

UNIT 2 FOURIER SERIES

CONTENTS

8.0 Introduction
9.0 Objectives
10.0 Main Content
10.1 Periodic Functions
10.2 Even and Odd Functions
10.3 Fourier Theorem
10.4 Evaluation of Fourier Coefficients
10.5 Application of Fourier Series in Forced Vibrations
10.6 Half-Range Expansions
10.7 Fourier Integral
10.8 Fourier Integrals of Even and Odd Functions
11.0 Conclusion
12.0 Summary
13.0 Tutor-Marked Assignment
14.0 References/Further Reading

8.0 INTRODUCTION

In this unit, we shall discuss basic concepts, facts and techniques in

connection with Fourier series. Illustrative examples and some important
applications of Fourier series to Partial differential equations will be
studied.
We will also study the concept of periodic functions, even and odd
functions and the conditions for Fourier expansion.

9.0 OBJECTIVES

At the end of this unit, you should able to:

• identify whether a given function is even, odd or periodic

• evaluate the Fourier coefficients
• derive and apply Fourier series in forced vibration problems
• use Fourier Integral for treating various problems involving
periodic function
• apply half range expansion to solutions of some problems.

18
PHY312 MATHEMATICAL METHODS OF PHYSICS I

10.0 MAIN CONTENT

3.1 Periodic Functions

A function f (x) is said to be periodic if it defined for all real x and if

there is some positive number T such that

f ( x + T ) = f ( x) (1)

This number T is then called a period of f (x) .

Periodic functions occur very frequently in many application of

mathematics to various branches of science. Many phenomena in nature
such as propagation of water waves, light waves, electromagnetic waves,
etc are periodic and we need periodic functions to describe them.
Familiar examples of periodic functions are the sine and cosine
functions.

Example 1

Find the period of Tan x.

Solution: Suppose T is its period

f ( x + T ) = Tan ( x + T ) = Tanx
so that

Tan ( x + T ) − Tanx = 0
using trigonometric identity, we have

Tan T (1 − Tan 2 x)
=0
1 − TanxTanT

This implies that

Tan T = 0 If and only if 1 − Tan 2 x ≠ 0
T = Tan −1 0
Hence T =π

3.2 Even and Odd Functions

A function f (x) defined on interval [a, b] is said to be a even function if

f (− x ) = f ( x ) (2)

19
PHY312 MATHEMATICAL METHODS OF PHYSICS I

It is odd otherwise, that is

f (− x) = − f ( x) (3)

Example 2

Let f ( x) = sin x
Then f (− x) = − f ( x) i.e. sin( − x) − = − sin x

Thus it is obvious that sine function is always an odd function while

cosine function is an even function.

3.11 Fourier Theorem

According to the Fourier theorem, any finite, single valued periodic

function f (x) which is either continuous or possess only a finite number
of discontinuities (of slope or magnitude), can be represented as the sum
of the harmonic terms as
1
f ( x) = a 0 + a1 cos x + a 2 cos 2 x + . . . + a n cos nx
2
+ b1 sin x + b2 sin 2 x + . . . + bn sin nx

∞
a 0 + ∑ (a n cos nx + bn sin nx )
1
= (4)
2 n =1

3.12 Evaluation of Fourier Coefficients

Let us assume that f (x) is a periodic function of period 2π which can be

represented by a trigonometric series
∞
f ( x ) = a0 + ∑ (a n cos nx + bn sin nx) (5)
n =1

Given such a function f (x) we want to determine the coefficients of

a n and bn in the corresponding series in eq. (5).

We first determine a0 . Integrating on both sides of eq. (4) from − π to

π , we have
π  ∞
π 
∫π f ( x)dx = ∫ a 0 + ∑ ( a n cos nx + bn sin nx)  dx
− −π
 n =1 

If term-by-term integration of the series is allowed, then we obtain

π π ∞ π π
∫π f ( x)dx = a0 ∫ dx + ∑  a n ∫ cosnxdx + bn ∫ sinnxdx 
n =1  
− −π − π −π

20
PHY312 MATHEMATICAL METHODS OF PHYSICS I

The first term on the right equals 2πa0 . All other integrals on the right
are zero, as can be readily seen by performing the integration. Hence our
first result is

1 π
a0 =
2π ∫ π f ( x)dx
−
(6)

We now determine a1 , a2 , . . . by a similar procedure. We multiply

Eq. (5) by cos mx , where m is any fixed positive integer, and then
integrate from − π to π ,

π  ∞
π 
∫π f ( x ) cos mxdx = ∫ a 0 + ∑ (a n cos nx + bn sin nx) cos mxdx (7)
− −π
 n =1 

Integrating term-by-term, we see that the right-hand side becomes

π π ∞ π
a 0 ∫ cos mxdx + ∑ a n ∫ cos nx cos mxdx + bn ∫ sin nx cos mxdx 
−π  −π
n =1 
−π 

The first integration is zero. By applying trigonometric identity, we

obtain

π 1 π 1 π
∫ πcos nx cos mxdx = 2 ∫ π cos(n + m) xdx + 2 ∫ π cos(n − m) xdx
− − −

π 1 π 1 π
∫ πsin nx cos mxdx = 2 ∫ π sin(n + m) xdx + 2 ∫ π sin(n − m) xdx.
− − −

Integration shows that the four terms on the right are zero, except for the
last term in the first line which equals π when n = m. since in eq. (7)
this term is multiplied by a m , the right-hand side in eq. (7) is equal
to a mπ , and our second result is

1 π

π ∫−π
am = f ( x) cos mxdx m = 1, 2 . . . (8)
We finally determine b1, b2, ...... in eq.(5) by sin mx , where m is any fixed
positive integer, and the integrate from − π to π , we have

∞
π π  
∫ π f ( x) sin mxdx = ∫ π a + ∑ (a
− −
0
n =1
n cos nx + bn sin nx) sin mxdx

(9)

Integrating term-by-term, we see that the right-hand side becomes

21
PHY312 MATHEMATICAL METHODS OF PHYSICS I

π ∞ π π
a 0 ∫ sin mxdx + ∑ a n ∫ cos nx sin mxdx + bn ∫ sin nx sin mxdx
−π  −π
n =1 
−π 

The first integral is zero. The next integral is of the type considered
before, and we know that it is zero for all n = 1, 2,…. For the integral we
obtain
π 1 π 1 π
∫ πsin nx sinmxdx = 2 ∫ π cos(n − m) xdx − 2 ∫ π cos(n + m) xdx
− − −

The last term is zero. The first term on the right is zero when n ≠ m and
is π when n = m . Since in eq. (9) this term is multiplied by bm , the right-
hand side in eq. (6) is equal to bmπ , and our last result is

1 π

π ∫π
bm = f ( x) sin mxdx m = 1, 2 . . .
−

Writing n in place of m, we altogether have the so-called Euler

formulas

1 π

π ∫π
(a) a0 = f ( x)dx
−

1 π
(b) an =
π ∫ π f ( x) cos nxdx
−
n = 1, 2 . . . (10)
1 π

π ∫π
(c) bn = f ( x) sin nxdx
−

Example 3 Square wave

Find the Fourier coefficients of the periodic function

− k when −π < x < 0

f ( x) =  and f ( x + 2π ) = f ( x)
k when 0< x <π

Functions of this type may occur as external forces acting on mechanical

systems, electromotive forces in electric circuits, etc

Solution: From eq. (10a) we obtain a 0 = 0 . This can also be seen without
integration since the area under curve of f(x) between − π and π is zero.
From eq. (10b)

1 π 1 0 π
(−k ) cos nxdx + ∫ k cos nxdx 
π ∫π ∫
an = f ( x) cos nxdx =
− π  −π 0 

22
PHY312 MATHEMATICAL METHODS OF PHYSICS I

π
1  sin nx sin nx 
0

= − k +k =0
π  n −π n 0 

Because sin nx = 0 at − π , 0 and π for all n = 1, 2….Similarly, from Eq.

(10c) we obtain
1 π 1 0 π
(−k ) sin nxdx + ∫ k sin nxdx 
π ∫π ∫
bn = f ( x) sin nxdx =
− 
π − π 0 

π
1  cos nx cos nx 
0

= − k +k =0
π  n −π n 0 
Since cos(−α ) = cos α and cos 0 = 1 , this yields
k
[cos 0 − cos(−nπ ) − cos nπ + cos 0] = 2k (1 − cos nπ )
bn =
nπ nπ
Now, cos π = −1, cos 2π = 1, cos 3π = −1 etc, in general
− 1 for odd n,
cos nπ =  and thus
1 for even n,
2 for odd n,
1 − cos nπ = 
0 for even n,

Hence the Fourier coefficients bn of our function are

4k 4k 4k
b1 = , b2 = 0, b3 = , b4 = 0, b5 =
π 3π 5π
and since the a n are zero, the corresponding Fourier series is
4k  1 1 
 sin x + sin 3 x + sin 5 x + ........ (11)
π  3 5 
The partial sums are
4k 4k  1 
S1 = sin x, S2 =  sin x + sin 3 x , etc,
π π  3 

Furthermore, assuming that f (x) is the sum of the series and

setting x = π 2 , we have

π  4k  1 1 
f =k = 1 − + − +.....
2 π  3 5 
1 1 1 π
or 1 − + − + −......... =
3 5 7 4

23
PHY312 MATHEMATICAL METHODS OF PHYSICS I

SELF-ASSESSMENT EXERCISE 1

5. Define the periodic function. Give five examples.

6. Find the smallest positive period T of the following functions.

a. sin x

7. Are the following functions odd, even, or neither odd nor even?
a. e x
b. x sin x
8. Find the Fourier series of the following functions which are
assumed to have the

d. period 2π
e. f ( x) = x 2 4 −π < x < π
f. f ( x ) = sin x −π < x < π

3.13 Application of Fourier Series in Forced Vibrations

We now consider an important application of Fourier series in solving a

differential equation of the type

d 2x dx
m 2
+ Γ + kx(t ) = F (t ) (12)
dt dt

For example, the above equation would represent the forced vibrations
of a damped oscillator with Γ representing the damping constant, F(t)
the external force and m and k representing the mass of the particle and
the force constant respectively. We write eq. (12) in the form

d 2x dx
2
+ 2K + ω 02 x(t ) = G (t ) (13)
dt dt
Γ k F (t )
Where K = , ω 02 = and G (t ) = . The solution of the
2m m m
homogeneous part of eq. (13) can be readily obtained and is given by

x(t ) = A1e = Kt cos [ (ω 2

0 )
− K 2 t +θ ] for ω 02 > K 2 (14)
x(t ) = ( A2 t + B )e − Kt
for ω >K
2
0
2
(15)

In order to obtain the solution of the inhomogeneous part of eq. (13), we

first assume F(t) to be a sine or cosine function; for definiteness we
assume

24
PHY312 MATHEMATICAL METHODS OF PHYSICS I

G (t ) = b sin ωt (16)

The particular solution of eq. (13) can be written in the form

x(t ) = C sin ωt + D cos ωt (17)

The values of C and D can readily be obtained by substituting eq. (17) in

eq. (13),
and comparing coefficients of sin ωt and cos ωt we obtain

2ωK
D = − b
(ω 2
0 −ω 2 ) 2
+ 4ω 2 K 2

(18)
ω 02 − ω 2
C = − b
(ω 2
0 −ω 2 ) 2
+ 4ω 2 K 2

Now, if G(t) is not a sine or cosine function, a general solution of eq. (13)
is difficult to obtain. However, if we make a Fourier expansion of G(t)
then the general solution of eq.

(13) can easily be written down. As a specific example, we assume

G (t ) = αt (19)

The Fourier expansion of G (t ) can readily be obtained as

∞
G (t ) = ∑ bn sin nωt (20)
n =1

Proceeding in a manner similar to that described above we obtained the

following solution for the inhomogeneous part of eq. (13)
∞
x(t ) = ∑ [C n sin nωt + Dn cos nωt ] (21)
n =1

Where
2nωK
Dn = − bn
(ω 2
0 − n ω2
2
) 2
+ 4 n 2ω 2 K 2
(22)
(ω02 − n 2ω 2 )
Cn = − bn
(ω 2
0 − n 2ω 2 )2
+ 4 n 2ω 2 K 2
thus, if G(t) is a periodic function with period T then eq. (21) will be
valid for all values
of t.

3.14 Half-Range Expansions

In various physical and engineering problems there is a practical need

for applying Fourier series to functions f (t ) which are defined merely

25
PHY312 MATHEMATICAL METHODS OF PHYSICS I

on some finite interval. The function f (t ) is defined on an interval

0< t < l and on this interval we want represent f (t ) by a Fourier series.

A half-range Fourier series for a function f (x) is a series consisting of

the sine and cosine terms only.

Such functions are defined on an interval (0, l ) and we then obtain a

Fourier cosine series which represents an even periodic function f1 (t ) of
period T = 2l so that
∞
nπ
f (t ) = a 0 + ∑ a n cos t 0< t < l
n =1 l
(23)
and the coefficients are
1 l 2 l nπ
l ∫0 ∫
a0 = f (t )tdt , an = f (t ) cos tdt n = 1, 2 . . .
l 0 l
(24)

Then we obtain a Fourier sine series which represents an odd periodic

function f 2 (t ) of period T = 2l so that
∞
nπ
f (t ) = ∑ bn sin t 0< t < l
n =1 l
(25)
and the coefficients are

2 l nπ
bn =
l ∫0
f (t ) sin
l
tdt n = 1, 2 . . .
(26)

The series in eqs.(23) and (25) with the coefficients in eqs.(24) and (26)
are called half-range expansions of the given function f (t )

Example 4

Find the half-range expansions of the function

 2k l
 l t when 0<t <
2
f (t ) = 
 2k (l − t ) when l
<t <l
 l 2

26
PHY312 MATHEMATICAL METHODS OF PHYSICS I

Solution: From eq. (24) we obtain

1  2k l / 2 2k l  k
a 0 =  ∫ tdt + ∫ (l − t ) dt  =
l l 0 l l/2  2
2  2k l / 2 nπ 2k l nπ 
a n =  ∫ t cos tdt + ∫ (l − t ) cos tdt 
l l 0 l l l/2 l 

Now by integration by part

l/2 nπ lt nπ l / 2 1 l / 2 nπ
∫
0
t cos
l
tdt =
nπ
sin
l
t0 −
nπ ∫0
sin
l
tdt

l2 nπ l2  nπ 
= sin + 2 2  cos − 1
2nπ 2 nπ  2 
Similarly,
l nπ l2 nπ l2  nπ 
∫ (l − t ) cos tdt = − sin − 2 2  cos nπ − cos 
l/2 l 2nπ 2 nπ  2 

By inserting these two results we obtain

4k  nπ 
un = 2 2 
2 cos − cos nπ − 1
nπ  2 
Thus,
a 2 = −16k 2 2 π 2 , a 6 = −16k 6 2 π 2 , a10 = −16k 10 2 π 2 , ...

