MATHEMATICS AND PHYSICS FOR

CHEMISTS 1.
R.L.JACOBS.

INTRODUCTION.
These notes summarise the Mathematics for your degree course. The Physics
will be given by a di↵erent lecturer in a di↵erent style. The notes contain the
basic outline of the course but more material and a fuller treatment will be
given in the lectures so do not feel that you can dispense with the lectures.
Remember Mathematics is a language and an active discipline. You cannot
learn it by reading alone; you have to participate. You also have to tackle
the problem sheets which will be handed out later.

The course will be taught by lectures, tutorials and an office hour when
I will be available to talk to students individually or in small groups on a
drop-in basis. If you decide to come to the office hour please come early
because I do not want to hang around if there are no takers. In the lectures
I will project the notes onto a screen and talk about them and write addi-
tional material on the board. I will also work out some problems which are
presented in the notes but not solved there.

If the course is too easy or too difficult let me know through the class reps.
I will take comments and constructive criticisms seriously.

You should be familiar with the following topics: basic algebraic operations
of addition, subtraction, multiplication and division, relationships such as
equality and inequality, powers, simple functions such as polynomials, the so-
lution of quadratic equations, rational expressions, trigonometric functions,
the exponential function, the natural logarithm, graphs. If you are rusty on
any of these topics feel free to ask me for help.

1
 TOPIC A. SETS AND FUNCTIONS.
1. SETS.

A set is a collection of objects with a rule for deciding whether an

object belongs to the set
eg. = {the set of students in this class}
The rule here is that a member of the set is human and is registered as
a first year Chemistry student.

Another example is = {n : where n is an integer greater than 0}.

Observe the notation. A general member of the set lies before the colon
and the rule is after the colon. Another way of representing this set is
= {1, 2, 3, 4, 5, . . .} where the set is defined by a list.

Another example is the set of all integers, positive, negative or zero

= {. . . 2, 1, 0, 1, 2, 3, . . .}.

Another important set is the empty set i.e. the set with no mem-
bers which is written ; .

An important property of a set is the number of elements. is a

finite set with n =? elements. and are infinite sets. You should
always be clear in your mind whether a given set is finite or infinite.

Some more notation:

x2 means that x is a member or element of set ,
⇢ means that set is included in set or alternatively is a
subset of ,
= [ means that every member of is a member of and every
member of is also a member of or alternatively is the union of
and of ,
= \ means that is a set consisting of members which belong
to both and simultaneously or alternatively is the intersection
of and .

These ideas can be conveniently illustrated by Venn diagrams which

will be explained in the lectures.

2
 There are some more sets which are important:

is the set of all real numbers whether positive, negative or zero

including integers, fractions and irrational numbers such as e and ⇡.
These numbers can be understood as points on a line (the x-axis) re-
ferred to an origin O.
Note that ⇢ and ⇢ .

There are also intervals of several di↵erent types:

O = (a, b). This is called an open interval and is a set of points in a
segment of the x-axis between a and b not including the end points.
C = [a, b]. This is called a closed interval and is a set of points in a
segment of the x-axis between a and b including the end points.
There are also mixed intervals such as 1 = [a, b) which includes a but
not b and 2 = (a, b] which includes b but not a.
All of these intervals are subsets of so O ⇢ etc.
Notice that we have a square bracket if the end-point is included and
a round bracket if it is not.

If we write {a, b} for the set containing two points a and b only
then (a, b) [ {a, b} = [a, b] but (a, b) \ {a, b} = ;.
Also 0 2/ (0, 1) but 1 2 [0, 1] and 1 2 (0, 2) and = ( 1, 1).

Sets are very important in defining functions and in probability theory

and statistics which are subjects of great importance in biochemistry
and pharmaceuticals.

2. FUNCTIONS: DEFINITIONS .

Consider two sets, a source set and a target set . Then a function
or mapping is a rule which associates an element (or several elements)
of with each element of . In other words if you are given x 2
then we can calculate an element y 2 (or perhaps more than one
element of ). A function is not necessarily expressed by an algebraic
formula. The rule involved may be expressed in some other way and
we shall see some examples..

We write y = f (x) and sometimes refer to x as the input or in-

dependent variable and y as the output or dependent variable. The
letter f denotes the function. We also write f : ! .

3
The source set is called the domain of f . The subset of onto
which maps is called the range of f . In other words for each ele-
ment of the range we can find an element of the domain which maps
onto it but there are perhaps some elements of the target for which we
can find no corresponding element of .
Note that the range and the target are not necessarily the same.

Some examples:

1. f : ! where f (x) = x2 + 5 .
Here the domain is and the target is also but the range is
[5, 1) .
2. f : ( 1, 0) [ (0, 1) ! where f (x) = 1/x .
Here the domain is ( 1, 0) [ (0, 1) , the target is and the
range is ( 1, 0) [ (0, 1) .
3. f : ! where f (x) = [x] and [x] is the largest integer less
than or equal to x. This is called the staircase function and a
graph will be presented.
Notice that we have not used an algebraic formula to define f .
4. f : ! where f (x) = sin x.
Here the domain is , the target is and the range is [ 1, 1] .
5. f : [0, 1) ! where if we are given x 2 [0, 1) we calculate y
2
from y = x .
Note that x must be positive
p or zero and that for each value of x
we get two values of y = ± x . This is a two-valued function.
The target and the range are both . Two-valued functions
are very awkward so we make this into a one-valued function by
eliminating the negative values of y and in this case the range
becomes [0, 1).
6. f : [ 1, 1] ! where if we are given x 2 [ 1, 1] we calculate y
from sin y = x .
Note that for each value of x we get many values of y = sin 1 x .
Again we make this into a one-valued function by keeping only
the values of y in the range [ ⇡/2, ⇡/2] .
We now have f : [ 1, 1] ! [ ⇡/2, ⇡/2].

It is always desirable to have one-valued functions and we shall

assume our functions are one-valued from now on.

4
3. SPECIAL FUNCTIONS.

Here are some special types of functions:

1. Even functions have the property f ( x) = f (x)

provided both x and x are in the domain of f .
2. Odd functions have the property f ( x) = f (x)
provided both x and x are in the domain of f .

You can see from a graph whether a function is even or odd.

