SPH 205/308

Mathematical Physics

Vector algebra: dot and cross products, scalar and vector triple products. Vector calculus:
scalar and vector fields, gradient, divergence and curl, curvilinear coordinates, line integrals,
Greens theorem in a plane. Series solutions of second order ordinary differential equations:
ordinary and singular points, Legendre, Bessel, Hermite and Laguerre equations, polynomial
solutions, Rodrigues formulas and generating functions. Partial differential equations:
Poissons and Laplace equations, separation of variables, variational methods, operator
technique.

References
1. Stephenson, G. (1974). Mathematical Methods for Science Students. 2nd edition.
(Longman, Singapore).
2. Riley, K., Hobson, M., and Bence, S. (1999). Mathematical Methods for Physics and
Engineering. (University Press, Cambridge).

TITLE: SPH 205/308

YEAR: 2014/2015
INSTRUCTOR: MR. P.K. KARIUKI
OFFICE HOURS: WEDNESDAY 10 11 AM

SEMESTER: I
OFFICE LOCATION: MATH 4, RM 2

COURSE REQUIREMENTS: ATTEND 100% OF LECTURE HOURS,

DO TWO CATs (30%),
AND SIT ONE EXAM (70%)

NB: FAILURE TO ATTEND MORE THAN 67% OF LECTURE HOURS WITHOUT

PERMISSION FROM THE REGISTRAR (ACADEMIC), THROUGH THE CHAIRMAN
PHYSICS DEPARTMENT, WILL LEAD TO A RETAKE OF THE UNIT.

VECTOR ALGEBRA

1.1

Introduction

A vector is a quantity having both magnitude and direction. Examples of vectors are
displacement, velocity, acceleration, momentum, and force.

Any vector may be represented algebraically by A or A and its magnitude by A or A .

Graphically, a vector is represented by an arrow pointing in the direction of the vector and
whose length is proportional to the magnitude of the vector:

A unit vector denoted generally as e is a dimensionless vector having magnitude equal to

unity.
Any vector A may be represented in terms of a unit vector thus:
A e A A

(1)

where e A points in the same direction as A . Graphically we have:

1.2

Addition of Vectors

Any two vectors A and B may be added together to obtain a third (resultant) vector C whose
effect is equivalent to that of the two vectors combined. Addition of vectors is illustrated
below:

Vector addition is commutative, that is

CABB A

(2)

Three or more vectors may be added in a similar manner as shown

The addition of three or more vectors is associative, that is

A B C A B C
1.3

(3)

Product of Two Vectors: Scalar and Vector Products

Any two vectors A and B may be multiplied in such a way as to form a scalar quantity. This
is known as the scalar or dot product of the vectors and is defined as
A B AB cos

(4)

where is the angle between A and B .

An example of a scalar quantity obtained in this way is work, defined as
W F r Fr cos

(5)

The scalar product of a vector A with itself is given by

A A A2 cos0 A2

(6)

Therefore the magnitude of A may be written as

A AA

(7)

Two vectors A and B may also be multiplied to form a vector quantity. The new vector C is
referred to as the cross or vector product and is defined as
C A B eC AB sin

(8)

where eC is a unit vector in the direction of C .

By convention, this direction is taken to be the direction in which a screw rotated from vector
A to B would advance

The vector C is perpendicular to the plane containing vectors A and B .

An example of a cross product is the torque (moment of a force) defined as
r F e rF sin

(9)

The cross product of a vector with itself gives a vector O having zero magnitude and no
directional property.
A A O

(10)

NB: The scalar product is commutative while the vector product is not, that is
AB BA

(11)

A B B A

(12)

1.4

Cartesian Coordinate System

A coordinate system is a means of locating the position of a point in space.

In the Cartesian coordinate system, the position vector r of a point is specified in terms of
the unit vectors i , j , and k which point along the x , y , and z axis respectively.

The unit vectors are mutually perpendicular and the position vector for point Px, y, z is
given by
r ix jy kz

(13)

The magnitude of r is obtained by help of equation (7) thus

r r r

(14)

But using equation (13)

r r ix jy kz ix jy kz

r r i ix 2 i jxy i kxz

j iyx j jy 2 j kyz
k izx k jzy k kz 2

Using
5

(15)

i i j j k k 1

(16)

i j i k j k ... 0

(17)

we have
r r x2 y2 z 2

and from equation (14)

r x2 y 2 z 2

(18)

Equation (18) is an extension of the Pythagoras theorem to three dimensions.

In general any vector A may be expressed in the Cartesian coordinate system as
A iAx jAy kAz

(19)

where Ax , Ay and Az are the components of A along the x , y , and z axis respectively.
The magnitude of A is given by
A A A A Ax 2 Ay 2 Az 2

(20)

The sum of the vectors A and B is given by

A B iAx jAy kAz
iBx jB y kBz

i Ax Bx j Ay B y k Az Bz

(21)

The dot product of the two vectors in the Cartesian system is

A B iAx jAy kAz iBx jB y kBz

A B Ax Bx Ay B y Az Bz

where we have again applied equations (16) and (17).