And a n = 0 when n ≠ 2, 6, 10, 14, ... Hence the first half-range expansion
of f (t ) is
k 16k  1 2π 1 6π 
f (t ) = − 2  2 cos t + 2 cos t + ...
2 π 2 l 6 l 

This series represents the even periodic expansion of the function f (t ) .

Similarly from eq. (26)

8k nπ
bn = sin
nπ 2 2
2
and the other half-range expansion of f (t ) is

8k  1 π 1 3π 1 5π 
f (t ) = 2  2
sin t − 2 sin t + 2 sin t − +...
π 1 l 3 l 5 l 

This series represents the odd periodic extension of f (t ) .

27
PHY312 MATHEMATICAL METHODS OF PHYSICS I

Example 5

Find a Fourier sine series for

0 x≤2
f ( x) =  on (0, 3).
2 x>2

Solution: Since the function is odd, then a 0 = 0

2 l nπ
Then bn = ∫ f ( x) sin xdx
l 0 l
2 3 nπ
= ∫ f ( x) sin xdx
3 0 3
2 2 nπ 2 3 nπ
= ∫ 0.sin xdx + ∫ 2 sin xdx
3 0 3 3 2 3
Now by integration, we have
4  2nπ 
bn =  cos − cos nπ 
nπ  3 
The series thus becomes
∞
4  2 nπ  nπx
f ( x) = ∑  cos − ( −1) n  sin
n =1 nπ  3  3

So that
4 1 πx 3 2πx 2 3πx 
f ( x) =  sin − sin + sin − +...
π 2 3 4 3 3 3 

Example 6

Find the Fourier cosine series for

f ( x) = e x on (0, π )

Solution: Since f (x) is an odd function, then

b0 =
1 l x
∫
l 0
1
e dx = e π − 1
π
( )
Also
nπx
bn =
2
π∫ 0
π
e x cos
π
dx =
2 1  π

π 1 + n2 
(
 e cos nπ − 1 )

Thus the series becomes

(e ) [(−1) ]
∞
2 2 1
ex =
π
π
−1 + ∑
π 1+ n n =1
2
n
e π − 1 cos nx

28
PHY312 MATHEMATICAL METHODS OF PHYSICS I

SELF-ASSESSMENT EXERCISE 2

1. Find the Fourier sine series for

f ( x) = e x on (0, π )

9. Find the Fourier series for

f ( x) = x on 0 < x < 2
consisting of (a) sine series only (b) cosine series only

3.15 Fourier Integral

Fourier series are powerful tools in treating various problems involving

periodic functions. When the fundamental period is made infinite, the
limiting form of the Fourier series becomes an integral which is called
Fourier Integral.

3.15.1 Definition

Let f (x) be defined and single valued in the interval [− L, L ] . If

f (x) satisfies the following conditions:

(i) f (x) is periodic and of period 2L

(ii) f (x) and f ′(x) are piecewise continuous
∞
(iii) ∫−∞
f ( x) dx is convergent, then f (x) can be expressed as
∞
f ( x) = ∫ ( A(α ) cos αx + B (α ) sin αx )dx
0

(27)
1 ∞
A(α ) = ∫ f ( x) cos αxdx
π −∞

(28)
1 ∞
B (α ) = f ( x) sin αxdx
π∫ −∞

(29)

3.16 Fourier Integrals of Even and Odd Functions

It is of practical interest to note that if a function is even or odd and can

be represented by a Fourier integral, and then this representation will be
simpler than in the case of an arbitrary function. This follows
immediately from our previous formulas, as we shall now see.

If f (x) is an even function, then B (α ) = 0

2 ∞
A(α ) = ∫ f ( x) cos αxdx (30)
π 0

29
PHY312 MATHEMATICAL METHODS OF PHYSICS I

and eq. (27) reduces to the simpler form

∞
f ( x) = ∫ A(α ) cosαxdx ( f even) (31)
0

Similarly, if f (x) is odd, then A(α ) = 0 in eq. (28), also

2 ∞
B(α ) = f ( x ) sin αxdx
π∫ 0

(32)
and
∞
f ( x ) = ∫ B (α ) sin αxdx ( f odd ) (33)
0

These simplifications are quite similar to those in the case of a Fourier

series discussed.

Example 7

Find the Fourier Integral of f ( x) = x 2 −π ≤ x ≤ π

Solution:
1 ∞
A(α ) = f ( x) cos αxdx
π∫ −∞

1 ∞
x 2 cos αxdx
π∫
=
−∞

Using integration by parts, we obtain

π
2 x 1 
A(α ) =  cos αx − 2 sin αx  = 0
πα α α  −π
Also
1 ∞
B (α ) = ∫ f ( x) sin αxdx
π −∞

1 ∞
x 2 sin αxdx
π∫
=
−∞

So that
π
1  x2 2  2π
B (α ) = −  cos αx − 3 cos αx  = (−1)α
π α α  −π α

30
PHY312 MATHEMATICAL METHODS OF PHYSICS I

From eq. (27)

∞
f ( x) = ∫ ( A(α ) cosαx + B (α ) sin αx )dx
0
and
∞ 2π  2π
f ( x ) = ∫  0 • cos αx + (−1)α sin αx dx = 2 (−1)α
0
 α  α
Hence
2π ∞ 2π
f ( x) = x 2 = (−1)α ∫ sin αxdx = (−1)α
α 0 α 2

11.0 CONCLUSION

In this unit, you have studied the concept of periodic functions,

representations of functions by Fourier series, involving sine and cosine
function are given special attention. We also use the series expansion in
the determination of Fourier coefficients and the half-range expansions.

12.0 SUMMARY

In this unit, you have studied:

• Even and odd functions

• Fourier Integral representations and Fourier series expansion.
• Application of Fourier Integral technique in the simplification of
even and odd functions.

13.0 TUTOR- MARKED ASSIGNMENT

1. Find the smallest positive period T of the following functions

c. (i) sin 2πx
2πnx
d. (ii) cos
k

2. Find the Fourier series for

0 −5 < x < 0
f ( x) =  where f (x) has period 10
3 0< x<5

10. Find the Fourier series for

f ( x) = x 2 for 0 < x < 2π

11. Find the Fourier series of function

f ( x ) = x + π when − π < x < π and f ( x + 2π ) = f ( x)

31
PHY312 MATHEMATICAL METHODS OF PHYSICS I

12. Expand the function

T T
f (t ) = t 2 −
< x < in a Fourier series to show that
2 2
T π
2 2
 1 1 
f (t ) = t 2 = 2 
− 4 cos ωt − cos 2ωt + cos ωt − ...
4π  3  4 9 
take ω = 2π T

13. Represent the following functions f (t ) by a Fourier cosine series

π
(a) f (t ) = sin t (0 < t < l )
l
(b) f (t ) = e t (0 < t < l )

14. Find the Fourier integral representation of the function

1 when x < 1,
f ( x) = 
0 when x > 1.

14.0 REFERENCES/FURTHER READING

Puri, S.P. (2004). Textbook of Vibrations and Waves. Macmillan India

Ltd.

Ghatak, A.K.; Goyal, I.C. & Chua, S.J. (1995). Mathematical Physics.
Macmillan India Ltd.

Carslaw, H.S. (1950). Introduction to the Fourier Series and Integral.

New York:

Dover Publications.

32
PHY312 MATHEMATICAL METHODS OF PHYSICS I

MODULE 2 APPLICATION OF FOURIER TO PDES

(LEGENDRE POLYNOMIALS AND
BESSEL FUNCTIONS)

Unit 1 Legendre Polynomials

Unit 2 Bessel Functions

UNIT 1 LEGENDRE POLYNOMIALS

CONTENTS

1.0 Introduction
2.0 Objectives
3.0 Main Content
3.1 Legendre Equation
3.2 The Polynomial Solution of the Legendre’s Equation
3.3 The Generating Function
3.4 Rodrigue’s Formula
3.5 Orthogonality of the Legendre Polynomials
3.6 The Angular Momentum Problem in Quantum Mechanics
3.7 Important Integrals Involving Legendre Functions
4.0 Conclusion
5.0 Summary
6.0 Tutor- Marked Assignment
7.0 References/Further Reading

2.0 INTRODUCTION

In this unit, you will be introduced to the polynomial solutions of the

Legendre equation, the generating function as well as the orthogonality
of Legendre polynomials. Also we shall consider some important
integrals involving Legendre functions which are of considerable use in
many areas of physics.

7.0 OBJECTIVES

At the end of this unit, you should be able:

• derive the polynomial solution of the Legendre equation

• use the generating functions to derive some important identities
• determine the orthogonality of the Legendre polynomials.

33
PHY312 MATHEMATICAL METHODS OF PHYSICS I

8.0 MAIN CONTENT

3.1 Legendre Equation

The equation

(1 − x 2 y ′′( x) − 2 xy ′( x) + n( n + 1) y ( x) = 0 (1)
where n is a constant is known as the Legendre’s differential equation.
In this unit we will discuss the solutions of the above equation in the
domain − 1 < x < 1 . We will show that when
n = 0, 1, 2, 3, . ..
one of the solutions of eq. (1) becomes a polynomial. These polynomial
solutions are known as the Legendre polynomials, which appear in
many diverse areas of physics and engineering.

3.6 The Polynomial Solution of the Legendre’s Equation

If we compare eq. (1) with homogeneous, linear differential equations of

the type

y ′′( x) + U ( x) y ′( x) + V ( x) y ( x) = 0 (2)

we find that the coefficients

2x n(n + 1)
U ( x) = − and V ( x) = (3)
1− x 2
1− x2

are analytical at the origin. Thus the point x = 0 is an ordinary point and
a series solution of eq. (1) using Frobenius method should be possible.
Such that

y ( x) = C 0 S n ( x) + C1Tn ( x)
where
n(n + 1) 2 n(n − 2)(n + 1)(n + 3) 4
S n ( x) = 1 − x + x −... (4a)
2! 4!
And
(n − 1)(n + 2) 3 (n − 1)(n − 3)(n + 2)(n + 4) 5
Tn ( x) = x − x + x − . . . (4b)
3! 5!

If n ≠ 0. 1, 2, . . . both eqs. (4a) and (4b) are infinite series and converge
only if x < 1.

It may be readily seen that when

n = 0. 2, 4, . . .

34
PHY312 MATHEMATICAL METHODS OF PHYSICS I

The even series becomes a polynomial and the odd series remains an
infinite series. Similarly for

n = 1, 3, 5, . . .
the odd series becomes a polynomial and the even series remains an
infinite series.
Thus when
n = 0, 1, 2, 3, . . .
one of the solutions becomes a polynomial. The Legendre polynomial,
or the Legendre function of the first kind is denoted by Pn (x) and is
defined in terms of the terminating series as below:

 S n ( x)
 S (1) for n = 0, 2, 4, 6, . ..
 n
Pn ( x) =  (5)
Tn ( x) for n =1, 3, 5, 7, . ..
 Tn (1)
Thus,
P0 ( x) = 1, P1 ( x) = x, P2 ( x) =
2
(
1 2
3x − 1 ,)
P3 ( x) =
1 3
2
( ) 1
(
5 x − 3 x , P4 ( x) = 35 x 4 − 30 x 2 + 3 ,
8
) (6)
1
( )
P5 ( x) = 63 x 5 − 70 x 3 + 15 x ,. . .
8

Obviously, Pn (1) = 1 (7)

Higher order Legendre polynomials can easily be obtained by using the

recurrence relation

nPn ( x) = ( 2n − 1) x Pn−1 ( x) − ( n − 1) Pn− 2 ( x)

Since for even values of n the polynomials Pn (x) contain only even
powers of x and for odd values of n the polynomials contain only odd
powers of x, we readily have

Pn (− x) = (−1) n Pn ( x) and obviously (8)

Pn (−1) = (−1) n
(9)

3.7 The Generating Function

The generating function for the Legendre polynomials is given by

∞
G ( x, t ) = (1 − 2 xt + t 2 ) −1 2 = ∑ Pn ( x)t n ; − 1 ≤ x ≤ 1, t < 1 (10)
n =0

35
PHY312 MATHEMATICAL METHODS OF PHYSICS I

Let us assume that

∞
G ( x, t ) = (1 − 2 xt + t 2 ) −1 2 = ∑ K n ( x)t n (11)
n=0

Where K n (x) is a polynomial of degree n. Putting x = 1 in eq. (11), we

obtain
∞

∑K
n =0
n ( x)t n = (1 − 2t + t 2 ) −1 2

= (1 − t ) −1
= 1+ t + t 2 + t 3 + . . . + t n + . . .