Many functions are neither even nor odd but all can be decom-
posed into a sum of an even function and an odd function (pro-
vided the domain is symmetrically disposed with respect to the
origin):
1h i 1h i
f (x) = f (x) + f ( x) + f (x) f ( x)
2 2
The first term is even and the second odd.

Thus if for example f (x) = ex we have

f (x) = cosh x + sinh x

where cosh x = 12 (ex + e x ) which is even and sinh = 12 (ex e x)

which is odd.
3. A function with domain is periodic if it repeats after some period T
i.e. if f (x + T ) = f (x).
Note that if f is periodic with period T then f (x + nT ) = f (x)
for any integer n.

Here are some examples of periodic functions:

f (x) = sin x and g(x) = cos x are periodic with period 2⇡.
For these functions the domain is and the range is [ 1, 1].

h(x) = tan x is periodic of period ⇡.

The range is but the domain is excluding the points n⇡/2
where n is an integer.

5
4. There are four types of monotonic functions:
f (x) is monotonic increasing if when x1 < x2 then f (x1 )  f (x2 ).
f (x) is strictly monotonic increasing if when x1 < x2 then f (x1 ) < f (x2 ).
f (x) is monotonic decreasing if when x1 < x2 then f (x1 ) f (x2 ).
f (x) is strictly monotonic decreasing if when x1 < x2 then f (x1 ) > f (x2 ).

Examples:
f : ! where f (x) = 2x + a is strictly monotonic increasing.

f : [0, 1] ! where f (x) = x2 is strictly monotonic increasing.

f: ! [ 1, 1] where f (x) = cos x is not monotone.

f : [0, ⇡] ! [ 1, 1] where f (x) = cos x is strictly monotonic decreasing.

f : ! [0, 1] where f (x) = 0 if x < 0 and f (x) = 1 if x 0 is mono-

tonic increasing. This is called the step or Heaviside function.
5. We now define inverse functions.

If f : ! is strictly monotonic with the domain and

the range then its inverse exists.

The inverse of f is a function f 1 : ! with the following property:

If y 2 and x is the corresponding value in such that y = f (x)
then f 1 (y) = x. Now is the domain and is the range.

Another way of saying the same thing is that

1 1
f (f (x)) = x or f (f (y)) = y.

You can understand the inverse function graphically. If you draw

a graph of f (x) vs x and then swap around the x and y axes you
get a graph of f 1 (y).
You can also get f 1 by solving the equation y = f (x) for x in
terms of y.

1
Warning: It is a very bad mistake to put f (y) = 1/f (y).

6
Examples:

1. f : ! and f (x) = ax + b with a 6= 0 then f is strictly

monotonic and f 1 (y) = (y b)/a
2. f : [0, 1) ! [0, 1) and f (x) = x2 then f is strictly monotonic
p
and f 1 (y) = y.
It is necessary to restrict the domain of f to positive values of x
so that f is strictly monotonic. If we do not restrict the values of
x then f 1 is two-valued and if we do not restrict the values of y
then we have to take the square root of a negative number which
is not possible (if we only have real numbers).
3. f : ! (0, 1) and f (x) = ex then f is strictly monotonic and
f 1 (x) = ln y
Note that f 1 : (0, 1) ! so ln y is only defined for positive
arguments.
4. f : [ ⇡/2, ⇡/2] ! [ 1, 1] and f (x) = sin x then f is strictly
monotonic and f 1 (y) = sin 1 x.
Note that f 1 : [ 1, 1] ! [ ⇡/2, ⇡/2] so that f 1 = sin 1 x is
only defined for x 2 [ 1, 1]. Also if we do not restrict the domain
of f to [ ⇡/2, ⇡/2] then f 1 is multi-valued.

7
 TOPIC B CALCULUS
1. LIMITS

Limits are a key concept in calculus. They are concerned with what
happens to a function when the argument approaches a value without
ever reaching it. They are also useful in defining functions when the
argument gets very large either positively or negatively.
sin x
Example: Consider f (x) = . This function is not defined at
x
x = 0 because to evaluate it you have to try to evaluate 0 divided by
0 which is not allowed. What we have to do is to see what happens as
x approaches 0. If you use your calculator to evaluate f(0.1), f(0.01),
f(0.001), f(0.0001) you will soon see that f (x) approaches 1 as x ap-
proaches 0.

Here are some important results:

(a) We now have the idea of a limit:

If f (x) approaches a constant c as x approaches a then c is called
the limit of f (x) as x tends to a.
The notation is x!a lim f (x) = c.
(b) The example above suggests but does not prove that
sin x
lim = 1.
x!0 x

The proof is actually quite difficult so we just take it for granted.

But nevertheless you are expected to know this result.

(c) Sometimes limits do not exist.

Consider for example lim 1/x.
x!0

(d) We can also define right and left sided limits:

A right sided limit is one where the variable x approaches a from
the right i.e. x is always greater than a as it approaches a.
A left sided limit is one where the variable x approaches a from
the left i.e. x is always less than a as it approaches a.

8
 We write
limx# f (x) = c1 for the limit from the right and
limx" f (x) = c2 for the limit from the left.
It may be that c1 6= c2 in which case x!a
lim f (x) does not exist.
However if c1 = c2 then the limit does exist and lim f (x) = c1 = c2 .
x!a

Example:
Suppose f (x) = sgn(x)
(i.e. f (x) = 1 if x > 0 and f (x) = 1 if x < 1)
then lim f (x) = 1 and lim f (x) = 1.
x"0 x#a
In this case lim f (x) does not exist.
x!a

(e) Finally if lim f (x) = f (a) then f (x) is said to be continuous at

x!a
x = a.
Note that if a function is continuous at evey point then the func-
tion is called a continuous function.

Some examples:

1. Suppose f (x) =| x |. Then lim f (x) = 0 and lim f (x) = 0.

x"0 x#0
So the limit exists and lim f (x) = 0.
x!0
Further f (0) = 0 so the function is continuous at x = 0.
2. Suppose f (x) = 0 if x < 0 and f (x) = 1 if x 0.
Then lim f (x) = 0 and lim f (x) = 1.
x"0 x#0
So the limit lim f (x) does not exist.
x!0
This function is called the Heaviside function and is quite im-
portant because it represents a switch.
3. Suppose f (x) = 1/x.Then lim f (x) = 1 and lim f (x) = 1.
x"0 x#0
So the limit does not exist.
(However when we say that a function tends to 1 we really mean
the function increases indefinitely as x approaches a.)