The cross product is given by

(22)

A B iAx jAy kAz iBx jB y kBz

A B i iAx Bx i jAx B y i kAx Bz

j iAy Bx j jAy B y j kAy Bz

k iAz Bx k jAz B y k kAz Bz

(23)

using
i i j j k k 0

(24)

and
i j k
j k i

j i k
k j i

(25b)

k i j

i k j

(25c)

(25a)

we have
A B kAx B y jAx Bz kAy Bx iAy Bz jAz Bx iAz B y

or

A B i Ay Bz Az B y j Az Bx Ax Bz k Ax B y Ay Bx

(26)

Equations (26) may be written compactly in terms of determinants of second order

A B i

Ay
By

Az
A
j z
Bz
Bz

Ax Ax
k
Bx
Bx

Ay
By

(27)

or in terms of a determinant of third order

i

A B Ax

Ay

Az

Bx

By

Bz

Example I
Given vectors A 4i 2 j and B 5 j , determine
(a) A B
(b) A B
7

(28)

Solution
(a) A B 4i 2 j 5 j

4 5 i j 2 5 j j
10

(b) A B 4 2 0
0 5 0

i 2 0 5 0 j0 0 4 0 k4 5 2 0
20k

or

A B 4 5 i j 2 5 j j
20k

1.5

Scalar Triple Product

Any three vectors may be multiplied as A B C to form a scalar quantity known as a

scalar triple product.
Writing the vectors in terms of Cartesian coordinates and using equation (28) we have

A B C iAx jAy kAz Bx

By

Bz

Cx

Cy

Cz

iAx jAy kAz i B y C z Bz C y j Bz C x Bx C z k Bx C y B y C x

Ax B y C z Bz C y Ay Bz C x Bx C z Az BxC y B y C x

(29)
Writing equation (29) in terms of the third order determinant we have
Ax

A B C Bx

Ay

Az

By

Bz

Cx

Cy

Cz

Using the following property of determinants

(30)

Ax

Ay

Az

Bx

By

Bz

Cx

Cy

Cz

Bx

By

Bz C x

Cy

C z Ax

Ay

Az

Cx

Cy

Cz

Ay

Az

By

Bz

Ax

Bx

we see that
A B C B C A C A B

Example II
Given the vectors A 3i , B 4 j 5k , and C i j 2k , determine A.B C

Solution

B C 4 j 5k i j 2k

4 j i 4 j j 8 j k 5 k i 5 k j 10 k k

8 5i 5 j 4k
3i 5 j 4k

A B C 3i 3i 5 j 4k

9 i i 15 i j 12 i k

or
Ax

A.B C Bx

Ay

Az

By

Bz

Cx

Cy

Cz

3 0 0
0 4 5
1 1 2

34 2 5 1 05 1 0 2 00 1 4 1

(31)

1.6

Vector Triple Product

Three vectors may also be multiplied as A B C or A B C , to form a vector known as a

vector triple product.
To evaluate A B C we write
F B C

(32)

Then
i

iFx jFy kFz Bx

By

Bz

Cx

Cy

Cz

By
Cy

Bz
B
j z
Cz
Cz

Bx Bx
k
Cx
Cx

By
Cy

(33)

Consequently

A B C A F Ax

Ay

Az

Fx

Fy

Fz

Ax

Ay

By

Bz

Bz

Bx

Az
Bx B y

Cy

Cz

Cz

Cx

Cx

(34)

Cy

Expanding equation (34) we have

A B C i Ay BxC y By C x Az Bz C x BxC z

j Az B y C z Bz C y Ax Bx C y B y C x

k Ax Bz C x Bx C z Ay B y C z Bz C y

(35)

Rearranging equation (35) we have

A B C i Bx Ay C y Az C z C x Ay B y Az Bz

j B y Ax C x Az C z C y Ax Bx Az Bz

10

k Bz Ax C x Ay C y C z Ax Bx Ay B y

(36)

Further we may write equation (36) as

A B C i Bx Ax C x Ay C y Az C z C x Ax Bx Ay B y Az Bz

j B y Ax C x Ay C y Az C z C y Ax Bx Ay B y Az Bz
k Bz Ax C x Ay C y Az C z C z Ax Bx Ay B y Az Bz

(37)

Writing equation (37) compactly we have

A B C BA C CA B

(38)

Equation (38) is known as the Back cab rule.

Similarly it can be shown that

A B C BC A AC B

(39)

A B C A B C

(40)

Consequently

Example III
Given the vectors A 3i , B 4 j 5k , and C i j 2k , determine A B C

11

VECTOR CALCULUS

2.1

Scalar and Vector Fields

Consider a region R is space in which a scalar x, y, z is defined at every point in the region.
Then a scalar field is said to exist in that region.
Similarly if at every point in a region R , a vector Ax, y, z is defined, then a vector field exists
in that region.
An example of a scalar field is the gravitational potential near the surface of the earth V mgz
.
The gravitational force field F kmg associated with this gravitational potential is an
example of a (position independent) vector field.

2.2

Gradient of a Scalar Field

The gradient of a scalar field x, y, z , which is denoted by grad , is defined as

grad i

j
k
x
y
z

(1)

where x , y , and z , are the partial derivatives of x, y, z with respect to x , y ,

and z , respectively.
These partial derivatives are defined as

x x, y, z x, y, z
lim
x x 0
x

(2a)

x, y y, z x, y, z
lim
y y 0
y

(2b)

x, y, z z x, y, z
lim
z z 0
z

(2c)

The gradient of a scalar field x, y, z is a vector quantity.