Equating the coefficients of t n from both sides, we have

K n (1) = 1 (12)

Now, if we can show that K n (x) satisfies eq. (1), then K n (x) will be
identical to Pn (x) . Differentiating G(x, t) with respect to x and t, we
obtain
∂G
(1 − 2 xt + t 2 ) = ( x − t )G ( x, t ) (13)
∂t
and
∂G ∂G
t = (x − t) (14)
∂t ∂x
Using eqs. (11), (13) and (14), we have
∞ ∞
(1 − 2t + t 2 )∑ nK n ( x)t n−1 = ( x − t )∑ K n ( x)t n (15)
n=0 n=0

and
∞ ∞
t ∑ nK n ( x)t n−1 = ( x − t )∑ K n′ ( x)t n (16)
n=0 n=0

Equating the coefficient of t n −1 on both sides of eqs. (15) and (16), we

get
nK n ( x) − ( 2n − 1) xK n −1 ( x ) + ( n − 1) K n − 2 ( x) = 0 (17)
and
xK n′ −1 ( x) − K n′ − 2 ( x) = ( n − 1) K n−1 ( x) (18)

Replacing n by n+1 in Eq. (18), we obtain

xK n′ ( x) − K n′ −1 ( x) = nK n ( x) (19)

We next differentiate Eq. (17) with respect to x and eliminate K n′ −2 with

help of Eq. (18) to obtain

36
PHY312 MATHEMATICAL METHODS OF PHYSICS I

K n′ ( x ) − xK n′ −1 ( x) − nK n−1 ( x) = 0 (20)

If we multiply eq. (19) by x and subtract it from eq. (20), we would get

(1 − x 2 ) K n′ − n( K n−1 − xK n ) = 0 (21)

Differentiating the above equation with respect to x, we have

(1 − x 2 ) K n′′ − 2 xK n′ − n( K n′ −1 − xK n′ − K n ) = 0 (22)

Using eqs. (19) and (22), we obtain

(1 − x 2 ) K n′′( x) − 2 xK n′ ( x) − n(n + 1) K n ( x) = 0 (23)

which shows that K n (x) is a solution of Legendre equation. In view of

eqs. (7) and (12) and the fact that K n (x) is a polynomial in x of degree n,
it follows that K n (x) is nothing but Pn (x) . eq. (17) gives the recurrence
relation for Pn (x)
nPn ( x) = ( 2n − 1) x Pn−1 ( x) − ( n − 1) Pn− 2 ( x) (24)

3.8 Rodrigues’ Formula

Let
φ ( x ) = ( x 2 − 1) n (25)

Differentiating eq. (25), we get

dφ
= 2nx( x 2 − 1) n−1
dx
or
d 2φ dφ
(1 − x 2 ) 2
+ 2 x(n − 1) + 2nφ = 0
dx dx

Differentiating the above equation n times with respect to x, we would

get

d 2φn dφ
(1 − x 2 ) 2
+ 2 x n + n( n + 1)φn = 0 (26)
dx dx
where
φn =
d nφ d n
dx n
= n x2 −1
dx
n
[( )] (27)

37
PHY312 MATHEMATICAL METHODS OF PHYSICS I

This shows that φn (x) is a solution of the Legendre’s equation. Further, it

is obvious from eq. (27) that φn (x) is a polynomial of degree n in x.
Hence φn (x) should be a constant multiple of Pn (x) , i.e.

[(
d n x2 −1 )]
n

= CPn ( x) (28)
dx n
dn
dx n
x[(
2
− 1
n
=)]dn
dx n
[
(x + 1)n (x − 1)n ]
= n!( x − 1) n + n ( x + 1)n(x − 1)
n! n −1

1!
n(n − 1) n!
+ (x + 1)2 n(n − 1)( x − 1) n−2 + . . . + ( x + 1) n n! (29)
2! 2!

It may be seen that all terms on the right hand side of eq. (29) contain a
factor (x-1) except for the last term. Hence

dn 2
dx n
(
x −1
n
) = 2 n n! (30)
x =1

Using Eqs. (7), (28) and (29), we obtain

C = 2 n n! (31)

Therefore
Pn ( x ) =
1 dn 2
n
2 n! dx n
x −1
n
( ) (32)

This is known as the Rodrigues formula for the Legendre polynomials.

For example
P2 ( x ) =
1 d2 2
2
2 2! dx 2
x −1
2
( )
1
(
= 3x 2 − 1
2
)

Which is consistent with eq. (6)

3.9 Orthogonality of the Legendre Polynomials

Since the Legendre’s differential equation is of the Sturm-Liouville form

in the interval − 1 ≤ x ≤ 1, with Pn (x) satisfying the appropriate boundary
conditions at x = ±1 . The Legendre polynomials form an orthogonal set
of functions in the interval − 1 ≤ x ≤ 1, i.e

38
PHY312 MATHEMATICAL METHODS OF PHYSICS I

1
∫−1
p n ( x) p m ( x)dx = 0 m≠n (33)

The Orthogonality of the Legendre polynomials can be proved as

follows: Pn (x) satisfies eq. (1) which can be written in the Sturm-
Liouville form as

d  2

dx 
(
dP ( x) 
)
x − 1 n  + n( n + 1) Pn ( x) = 0
dx 
(34)

Similarly
d  2

dx 
(
dP ( x) 
)
x − 1 m  + m( m + 1) Pm ( x) = 0
dx 
(35)

Multiply eq. (34) by Pm (x ) and eq. (35) by Pn (x) and subtracting eq. (35)
from eq. (34), we get

d
dx
[
(1 − x 2 )(Pn′( x) Pm ( x) − Pm′ ( x) Pn ( x) ) ]
= (m − n)(n + m + 1) Pn ( x) Pm ( x)

Integrating the above equation from x = -1 to x = 1, we get

(1 − x 2 )[(Pn′ ( x) Pm ( x) − Pm′ ( x) Pn ( x) )] −1
+1

1
= (m − n)(n + m + 1) ∫ Pn ( x) Pm ( x)dx
−1

Because of the factor (1 − x 2 ) the left hand side of the above equation
vanishes; hence

1
∫−1
Pn ( x) Pm ( x )dx for m ≠ n

To determine the value of the integral

1
∫−1
Pm2 ( x )dx

we square both sides of eq. (10) and obtain

∞ ∞
(1 − 2 xt + t 2 ) −1 = ∑∑ Pm ( x) Pn ( x)t m + n (36)
m =0 n =0

Integrating both sides of the above equation with respect to x from -1 to

+1 and using eq. (33), we get

39
PHY312 MATHEMATICAL METHODS OF PHYSICS I

∞
1 1 1+ t
∑t ∫
1 1
2n
Pn2 ( x)dx = ∫ dx = ln
n =0
−1 −1 1 − 2 xt + t 2
t 1− t
 1 1 1 2n 
= 21 + t 2 + t 4 + . . . + t + ... 
 3 5 2 n + 1 

Equating the coefficients of t 2 n on both sides of the above equation, we

have

1 2
∫−1
Pn2 ( x)dx =
2n + 1
n = 0, 1, 2, 3, .... (37)

Thus we may write

1 2
∫ Pn ( x) Pm ( x)dx = δ nm
−1 2n + 1
where
0 if n ≠ m
δ nm = 
1 if n = m

Example

We consider the function cos πx 2 and expand it in a series (in the

domain − 1 < x < 1 ) up to the second power of x:

πx 2
cos = ∑ C n Pn ( x)
2 n=0

Now
2 n + 1 +1 πx
Cn =
2 ∫−1
cos Pn ( x) dx
2

Substituting for Pn (x) from eq. (6) and carrying out brute force
integration, we readily get

2 10  12 
C0 = ; C1 = 0; C 2 = 1 − 
π π  π2 
Thus
πx 2 10  12  3 x 2 − 1 
cos = + 1 −  
2 π π  π 2  2 

40
PHY312 MATHEMATICAL METHODS OF PHYSICS I

3.6 The Angular Momentum Problem in Quantum

Mechanics

In electrostatics the potential Φ satisfies the Laplace equation

∇ 2Φ = 0 (38)

We wish to solve the above equation for a perfectly conducting sphere

(of radius a), place in an electric field which is in the absence of the
sphere was of uniform magnitude E0 along z-direction. We assume the
origin of our coordinate system to be at the centre of the sphere. Because
the sphere is a perfect conductor, the potential on its surface will be
constant which, without any loss of generality, may be assumed to be
zero. Thus, eq. (35) is said to be solved subject to the boundary
condition

Φ(r = a) = 0 (39)

At a large distance from the sphere the field should remain unchanged
and thus
E (r → ∞) = E 0 zˆ
Since
E = −∇Φ
we have
Φ ( r → ∞) = − E0 z + C
= − E 0 r cos θ + C (40)

Where C is a constant. Obviously, we should use the spherical system of

coordinates so that

1 ∂  2 ∂Φ  1 ∂  ∂Φ 
∇ 2Φ = r + 2  sin θ 
r ∂r  ∂r  r sin θ ∂θ 
2
∂θ 
1 ∂ 2Φ
+ 2 2 =0 (41)
r sin θ ∂φ 2

From the symmetry of the problem it is obvious that Φ would be

independent of the azimuthal coordinate φ so that eq. (41) simplifies to

1 ∂  2 ∂Φ  1 ∂  ∂Φ 
r + 2  sin θ =0 (42)
r 2
∂r  ∂r  r sin θ ∂θ  ∂θ 

Separation of variables

Φ = R ( r )Θ(θ )
will yield

41
PHY312 MATHEMATICAL METHODS OF PHYSICS I

1 d  2 dR  1 d  dΘ 
r =  sin θ  = a cons tan t (= λ ) (43)
R dr  dr  Θ sin θ dθ  dθ 

Changing the independent variable from θ to µ by the relation

µ = cosθ
In the angular equation, we get
d 2Θ dΘ
(1 − µ 2 ) − 2µ + λΘ = 0 (44)
dµ 2
dµ

In order that the solution of eq. (44) does not diverge

at µ = ±1(θ = 0 and π ) , we must have
λ = l (l + 1); l = 0, 1, 2, . . .
and then
2l + 1
Θ(θ ) = Pl (cosθ ) (45)
2

Thus the radial equation can be written as

1 d  2 dR 
r  = l (l + 1)
R dr  dr 
or
d 2R dR
r2 2
+ 2r − l (l + 1) R = 0 (46)
dr dr

The above equation is the Cauchy’s differential equation and its solution
can readily be written as

Bl
R = Al r l +
r l +1

Hence the complete solution of eq. (42) is given by

∞ ∞
Bl
Φ ( r ,θ ) = ∑ Al r l Pl (cosθ ) + ∑ P (cosθ )
l +1 l
l =0 l =0 r

[
= Ao Po (cosθ ) + A1rP1 (cosθ ) + A2 r 2 P2 (cosθ ) + . . . ]
Bo B
+ Po (cos θ ) + 21 P1 (cos θ ) + . . .
r r

Applying the boundary condition given by eq. (40), we get

A0 = C , A1 = − E o , A2 = A3 = . . . = 0
Thus
 B   B 
Φ ( r ,θ ) =  C + 0  P0 (cosθ ) +  − E 0 r + 21  P1 (cosθ )
 r   r 

42
PHY312 MATHEMATICAL METHODS OF PHYSICS I

B2
+ P2 (cosθ ) + . . .
r3
Applying the condition at r = a [see eq. (39)], we get

 B   B 
 C + 0  +  − E 0 a + 21  P1 (cosθ )
 a   a 
B2 B
+ 3 P2 (cosθ ) + 42 P3 (cosθ ) + . . . = 0
a a

Since the above equation has to be satisfied for all values of θ and
since Pn (cosθ ) form a set of orthogonal functions, the coefficients
of Pn (cosθ ) should be zero giving

B0 = − aC , B1 = E 0 a 3
B2 = B3 = B4 . . . = 0
Thus
 a  a3 
Φ (r ,θ ) = C 1 +  − E0 1 + 3 r cosθ (47)
 r  r 

The 1 r potential would correspond to a charged sphere and, therefore,

for an uncharged sphere we must have C = 0 giving

 a3 
Φ (r ,θ ) = − E0 r cosθ 1 + 3  (48)
 r 

This is the required solution to the problem. One can easily determine
the components of the electric field as:

∂Φ  a3 
Er = − = E0 cosθ 1 + 2 3 
∂r  r 
1 ∂Φ  a3 
Eθ = − = E0 sin θ 1 − 3 
r ∂θ  r 
1 ∂Φ
Eφ = − =0
r sin θ ∂φ

3.7 Important Integrals Involving Legendre Functions

We give below some important integrals involving Legendre functions

which are of considerable use in many areas of physics.

[x + ( x ]
∞ n
1
− 1)1 2 cosθ dθ
π∫
Pn ( x) = 2
(49)
0

43
PHY312 MATHEMATICAL METHODS OF PHYSICS I

π n

∫0 (cosφ + i sin φ cosθ ) dθ

1
Pn (cos φ ) = (50)
π
1 23 2
∫−1 (1 − x) −1 2 Pn ( x)dx = (51)
2n + 1
1 1
[ 2
]
(n + m)!
∫−1 (1 − x 2 ) Pn ( x) dx = m(n − m)!
m
(52)
1
[ ]
2 1 (n + m)!
∫−1 Pn ( x) dx =  1  (n − m)!
m
(53)
n + 
 2

SELF-ASSESSMENT EXERCISE

3. Show that (n + 1) Pn ( x ) = Pn′+1 ( x) − Pn′−1 ( x)

4. Using the Rodrigue’s formula show that
1
Pn′ ( x) = n(n + 1)
2

9.0 CONCLUSION

The concept of generating function for the Legendre polynomials allows

us to readily derive some important identities.

We have also established in this unit, relationship between

Orthogonality of the Legendre polynomials and the generating function.

10.0 SUMMARY

This unit deals with Legendre functions and its applications to physical
problems especially in quantum mechanics.