9
2. CALCULATION OF LIMITS.

If both lim f (x) and lim g(x) exist and we know their values then we
x!a x!a
can calculate the following combinations very easily (except in certain
indeterminate cases which will be discussed below):

lim [f (x) ± g(x)] = x!a

(a) x!a lim f (x) ± x!a
lim g(x).

lim [f (x) ⇥ g(x)] = x!a

(b) x!a lim f (x) ⇥ x!a
lim g(x).
!
f (x) limx!a f (x)
(c) lim = provided lim g(x) 6= 0.
x!a g(x) limx!a g(x) x!a

(d) lim f (g(x)) = f (lim g(x)) provided f (x) is continuous in a

x!a x!a
range including the point x = lim g(x).
x!a

Some examples:

1. lim xµ = aµ
x!a
but care is needed and we require that µ be positive when a = 0.
cos x cos x
2. lim = 1 and lim = 1
x#0 x x"0 x
sin ax
3. lim = a
x!0 x

The indeterminate forms involve situations where two of the limits

are infinite or the limit of a function in the denominator is zero and we
have to use a variety of methods to deal with them.

10
Examples:

1. In some cases algebraic manipulation works.

x2 1
In the case lim we notice that both numerator and de-
x!1 x 1
nominator tend to 0. If we factorise numerator and denominator
we get:
x2 1 (x 1)(x + 1)
lim = lim = lim (x + 1) = 2.
x!1 x 1 x!1 x 1 x!1

2. Here we use trigonometric manipulation and also result (b) above

1 cos 2x
to evaluate lim
x!0 x
We use the trig identity 1 cos 2x = 2 sin2 x and then get
 ✓ ◆✓ ◆
2 sin2 x sin x sin x
lim = lim 2x = 0 ⇥ 1 ⇥ 1 = 0.
x!0 x x!0 x x
3. Here we use trigonometric manipulation and also results (b) and
1 cos x
(d) above to evaluate lim
x!0 x2
We use the trig identity 1 cos x = 2 sin2 x/2 and then get
" ! !#
2 sin2 x/2 1 sin x/2 sin x/2 1 1
lim 2
= lim = ⇥1⇥1=
x!0 x x!0 2 x/2 x/2 2 2
x3 + x2 + 1 1+x 1+x 3
1
4. x!1
lim = lim = .
3 x3 + 2 x x!1 3 + 2x 2 3
(1 + x)1/2 1
5. To evaluate lim you multiply inside the limit by
x!0 x
(1 + x)1/2 + 1
to get
(1 + x)1/2 + 1
(1 + x) 1 x 1
lim = lim = .
x!0 x((1 + x)1/2 + 1) x!0 x[(1 + x)1/2 + 1] 2
This can also be evaluated using Taylor series which we will discuss
later.
x2 + 1 x 1+x 3 0
6. lim 3
= lim 2
= =0
x!1 3 x + 2 x x!1 3 + 2 x 3

11
7. We now anticipate a result of calculus. There is a procedure
called L’Hopital’s Rule which helps you to evaluate limits of
f (x)
the form lim where both f (x) and g(x) are continuous at
x!a g(x)
x = a andf (a) = 0 and g(a) = 0 and g 0 (a) 6= 0.

The rule states that

f (x) f 0 (a)
lim = 0 .
x!a g(x) g (a)
1
If f (x) = (1 + x) 2 1 and g(x) = x and a = 0 then we have the
1
same limit as example (5). We now have f 0 (x) = 12 (1 + x) 2 and
g 0 (x) = 1 so f 0 (0) = 12 and g0(0) = 1 and therefore

(1 + x)1/2 1 1/2 1
lim = = .
x!0 x 1 2

12
3. DIFFERENTIAL CALCULUS.

We calculate the slope of the secant of the function f (x) between the
points x0 and x0 + x from the formula

f (x0 + x) f (x0 )
SLOPE = .
x
We get the slope of the tangent to the curve at x = x0 (otherwise
known as the derivative) by taking the limit of the slope of the secant
as x ! 0.

********************************************************************

This gives the definition of the derivative of f (x) at the point x

d f (x0 + x) f (x0 )
f (x0 ) = lim .
dx x!0 x

********************************************************************
YOU MUST COMMIT THIS DEFINITION TO MEMORY.

There is an alternative more concise notation where we write f 0 (x0 )

for the derivative.

If the limit does not exist or is ±1 then the derivative does not exist.
In other words the graph of the function does not have a properly de-
fined tangent at the point.

We now try to calculate the derivative of some functions:

(a) Suppose f (x) = x2 .

(x + x)2 x2
Then f 0 (x) = lim
x!0 x
(x2 + 2 x x + x2 ) x2
= lim
x!0 x
2 x x + x2
= lim = 2 x.
x!0 x

13
 (b) Suppose f (x) =| x |. We now calculate limits from the right
(where x > 0) and from the left (where x < 0) both at x = 0.

|0+ x| |0|
Limit from right = lim =1
x#0 x
because the numerator and denominator are both positive.

|0+ x| |0|
Limit from left = lim = 1
x"0 x
because the numerator is positive but the denominator is negative.

These are di↵erent so the derivative does not exist at x = 0 and

the function is not di↵erentiable at that point.

However you can calculate the derivative at any point x > 0 and
you get f 0 (x) = 1 and you can also calculate the derivative at any
point x < 0 and you get f 0 (x) = 1.

We now present a table of derivatives which you MUST memorise. If

you do not memorise them you are not taking the course seriously. It
is a good exercise to prove them for yourself.

f (x) = const ) f 0 (x) = 0

f (x) = x↵ ) f 0 (x) = ↵ x↵ 1
f (x) = ex ) f 0 (x) = ex
f (x) = sin x ) f 0 (x) = cos x
f (x) = cos x ) f 0 (x) = sin x
f (x) = tan x ) f 0 (x) = sec2 x
f (x) = tan 1 x ) f 0 (x) = 1/(1
q
+ x2 )
f (x) = sin 1 x ) f 0 (x) = 1/ (1 x2 )

14
You also need to know how to di↵erentiate combinations of functions.
In what follows u(x) and v(x) are any di↵erentiable functions of x.