Example I
Determine the gradient of the function xyz3 .

12

Solution
Differentiating with respect to x we have

x x y 2 z 3 xy 2 z 3

lim
x x 0
x
xy 2 z 3 xy 2 z 3 xy 2 z 3
x 0
x

lim

xy 2 z 3
y2 z3
x 0
x

lim

or writing the partial derivative in terms of the ordinary derivative we have

dx
y2 z3
y2 z3
x
dx

Similarly

x y y 2 z 3 xy 2 z 3
lim
y y 0
y
xy 2 z 3 2 xyyz 3 xy 2 z 3 xy 2 z 3
y 0
y

lim

2 xyyz 3 xy 2 z 3
lim 2 xyz3 xyz 3
y 0
y 0
y

lim

2xyz3

Alternatively

dy 2
xz 3
2xyz 3
y
dy

Finally

xy 2 z z 3 xy 2 z 3
lim
z z 0
z

xy 2 z 3 3z 2 z 3zz 2 z 3 xy 2 z 3
z 0
z

lim

lim xy 2 3z 2 3zz z 3xy 2 z 2

z 0

or
13


dz3
xy 2
3xy 2 z 2
z
dz

Therefore
grad iy 2 z 3 j 2 xyz3 k3xy 2 z 2 .

The gradient of a scalar field x, y, z as defined in equation (1) may be written as

grad i j k
y
z
x

(3)

where (pronounced del or nabla) is a differential operator defined as

j k
x
y
z

(4)

To illustrate the geometrical property of the gradient of a scalar field x, y, z , consider the
infinitesimal displacement dr given by
dr idx jdy kdz

(5)

Suppose dr is an infinitesimal displacement between two nearby surfaces defined by

x, y, z c1 and x, y, z c2 where c1 and c2 are constants.

In the diagram is the gradient of x, y, z at point P , and is the angle between and dr .
The total differential of at point P is given by

14

dx
dy
dz
x
y
z

(6)

Equation (6) may be written as

i dx jdy kdz
d i
j
k
y
z
x

or using equations (3) and (5)

d dr

(7)

The space rate of change d dr , of is

d
dr

dr
dr

(8)

But dr dr n , where n is the unit vector in the direction of dr . Therefore

d
n
dr

(9)

From the definition of the dot product, we have

d
n cos cos
dr

(10)

Now suppose dr is in the direction of , then 0 and cos 1 . In this case d dr has it
maximum value given by the magnitude of the gradient thus
d

dr max

(11)

Thus the gradient of a scalar field is a vector whose magnitude and direction is that of the
maximum space rate of change of the scalar field.

Example II
The gravitational potential near the surface of the earth is given by mgz .Show that the
gravitational force F kmg is related to by F .

15

Solution

i
j
k
x
y
z
mgz

0
x
x

mgz

0
y
y
mgz
mg
z

Therefore
kmg F

2.3

Divergence of a Vector Field

The divergence of a vector field A is defined as

divA A

(12)

where is the del operator defined in equation (4).

In Cartesian coordinates

divA i j k iAx jAy kAz

y
z
x

(13a)

or
divA

Ax Ay Az

x
y
z

The divergence of a vector field is a scalar function.

Example III
Determine the divergence of the vector field defined by
A 2 xyi j xz 3k

16

(13b)

Solution
Ax
dx
2y
2y
x
dx

Ay
y

d1
0
dy

Az
dz3
x
3xz 2
z
dz

Therefore
divA 2 y 3xz 2 .

To investigate the geometrical interpretation of the divergence of a vector field A , consider

an infinitesimal volume element xyz , which is located in the vector field as shown
Az x, y, z z

Ax x, y, z

x
Ay x, y, z

Ay x, y y, z

z
y
Ax x x, y, z

Az x, y, z

The flux of a vector field A through an infinitesimal surface vector element s , of a closed
surface, is defined as
s
A s A n

where n is a unit vector pointing perpendicularly outward from the closed surface.
The flux of vector A through the rear face of the volume element is then given by

17

(14)

 backface Ax x, y, z yz

(15)

and the flux through the front face is

frontface Ax x x, y, z yz

(16)

In obtaining equations (15) and (16) we assume that the infinitesimal area yz is small
enough such that the components Ax x, y, z and Ax x x, y, z of A along the x- axis do not
vary appreciably over the surface element.
Also we note that the components of A along the y- , and z- axes do not contribute to the flux
through the back and front faces since these components are parallel to the two faces.
Similarly the flux of vector A through the left and right faces of the volume element is given
respectively by
(17)
leftface Ay x, y, z xz
rightface Ay x, y y, z xz

(18)

Lastly, for the flux through the bottom and top faces we have
bottomface Az x, y, z xy

(19)

topface Az x, y, z z xy

(20)

The net flux of vector A through the volume element is then

net Ax x x, y, z Ax x, y, z yz Ay x, y y, z Ay x, y, z xz
Az x, y, z z Az x, y, z xy

(21)

Dividing equation (21) by we obtain the net flux of vector A per unit volume

net Ax x x, y, z Ax x, y, z Ay x, y y, z Ay x, y, z

x
y

Az x, y, z z Az x, y, z
z

(22)

Since the volume element is very small, x , y , and z , all tend to zero, and we may use the
definition of the partial derivative to write equation (22) as
net Ax Ay Az

x
y
z

18

(23)

Thus the divergence of a vector field A at a point is the net outward flux of A per unit volume
as the volume becomes infinitesimal.