11.0 TUTOR-MARKED ASSIGNMENT

4. Show that
(1 − x 2 ) Pn′( x) = nPn −1 ( x) − nxPn ( x)
n(n + 1)
=− [Pn+1 ( x) − Pn−1 ( x)]
2n + 1

5. Determine the coefficients

C 0 , C1 , C 2 , C3 , in the expansion
 nx  3
sin   = ∑ C n Pn ( x) −1 < x < 1
 2  n=0

44
PHY312 MATHEMATICAL METHODS OF PHYSICS I

6. Consider the function

0 −1 ≤ x < 0
f ( x) = 
1 0 < x ≤1
Show that
1 1 ∞
f ( x) = − ∑ [Pn+1 (0) − Pn−1 (0) −]Pn ( x) −1 < x < 1
2 2 n =1

4. Show that the generating function

∞
1
= ∑ Pn ( x)u n
1 − 2 xu + u 2 n =0

Hint: Start from the binomial expansion of 1 1 − v , set v = 2 xu − u 2 ,

multiply the powers of 2 xu − u 2 out, collect all the terms involving u n ,
and verify that the sum of these terms is Pn ( x)u n .

7.0 REFERENCES/FURTHER READING

Ghatak, A.K.; Goyal, I.C. & Chua, S.J. (1995). Mathematical Physics.
Macmillan India Ltd.

Erwin, K. (1991). Advanced Engineering Mathematics. John Wiley &

Sons, Inc.

45
PHY312 MATHEMATICAL METHODS OF PHYSICS I

UNIT 2 BESSEL FUNCTIONS

CONTENTS

3.0 Introduction
4.0 Objectives
3.0 Main Content
3.1 Bessel Differential Equation
3.2 Series Solution and Bessel Function of the First Kind
3.3 Recurrence Relations
3.4 The Generating Function
3.4.1 Derivation of Recurrence Relation from the
Generating Function
3.5 Some Useful Integrals
3.6 Spherical Bessel functions
3.7 Bessel Function of the Second Kind
3.8 Modified Bessel functions
4.0 Conclusion
5.0 Summary
6.0 Tutor-Marked Assignment
7.0 References/Further Reading

1.0 INTRODUCTION

In this unit we shall consider the series solution as well as Bessel

functions of the first and second kinds of order n.
We will also be introduced to some integrals which are useful in
obtaining solutions of some problems.

2.0 OBJECTIVES

At the end of this unit, you should be able to:

• derive the solution of Bessel function of the first kind

• prove a relationship between the recurrence relation and the
generating functions
• derive the solution of Bessel function of the second kind.

46
PHY312 MATHEMATICAL METHODS OF PHYSICS I

3.0 MAIN CONTENT

3.1 Bessel Differential Equation

The equation

d2y dy
x2 2
+ x ( x 2 − n 2 ) y ( x) = 0 (1)
dx dx

Where n is a constant known as Bessel’s differential equation.

Since n 2 appears in eq. (1), we will assume, without any loss of
generality, that n is either zero or a positive number. The two linearly
independent solutions of eq. (1) are

J n (x ) and J −n (x)
Where J n (x) is defined by the infinite series
∞ n+ 2 r
1  x
J n ( x) = ∑ (−1) r   (2)
r =0 r!Γ(n + r + 1)  2 
or
xn  x2 x4 
J n ( x) =  2(2n + 2) 2.4(2n + 2)(2n + 4) − .. .
1 − + (3)
2 Γ(n + 1) 
n

where Γ(n + r + 1) represents the gamma function.

3.4 Series Solution and Bessel Function of the First Kind

If we use eq. (1) with the homogeneous, linear differential equation of

the type

y ′′( x) + U ( x) y ′( x ) + V ( x) y ( x) = R ( x) (4)
we find the coefficients
1 n2
U ( x) = and V ( x) = 1 −
x x2
are singular at x = 0. However, x = 0 is a regular singular point of the
differential equation and a series solution of eq. (1) in ascending powers
of x. Indeed, one of the solutions of eq. (1) is given by

 x2 x4 
J n ( x) = C0 x n 1 − + − .. . (5)
 2(2n + 2) 2.4(2n + 2)(2n + 4) 

and where C0 is an arbitrary constant. This solution is analytic at x = 0

for n ≥ 0 and converges for all finite values of x. If we choose

C 0 = 2 n Γ(n + 1) (6)

47
PHY312 MATHEMATICAL METHODS OF PHYSICS I

then the eq. (5) is denoted by J n(x) and is known as the Bessel function
of the first kind of order n.

∞ n+ 2 r
1  x
J n ( x) = ∑ (−1) r  
r =0 r!Γ(n + r + 1)  2 
n n+ 2 n+4
1  x 1 x 1 x
=   −   +   − ... (7)
Γ(n + 1)  2  1!Γ(n + 2)  2  2!Γ(n + 3)  2 

In particular
J 0 ( x) = 1 −
(x 2)2 + (x 2)4 − (x 2)6 + ... (8)
(1!) 2 (2!) 2 (3!) 2

J 1 2 ( x) =
(x 2)1 2 − (x 2)5 2 + (x 2)9 2 +. . .
Γ(3 2 ) 1!Γ(5 2) 2!Γ(7 2 )
2  x3 x5 
= x − + −. . .
πx  3! 5! 
2
= sin x (9)
πx

It follows immediately from eqs. (7) and (8) that

J n ( 0) = 0 for n > 0
and
J n ( 0) = 1
If n ≠ 0, 1, 2, 3, . . . then

J −n ( x) = ∑
∞
(x 2)− n+ 2r (10)
r =0 r!Γ(− n + r + 1)

Example 1

In this example we will determine the value of J −1 2 ( x) from eq. (10).

Thus
J −1 2 ( x) =
( x 2)
−1 2
−
( x 2)
32
+
( x 2)
72
+. . .
Γ(1 2) 1!Γ(3 / 2) 2!Γ(5 2 )
2  x2 x4 
= 1− + −. . .
πx  2! 4! 
2
= cos x
πx

Which is linearly independent of J 1 2 ( x ) [see eq. (9)] and it can be

verified that J −1 2 ( x) does in fact satisfy eq. (1) for n = ½. Thus

48
PHY312 MATHEMATICAL METHODS OF PHYSICS I

2
J 1 / 2 ( x) = sin x (11)
πx
and
2
J −1 / 2 ( x) = cos x (12)
πx

Using the above two equations and the recurrence relation [see Eq. (21)]
2n
J n +1 ( x) = J n ( x) − J n −1 ( x) (13)
x

We can readily obtain closed form expression for J ±3 / 2 ( x) ,

J ± 5 / 2 ( x ) , J ±7 / 2 ( x ) , …
2  sin x 
J 3 / 2 ( x) =  − cos x  (14)
πx  x 
2 1 
J −3 / 2 ( x ) =  − cos x − sin x  (15)
πx  x 
2  (3 − x 2 ) 3 
J 5 / 2 ( x) =  sin x − cos x  (16)
πx  x 2
x 
2  (3 − x 2 ) 3 
J −5 / 2 ( x ) =  cos x − sin x  (17)
πx  x 2
x 
etc.

Next, we will examine eq.(10) when n is a positive integer. To be

specific we assume n = 4; then the first, second, third and fourth terms in
the series given by eq. (10) will contain the terms

1 1 1 1
, , , and
Γ(−3) Γ(−2) Γ(−1) Γ(0)
respectively and all these terms are zero. In general the first n terms of
the series would vanish giving

J −n ( x) = ∑
∞
(x 2)− n+ 2r (18)
r =n r!Γ(− n + r + 1)

If we put r = k+n, we would obtain

∞
J − n ( x) = ∑ (−1) k + n
(x 2)n+2 k
k =0 (k + n)!Γ(k + 1)
∞
= ∑ ( −1) k (x 2)n+ 2k
k =0 k!Γ( k + n + 1)
= (−1) J n ( x)
n
(19)

49
PHY312 MATHEMATICAL METHODS OF PHYSICS I

Thus for n = 0, 1, 2, 3, . . ., J −n ( x) does not represent the second

independent solution of eq. (1). The second independent solution will be
discussed later.

3.5 Recurrence Relations

The following are some very useful relations involving J n (x) :

xJ n′ ( x) = nJ n ( x) − xJ n +1 ( x) (20a)
= xJ n−1 ( x) − nJ n ( x) (20b)
Thus
2n
J n−1 ( x) + J n+1 ( x) = J n ( x) (21)
x
Also
d n
dx
[ ]
x J n ( x) = x n J n−1 ( x) (22)

In order to prove eq. (20a) w.r.t x to obtain

n + 2 r −1
∞
( n + 2r )  x  1
xJ n′ ( x) = ∑ (−1) r   x (23)
r =0 r!Γ(n + r + 1)  2  2
or
∞ n+ 2 r
1 x
xJ n′ ( x) = n∑ (−1) r  
r =0 r!Γ(n + r + 1)  2 
∞ n + 2 r −1
1 x
+ x ∑ (−1) r
 
r =0 (r − 1)!Γ(n + r + 1)  2 
∞ n + 2 r −1
1 x
= nJ n ( x) − x ∑ (−1) r
  (24)
r =0 r!Γ(n + r + 1)  2 
or
xJ n′ ( x) = nJ n ( x ) − xJ n+1 ( x) (25)

Which proves eq. (20a). eq. (23) can also be written as

∞
2(n + r )( x / 2) n + 2 r −1 x
xJ n′ ( x) = ∑ (−1) r
r =0 r!Γ(n + r + 1) 2
∞
( x / 2) n + 2 r
− n∑ (−1) r
r =0 r!Γ(n + r + 1)
∞ n −1+ 2 r
( x / 2)
= x ∑ (−1) r − nJ n ( x)
r =0 r!Γ(n + r )
or
xJ n′ ( x) = xJ n−1 ( x ) − nJ n ( x) (26)

50
PHY312 MATHEMATICAL METHODS OF PHYSICS I

Which proves eq. (20b). From eq. (25) we readily obtain

d −n
dx
[ ]
x J n ( x) = x − n J n+1 ( x) (27)

Further, adding eqs. (25) and (26) we get

J n−1 ( x ) − J n+1 ( x ) = 2 J n′ ( x) (28)

Using eq. (21) we may write

2
J 2 ( x) = J 1 ( x) − J 0 ( x) (29)
x
4
J 3 ( x ) = J 2 ( x) − J 1 ( x)
x
 8  4
=  2 − 1 J 1 ( x ) − J 0 ( x ) (30)
x  x
6
J 4 ( x) = J 3 ( x ) − J 2 ( x)
x
 48 8   24 
=  3 −  J 1 ( x) −  2 + 1 J 0 ( x ) (31)
x x x 
etc.

The proof of eq. (22) is simple

d n
dx
[ ]
x J n ( x) = x n J n′ ( x) + nx n−1 J n ( x)

 n 
= x n  J n−1 ( x) − J n ( x)  + nx n−1 J n ( x)
 x 
= x J n−1 ( x)
n
[Using eq. (20b)] (32)

Now using eq. (20a)

J 0′ ( x) = − J 1 ( x) (33)

Therefore
∫ J ( x)dx = − J
1 0 ( x) + Constant (34)
or
∞
∫
0
J 1 ( x )dx = 1 [Because J 0 (0) = 1 ] (35)

Equation (32) gives us

∫x J n−1 ( x) dx = x n J n ( x)
n
(36)

51
PHY312 MATHEMATICAL METHODS OF PHYSICS I

Example 2

In this example we will evaluate the integral

∫x
4
J 1 ( x)dx
in terms of J 0 ( x) and J 1 ( x) . Since
d p
dx
[ ]
x J p ( x) = x p J p −1 ( x) [see eq. (22)]
we have
∫x J p −1 ( x )dx = x p J p ( x)
p

Thus
∫x
4
[
J 1 ( x) dx = ∫ x 2 x 2 J 1 ( x) dx ]
= x2 [x J2
2 ( x)]− ∫ 2 x J
3
2 ( x) dx
= x 4 J 2 ( x) − 2 x 3 J 3 ( x)
4 
= x 4 J 2 ( x) − 2 x 3  J 2 ( x) − J 1 ( x)
x 
 2 
= ( x 4 − 8 x 2 ) J 1 ( x) − J 0 ( x)  + 2 x 3 J 1 ( x)
x 
( )
= 4 x 16 x J 1 ( x) − ( x 4 − 8 x 2 ) J 0 ( x)
3

plus, of course, a constant of integration.

3.4 The Generating Function

Bessel functions are often defined through the generating function G(z,t)
which is given by the following equation

 z  1 
G ( z, t ) = exp   t −  (37)
 2  t 

For every finite value of z, the function G(z,t) is a regular function of t

for all (real or complex) values of t except at point t = 0. Thus it can be
expanded in a Laurent series

 z  1  +∞
exp   t −  = ∑ t n J n ( z ) (38)
 2  t  n = −∞

In the above equation, the coefficient of t n is defined as J n (z ) ; we will

presently show that this definition is consistent with series given by eq.
(3). Now, for any finite value of z and for 0 < t < ∞ we may write

52
PHY312 MATHEMATICAL METHODS OF PHYSICS I

n
 zt  ∞ 1  zt 
exp   = ∑  
 2  n =0 n!  2 
2 3
z 1  z  t2  z  t3
= 1+ +  +  +. . . (39)
2 1!  2  2!  2  3!
and
 z  ∞ (−1)  z 
n n

exp −  = ∑  
 2t  n=0 n!  2t 
2 3
z1 z 1 z 1
= 1− +  −  +. . . (40)
2 t  2  2!t 2  2  3!t 3
Thus the generating function can be expressed as a series of the form

 z  1  +∞
G ( z , t ) = exp   t −  = ∑ An ( z )t n (41)
 2  t  n= −∞
or
+∞  z 1  z  2 t 2  z 3 t 3 
∑ An ( z )t = 1 +
n
+  +  + . . .
n = −∞  2 1!  2  2!  2  3! 
 z 1  z 2 1  z 3 1 
× 1 − +  2
−  3
+ . . . (42)
 2 1!t  2  2!t  2  3!t 

On the other hand, the coefficient of t 0 will be given by

2 4 6
z 1 z 1 z 1
A0 ( z ) = 1 −   2
+  2
−  2
+.. . (43)
 2  (1!)  2  (2!)  2  (3!)