1. SUM RULE:
f (x) = u(x) + v(x) ) f 0 (x) = u0 (x) + v 0 (x).
2. PRODUCT RULE:
f (x) = u(x) v(x) ) f 0 (x) = u0 (x) v(x) + u(x) v 0 (x).
3. QUOTIENT RULE:
f (x) = u(x)/v(x) ) f 0 (x) = (u0 (x) v(x) u(x) v 0 (x))/v(x)2 .
4. CHAIN RULE:
f (x) = u(v) where v = v(x) ) f 0 (x) = u0 (v) v 0 (x).
df du dv
Or alternatively = .
dx dv dx
This enablespyou to di↵erentiate a function of a function
e.g. f (x) = 1 + x2 where u(v) = v 1/2 and v(x) = 1 + x2 .
5. INVERSE FUNCTION RULE:
Suppose y = f 1 (x) then we first write x = f (y), then calculate
dx/dy = f 0 (y) and finally write dy/dx = (f 0 (y)) 1 .
! 1
dy dx
In summary = .
dx dy
6. IMPLICIT FUNCTIONS:
Suppose x and y are related by an implicit equation f (x, y) = 0
which we cannot solve for y in terms of x or for x in terms of
dy
y. Then it is still possible to calculate in terms of partial
dx
derivatives. We shall come to this later.

With the help of the rules above you should be able to di↵erentiate any
combination of simple functions no matter how complicated.
You MUST remember these rules.

15
4. EXTREME VALUES, ASYMPTOTES AND CURVE SKETCHING.

(a) EXTREME VALUES.

One of the points of di↵erential calculus is to

be able to tell when functions have extremes
or stationary points or turning points.
In the graph the function has a maximum at A
and a minimum at B. At these points the
tangent is horizontal so the slope is zero. The slope
is also zero at the point C where the tangent crosses
the curve. C is not a maximum or minimum but it is a
point of inflexion with horizontal tangent. The
point D is also a point of inflexion because the tangent
crosses the curve but the tangent is not horizontal.

In order to locate and distinguish these points we need to know

how to di↵erentiate the function once or twice or many times de-
pending on the circumstances. If you di↵erentiate twice you get
the second derivative written in one of these ways:
!
d df d2 f
= = f 00 (x) = f (2) (x).
dx dx dx2

Third and higher derivatives are usually written f (n) (x) with n = 3
or more to save counting dashes. You then have the following rules
for a point P on the curve with coordinates (x, y):
1. If f 0 (x) = 0 and f 00 (x) > 0 then there is a minimum at P –
type B.
2. If f 0 (x) = 0 and f 00 (x) < 0 then there is a maximum at P –
type A.
3. If f 0 (x) = 0 and f 00 (x) = 0 you need more information to
determine the nature of P .
4. If f 0 (x) = 0 and f 00 (x) = 0 but f (3) (x) 6= 0 then there is a
point of inflexion with horizontal tangent at P – type C.
5. If f 0 (x) 6= 0 and f 00 (x) = 0 and f (3) (x) 6= 0 there is a point of
inflexion at P but the tangent is not horizontal – type D.

16
 (Other cases rarely arise and are more complicated. For the
record:
If the first derivative which is not 0 is of even order then the func-
tion has a maximum or minimum;
But if the first derivative which is not 0 is of odd order then you
have a point of inflexion with horizontal tangent;
Finally if f 0 (x) 6= 0 and all other derivatives up to but not includ-
ing an odd one are zero then you have a point of inflexion but the
tangent is not horizontal.)

Examples:

Find all the extreme values of the following functions and de-
termine their nature. Find also all points of inflection. Sketch the
curves.
1. f (x) = 2 x2 + 3 x + 1
2. f (x) = x3 2 x2 x + 2
3. f (x) = sin2 x sin x + 2

(b) ASYMPTOTES.

Asymptotes describe the behaviour of a curve y = f (x) when the

values of the input x or the output y are very large and they are
a help in curve sketching. We will only discuss horizontal and
vertical asymptotes. (There are other kinds.)

If lim f (x) = c then the line y = c is a horizontal asymptote.

x!±1

lim f (x) = ±1 then the line x = a is a vertical asymptote.

If x!a

Example:
3x
Suppose f (x) = .
x+4
Then lim f (x) = 3 and also lim f (x) = ±1.
x!±1 x! 4
Thus the line y = 3 is a horizontal asymptote and the line x = 4
is a vertical asymptote. Notice that the vertical asymptotes arise
from the zeroes of the denominator.

17
 Also f (x) = 0 at x = 0. Further at a large positive value of
x say x = 100 we have f (x) < 3 and at a large negative value of
x say x = 100 we have f (x) > 3.

This enables us to draw a rough sketch of the graph of y = f (x).

(c) CURVE SKETCHING.

We now present a method for sketching more general curves where

there are extreme points as well as asymptotes.

The information you need is as follows:

1. The location and type of all extreme points including points
of inflection.
2. All asymptotes including information as to the direction from
which the curve approaches them.
3. The location of all intercepts with the horizontal and vertical
axes.
Examples:
Sketch graphs of the following functions:
x2 1
1. f (x) = 2
x 4
x
2. f (x) = 2
x 1

5. POLAR CO-ORDINATES.

You can specify the position of a point P in the plane either in

terms of Cartesian co-ordinates (x, y) or in terms of polar co-
ordinates (r, ✓) where r is the distance from the origin O to the
point P and ✓ is the angle between the x-axis and the line OP
(see diagram).

Any equation which defines a curve in terms of Cartesians can

also be written in terms of polars and vice versa. It is sometimes
much simpler to use polars than Cartesians particularly when the
problem has a natural centre to which we can refer the position of

18
a point. This is the case in many Quantum Mechanical problems
concerning atoms and molecules which come up in Chemistry.

Again referring to the diagram we notice that

x = r cos ✓ and y = r sin ✓ (1)
q y
r = x2 + y 2 and tan ✓ = . (2)
x
Note that r is always positive and there are many values of ✓ that
satisfy the equation (2). ✓ must be chosen carefully.

(a) If x 0 then ✓ is chosen in the range [ ⇡/2, ⇡/2] or to put

it another way ⇡/2  ✓  ⇡/2.

(b) If x < 0 and y 0 then ✓ is chosen in the range (⇡/2, ⇡] or

to put it another way ⇡/2 < ✓  ⇡.

(c) If x < 0 and y < 0 then ✓ is chosen in the range ( ⇡, ⇡/2)

or to put it another way ⇡ < ✓ < ⇡/2.