Example IV
Determine the divergence of the gravitational force field F kmg and interpret the result.
Solution
Fx 0
Fy 0
Fz mg
Fy Fz
Fx
0

x
y
z
F 0

This results is consistent with the fact that the vector field does not spread (diverge).

2.4

The Laplacian

The divergence of the gradient of a scalar field is also defined thus

divgrad

i j
k i
j
k
y
z x
y
z
x

x x y y z z

2
x 2

2
y 2

2
z 2

(24)

where 2 x 2 , 2 y 2 , and 2 z 2 is the second partial derivative of with respect to x,

y, and z respectively.
The divergence of the gradient may also be written as
divgrad 2

where
19

(25)

2
x 2

2
y 2

(26)

z 2

is called the Laplacian operator.

This operator occurs in many equations of mathematical physics such as the heat equation
and the wave equation.

Example V
Determine the Laplacian of the scalar function xyz3 .
Solution

xyz yz

2 xyz3
x

x x

1
0
x

2 xyz3

1

xyz3 xz 3
0
2
y y
y
y

xyz 3xy z

2 xyz3
z 2

z z

6 xyz

Therefore

2 xyz3 6 xyz

2.5

Curl of a Vector Field

The Curl of a vector field A is defined as

curlA A

(27)

where is the del operator.

In Cartesian coordinates we have

curlA i j
k i Ax jAy kAz
y
z
x

i j

Axy i k Axz j i Ayx

20

Ay
Ax
Az
j k
k i
k j
y
z
z
k

Ay

Ay
A
A
A
A
j z k x i z j x i
x
x
y
y
z
z

(28)

Rearranging equation (28) yields

A Ay Ax Az Ay Ax
j

curlA i z

k
z z
x x
y
y

(29)

Equation (29) may be written compactly as

i

curlA
x
Ax

y
Ay

z
Az

(30)

Example VI
Determine the curl of the vector
A i2 xy j xz 3k

Solution
Using equation (30) we have
j
i
k
xz 3 1

curlA
i

x y z
z
y
2 xy 1 xz 3

j 2 xy xz k 1 2 xy
x
x
y

curlA z 3 j 2 xk

2.6

Cylindrical Polar Coordinate System

It is usually convenient to express , A , A ,and 2 , of scalar and vector fields ,

and A , respectively, in terms of coordinate systems other than the Cartesian coordinate
system.
One such coordinate system is the cylindrical polar coordinate system.

21

z
Px, y, z

ez

e
e

z
k
j

As shown in the figure above, the position of a point P having Cartesian coordinates x , y , z
may be expressed in terms of cylindrical polar coordinates , , z where
x cos

(31.a)

y sin

(31.b)

zz

(31.c)

The position vector of P may therefore be written as

r i cos j sin kz
Taking the partial derivative of r with respect to , , and z respectively we have

(32)

r
e i cos j sin

(33a)

r
e i sin j cos

(33b)

r
e z k
z

(33c)

The vectors e , e and e z point in the direction of increasing , , and z respectively. Their
magnitudes are
e e e

22

cos sin 1
2

(34a)

e e e 2 sin 2 cos2
e z k 1

(34b)
(34c)

Thus e and e z are unit vectors. Since e too is a unit vector, we may write the unit vectors
e , e , ez of the cylindrical polar coordinate system as
e

r
i cos j sin

(35a)

1 r
i sin j cos

(35b)

ez

r
k
z

(35c)

The unit vectors in equations (35) may also be written as

e

1 r
h

(36a)

1 r
h

(36b)

ez

1 r
hz z

(36c)

where h 1, h , hz 1 are known as the scale factors for the cylindrical polar coordinate
system.
Example VII
Express the cylindrical polar coordinates , in terms of the Cartesian coordinates x, y .
Solution
From equation (31a) and (31b) we have
x 2 2 cos2

y 2 2 sin 2

Consequently

x 2 y 2 2 cos2 sin 2 2

23

and

y2

Also dividing equation (31b) by (31a) we get

y sin

tan
x cos

Therefore
y
x

tan1

2.7

Line Integrals

In physical problems, one often encounters integrals of the form

I A dr

(37)

where is a A vector field, dr is an infinitesimal displacement, and C is path along which the
integral is to be evaluated.
Examples of such integrals are the total work done W in moving a particle in a force field F
W F dr
C

and Amperes law:

C B dr oi
where B is the magnetic flux density, i is the current, o and is the permeability of free space.
In Cartesian coordinates we write equation (37) as

I A dr iAx jAy kAz idx jdy kdz

C

Ax dx Ay dy Az dz
C

(38)

To illustrate the evaluation of these line integrals, consider the line integral
I Px, y dx
C

24

(39)

The path C is specified by a functional relation y f x . Then if the limits of the integration
are points Ax1 , y1 and Bx2 , y2 we have
B

x2

x1

I Px, y dx Px, f x dx
C

Px, f x dx

(40)

This is illustrated below

Bx2 , y2
y f x

Ax1 , y1
x

The line integral may also be written as

B

I Px, y dx Pg y , y
C

dx
dy
dy

y2

Pg y , y dy dy
dx

y1

where g y f 1 is the inverse function of f x .