Comparing the above equation with eq. (8), we find

A0 ( z ) = J 0 ( z )

Similarly, the coefficient of t n on the right hand side of eq. (42) will be
given by
n n+2 n+ 4
z 1 z 1 z 1
An ( z ) =   +  +  −. . .
 2  n!  2  (n + 1)!1!  2  (n + 2)!

which when compared with eq. (7) gives us

An ( z ) = J n ( z )
Proving
 z  1   +∞
exp   t −   = ∑ t n J n ( z )
 2  t   n= −∞
In the above equation, if we replace t by -1/y, we obtain
z  1  + ∞ +∞
exp   y −  = ∑ (−1) n y − n J n ( z ) = ∑ y n J n ( z )
2  y  n= −∞ n = −∞

53
PHY312 MATHEMATICAL METHODS OF PHYSICS I

Thus
J n ( z ) = (−1) n J −n ( z )

3.8.1 Derivation of the Recurrence Relations from the

Generating Function

Differentiating eq. (38) w.r.t z, we obtain

1  1  z  1   +∞ n
 t −  exp   t −   = ∑ t J n′ ( z ) (44)
2 t   2  t   n= −∞
Thus
+∞ +∞ +∞

∑t
n = −∞
n +1
J n ( z ) − ∑ t n−1 J n ( z ) =
n = −∞
∑t
n = −∞
n
2 J n′ ( z )

Comparing the coefficients of t n , we obtain

J n−1 ( z ) − J n+1 ( z ) = 2 J n′ ( z )

Similarly, if we differentiate eq. (38) w.r.t t we will obtain

z 1  +∞ n +∞
1 + 2  ∑ t J n ( z ) = ∑ nt J n ( z )
n −1

2  t  n= −∞ n = −∞

Comparing the coefficients of t n −1 , we get

z[J n−1 ( z ) − J n+1 ( z )] = 2nJ n ( z )

3.9 Some Useful Integrals

cos[x sin θ − nθ ]dθ

1 π

π∫
Using J n ( z ) =
0

2 π /2
cos( x sin θ )dθ
π∫
J 0 ( z) = (45)
0

Thus
∞ 2  ∞ −αx e ix sin θ + e − ix sin θ 
π /2
∫ e −αx J 0 ( x)dx = dx  dθ
π ∫0  ∫0
e
0 2 
1 π / 2  1 1 
= ∫ 
π
+  dθ
0
α − i sin θ α + i sin θ 
2α π / 2 dθ
= ∫
π 0 α + sin 2 θ
2
(46)
or
∞ 1
∫ e −αx J 0 ( x)dx = (47)
0
1+α 2

54
PHY312 MATHEMATICAL METHODS OF PHYSICS I

where in evaluating the integral on the right hand side of eq. (46), we
have used the substitution y = α cot θ . By making α → 0 , we get
∞
∫0
J 0 ( x)dx = 1 (48)

From eq. (28), we have

2 J n′ ( x) = J n−1 ( x) − J n+1 ( x)
Thus
∞ ∞ ∞
2 ∫ J n′ ( x) = ∫ J n−1 ( x) − ∫ J n+1 ( x)
0 0 0

But
∞ ∞
∫ J n′ ( x) = J n ( x) 0
0
= 0 for n > 0
Thus
∞ ∞
∫0
J n+1 ( x) = ∫ J n−1 ( x)
0
n>0 (49)
Since
∞
∫0
J 1 ( x) = 1 [see eq. (35)]
and
∞
∫0
J 0 ( x) = 1 [see eq. (48)]

Using eq. (49), we have

∞
∫0
J n ( x) = 1 n = 0, 1, 2, 3, ... (50)

Replacing α by α + iβ in eq. (47), we get

∞ 1
∫ e −(α +iβ ) x J 0 ( x )dx = (51)
0
(α + iβ ) 2 + 1
which in the limit of α → 0 becomes
∞ 1
∫ e −iβx J 0 ( x)dx = (52)
0
1− β 2

For β < 1, the right hand side is real and we have

∞ 1
∫ J 0 ( x) cos βxdx = (53)
0
1− β 2
and
∞
∫ J 0 ( x) sin β xdx = 0
0

Similarly, β > 1, the right hand side of Eq. (52) is imaginary and we have
∞
∫ J 0 ( x) cos βxdx = 0
0

55
PHY312 MATHEMATICAL METHODS OF PHYSICS I

∞ 1
∫ J 0 ( x) sin βxdx = (54)
0
1− β 2

3.10 Spherical Bessel Functions

1
We start with the Bessel equation eq. (1)] with n = l + , i.e.
2
dy   1  
2
d2y
x 2
+ x  x −  l +   y ( x) = 0 (55)
dx 2 dx   2  
where
l = 0, 1, 2, . . .

The solutions of eq. (55) are

J 1 ( x) and J 1 ( x)
l+ −l −
2 2

If we make the transformation

1
f ( x) = y ( x) (56)
x
we would readily obtain
1 d  2 df   l (l + 1) 
x  + 1− f ( x) = 0 (57)
x 2 dx  dx   x 2 

The above equation represents the spherical Bessel equation. From eqs.
(55) and (56) it readily follows that the two independent solutions of
eq.(57) are

1 1
J 1 ( x ) and J 1 ( x)
x 2
l + x −l − 2

The spherical Bessel functions are defined through the equations

π
jl ( x) = J 1 ( x) (58)
2x l+
2
and
π
nl ( x) = (−1) l J 1 ( x) (59)
2x −l −
2

and represent the two independent solutions of eq. (57). Now, if we

define the function
u(x) = xf(x)

then eq. (57) takes the form

d 2 u  l (l + 1) 
+ 1− u ( x) = 0 (60)
dx 2  x 2 

56
PHY312 MATHEMATICAL METHODS OF PHYSICS I

The above equation also appears at many places and the general solution
is given by
u ( x) = c1[xJ l ( x)] + c2 [xnl ( x)] (61)
which also be written in the form
   
u ( x) = A1  x J 1 ( x) + A2  x J 1 ( x) (62)
l+ −l −
 2   2 
For l = 0 , the solutions of eq. (60) are
sin x and cos x

Thus, for l = 0 the two independent solutions of eq.(57) are

sin x cos x
and
x x

Indeed if we use the definitions of jl (x) and nl (x) given eqs. (58) and (59)
respectively, we would readily obtain
sin x
j0 ( x) = (63)
x
cos x
n0 ( x) = (64)
x
sin x cos x
j1 ( x) = 2 − (65)
x x
cos x sin x
n1 ( x ) = − etc (66)
x2 x

Further, if we multiply the recurrence relation [Eq. (21)]

2n
J n +1 ( x ) = J n ( x) − J n −1 ( x)
x
π l
by and assume n = l − , we would get
2x 2
(2l − 1)
jl ( x) = nl −1 ( x) − nl − 2 ( x ) (67)
x
using which we can readily obtain analytic expression for j2 ( x), j3 ( x), . ..
etc.
Similarly,
(2l − 1)
nl ( x) = nl −1 ( x) − nl − 2 ( x) (68)
x

3.11 Bessel Functions of the Second Kind: Yn

The Bessel functions of the second kind, denoted by Yn (x) , are solutions
of the Bessel differential equation. They have a singularity at the origin
(x = 0). Yn (x) is sometimes also called the Neumann function. For non-
integer n, it is related to J n (x ) by:

57
PHY312 MATHEMATICAL METHODS OF PHYSICS I

J n ( x ) cos µπ − J − n ( x)
Yn ( x) =
sin µπ
(69)
or
1 ∂ ∂ 
Yn ( x) =  J µ ( x ) − ( −) n J − µ ( x) (70)
n  ∂µ ∂µ  µ =n

We need to show now that Yn (x ) defined by eq.(70) satisfies Eq.(1)

where n is either zero or an integer. We know that

1  µ2 
J µ′′ ( x) + J µ′ ( x) + 1 − 2  J µ ( x) = 0 (71)
x  x 
for any value of µ . Differentiating the above equation with respect to µ ,
we get
d 2 ∂J µ ( x) 1 d ∂J µ ( x)  µ 2  ∂J µ ( x) 2 µ
+ + 1 − 2  = 2 J µ ( x) (72)
dx 2 ∂µ x dx ∂µ  x  ∂µ x

Similarly

d 2 ∂J − µ ( x ) 1 d ∂J − µ ( x)  µ 2  ∂J − µ ( x) 2 µ
+ + 1 − 2  = 2 J − µ ( x) (73)
dx 2 ∂µ x dx ∂µ  x  ∂µ x
From eqs. (72) and (73), it is easy to show that
d2 1 d  µ2 
S µ ( x) + S µ ( x) + 1 − 2  S µ ( x)
dx 2 x dx  x 
2µ
[
= 2 J µ ( x) − (−1) n J − µ ( x)
x
] (74)
where
∂ ∂
S µ ( x) = J µ ( x) − (−1) n J −µ ( x) (75)
∂µ ∂µ

Thus Yn (x ) is the second solution of Bessel’s equation for all real values
of n and is known as the Bessel function of the second kind of order n.
The general solution of eq.(1) can, therefore, be written as

y = C1 J n ( x) + C 2Yn ( x) (76)
where C1 and C 2 are arbitrary constants.

The expression for Yn (x ) for n = 0, 1, 2, . . . can be obtained by using eqs.

(2) and (70) and is given below

58
PHY312 MATHEMATICAL METHODS OF PHYSICS I

− n n −1 r
1  x (n − r − 1)!  x 2 
Yn ( x) = (ln( x / 2) + γ )J n ( x) −  
2
π π 2
∑ r!
 
r =0  4 
−n ∞
1  x ( x 2 / 4) r
−   ∑ (−1) r [ϕ (r ) + ϕ (r + n)] (77)
π 2 r =0 r!(n + r )!
m
Where ϕ (r ) = ∑ s −1 ; ϕ (0) = 0
s =1

and γ = Lim[ϕ (n) − ln n]

n →∞

Example 3

In this example we will solve the radial part of the Schrodinger equation
1 d  2 dR   2 µE l (l + 1) 
2 r + − 2  R ( x) = 0; l = 0, 1, 2, (78)
r dr  dr   h r 
in the region 0 < r < a subject to the following boundary conditions that
R (a) = 0 (79)

and R (r ) is finite in the region 0 < r < a . Equation (78) can be

conveniently written in the form

1 d  2 dR   l (l + 1) 
ρ  + 1 −  R( ρ ) = 0
ρ 2 dρ  dρ   ρ 2 
Where
ρ = kr; k = (2 µE / h 2 )1 / 2

Thus the general solution of the above equation is given by

R ( ρ ) = Aj l ( ρ ) + Bnl ( ρ ) (80)

But nl ( ρ ) diverges at ρ = 0 , therefore, we must choose B=0. The

boundary condition R(a)=0 leads to the transcendental equation
jl ( ka) = 0 (81)

Thus, for l = 0 , we have

ka = nπ ; n = 1, 2,. . . . (82)

Which will give allowed values of k. Similarly, for l = 1 , we get

tan ka = ka (83)

3.12 Modified Bessel Functions

If we replace x by ix in eq. (1), we obtain

x 2 y ′′ + xy ′ − ( x 2 + n 2 ) y = 0 (84)

59
PHY312 MATHEMATICAL METHODS OF PHYSICS I

The two solutions of the above equation will obviously be

J n (ix ) and Yn (ix)

As these functions are real for all values of n, let us define a real
function as

I n ( x) = i − n J n (ix) (85)

or
∞
( x / 4) n + 2 r
I n ( x) = ∑ (86)
r = 0 r! ( n + r + 1)!

This function will be the solution of eq. (84) and is known as the
Modified Bessel function of the first kind. For very large values of x
ex
I n ( x) ~ (87)
2πx

The other solution known as the Modified Bessel function of the second
kind is defined as

π I −n ( x) − I n ( x)
K n ( x) = (88)
2 sin nπ

For non-integer values of n, I n and I −n are linearly independent and as

such K n (x) is a linear combination of these functions [compare with eq.
(69) which gives the definition of Yn (x)] . When n is an integer, it can be
shown [see eq. (86)] that

I −n = I n (89)
and therefore K n (x) becomes indeterminate for n = 0 or an integer. As
in the case of Yn (x ) for n = 0 or an integer, we define K n (x) as
 π I − µ ( x) − I µ ( x) 
K n ( x) = Lim   (90)
µ →n 2
 sin µπ 
or
( −1) n  ∂I − µ ( x) ∂I µ ( x) 
K n ( x) =  −  (91)
2  ∂µ ∂µ  µ = n
For x very large
π −x
K n ( x) ~ e (92)
2x
From eq. (88) it follows that
K − n ( x) = K n ( x) (93)

60
PHY312 MATHEMATICAL METHODS OF PHYSICS I

Which is true for all values of n. recurrence relations for I n can be

derived from those of J n (x ) and Eq. (85). They are

xI n′ ( x ) = xI n −1 ( x) − nI n ( x)
(94) xI n′ ( x ) = nI n ( x) + xI n +1 ( x)
95) I n −1 ( x) + I n +1 ( x) = 2 I n′ ( x) (96)
and similarly
xK n′ = ( nK n − xK n −1 ) (97)
xK n′ = nK n + xK n +1 (98)
K n −1 + K n +1 = −2 K n′ (99)

Example 4

In this example we will consider the solutions of the equation

r2
d 2R
dr 2
+r
dR
dr
[( ) ]
+ k 02 n 2 (r ) − β 2 r 2 − l 2 R (r ) = 0 l = 0, 1, . . . (100)
n(r ) = n1 0<r<a
Where (101)
= n2 r>a
and n2 < n1; k 0 (ω / c) represents the free space wave number. The
quantity β represents the propagation constant and for guided modes
β 2 takes discrete values in the domain
k 02 n 22 < β 2 < k 02 n12 (102)

Thus, in the regions 0 < r < a and r > a , eq. (100) can be written in the
form

d 2R dR  2 r 2 
r 2
2
+r + U 2 − l 2  R(r ) = 0 0 < r < a. (103)
dr dr  a 
and
d 2R dR  2 r 2 
r2 2
+ r + W 2 + l 2  R (r ) = 0 r > a. (104)
dr dr  a 
where
[
U 2 = a 2 k 02 n12 − β 2 ] (105)
and
[
W 2 = a 2 β 2 − k 02 n22 ] (106)
so that
V 2 = U 2 + W 2 = a 2 k 02 (n12 − n 22 ) (107)
is a constant. The solutions of Eq. (103) are
 r  r
J l U  and Yl  U  (108)
 a  a

61
PHY312 MATHEMATICAL METHODS OF PHYSICS I

and the latter solution has to be rejected as it diverges at r = 0. Similarly,

the solutions of eq. (104) are

 r  r
K l W  and I l W 
 a  a
and the second solution has to be rejected because it diverges as
r → ∞ . Thus
 A  r
 J (U ) J l U  0<r<a
  a
and R(r ) =  l (109)
 A  r
K l W  r>a
 K l (W )  a
where the constants have been so chosen and R(r) is continuous at r = a.
Continuity of dR/dr at r = a gives us

J l′ (U ) K ′ (U )
U = WU l (110)
J l (U ) K l (U )

which is the fundamental equation determining the eigenvalues β / k 0 .