You can write the equation of a curve in Cartesians or in polars.

For example the equation of a circle centre O is x2 + y 2 = a2 .
We can get the corresponding equation in polars by substituting
equations (1) and get r = a which is much simpler and very
easy to interpret.

The method is general - you take equations (1) and substitute

into the Cartesian form and then simplify but you must remem-
ber that r is positive by its nature. If you have the polar version
of the equation you take equations (2) substitute into the polar
form and then simplify.

If you have the polar form then draw up a table of values of r

against ✓ at simple values of ✓ (and possibly some intermediate
values) and then connect up points. Simple values are:
0, ±⇡/6, ±⇡/4, ±⇡/3, ±⇡/2, ±2⇡/3, ±3⇡/4, ±5⇡/6, ±⇡, . . .

Example:
Draw graphs of the following polar equations r = a cos 2✓ and
r = a sin 3✓

19
5. INTEGRAL CALCULUS.

There are two ways of understanding the meaning of an integral. The

link between them is given by what is sometimes called the Funda-
mental Theorem of Calculus.

The first way of understanding integrals is in terms of in-

definite and definite integrals:

1. The indefinite integral of a function f (x) is defined to be the

function F (x) which when di↵erentiated gives back f (x).

Z
dF (x)
Thus = f (x) which is also written f (x) dx = F (x)+c.
dx
The arbitrary constant c is added because on di↵erentiating it
gives 0. You must always put in the arbitrary constant explicitly
when integrating. This form of integral is called the indefinite
integral.

At the end of this section you should be able to evaluate indefinite

integrals of polynomials, exponentials, trigonometric functions, ra-
tional functions and logarithms. You should be able to change the
variable of integration and you should know about integration by
parts which is a technique that enables you to integrate products
of these functions and some other examples.

2. Another form of integral is the definite integral. It is given in

terms of the function F (x) above as follows:
Z b
f (x) dx = F (b) F (a).
a

20
The second way of understanding an integral is as an area:

Consider the area A between the curve y = f (x), the x axis and the
two vertical lines x = a and x = b. (See diagram.)

You can calculate the area as follows (see diagram above):

1. Divide up the area to be evaluated into m thin strips of width

xm = (b a)/m called elements. Strip i is at position xi .
2. Then evaluate the area of each element ignoring the fact that
the element has a sloping and curving top. Area of strip i is
Ai ⇡ f (xi ) xm .
3. Add up these areas to get an approximation to the whole area.
P
A⇡ m i=1 f (xi ) xm .

4. Take the limit of the sum as the width of each element goes to 0.
(You must remember of course that the number of elements goes
up as the width goes down.)
m
X
A = lim f (xi ) xm .
m!1
i=1

5. This gives the exact area under the curve.

Note also that addition of areas is interpreted in an algebraic sense
so that areas below the a-axis are taken as negative.

21
*******************************************************************

The Fundamental Theorem of Calculus states that the definite

integral is equal to the area A obtained above:
Z b
f (x) dx = A
a

******************************************************************

We can now see that the name given to the variable of integration
has no significance. (It is called a dummy variable.) So we can write:
Z b Z b
f (x) dx = f (t) dt
a a

and we can also write:

d Zx
f (t) dt = f (x)
dx a
This follows from the fact that
Z x
dF (x)
f (t) dt = F (x) F (b) and = f (x).
a dx

The point of the Fundamental Theorem is that it is generalisable and

gives a method for finding many properties of geometrically extended
objects in addition to areas. For example we can use this method to
calculate by integration such quantities as lengths of curves, areas, vol-
umes, centroids, moments of inertia and many others.

We now go to the practical problem of how to calculate integrals.

There are basically only three methods for integration:
inspection, integration by parts and substitution.

22
1 INSPECTION

Some integrals can be evaluated by inspection. Try to manip-

ulate the integral into a form you recognise and then ask yourself
what function
Z F (x) returns f (x) when di↵erentiated.
Then f (x) dx = F (x) + c.

Some simple standard integrals and results which MUST be re-

membered so that they can be recognised:
Z
(a) a dx = a x + c
Z
xn+1
(b) xn dx = + c provided n 6= 1
Z
n+1
1
(c) dx = ln | x | +c
Z
x
1
(d) ekx dx = ekx + c
Z
k
1
(e) sin kx dx = cos kx + c
Z
k
1
(f) cos kx dx = sin kx + c
Z
k
1
(g) tan kx dx = ln | cos kx | +c
Z k
(h) sec x dx = ln | (sec x + tan x) | +c
Z
1 1 x
(i) dx = tan 1 + c
+x 2 a2 a a
(j) Partial Fractions.
This enables you to deal with ratios of polynomials. A simple
example follows:
Z Z ✓ ◆
1 1 1 1 1 x 1
dx = dx = ln +c
1 x2 2 x 1 x+1 2 x+1
(k) Linearity.
This rule enables you to split complicated functions into sim-
pler parts each of which can be treated by inspection.
Z ⇣ ⌘ Z Z
a f (x) + b g(x) dx = a f (x) dx + b g(x) dx

N.B. The use of modulus signs in (c), (g), (h) and (j) is designed
to ensure that the argument of the ln function is positive.

23
 Here are some examples of integration by inspection:
Z
1. x5 dx. We note that when we di↵erentiate x6 we get 6 x5
so
Z
the integral equals x6 /6 + c.
2. sin x cos x dx. We note that when we di↵erentiate sin2 x
we get 2 sin x cos x so the integral equals 1
2
sin2 x + c.
Z
3. (x3 + sin x + 2/x)d x = x4 /4 cos x + 2 ln x + c.
Z Z
x x 1+1
4. dx = dx
xZ ✓1 ◆ x 1
1
= 1+ dx = x + ln | x 1 | +c
x 1

2 CHANGE OF VARIABLE

The formula below is a consequence of the formula for change

of variable in di↵erentiation:
Z ⇣ ⌘ du Z
f u(x) dx =
f (u) du
dx
Here we are substituting one variable u for some complicated ex-
pression involving x. We use the formula to go from a complicated
integral to a simple integral which can be evaluated by inspection.
It is sometimes necessary to go from left to right and sometimes
necessary to go from right to left.
In the examples below the first three go L to R and the fourth
goes R to L.