Example VIII
Evaluate
I x y dx
C

from point A0,0 to B2,4 along the curve y x 2 .

Solution
The curve y x 2 and the integrand Px, y x y x x 2 vary with x as shown below

25

(41)

6
5
4
y

P
2
1
0
0

0.5

1.5

From equation (40) we have

I x y dx
C

B 2, 4

A0,0

2
x y dx x x dx

1 2
1
x x 3 4.67
2 0 3 0

Alternatively using equation (41) and since

x y1 2
dx
1
12
dy 2 y

we have
I x y dx
C

B 2, 4

A0,0

dx
12
x y dx y y dy

dy

1 4 y1 2 y
14
dy

1 y1 2 d y
2 0 y1 2
2 0
4

1
1
y y 3 2 4.67
2 0 3
0

26

Line integrals have the following important properties:

(i) A line integral changes sign on interchanging the lower and upper limits of integration
B

Px, y dx Px, y dx

(42)

(ii) For any point E which lies between the lower and upper limits of integration A and B
we have
B

Px, y dx Px, y dx Px, y dx

(43)

y
B

E
A

(iii) If y is not a single-valued function of x as illustrated below

y
B

y f 2 x

y f1 x

Then
B

Px, y dx Px, f1x dx Px, f 2 x dx

A

Line integrals are often taken along a closed path. This is indicated as
I Px, y dx
C

An arrow is used to indicate the direction of integration as clockwise or anti-clockwise.

27

(44)

Example IX
Evaluate the line integral I C x y dx from point A0,1 to point B0,1 where C is the

semicircle y 1 x 2 shown below.

y
A0,1

E 1,0

B0,1

Soln
Using property (iii)/equation (44) we have
y f1 x

1 x
2

and y f 2 x 1 x 2

therefore

12
12
1
0
I x y dx x 1 x 2 dx x 1 x 2 dx

C
0
1

Applying property (i)/equation (42) of line integrals in the second integral above gives

12
12
1
1
I x y dx x 1 x 2 dx x 1 x 2 dx

C
0
0

The integral then simplifies to

1

I 2 1 x 2
0

12

dx

To evaluate this integral we us cylindrical polar coordinates in place of Cartesian coordinates

and write, since the semicircle has a radius of one unit (i.e. 1 )
x cos

Differentiating with respect to we have

dx
sin
d

28

which implies
dx sin d

While the integrand becomes

1 x 1 cos sin
2 12

12

12

sin

Making these substitutions the integral becomes

1

I 2 1 x 2
0

12

dx 2

sin 2 d

since the point Ax 0, y 1 in Cartesian coordinates is equivalent to A 1, 2 in

cylindrical polar coordinates and Ex 1, y 0 in Cartesian coordinates is E 1, 0 in
cylindrical polar coordinates.
Using the trigonometric identity
sin 2

1
1 cos 2
2

we have
1

I 2 1 x 2
0

12

dx 2
0

1
sin 2
2
2

1 0
1 cos 2 d
2 2

0 2
2

1
sin 0 sin
2

I 2

which happens to be the area of a semicircle with radius 1 .

2.8 Greens Theorem in a Plane

Suppose the functions Px, y , Qx, y , and their derivatives are single-valued, finite and
continuous inside and on the boundary C of some region R in the xy-plane. Greens theorem
in a plane relates the line integral around the closed curve C to a double integral over the
enclosed region R
Q

Pdx Qdy x

29

P
dxdy
y

(45)

where the line integral is taken in the anti-clockwise direction. Greens theorem in a plane is
useful in deriving the area of geometrical shapes such the circle, ellipse and triangle. It is also
used in interpretation of the curl of a vector field.
y

30

SERIES SOLUTIONS OF ORDINARY DIFFERENTIAL EQUATIONS

3.1

Introduction

The general homogenous second-order linear ordinary differential equation (ODE) has the
form
d2y
dx

px

dy
q x y 0
dx

(1)

The equation has two linearly independent solutions y1 and y2 such that y2 is not a multiple of
y1 . For this to be the case the Wronskian W of the two solutions, define by
W y1

dy2
dy
y2 1
dx
dx

y1
dy1
dx

y2
dy2
dx

(2)

must be non-zero.
The general solution y of equation (1) is a linear combination of the two solutions
y c1 y1 c2 y2

where c1 and c2 are arbitrary constants.

Example I
The functions y1 sin x and y2 cos x are both solutions of the equation
d2y
dx2

y0

Show that y1 and y2 are linearly independent.

Solution
This ODE is obtained from equation (1) by setting px 0 and qx 1 . Now
dy1
cos x
dx

31

(3)

dy2
sin x
dx

and
W sin x sin x cos xcos x

W sin 2 x cos2 x 1

Since W 0 , y1 and y2 are linearly independent.

3.2

Ordinary and Singular Points

An ordinary point x0 of an ODE of the form given by equation (1) is one for which the
following conditions are satisfied
lim px finite

(4)

lim qx finite

(5)

x x0

x x0

If conditions (4) and (5) are not satisfied, then the point x0 is a singular point.
A regular singular point is a singular point which satisfies the following conditions
lim x x0 px finite

(6)

lim x x0 2 qx finite

(7)

x x0

x x0

If a singular point does not satisfy the conditions (6) and (7), it is called an irregular singular
point.