SELF-ASSESSMENT EXERCISE

3. Using
J 0 (2) = 0.22389, J 1 ( 2) = 0.57672, calculate J 2 ( 2), J 3 (2), and J 4 ( 2).

Hint: Use Eq. (21)

4. Show that
a 1 2 2  J n−1 (a ) J n +1 (a ) 
∫0
J n2 ( x)xdx =
2
a J n (a ) 1 −
 J n2 (a )



4.0 CONCLUSION

In this unit, we have considered Bessel function and spherical Bessel

function.

We have also established in this unit, relationship between the

recurrence relation and the generating function.

5.0 SUMMARY

This unit is on Bessel functions. It has a lot of application that arises in

numerous diverse areas of applied mathematics. This unit will be of
significant importance in the subsequent course in quantum mechanics.

62
PHY312 MATHEMATICAL METHODS OF PHYSICS I

6.0 TUTOR- MARKED ASSIGNMENT

1. Using

J 1 ( 2) = 0.57672, J 2 ( 2) = 0.35283 calculate J 3 ( 2), J 4 ( 2), and J 5 ( 2).

Hint: Use Eq. (21)

2. Using the integral

1 Γ(n + r + 1)Γ(m + 1)
∫ (1 − x ) x 2 n + 2 r +1 dx =
2 m
; m > −1, n > −1
0 2Γ(m + n + r + 2)
Prove that
m +1
2 x 1
J n + m +1 ( x) =   ∫ (1 − y ) y n +1 J n ( xy )dy
2 m

Γ(m + 1)  2  0

3. Hint: Use the expansion given by eq. (2) and integrate term by
term.

1
In problem 2 assume m = n = − , and use eq. (12) to deduce
2
2 1 cos xy
π∫
J 0 ( x) = dy
0
1− y2

4. Show that the solution of the differential equation

y ′′( x) + (ae x − b) y ( x) = 0
is given by y ( x) = AJ µ (ξ ) + BJ µ (ξ ); ξ = 2 ae x / 2 ; µ = 2 b

7.0 REFERENCES/FURTHER READING

Erwin, Kreyszig (1991). Advanced Engineering Mathematics. John

Wiley & Sons, Inc.

Arfken, G. (1990). Mathematical Methods for Physicists. New York:

Academic Press

63
PHY312 MATHEMATICAL METHODS OF PHYSICS I

MODULE 3 APPLICATION OF FOURIER TO PDES

(HERMITE POLYNOMIALS AND
LAGUERRE POLYNOMIALS)

Unit 1 Hermite Polynomials

Unit 2 Laguerre Polynomials

UNIT 1 HERMITE POLYNOMIALS

CONTENTS

8.0 Introduction
9.0 Objectives
10.0 Main Content
10.1 Hermite Differential Equation
10.2 The Generating Function
10.3 Rodrigues Formula
10.4 Orthogonality of Hermite Polynomials
10.5 The Integral Representation of the Hermite Polynomials
10.6 Fourier Transform of Hermite-Gauss Functions
10.7 Some Important Formulae Involving Hermite Polynomials
11.0 Conclusion
12.0 Summary
13.0 Tutor-Marked Assignment
14.0 References/Further Reading

4.0 INTRODUCTION

In this unit, we shall consider certain boundary value problems whose

solutions form orthogonal set of functions. It can also be seen in this unit
how the generating function can readily be used to derive the Rodrigues’
formula.

5.0 OBJECTIVES

At the end of this unit, you should be able to:

• define Hermite polynomials as the polynomial solutions of the

Hermite differential equation
• prove the Orthogonality of Hermite polynomials
• derive the Rodrigues’ formula which can be used to obtain
explicit expressions for Hermite polynomials
• solve the exercises at the end of this unit.

64
PHY312 MATHEMATICAL METHODS OF PHYSICS I

6.0 MAIN CONTENT

3.1 Hermite Differential Equation

The equation

y ′′( x) − 2 xy ′( x) + (λ − 1) y ( x) = 0 (1)
where λ is a constant is known as the Hermite differential equation.
When λ is an odd integer, i.e. when
λ = 2n + 1; n = 0, 1, 2, . . . (2)

One of the solutions of eq. (1) becomes a polynomial. These polynomial

solutions are called Hermite polynomials. Hermite polynomials appear
in many diverse areas, the most important being the harmonic oscillator
problem in quantum mechanics.

Using Frobenius method to solve eq.(1), and following the various steps,
we have

Step1: We substitute the power series

∞
y ( x) = ∑ C r x p + r (3)
r =0

in eq. (1) and obtain the identity

C 0 p ( p − 1) + C1 ( p + 1) px +
∞

∑ [C
r =2
r ( p + r )( p + r − 1) −C r − 2 ( 2 p + 2r − 3 − λ )] x r = 0

Step 2: Equating to zero the coefficients of various powers of x, we

obtain

(i) p=0 or p =1 (4a)

(ii) p ( p + 1)C1 = 0 (4b)
2 p + 2r − 3 − λ
(iii) Cr = C r −2 for r ≥ 2 (4c)
( p + r )( p + r − 1)

When p = 0, C1 becomes indeterminate; hence p = 0 will yield both the

linearly independent solutions of eq. (1). Thus, we get

2r − 3 − λ
Cr = C r −2 for r ≥ 2 (5)
r (r − 1)
which gives
1− λ
C2 = C0 .
2!
3−λ
C3 = C1 , . .
3!

65
PHY312 MATHEMATICAL METHODS OF PHYSICS I

(1 − λ )(5 − λ )
C4 = C0
4!
(3 − λ )(7 − λ )
C5 = C1 , . . . etc
5!

Because C 2 , C 4 ,. . . . are related to C 0 and C 3 , C 5 , . . . are related to

C1 , we can split the solution into even and odd series. Thus, we may
write

y ( x) = (C 0 + C 2 x 2 + C 4 x 4 + . . .) + (C1 x + C 3 x 3 + + . . .)
 1 − λ 2 (1 − λ )(1 − λ ) 4 
= C 0 1 + x + x + . . .
 2! 4! 
 (3 − λ ) (3 − λ )( 7 − λ ) 
+ C1  x + x3 + x5 + . . . (6)
 3 ! 5! 

It may be readily seen that when

λ = 1, 5, 9, . . .
the even series becomes a polynomial and the odd series remains an
infinite series. Similarly, for
λ = 3, 7, 11, . . .
the odd series becomes a polynomial and the even series remains an
infinite series. Thus, when

λ = 2n + 1; n = 0, 1, 2, . . .

One of the solutions becomes a polynomial. If the multiplication

constant C 0 or C1 is chosen that the coefficient of the highest power of x
in the polynomial becomes 2 n , then these polynomials are known as
Hermite polynomials of order n and are denoted by H n (x) . For example,
for λ = 9 (n = 4) , the polynomial solution
 4 
y ( x) = C 0 1 − 4 x + x 4 
 3 
If we choose
C 0 = 12
the coefficient of x 4 becomes 2 4 and, therefore
H 4 ( x) = 16 x 4 − 48 x 2 + 12

Similarly,

for λ = 7 (n = 3) , the polynomial solution is given by

 2 
y ( x) = C1  x − x 3 
 3 

66
PHY312 MATHEMATICAL METHODS OF PHYSICS I

Choosing
C1 = −12
we get
H 3 ( x ) = 8 x 3 − 12 x

In general

n!(2 x) n − 2 r
N
H n ( x) = ∑ (7)
r = 0 r!( n − 2 r )!

where
n
 2 if n is even
N =
n −1 if n is odd
 2

Using eq. (7) one can obtain Hermite polynomials of various orders, the
first few are given below:

H 0 ( x ) = 1; H 1 ( x) = 2 x; H 2 ( x ) = 4 x 2 − 2; 

H 3 ( x) = 8 x 3 − 12 x; H 4 ( x) = 16 x 4 − 48 x 2 + 12  (8)

Higher order Hermite polynomials can easily be obtained either by using

eq. (7) or by using the recurrence relation (see eq. 20)

3.5 The Generating Function

The generating function for Hermite polynomials is given by

∞
1
G ( x, t ) = e − t + 2 xt
=∑
2
H n ( x)t n (9)
n =0 n!

Expanding e −t and e 2 xt in power series, we have

2

1 4 1 6
e −t = 1 − t 2 + t − t +. . .
2

2! 3!
(2 x ) 2 2 ( 2 x ) 3 3
e 2 xt = 1 + (2 x )t + t + t +. .
2! 3!

Multiplying the above two series, we shall obtain a power series in t

with

67
PHY312 MATHEMATICAL METHODS OF PHYSICS I

1
Coefficient of t 0 = 1 = H 0 ( x)
0!
1
“ “ t=2 = H 1 ( x)
1!
1
“ “ t 2 = 2x 2 − 1 = H 2 ( x) etc
2!

It is also evident that the coefficient of t 2 in the multiplication of the two

series will be a polynomial of degree n and will contain odd powers
when n is odd and even powers when n is even. In this polynomial, the
coefficient of x n can easily be seen to be (2 n / n!) . We then assume that

∞
1
G ( x, t ) = e − t + 2 xt
=∑
2
K n ( x)t n (10)
n=0 n!

Where K n (x) is a polynomial of degree n. Differentiating eq. (10) with

respect to t, we get
∞ ∞
n 1
( 2 x − 2t )e −t + 2 xt
=∑ K n ( x)t n −1 = ∑ K n ( x)t n −1
2

n = 0 n! n = 0 ( n − 1)!

or
∞ ∞
1 1
2( x − t ) ∑ K n ( x)t n = ∑ K n +1 ( x)t n (11)
n = 0 n! n = 0 n!

Comparing the coefficients of t n on both sides of eq. (11), we obtain

2 xK n ( x) − 2nK n −1 ( x) = K n +1 ( x ) (12)

We next differentiate eq.(10) with respect to x to obtain

∞ ∞
1 1
2t ∑ K n ( x)t n = ∑ K n′ ( x)t n (13)
n = 0 n! n = 0 n!

Comparing the coefficients of t n on both sides of eq. (11), we get

K n′ ( x) = 2nK n −1 ( x) (14)

If we replace n by (n+1) in eq.(14), we would get

K n′ +1 ( x) = 2(n + 1) K n ( x) (15)

Differentiating eqs.(14) and (12) with respect to x, we obtain

respectively

K n′′ ( x) = 2nK n′ −1 ( x) (16)

and

68
PHY312 MATHEMATICAL METHODS OF PHYSICS I

2 xK n′ ( x) + 2 K n ( x) − 2nK n′ −1 ( x) = K n′ +1 ( x) (17)

Subtracting eqs. (17) and (16) and using (15), we get

K n′′ ( x ) − 2 xK n′ ( x) + 2nK n ( x) = 0 (18)

which shows that K n (x) is a solution of the Hermite equation (1) with
λ = 2n + 1 , i.e. of the equation
y ′′( x) − 2 xy ′( x) + 2ny ( x) = 0 (19)

Since, as discussed before, K n (x) is also a polynomial of degree n (with

coefficient of x n equal to 2 n ), K n (x) is, therefore, nothing but H n (x) .
Equations (12) and (14), thus, give recurrence relations for H n (x)
2 xH n ( x) = 2nH n −1 ( x) + H n +1 ( x) (20)
and
H n′ ( x) = 2nH n −1 ( x) (21)
3.6 Rodrigues Formula

In the preceding section we have shown that

∞
1
G ( x, t ) = e − t + 2 xt
=∑
2
H n ( x)t n (22)
n = 0 n!

One can rewrite the generating function G(x, t) in the form

G ( x, t ) = e x e − ( t + x )
2 2

It may be easily seen that

∂ nG n ∂
n
= x2
− e −(t − x )
2
e ( 1) (23)
∂t n
∂x n

From eq. (22) it follows that

∂ nG
= H n ( x) (24)
∂t n t =0

Using eqs (23) and (24), we obtain

d n − x2
H n ( x) = (−1) n e x
2
e (25)
dx n
which is known as Rodrigues formula for Hermite polynomials. For
example,
d 2 −x2 2 d
H 2 ( x) = e x e = ex (−2 xe − x )
2 2

2
dx dx
[
− x2
= e − 2e + 4 x e
x2 2 − x2
]
= 4x 2 − 2

69
PHY312 MATHEMATICAL METHODS OF PHYSICS I

Which is consistent with eq. (8). Similarly, we can determine other

Hermite polynomials by elementary differentiation of eq. (25).