Examples:
Z
x
1. We wish to evaluate dx .
1 + x2
du
We put u = 1 + x2 and then = 2 x.
dx Z
1 1
Consequently the integral equals du = ln u + c.
2 u
Z
2. sin ✓ cos ✓ d✓.
Z
2
3. x ex dx.
Z
1
4. p dx.
1 + x2

24
 One very commonly used substitution is called the t-substitution.
It is used for integrating rational expressions which involve trigono-
metric functions in numerator and denominator and reduces these
expressions to ratios of polynomials.
The substitution is
x
t = tan .
2
It is a matter of easy trigonometry to show the following:
2t 1 t2 2t 2 dt
sin x = , cos x = , tan x = , dx =
1 + t2 1 + t2 1 t2 1 + t2
You should remember these.

Example: Z
1
Evaluate the following integral dx.
1 + cos x
Using the t-substitution we find the integral is equal to
Z Z
1 t2 1 2 dt x
(1 + ) = 1 dt = t + c = tan + c
1 + t2 1+t 2 2

3 INTEGRATION BY PARTS

The formula below is a consequence of the formula for di↵erenti-

ating a product:
Z Z
dv du
u(x) dx = u(x) v(x) v(x) dx
dx dx
When we use it we must factorise the integrand (the function to
dv
be integrated) into two factors u(x) and . The second factor
dx
must be easy to integrate and the first factor must give something
simpler when di↵erentiated.
Examples:
Z
dv
1. x sin x dx. Here we put u(x) = x and = sin x
dx
du
Then = 1 and v(x) = cos x.
dx
We
Z now substitute in the Zrhs of the formula and get
x sin x dx = x cos x + cos x dx = x cos x + sin x + c.

25
 Z
dv
2. ln x dx. Here we put u(x) = ln x and =1
dx
du 1
Then = and v(x) = x.
dx x
We
Z now substitute into
Z the formula and get
ln x dx = x ln x 1 dx = x ln x x + c.
Z
dv
3. ex sin x dx. Here we put u(x) = sin x and = ex .
dx
du
Then = cos x and v(x) = ex . We now substitute into
dx
the formula and get
Z Z
ex sin x dx = ex sin x ex cos x dx. (1)

The last term is as bad as what we started with so we have

to do integration by parts a second time on the last term.
dv
We now put u(x) = cos x = ex .
dx
du
Then = sin x and v(x) = ex .
dx
We now substitute into the formula and get
Z Z
x x
e cos x dx = e cos x + ex sin x dx. (2)

We combine equations (1) and (2) and get

Z Z
ex sin x dx = ex sin x ex cos x ex sin x dx.

The last term is the negative of what we started with at the

very beginning so we take it over to the lhs and divide by 2
and (remembering to put in the arbitrary constant) we get
Z
ex sin x dx = 12 (ex sin x ex cos x) + c.

This is quite a lot of work but we get two for the price of
one because substituting this last into equation (2) we get
Z
ex cos x dx = 12 (ex sin x + ex cos x) + c.

26
TOPIC C. SEQUENCES AND SERIES.
I INFINITE SEQUENCES.

An infinite sequence is a function defined on the domain which takes

on values which are usually in the range .
i.e. it takes real values on the set of integers greater than 0
i.e. on the set {1, 2, 3, 4, . . .}

The sequence is usually written

a1 , a 2 , a 3 , a 4 , a 5 . . . or sometimes {an }1
n=1
The subscript n on an is the index which tells you how far you are
along in the sequence.

There is a rule or formula which tells you how to calculate the terms
in the sequence such as e.g. an = 1/n or bn = ln n.

1. A sequence converges if n!1 lim an exists and is finite.

If the limit does not exist or is infinite then the sequence diverges.
In the two examples above an converges but bn diverges.

2. A sequence is increasing if the values of an are always increasing

i.e. if an+1 an for all n.
A sequence is decreasing if the values of an are always decreasing
i.e. if an+1  an for all n.

3. A sequence is bounded above if there is a number c such that

an  c for all n.
A sequence is bounded below if there is a number c such that
an c for all n.

4. We have the following important results:

An increasing sequence that is bounded above is convergent.
A decreasing sequence that is bounded below is convergent.
The sequence an = 1/n is bounded below by 0 and decreasing
so it converges.
The sequence bn = ln n is increasing but it is not bounded
above so it diverges.

27
 Examples:
Use the four statements above to categorise the following sequences:
an = ( 1)n , bn = ( 1)n /n, cn = 1 1/n2 , dn = n

II FINITE AND INFINITE SERIES.

A finite series is the sum of all the terms of an infinite sequence

up to a certain value n = N at which it ceases
N
X
i.e. SN = an = a1 + a2 + a3 + a4 + . . . + aN
n=1
This is called a partial sum of the series or the sum to N terms.

An infinite series is the sum of all the terms in an infinite sequence

and is evaluated as follows:
N
X
S = lim SN = lim an
N !1 N !1
n=1

The series is convergent if the limit exists and is finite but divergent
otherwise.

III TWO IMPORTANT SERIES.

An arithmetic series has terms of the form an = a + b n.

The constant a + b is the first term and b is called the common di↵er-
ence. It is easy to find the sum to N terms and the proof is by induction
N
X
(a + bn) = a N + b N (N + 1)/2
n=1

This series is always divergent.

A geometric series has terms of the form an = a rn 1 .

The constant a is the first term and r is the common ratio. It is easy
to find the sum to N terms and the proof is by induction
N
X 1 rN
a rn 1
=a
n!1 1 r

28
 This series is convergent if | r |< 1 and divergent otherwise. This
demonstrates that a series can be covergent or divergent depending on
the value of some parameter, r in this case.

Both of these cases are fairly obvious because we can calculate the
sum to N terms exactly but in most cases we cannot. We need some
way of testing whether a series converges or not in these cases.

To illustrate these points consider the convergence or divergence of

the following series and calculate a few partial sums to illustrate your
results: 1 1 1
X X X
( 1)n 1
(2)n 1
(1/2)n 1

n=1 n=1 n=1

IV TESTS FOR CONVERGENCE AND DIVERGENCE.

1
X
The following tests give you information about whether a series an
i=1
converges or diverges.

(a) Limit Test.

lim an 6= 0 then the series diverges.

If n!1
Conversely if the series converges then lim an = 0.
n!1

This is a very easy test to apply so you should always carry out
this test right at the start.