Example II
Consider the Legendre equation

1 x ddx y 2x dy
l l 1 y 0
dx
2

where l is a constant. Show that:

(a) the point x 0 is an ordinary point.
(b) the point x 1 is a regular singular point.

32

Solution

(a) We divide Legendre equation by 1 x 2 to put it in the general form of equation (1) thus
d2y
dx

2x

dy

1 x

dx

l l 1

1 x
2

y0

then
px

and
qx

2x

1 x
2

l l 1

1 x
2

Now
lim px lim

x 0

x 0

2x
1 x

lim 2 x

x 0

lim 1 x 2
x 0

lim 2 x

x 0

0 finite

and
lim qx lim

x 0

x 0

l l 1
1 x2

l l 1
l l 1 finite
lim 1 x 2
x 0

Therefore x 0 is an ordinary point for the Legendre equation.

(b)
lim px lim

x 1

x 1

2x
1 x

lim qx lim

x 1

x 1

lim 2 x

x 1

lim 1 x

x 1

l l 1
1 x2

2
lim 1 x 2

not finite

x 1

l l 1
lim 1 x 2

not finite

x 1

Therefore x 1 is a singular point for the Legendre equation. However

lim x 1 px lim

x 1

x 1

x 1

2x

1 x
2

1 x 2 x
x 1 1 x 1 x

lim

33

lim 2 x

x 1

lim 1 x

x 1

1 finite

and

x 12 l l 1 lim x 1x 1l l 1
x 1
1 x 2 x 1 1 x 1 x

lim x 12 qx lim

x 1

x 1l l 1
x 1l l 1 xlim
1
0 finite
x 1
1 x
lim 1 x

lim

x 1

Therefore x 1 is a regular singular point for the Legendre equation.

Example III
The Bessel equation is given by
x2

d2y
dy
x
x2 v2 y 0
2
dx
dx

where v is a constant. Show that the point x 0 is a regular singular point for this equation.
Solution
Dividing the Bessel equation by x 2 , we have
d2y
dx2

1 dy v 2
1 2 y 0
x dx
x

Therefore
px

1
x

v2
qx 1 2
x

Now
lim px lim

x 0

x 0

1
x

v2
lim qx lim 1 2
x 0
x 0
x

since px and qx do not satisfy equations (4) and (5), the point x 0 is a singular point.
34

But
lim xpx lim

x 0

x 0

x
x

and

v2
lim x 2 qx lim x 2 1 2 lim x 2 v 2 v 2 finite

x 0
x 0
x x 0

since px and qx satisfy equations (6) and (7), the point x 0 is a regular singular point.

3.3

Series Solutions near an Ordinary Point

Suppose x x0 is an ordinary point of an ODE of the form given by equation (1). In this case,
we may seek a solution in the form of a power series about x0
y a0 a1 x x0 a2 x x0 2 a3 x x0 3

(8)

where a0 , a1, a2 , a3 ,.....are coefficients (constants). Equation (8) contains an infinite number of
terms. This equation may be written more compactly as
y

an x x0 n

(9)

n0

If x 0 is an ordinary point, equation (9) simplifies to

y

an x n

(10)

n0

3.4

The Simple Harmonic Oscillator Equation

Consider the simple harmonic oscillator (SHO) equation

d2y
dx2

y0

(11)

which is obtained from equation (1) by setting px 0 and qx 1 . For this equation, all
points are ordinary points (i.e. the equation has no singular points). We may therefore seek a
solution of the form given equation (10) about the point x 0 . Then since
dxn
dy d
na x n 1
an x n an
dx n
dx dx n 0
n 0
n0

35

(12)

where we have used dxn dx nxn 1 .

Further applying equation (12) we have
d2y
dx2

d dy d
d n 1
n 1
x
nn 1an x n 2
nan x nan
dx dx dx n 0
dx
n 0
n 0

(13)

where we have used dxn 1 dx n 1x n 2 . Substituting equations (10) and (13) in equation
(11) yields

n0

n0

nn 1an x n 2 an x n 0

(14)

Evaluating the first two terms in the first series in equation (14) gives
0 1 a0 x 2 1 0 a1 x 1

n2

n0

nn 1an x n 2 an x n 0

(15)

Considering these two terms in equation (15) we see that the coefficients a0 and a1 can take
any value and the equation will still be satisfied (because of the zeroes present in the terms).
Consequently these two coefficients are called arbitrary constants.
Substituting m 2 for n , the third term in equation (15) may be written as

n2

m 22

m0

nn 1an x n 2 m 2m 2 1am 2 x m 2 2 m 2m 1am 2 x m

(16)

replacing the dummy index m in equation (16) by n we have

n2

n0

nn 1an x n 2 n 1n 2an 2 x n

(17)

Neglecting the first two zero terms in equation (15) and substituting equation (17) into (15)
we get

n0

n0

n 1n 2an 2 x n an x n 0

or

n 1n 2an 2 an x n 0

n0

36

(18)

For equation (18) to be satisfied for all x , we must have

n 1n 2an 2 an 0
where n 0,1,2,3, . Thus the coefficients must be related by
an 2

an

n 1n 2

(19)

Equation (19) is called the recurrence relation of the SHO equation (11). From this relation
we have for n 0,1,2, and 3 ,
a2

a3

a4

a0

0 10 2
a1

1 11 2

a0
2!