3.7 Orthogonality of Hermite Polynomials

The Hermite polynomials satisfy eq.(1) for λ = 2n + 1 . Thus, we have

d 2Hn dH n
2
− 2x + 2nH n ( x) = 0 (26)
dx dx

In order to derive the Orthogonality condition we transform eq. (26) to

the Sturm-Liouville form by multiplying it by

[
exp − ∫ 2 xdx = e − x ] 2
(27)
to obtain
d  − x 2 dH n 

dx 
e
dx  = −2n e − x H n ( x )
2
[ ] (28)

Similarly
d  − x 2 dH m 

dx 
e
dx  = −2m e − x H m ( x)
2
[ ] (29)

We multiply eq.(28) by H m (x) and eq.(29) by H n (x) , subtract them and

integrate the resulting equation with respect to x from − ∞ to ∞ to obtain

+∞ d  2 dH n  d  2 dH m  
∫  H m ( x ) e − x  − H n ( x ) e − x dx
−∞
 dx  dx  dx  dx  
+∞
= 2(m − n) ∫ e − x H m ( x) H n ( x)dx
2

−∞

Now
+∞ d   − x 2 dH n   2 dH m  
LHS= ∫  H m ( x) e  − H n ( x ) e − x dx
−∞ dx   dx   dx  
+∞
 2 dH 2 dH
m 
=  H m ( x )e − x n
− H n ( x )e − x
 dx dx  −∞
=0
Thus
+∞
∫ e − x H m ( x) H n ( x)dx = 0; m≠n
2
(30)
−∞

which shows that the Hermite polynomials are Orthogonal with respect
to the weight function e − x . Thus if we define the functions
2

φ n ( x) = N n e − x / 2 H n ( x ); n = 0, 1, 2, . . .
2
(31)
then eq. (30) assumes the form
+∞
∫ φ m ( x )φ n ( x )dx = 0; m≠n (32)
−∞

70
PHY312 MATHEMATICAL METHODS OF PHYSICS I

3.5 The Integral Representation of the Hermite Polynomials

The integral representation of the Hermite polynomial is given by

2 n ( −i ) n +∞
H n ( x) = ∫ t n e −t + 2 ixt
2 2
ex dt (33)
π −∞

In order to prove the above relation we start with the relation

1 +∞
e−x = ∫ e −t + 2 ixt
2 2 2
ex dt
π −∞

which can easily be obtained from the well known formula

+∞
−αt + βt π β 2 
∫−∞ =
2
e dt exp  
α  4α 
by assuming α = 1 and β = 2ix . Now according to the Rodrigues formula
d n − x2
H n ( x) = (−1) n e x
2
e
dx n
1 d n + ∞ −t 2 + 2ixt
π dx n ∫−∞
= (−1) n e x
2
e dt

1 x 2 +∞
= (−1) n e ∫ (2i ) n t n e −t + 2ixt dt
2

π −∞

from which eq. (50) readily follows.

3.6 Fourier Transform of Hermite-Gauss Functions

In this section we will show that

e−x
2
/2
H n ( x) =
i n
1
2π
∫ −∞
+∞
[e k2 / 2
]
H n (k ) e ikx dk (34)

Implying that the Fourier transform of the Hermite-Gauss function is a

Hermite-Guass function. In order to prove eq. (34) we start with the
generating function

∞
1
G ( x, t ) = e 2 kt −t = ∑
2
H n ( k )t n
n = 0 ,1, . . . n!

 1 
We multiply the above by  ikx − k 2  and integrate over k to obtain
 2 
∞
+∞  1 2  t n +∞
e −t ∫−∞  2
− + + = ∑ ∫ H n ( x)e − k / 2 e ikx dk
2 2
e k ( 2t ix ) k  dk (35)
−∞
n = 0 n!

71
PHY312 MATHEMATICAL METHODS OF PHYSICS I

Now
 (2t + ix ) 2 
LHS = e −t 2π exp 
2


 2 
= 2π e t + 2ixt
e−x
2 2
/2

H n ( x)
= 2π e − x ∑
2
/2
(it ) n
n n!

Comparing coefficients of t n on both sides of eq. (35), we get eq. (34).

3.7 Some Important Formulae Involving Hermite

Polynomials
n
n!
H n ( x + y ) = 2 −n / 2 ∑ H n − p ( x 2 ) H p ( y 2) (36)
p p!(n − p )!
nπ 
n/2
 2n  
H n ( x) → ( 2 n + 1) x −
2
2   e x / 2 cos  (37)
n→∞
  e  2 
(2 s)! s
H 2 s − 2 n ( x) 
x 2s = 2s ∑ 
2 n =0,1, ... n!(2 s − 2n)! 
 s = 0, 1, 2, . . . (38)
(2 s + 1)! s H 2 s +1− 2 n ( x) 
x 2 s +1
=
2 2 s +1 ∑
n = 0 , 1, . . . n! ( 2 s + 1 − 2n )!


SELF-ASSESSMENT EXERCISE

Using the generating function for H n (x) , show that

∞
1 1
(a) cosh 2 x = ∑ H 2 n ( x)
e n = 0 , 1, . . . ( 2 n)!
∞
1 1
(b) sinh 2 x = ∑ H 2 n+1 ( x)
e n = 0, 1, . . . ( 2n + 1)!
∞
1
(c) e cos 2 x = ∑
n = 0 , 1, . . .
(−1) n
(2n)!
H 2 n ( x)
∞
1
(d) e sin 2 x = ∑
n = 0 , 1, . . .
( −1) n
( 2n + 1)!
H 2 n +1 ( x)

Hint: To obtain (a) and (b) substitute t = 1 and t = -1in eq. (9) add and
subtract the resulting equations. Similarly for (c) and (d), substitute t = I
and equate real and imaginary parts.

Prove that
+∞ (2n)! 2
∫ e − x H 2 n (α , x)dx = n (α − 1) n
2

−∞ n!

72
PHY312 MATHEMATICAL METHODS OF PHYSICS I

Hint: Replace x by αy in eq.(9), multiply the resulting equation

by e − y and integrate with respect to y.
2

10.0 CONCLUSION

Here, in this unit, we have dealt with the Hermite polynomials which are
Orthogonal with respect to the weight function e − x . We have also
2

established that the Fourier transform of the Hermite-Gauss function is a

Hermite-Guass function.

11.0 SUMMARY

This unit was on the Hermite polynomials. It has a lot of application in

linear harmonic oscillator problem in quantum mechanics. The unit will
be of immense importance in the subsequent course in classical
mechanics.

12.0 TUTOR- MARKED ASSIGNMENT

1. If two operators are defined as

1  d 
a= x + 
2 dx 
1  d 
a= x− 
2 dx 
Show that
aφ n ( x) = nφ n −1 ( x)
a φ n ( x) = nφ n −1 ( x)

2. Prove that
+∞ 1  2
H n  x + x0 e − x / 2 dx = n x 0n
∫
−∞
 2 
  1  
2

Hint: Multiply Eq. (9) by −  x + x0   and integrate over x.

  2  

7.0 REFERENCES/FURTHER READING

Erwin, K. (1991). Advanced Engineering Mathematics. John Wiley &

Sons, Inc.

Arfken, G. (1990). Mathematical Methods for Physicists. New York:

Academic Press.

73
PHY312 MATHEMATICAL METHODS OF PHYSICS I

UNIT 2 LAGUERRE POLYNOMIALS

CONTENTS

1.0 Introduction
2.0 Objectives
3.0 Main Content
3.1 Laguerre Differential Equation
3.2 The Generating Function
3.3 Rodrigues Formula
3.4 Orthogonality of Hermite Polynomials
3.5 The Integral Representation of the Laguerre Polynomials
3.6 Some Important Results Involving Laguerre Polynomials
3.7 The Second Solution of the Laguerre Differential Equation
3.8 Associated Laguerre Polynomials
4.0 Conclusion
5.0 Summary
6.0 Tutor-Marked Assignment
7.0 References/Further Reading

1.0 INTRODUCTION

In the previous unit, you came across solutions of orthogonal set of

functions. This unit which is the last one in this book will examine
critically how a Laguerre differential equation can be transformed to
Sturm-Liouville form.

It shows that Laguerre polynomials and the associated functions arise in

many branches of physics, e.g. in the hydrogen atom problem in
quantum mechanics, in optical fibers characterised by parabolic
variation of refractive index, etc.

We also show that Laguerre polynomials are orthogonal in the interval

0 ≤ x ≤ ∞ with respect to the weight function e − x .

2.0 OBJECTIVES

At the end of this unit, you should be able to:

• use Frobenius method to obtain the polynomial solution of the

Laguerre differential equations
• determine the Orthogonality of the Laguerre polynomials
• derive the Rodrigues formula
• derive the second solution of the Laguerre differential equation.

74
PHY312 MATHEMATICAL METHODS OF PHYSICS I

3.0 MAIN CONTENT

3.1 Laguerre Differential Equation

The equation
xy ′′( x) − (1 − x) y ′( x) + ny ( x ) = 0
(1)
where n is a constant known as the Laguerre differential equation.
When n = 0, 1, 2, . . . (2)

One of the solutions of eq. (1) becomes a polynomial. These polynomial

solutions are known as the Laguerre polynomials.

Using Frobenius method to solve eq.(1), and following the various steps,
we have

Step We substitute the power series

∞
y ( x) = ∑ C r x p + r , C0 ≠ 0
r =0

Eq. (1) and obtain the identity

∞ ∞

∑C
r =0
r ( p + r ) 2 x p + r −1 − ∑ C r ( p + r − n) x p + r = 0
r =0

or
[ ]
∞
C 0 p 2 x p −1 − ∑ C r ( p + r ) 2 − C r −1 ( p + r − n − 1) x p + r −1 = 0 (3)
r =1

Step 2 Equating to zero the coefficients of various powers of x in the

identity (3), we obtain

(i) p2 = 0 INDICIAL EQUATION (4)

p + r − n −1
(ii) Cr = C r −1 r ≥ 1 RECURRENCE RELATION (5)
( p + r)2
Substituting p = 0 in eq. (5), we get
r − n −1
Cr = C r −1 r ≥1
r2
which gives
n n(n − 1)
C1 = − C0 C2 = C0
(1!) 2 (2!) 2
n(n − 1)(n − 2)
C3 = C0 etc
(3!) 2
n! ( −1) n
C n = ( −1) n =
( n!) 2 n!
and
C n+1 = C n + 2 = 0 . . . . = 0

75
PHY312 MATHEMATICAL METHODS OF PHYSICS I

Therefore one of the solutions of eq. (1) can be written as

 n n(n − 1) 2 n x 
n
y ( x) = C 0 1 − 2
x + x − . . . + ( − 1)  (6)
 (1!) (2!) 2 n! 
which is a polynomial of degree n. If the multiplication constant C 0 is
chosen to be unity so that the constant term becomes unity, the
polynomial solution given by eq. (6) is known as Laguerre Polynomial
of degree n and denoted by Ln (x) . Thus
n n( n − 1) 2 n x
n
Ln ( x) = 1 − x + x − . . . + ( −1)
(1!) 2 ( 2!) 2 n!
or
n
n!
Ln ( x ) = ∑ (−1) n xr (7)
r =0 (n − r )!(r!) 2
with
Ln (0) = 1 (8)

The first four Laguerre polynomials can be written as:

L0 ( x) = 1,
L1 ( x) = 1 − x,
1 2
L2 ( x) = 1 − 2 x + x , (9)
2
3 1
L3 ( x) = 1 − 3 x + x 2 − x 3 , . . .
2 6

Higher order polynomials can easily be obtained either by using eq.(7)

or by using the recurrence relation [see eq. (20)].

3.2 The Generating Function

The generating function for Laguerre polynomials is given by

1  xt  ∞
G ( x, t ) = exp −  = ∑ Ln ( x)t ; t <1
n
(10)
1− t  1 − t  n =0

We expand the left hand side of eq. (10) to obtain

 xt 
(1 − t ) −1 exp − 
 1− t 
x 2 t 2 (1 − t ) 3
= (1 − t ) −1 − xt (1 − t ) − 2 + − ..
2!
x 2t 2
= (1 + t + t 2 + . . .) − xt (1 + 2t + 3t 2 + . . .) + (1 + 3t + 6t 2 + . . .) − . . . (11)
2!