Examples:
1 ✓
X ◆ 1 ✓ ◆
X 1 ✓
X ◆
1 1 1
1
n=1 n2 n=1 n n=1 n2

29
(b) Ratio Test.

| an+1 |
Calculate t = n!1
lim .
| an |
If t < 1 then the series converges.
If t > 1 then the series diverges.
If t = 1 then you do not know.

This is also a very easy test and it should be the second one
you try.

Examples:
1 ✓
X ◆ 1 ✓
X ◆ 1 ✓ n
X ◆ 1 ✓ ◆
X 1 ✓
X ◆
1 1 3 +1 1 1
n
2 +1 n! n
2 +1 n n2
n=1 n=1 n=1 n=1 n=1

(c) Integral Test.

Suppose you can find a function f (x) from to such that

an = f (n) and f (x) is continuous, positive and decreasing.
Z 1
Then the series is convergent if f (x) dx is finite.
Z 1 1
It is divergent if f (x) dx is not finite or does not exist.
1

Examples:
1 ✓
X ◆ 1 ✓ n
X ◆ 1 ✓ ◆
X 1 ✓
X ◆
1 3 +1 1 1
n
2 +1 n
2 +1 n n2
n=1 n=1 n=1 n=1

30
(d) Alternating Series.

Consider a series with an = ( 1)n bn where bn 0 for

every value of n.
Then the series converges if the sequence bn is decreasing and
lim bn = 0.
n!1

Examples:
1 ✓
X ◆ 1 ✓
X ◆
11 n
( 1)n ( 1)n 1
2
n=1 n n=1 n +1

31
V POWER SERIES AND TAYLOR SERIES.

(a) Power Series and Radius of Convergence.

These are functions of x defined by the formula:

1
X
cn (x b)n = c0 + c1 (x b) + c2 (x b)2 + c3 (x b)3 + . . .
n=0

Many functions including sines and cosines can be written in terms

of power series as we shall see.
We shall mostly use power series in which b = 0 which are there-
fore of the form:
1
X
cn xn = c0 + c1 x + c2 x2 + c3 x3 + . . .
n=0

One of the important properties of a power series is the radius of

convergence which determines the range of values of x for which
the series converges. We find this by the ratio test:
!
| cn+1 (x b)n+1 | | cn+1 |
t = lim = lim |x b|
n!1 | cn (x b)n | n!1 | c |
n

By the ratio test the series converges if t < 1. In other words if

| cn |
|x b |< R where R = n!1
lim
| cn+1 |
R is called the radius of convergence because it defines the region
of the x-axis centred at b in which the series converges. The series
diverges outside this region because there t > 1. We can’t say
anything about the ends of the region.

32
 Examples:

Calculate the radius of convergence of

1
X xn x2 x3
=x+ + + ...
n=1 n 2 3
Calculate also the radius of convergence of
1
X xn x x2 x3
=1+ + + + ...
n=0 n! 1! 2! 3!

(b) Taylor Series.

If you are given a function such as sin x or cos x and you can
di↵erentiate it as many times as you like then you can get an ex-
pansion of the function as a Taylor series about the point at which
the derivatives are evaluated. All the derivatives of the Taylor se-
ries match up to the derivatives of the original function at the
point about which the expansion is made.
Suppose the expansion is made about the point x = 0 and the
function is f (x) then

f (1) (0) f (2) (0) 2 f (3) (0) 3 f (4) (0) 4

f (x) = f (0) + x+ x + x + x +...
1! 2! 3! 4!
or in summation notation
1
X f (n) (0) n
f (x) = x .
n=0 n!

You should check that the nth derivative of the series at x = 0 is

equal to the nth derivative of the function at the point x = 0

*****************************************************
There is an important condition. The function is only equal to the
series if the series converges so you should check for convergence.
If the series doesn’t converge it is useless.
*****************************************************

33
If you need to expand about the point x = b you replace the 0’s
in the above expression by b’s and the x’s by (x b)’s. This gives
1
X f (n) (b)
f (x) = (x b)n .
n=0 n!

The derivatives are now evaluated at x = b.

Again you must check convergence.

Examples:

Derive the following Taylor series for the functions given and cal-
culate the radius of convergence:
1
X xn
(1) ex =
n=0 n!
1
X x2 n 1
(2) sin x = ( 1)n+1
n=1 (2 n 1)!
1
X x2 n
(3) cos x = ( 1)n
n=0 (2 n)!
1
Xxn
(4) ln(1 x) =
n=1 n
(5) Use the series in (1), (2) and (3) above to calculate the
functions at x = 0.1 and x = 1 and check with your calculator
using radian mode. Also check the series in (4) at x = 0.1.
Why can’t you use x = 1 in case (4)?

34
(c) L’Hopital’s Rule.

Suppose you have two continuous functions f (x) and g(x) and
suppose also that f (a) = g(a) = 0.
f (x)
How do you then evaluate lim ?
x!a g(x)

L’Hopital’s rule states that provided g 0 (a) 6= 0 then

f (x) f 0 (a)
lim = 0
x!a g(x) g (a)
which is well defined.

The proof goes as follows:

We can expand both functions in Taylor series about x = a thus

f 00 (a) f (3) (a)

f (x) = f (a) + f 0 (a) (x a) + (x a)2 + (x a)3 + . . .
2! 3!
g 00 (a) g (3) (a)
g(x) = g(a) + g 0 (a) (x a) + (x a)2 + (x a)3 + . . .
2! 3!
Remember f (a) = g(a) = 0, take the ratio and divide numerator
and denominator by (x a) to get
f 00 (a) f (3) (a)
f (x) f 0 (a) + 2!
(x a) + 3!
(x a)2 + . . .
lim
x!a g(x)
= lim
x!a 0 g 00 (a) g (3) (a)
g (a) + 2!
(x a) + 3!
(x a)2 + . . .
0
f (a)
= 0
g (a)

because all terms proportional to (x a) go to 0.

If f 0 (a) = g 0 (a) = 0 but g 00 (a) 6= 0 also then we can rescue the

argument and the result is
f (x) f 00 (a)
lim = 00
x!a g(x) g (a)
Note that you are required to know the results for exam-
ination purposes but you will not be expected to repro-
duce the proofs which are given here to help you under-
stand the background.