(20a)

a1
3!

(20b)

a2

a
1 a0

4!
3 4 2!

(20c)

a3

a
1 a1

5!
4 5 3!

(20d)

2 12 2

a5

3 13 2

All other coefficients, for n 3 , may be obtained similarly in terms of the arbitrary constants
a0 and a1 , using the recurrence relation (equation (19)). Using equation (20) in equation (10),
we get
y a0 x 0 a1 x1

a0 2 a1 3 a0 4 a1 5
x x
x x
2!
3!
4!
5!

(21)

Factoring out a0 and a1 in equation (21), we have

x2 x4
x3 x5
y a0 1

a1 x

2! 4!
3! 5!

(22)

Equation (22) is the general solution of the SHO equation (11). This solution may be written
in closed form by recognizing the two series in equation (22) as the series for the cosine and
sine functions respectively. Therefore the solution may be written as
y a0 cos x a1 sin x

NB: In general for a SHO equation of the form

37

(23)

d2y
dx

2y 0

(24)

where is some constant, then following the above derivation, the solution is found to be
y a0 cos x a1 sin x

(25)

where again a0 and a1 are arbitrary constants.

3.5

Legendres Equation

Consider Legendres equation

1 x ddx y 2x dy
l l 1 y 0
dx
2

(26)

where l is a constant. As noted earlier, this equation is obtained from equation (1) by setting
px 2 x 1 x 2 and qx l l 1 1 x 2 . Since x 0 is an ordinary point for this ODE, we
seek a solution of the form given by equation (10), that is

an x n

n 0

Then as in equation (12)

2x

dxn

dy
d
2 x nan x n 1 2 nan x n
2 x an x n 2 x an

dx
dx n 0
n 0 dx
n0
n0

(27)

Also as in equation (13)

d

1 x ddx y 1 x dxd dy
1 x na x
dx
dx
2

1 x2

n0

nn 1a x

n0

n2

n 1

n0

n0

2
1 x

na dxd x

n0

n 1

nn 1an x n 2 nn 1an x n
(28)

Using equations (10), (27), and (28) in equation (26) yields

n0

n0

n0

n0

nn 1an x n 2 nn 1an x n 2 nan x n l l 1 an x n 0

Writing the first two terms in the first sum in equation (29) explicitly thus
38

(29)

n0

n2

nn 1an x n 2 0 1 a0 x 2 1 0 a1 x 1 nn 1an x n 2

equation (29) becomes

0 1 a0 x 2 1 0 a1 x 1

n2

n0

n0

n0

nn 1an x n 2 nn 1an x n 2 nan x n l l 1 an x n 0

(30)
From equation (30), we see that a0 and a1 are arbitrary constants which can take any value and
still satisfy the equation. As in equation (17), the third term in equation (30) can be written as

nn 1an x n2

n2

n 1n 2an2 x n

n 0

Therefore since the first two terms in (30) vanish, we obtain

n0

n0

n0

n0

n 1n 2an 2 x n nn 1an x n 2 nan x n l l 1 an x n 0

(31)

This equation may be written as

n0

n0

n 1n 2an 2 x n nn 1 2n l l 1an x n 0

(32)

Also writing nn 1 2n nn 1 , equation (32) becomes

n 1n 2an 2 nn 1 l l 1an x n 0

(33)

n0

For equation (33) to hold for all x , we must have

n 1n 2an 2 nn 1 l l 1 0
which yields the recurrence relation for Legendres equation thus
an 2

nn 1 l l 1 a
n 1n 2 n

(34)

From this relation, we have for n 0,1,2, and 3 respectively

a2

00 1 l l 1 a l l 1 a
0
0 10 2 0
2!

39

(35a)

a3

11 1 l l 1 a 2 l l 1 a
1
1 11 2 1
3!

l 1l 2
a1
3!

a4

22 1 l l 1 a 6 l l 1 l l 1 a
0
2 12 2 2
3 4
2!

a5

(35b)

l l 1 2
l l 1l 2l 3
l l 6 a0
a0
4!
4!

(35c)

33 1 l l 1 a 12 l l 1 l 1l 2 a
1
3 13 2 3
45
3!

l 3l 4l 1l 2 a

5!

(35d)

Using equation (35) in equation (10), we see that the general solution for Legendres equation
is
y a0 x 0 a1 x1

l 1l 2 a x3 l l 1l 2l 3 a x 4
l l 1
a0 x 2
1
0
2!
3!
4!

l 3l 4l 1l 2 a x5 ....

(36)

5!

Factoring out the arbitrary constants a0 and a1 in equation (36) we obtain

l l 1 2 l l 1l 2l 3 4
y a0 1
x
x ...
2!
4!

l 1l 2 x3 l 1l 2l 3l 4 x5 ...

a1 x

3!
5!

(37)

We may also write the solution as

y y1 y2

(38)

where
l l 1 2 l l 1l 2l 3 4
y1 a0 1
x
x ...
2!
4!