76
PHY312 MATHEMATICAL METHODS OF PHYSICS I

The right hand side of eq.(11) can be written as a power series in t with
Coefficient of t 0 = 1 = L0 ( x)
“ “ t =1− x = L1 ( x)
“ “ t = 1 − 2x + x / 2
2 2
= L2 ( x)

etc. It is also evident that the coefficient of t 2 on the right hand side of
eq.(11) will be a polynomial of degree n and that the constant term in
this polynomial will be unity. We can then assume that

1  xt  ∞
G ( x, t ) = exp−  = ∑ K n ( x)t
n
(12)
1− t  (1 − t )  n =0

where K n (x) is a polynomial of degree n. Differentiating eq. (12) with

respect to t, we get

(1 − x − t )  xt  ∞
 = ∑ nK n ( x)t
n −1
exp  −
(1 − t ) 3  (1 − t )  n = 0

or
∞ ∞
(1 − x − t )∑ K n ( x)t n = (1 − 2t + t 2 )∑ nK n ( x)t n−1
n=0 n =1

Comparing the coefficients of t n on both sides of the above equation,

we get
( n + 1) K n+1 ( x) − ( 2n + 1 − x) K n ( x ) + nK n−1 ( x) = 0; n ≥1 (13)

We next differentiate eq.(12) with respect to x to obtain

∞ ∞
− t ∑ K n ( x)t n = (1 − t )∑ K n′ ( x)t n (14)
n=0 n =0

Comparing the coefficients of t n on both sides of the above equation,

we get
K n′ ( x ) − K n′ −1 ( x) = − K n−1 ( x) (15)

If we replace n by (n+1) in the above equation, we would get

K n′ +1 ( x) = K n′ ( x) − K n ( x) (16)

Differentiating eq. (13) with respect to x, we obtain

(n + 1) K n′ +1 ( x) − (2n + 1 − x ) K n′ ( x ) + K n ( x) + nK n′ −1 ( x ) = 0 (17)

Substituting K n′ −1 ( x) and K n′ +1 ( x) from eqs. (15) and (16) respectively in

eq. (17), we get
xK n′ ( x) = nK n ( x) − nK n−1 ( x) (18)

77
PHY312 MATHEMATICAL METHODS OF PHYSICS I

Differentiating the above equation with respect to x and using eq. (15),
we have
xK n′′ ( x) + K n′ ( x ) = −nK n −1 ( x) (19)

Subtracting eq. (18) from eq. (19), we get

xK n′′( x) + (1 − x) K n′ ( x) + nK n ( x) = 0

Showing that K n (x) is a solution of the Laguerre equation, i.e. of the

equation
xy ′′( x) − (1 − x) y ′( x ) + ny ( x) = 0

Hence K n (x) nothing but Ln (x) . Equations (13) and (18) give the
following recurrence relations respectively:
( n + 1) Ln+1 ( x) = ( 2n + 1 − 1) Ln ( x) − nLn−1 ( x) (20)
xLn′ ( x) = nLn ( x) − nLn−1 ( x) (21)
We also have
Ln ( x ) = Ln′ ( x) − Ln′ +1 ( x) (22)

3.3 Rodrigues Formula

In the preceding section we have shown that

1  xt  ∞
G ( x, t ) = exp −  = ∑ Ln ( x)t
n

1− t  1 − t  n=0

We can write the above equation as

∞
1  x(1 − t − 1) 
∑ L ( x)t
n =0
n
n
=
1− t
exp −
 1 − t 
or
∞
 1  x 
∑ L ( x)t n
= ex  exp −  (23)
1 − t  1 − t 
n
n =0

Differentiating eq. (23) n times with respect to t and then putting t = 0,

we will have
 ∂n  1  x  
n! Ln ( x) = e x  n  exp −  
 ∂t 1 − t  1 − t    t =0
 ∂ n  ∞ (−1) r x r 
= e x  n ∑ r +1 
 ∂t  r =0 (1 − t ) r! t =0

78
PHY312 MATHEMATICAL METHODS OF PHYSICS I

∞ (r + 1)(r + 2) . . .(r + n) 
= e x ∑ (−1) r 
 r =0 (1 − t ) r + n+1 r!  t =0
∞
(n + r )! r
= e x ∑ (−1) r x
r =0 (r!) 2
or
ex ∞ ( n + r )! 2
Ln ( x) = ∑
n! r =0
( −1) r
( r!) 2
x (24)

d n n −x d2  n ∞ r x 
r
( x e ) =  ∑
x ( − 1) 
dx n dx 2  r =0 r! 
∞
( n + r )( n + r − 1) . . .(r + 1) r
= ∑ ( −1) r x
r =0 r!
∞
(n + r )! r
= ∑ (−1) r x
r =0 (r!) 2
Thus
e x d n n −x
Ln ( x ) = (x e ) (25)
n! dx n

This is known as Rodrigues formula for the Laguerre polynomials. For

example, putting n = 2 in the Rodrigues’ formula, we have

e x d 2 2 −x
L2 ( x ) = (x e )
2! dx 2

ex d
= (2 x 2 e − x − x 2 e − x )
2! dx
ex d
= (2e − x − 4 xe − x + x 2 e − x )
2! dx
x2
= 1 − 2x +
2
Which is consistent with eq. (9). Similarly, we can determine other
Laguerre polynomials by elementary differentiation of the result
expressed by eq. (25).

3.4 Orthogonality of Hermite Polynomials

As Laguerre differential equation is not of the form of Sturm-Liouville

differential equation, its solutions Ln (x) , therefore, do not by themselves
form an Orthogonal set. However, in order to transform Laguerre
differential equation to the Sturm-Liouville form, we may write eq. (1)
as
(1 − x) n
y ′′( x) − y ′( x) + y ( x) = 0
x x

79
PHY312 MATHEMATICAL METHODS OF PHYSICS I

Multiplying the above equation by

 1− x 
p ( x) = exp  ∫ = xe − x
2

 (26)
 x 

We obtain
d  dy 
 p ( x)  + ne − x y ( x) = 0 (27)
dx  dx 

Thus for Laguerre polynomials, the Sturm-Liouville form is given by

d  dL ( x ) 
 p ( x) n  = − ne − x Ln ( x) (28)
dx  dx 

Similarly
d  dL ( x) 
 p ( x) m  = −me − x Lm ( x) (29)
dx  dx 

Multiply eq.(28) by Lm (x) and eq.(29) by Ln (x) and subtracting the

resulting equations, we obtain
d  dL ( x)  d  dL ( x ) 
Lm ( x)  p ( x ) n  − Lm ( x )  p ( x) m 
dx  dx  dx  dx 
= (m − n) Lm ( x) Ln ( x) (30)

The left hand side of eq.(30) is simply

d  dL ( x) dL ( x) 
 Lm ( x) p ( x) n − Lm ( x) p ( x) m  (31)
dx  dx dx 

Integrating eq.(30) and using eq.(31), we get

∞
∞   dL ( x) dL ( x) 
(m − n) ∫ e − x Lm ( x ) Ln ( x)dx =  p ( x) Lm ( x) n − Lm ( x) m 
0
  dx dx  0

Since p(x) = 0 at x = 0 and at x = ∞ , the right hand side vanishes and we

readily obtain
∞
∫ e − x Lm ( x) Ln ( x)dx = 0 for m ≠ n (32)
0

The above equation shows that the Laguerre polynomials are Orthogonal
in the interval 0 ≤ x ≤ ∞ with respect to the weight function e − x . We now
define the functions
φ n ( x) = N n Ln ( x)e − x / 2
2
(33)

The constant N n is chosen so that the functions φn (x) are normalised, i.e.
+∞
∫ φn2 ( x)dx = 1 for m=n (32)
0

80
PHY312 MATHEMATICAL METHODS OF PHYSICS I

3.5 The Integral Representation of the Laguerre Polynomials

The integral representation of the Laguerre polynomial is given by

Ln ( x ) =
e x + ∞ −t n
n! ∫0
[
e t dt J 0 2( xt )1 / 2 dt ] (33)

In order to prove the above relation we start with the relation

∫0
+∞
[ ]
e −t t n dt J 0 2( xt )1 / 2 dt
+∞ ∞
( −1) r (tx) r
= ∫ e −t t n ∑ dt
0
r =0 ( r!) 2
∞
(−1) r ( x) r + ∞ −t n + r
=∑
(r!) 2 ∫0
e t dt
r =0
∞
( −1) r x r Γ( n + r + 1)
=∑
r =0 ( r!) 2
∞
(−1) r (n + r )! r
=∑ x (34)
r =0 (r!) 2

Using eqs. (24) and (33), we get

∫0
+∞
[ ]
e −t t n dt J 0 2( xt )1 / 2 dt = e − x n! Ln ( x) (35)
from which eq. (33) readily follows.

3.6 Some Important Results Involving Laguerre Polynomials

We give some important results involving Laguerre polynomials which

can be readily derived:

x
∫0
Ln ( x)dx = Ln ( x ) − Ln+1 ( x) [Use Eq. (22)] (36)

[ ]
∞
y n Ln ( x)
∑
n =0 n!
= e y J 0 2( xy )1 / 2 (37)

+∞
m −x 0 if m < n
∫0 x e Ln ( x)dx = (−1) n n! if m = n (38)
N
( N + 1)
∑ Ln ( x) Ln ( y ) = [LN ( x) LN +1 ( y ) − LN +1 ( x) LN ( y)] (39)
n =0 x− y 0
from which eq. (50) readily follows.

3.9 The Second Solution of the Laguerre Differential

Equation

Since the indicial equation [eq. (4)] has two equal roots, the two
independent solutions of eq. (1)

81
PHY312 MATHEMATICAL METHODS OF PHYSICS I

 ∂y 
( y ) p =0 and  
 ∂p  p =0
Now
 p−n ( p − n)( p − n + 1) 2
y ( x, p ) = x p 1 + x+ x +
 ( p + 1) ( p + 1) 2 ( p + 2) 2
2

( p − n)( p − n + 1)( p − n + 2) 3 
x + . . . (40)
( p + 1) ( p + 2) ( p + 3)
2 2 3

Thus,
y1 ( x) = y ( x, p = 0)
n(n − 1) 2 n( n − 1)( n − 2) 3
= 1 − nx + x − x .. (41)
(2!) 2 (3!) 2
and
∂y  p  p−n ( p − n)( p − n + 1) 2 
y 2 ( x) = =  x ln x 1 + x+ x + . . .
∂p   ( p + 1)
2
( p + 1) ( p + 2)
2 2


 1 2  p−n  1 2 2 2 
+ x p  −  x +  − − − 
 p − n p + 1  ( p + 1)  p − n p − n +1 p +1 p + 2 
2

 ( p − n)( p − n + 1) 2 
× x + . . .
 ( p + 1) ( p + 2)
2 2
 p = 0
 3n 2 − n − 1 2 
= y1 ( x) ln x + (2n + 1) x − 2
x + . . . (42)
 (2!) 

For example, for n = 0

y1 ( x) = 1 = L0 ( x)
and

x2 2! x 3 3! x 4
y 2 ( x) = ln x + x + + + +. . . (43)
( 2!) 2 (3!) 2 ( 4!) 2
Similarly, for n = 1
y1 ( x) = 1 − x = L1 ( x)
and

x2 x3
y 2 ( x) = (1 − x) ln x + 3 x − + −. . . (44)
(2!) 2 (3!) 2

3.10 Associated Laguerre Polynomials

Replace n by (n + k ) in eq. (1), it is obvious that Ln +k (x) will be a

solution of the following differential equation.

82
PHY312 MATHEMATICAL METHODS OF PHYSICS I

xy ′′ − (1 − x) y ′ + (n + k ) y = 0 (45)

Differentiating the above equation k times, it can easily be shown that

dk
y = k [Ln+ k (x)] (46)
dx
or a constant multiple of it is a solution of the differential equation
xy ′′ − ( k + 1 − x) y ′ + ny = 0 (47)

Where n and k are positive integers or zero. The above equation is

known as the Associated Laguerre Equation. Its polynomial solutions
[see eq.(45)] are denoted by Lkn (x) and are defined by

dk
Ln ( x ) = (−1) n [Ln+k ( x) ] (48)
dx k

This is known as the Associated Laguerre Polynomials. It is obvious

from eq. (48) that
Lkn (x) is polynomial of degree n in x and that
L0n ( x) = Ln ( x) (49)

Using eqs. (7) and (48), it follows that

n
(n + k )!
Lkn ( x) = ∑ ( −1) r xr (50)
r =0 ( n − r )!( r + k )!r!

We will define Lkn (x) for non-integer values of k, we may, therefore,

write the above equation as
n
Γ(n + k + 1)
Lkn ( x) = ∑ (−1) r xr (51)
r =0 (n − r )!Γ(r + k ) r!

Using the above equation, the first three polynomials can easily be
written as:
Lk0 ( x ) = 1
L1k ( x) = k + 1 − x (52)
1 1
Lk2 ( x ) = (k + 2)(k + 1) − (k + 2) x + x 2
2 2
Differentiating the Laguerre generating function [eq. (10)] k times with
respect to x, one can easily obtain the generating function for the
associated Laguerre polynomials. Thus
1  xt  ∞ k
g ( x, t ) ≡ − ∑ Ln ( x)t
n
k +1
exp (53)
(1 − t )  1 − t  n =0

Furthermore, from eq.(51)

83
PHY312 MATHEMATICAL METHODS OF PHYSICS I

Γ(n + k + 1)
Lkn (0) = (54)
n!Γ(k + 1)

SELF-ASSESSMENT EXERCISE

1. Show that
4
(−1) r Lkr ( x)
x 4 = Γ(5 + k ) 4!∑
r = 0 Γ ( r + k + 1)( 4 − r )!

2. Hint: Use eq. (51 Show that

Ln (0) = 1
Ln′ (0) = − n
1
Ln′′(0) = n(n − 1)
2
Hint: Use Eq. (7).

4.0 CONCLUSION

In this unit, we have established the relationship between Laguerre and

associated Laguerre polynomials. The generating function and some
important results involving Laguerre polynomials were also dealt with.

5.0 SUMMARY

This unit deals with Laguerre functions and its applications to physical
problems especially in Quantum mechanics.

7.0 TUTOR-MARKED ASSIGNMENT

1. Show that
n
Lγn+ k +1 ( x + y ) = ∑ Lγr ( x ) Lkn − r ( y ), n = 0, 1, 2, . . .
r =0

Hint: Use the generating function.

2. Show that
(−1) n H 2 n +1 ( x1 / 2 )
L1n/ 2 ( x) =
2 2 n+1 n! x1 / 2
(−1) n
L−n1 / 2 ( x ) = 2 n H 2 n ( x1 / 2 )
2 n!

Hint: Use the integral representation of Lkn (x) and H n (x) .

3. Using eq. (53), prove the identity

84
PHY312 MATHEMATICAL METHODS OF PHYSICS I

∂g
(1 − t ) + [x − (1 − t )(1 + k )]g = 0
∂t
and then derive the recurrence relation [eq. (56)].

4. Using eq.(53), prove the identity

∂g
(1 − t ) + tg ( x, t ) = 0
∂x
and hence derive the following relation
dLkn ( x ) dLkn−1 ( x)
− + dLkn −1 ( x) = 0
dx dx
n = 1, 2, . . .
5. Show that
x
∫0
Ln (t )dt = Ln ( x) − Ln+1 ( x)
Hint: Use the relation derived in problem 4.

7.0 REFERENCES/FURTHER READING

Erwin, K. (1991). Advanced Engineering Mathematics. John Wiley &

Sons, Inc.

Arfken, G. (1990). Mathematical Methods for Physicists. New York:

Academic Press.

85