(39a)

and

l 1l 2 x3 l 1l 2l 3l 4 x5 ...

y2 a1 x

3!
5!

40

(39b)

3.6

Legendre Polynomials

If l in Legendres equation is an integer, then from the recurrence relation (equation 34) we
see that the coefficients al 2 , al 4 , etc vanish and either series y1 or y2 terminates to yield a
polynomial of order l .
For example if l 0 , we have a2 a4 0 and the first series terminates to give
y1 a0

(40)

which is a polynomial of order 0.

Similarly if l 1 , we have a3 a5 0 and the second series terminates to a polynomial of
order 1 thus
y2 a1x

(41)

For l 2 , and a4 a6 0 the first series terminates to give

y1 a0 1 3x 2

(42)

For l 3 , and a5 a7 0 the second series terminates to yield

10

y2 a1 x x 3
6

(43)

The Legendre polynomials Pl x , where l 0,1,2,3, , are solutions of Legendres equation

which are normalized by choosing a0 or a1 such that Pl x 1 1 .
For example choosing a0 1, a1 1, a0 1 2 and a0 3 2 in equations (40), (41), (42) and
(43) respectively, yields the first four Legendre polynomials as follows.
P0 x 1

(44a)

P1 x x

(44b)

P2 x

1 2
3x 1
2

P3 x

1 3
5 x 3x
2

41

(44c)

(44d)

PARTIAL DIFFERENTIAL EQUATIONS

4.1

Introduction

A partial differential equation (PDE) is an equation involving the partial derivatives of a

dependent variable with respective to its independent variables.
Consider the dependent variable x, y, z, t (pronounced psi) of four independent variables
x, y, z, and t . The first partial derivatives of with respect to x and t , for example, are defined
by

x x, y, z, t x, y, z, t
lim
x x 0
x

(1)

and

x, y, z, t t x, y, z, t
lim
t t 0
t

(2)

respectively.
The second partial derivatives with respect to x and t are defined by
2
x

x x

(3)

t t

(4)

and
2
t

An example of a PDE is the (three-dimensional) wave equation

2

1 2

(5)

c 2 t 2

where c is a constant, t is the time variable, and 2 (pronounced del squared) is the
Laplacian operator defined in Cartesian coordinates by
2

2 y
x 2

2 y
y 2

2 y
z 2

Another example of a PDE is the heat (diffusion) given by

k 2

where k is a constant.
42

(6)

4.2

The Wave Equation

Consider the one-dimensional wave equation

2
x 2

1 2

(7)

c 2 t 2

where x, t is the vertical displacement of a segment of a string, located at point x , from its
equilibrium (i.e., rest position) at a given time t .
In the separation of variables method, we seek a solution to this equation of the form
X x T t

(8)

where X is a function of x only and T is function of t only. Therefore

2
x 2

2 XT
x 2

d2X

(9)

dx2

and
2
t 2

2 XT
t 2

d 2T

dt 2

(10)

Using equations (9), and (10) in equation (7) we have

T

d2X
dx2

X d 2T
c 2 dt 2

Rearranging this equation yields

1 d2X
1 d 2T

X dx2
c 2T dt 2

(11)

Equation (11) can only be satisfied if both sides of the equation are equal to a constant. For
convenience we choose this constant to be k 2 . Therefore
1 d2X
k 2
X dx2

(12)

and
1 d 2T
2

c T dt

k 2

Rearranging these ordinary differential equations we have

43

(13)

d2X
k2X 0
2
dx

(14)

and
d 2T
dt

c 2 k 2T 0

or
d 2T
dt

2T 0

(15)

where 2 c 2 k 2
Equations (14) and (15) are similar to equation (24) in topic 3. Their solutions are therefore
X x A cos kx B sin kx

(16)

T t C cost D sin t

(17)

and

where A, B, C, D are arbitrary constants.

Therefore from equations (8), (16) and (17) we see that the general solution to equation (7) is
x, t Acoskx B sin kxC cost D sin t

4.3

(18)

The Heat Equation

Consider the one-dimensional heat equation

k

2
x

(19)

where is the temperature of a material at point x at time t .

Again using the separation of variables method we assume a solution of the form
X x T t

(20)

where the variable X depends on x alone and the variable T depends on t alone. Substituting
this expression in equation (19) gives
kT

d2X
dx

44

dT
dt

(21)

Equation (21) may be rearranged thus

1 d2X
1 dT

X dx2
kT dt

(22)

For this equation to hold each side must equal a constant, which we conveniently take as 2 .
Consequently equation (22) results in the following ODEs
d2X
dx

2 X 0

(23)

dT
k2T 0
dt

(24)

The solution of the second order ordinary differential equation (23) is

X x Acosx B sin x

(25)

To solve the first order ordinary differential equation (24) we rearrange it as

dT
k2 dt
T

(26)

then integrating both sides of equation (26) with respect to T and t thus
1

T dT k dt

(27)

ln T k2t c

(28)

we obtain

where c is an integration constant. Writing c ln C , equation (28) becomes

ln T k2t ln C

or
ln T ln C ln

T
k2t
C

This implies that

T
exp k2t
C

or

T t C exp k2t

45

(29)

Substituting the solutions (25) and (29) in equation (20) gives the general solution of the heat
equation

x, t A cosx B sin x C exp k2t

46

(30)