35
 TOPIC D. FUNCTIONS OF SEVERAL
VARIABLES.
I PARTIAL DIFFERENTIATION.

In chemical Engineering we frequently encounter functions of sev-

eral variables:
(a) The entropy S of a homogeneous pure gas in a container of
volume V at temperature T is a function of two variables i.e.
S = S(V, T ).
(b) If the gas is a mixture of two components A and B of con-
centrations c and 1 c respectively then S is a function of
three variables S = S(V, T, c).
(c) The density of the atmosphere depends on where we are. It
depends on the co-ordinates (x, y, z) of the point where we
measure it i.e. ⇢ = ⇢(x, y, z) is a function of three spatial
variables.
We need to develop methods of calculus for dealing with functions
of several variables.
A function of one variable f (x) can be represented by a graph in
two dimensions. The derivative of a function of one variable f (x)
is defined by
df 1
= lim (f (x + h) f (x))
dx h!0 h
This derivative gives the slope of the function or graph at the
point x.

36
A function of two variables f (x, y) can be represented by a surface
in three dimensions x, y and z ⌘ f (x, y). The partial derivative
of the function f(x,y) with respect to one of the variables x when
the other variable y is kept constant is defined by
@f 1
= lim (f (x + h, y) f (x, y))
@x h!0 h
This gives the slope of the function or surface in the x-direction
at the point (x, y).

We can also di↵erentiate partially with respect to the other vari-

able
@f 1
= lim (f (x, y + k) f (x, y))
@y k!0 k
This gives the slope of the function or surface in the y-direction
at the point (x, y). Note that it is almost always the case that the
slope in the two directions is completely di↵erent.

It is possible to carry out these processes several times and calcu-

late second derivatives:
@ 2f @ 2f @ 2f
, and
@x2 @y 2 @x @y
and third or higher derivatives such as

@ 3f @ 3f
, etc.
@x3 @x @y 2

@ 2f
The symbol means that we first di↵erentiate with respect
@x @y
to y keeping x constant and then with respect to x keeping y
constant or vice versa. The result is the same irrespective of or-
der. Examples will be given in the problem sheets. These partial
derivatives have the same kind of applications as ordinary deriva-
tives viz. expansion of functions, sketching, location of stationary
points etc.

37
II TAYLOR’S THEOREM IN TWO OR MORE DIMENSIONS.

For a function of one variable f (x) under a wide set of conditions

we can write down a Taylor expansion about the point x0 :

df 1 d2 f 2 1 d3 f
f (x0 + h) = f (x0 ) + (x0 ) h + (x 0 ) h + (x0 ) h3 + · · ·
dx 2! dx2 3! dx3
or in an abbreviated notation:
1 00 1
f (x0 + h) = f (x0 ) + f 0 (x0 ) h + f (x0 ) h2 + f 000 (x0 ) h3 + · · ·
2! 3!
Many useful approximate formulas are based on these ideas.
The same idea can be used for expanding a function of two vari-
ables f (x, y) under similar conditions:
! !
@f @f 1 @ 2f 2 @ 2f @ 2f 2
f (x0 +h, y0 +k) = f (x0 , y0 )+ h+ k + h + 2 hk + k +· · ·
@x @y 2! @x2 @x @y @y 2

Notice the binomial coefficients in the second-order term.

In abbreviated notation the third order expansion is:
f (x0 +h, y0 +k) = f0 +(fx h + fy k)+ 2!1 (fxx h2 + 2 fxy hk + fyy k 2 )
+ 3!1 (fxxx h3 + 3 fxxy h2 k + 3 fxyy hk 2 + fyyy k 3 ) + · · ·
You should now be able to carry the expansion to any order. You
should also be able to expand a function of three or more variables.

There is an even more abbreviated notation which we shall use

which is based on vectors and matrices and enables us to!write
h
down the expansion very economically. We put h = and
k
! !
fx fxx fxy
rf = and Hf =
fy fyx fyy
and find that the first three terms of Taylor’s theorem can now be
written
1
f (x0 + h, y0 + k) = f0 + hT · rf + hT · Hf · h + · · ·
2
The two new quantities above are used frequently and are given
special names which you should remember:
the vector rf is called the vector gradient of f (or gradient or
gradf ) and the matrix Hf is called the Hessian matrix of f .

38
**************************************************************

From the first couple of terms in the Taylor expansion above we

can get an important approximation:
@f @f
f⇡ x+ y
@x @y

Where f = f (x0 + h, y0 + k) f (x0 , y0 ) and x = h and y = k.

This is the most fundamental and basic formula in partial

di↵erentiation and you MUST remember it.

**************************************************************

39
III ERRORS, CHAIN RULE AND CHANGE OF VARIABLES.

(a) Small Increments and Errors.

Suppose we change x by a small amount x and y by a
small amount y then the corresponding change in V (x, y)
is V = V (x + x, y + y) V (x, y). We then have the basic
formula
@V @V
V ⇡ x+ y.
@x @y
A similar result can be used to deal with functions of three or
more variables
@V @V @V
V ⇡ x+ y+ z.
@x @y @z
These can be used to estimate errors. Suppose we have a
cylinder of volume V = ⇡r2 h where r is the radius and h is
the height and suppose the measured radius has an error ± r
and the measured height has an error ± h. Then the absolute
value of the error in V is
@V @V
| V |=| || h | + | || r | .
@h @r
We have assumed the worst case where the errors add up.

40
(b) Chain Rule.
Suppose we have a function of two variables V (x, y) of two
variables x = x(t) and y = y(t) which are functions of a third
variable t. Then we can think of V as a function of t and use
the basic formula for V to evaluate the derivative of V .

We can for example have a particle moving in two dimensions

and with potential energy V (x, y). Then the two functions
x = x(t) and y = y(t) determine the path of the particle and
the derivative of V with respect t t is found by dividing by t
and taking the limit as t tends to 0. The derivative is
dV @V dx @V dy
= + .
dt @x dt @y dt
This is a simple generalisation of the chain rule for functions
of one variable. This can be generalised to functions of three
or more variables.

(c) Change of Variable.

Suppose z = f (x, y) is a function of two variables x and y
which in turn are functions of two other variables u and v so
x = x(u, v) and y = y(u, v). Then we can write z = F (u, v) ⌘
f (x(u, v), y(u, v)) We use the basic formula to write

@f @f
f= x+ y
@x @y
@x @x @y @y
and x= u+ v and y= u+ v.
@u @v @u @v
We now substitute the second and third of these into the first
and collect terms to get
! !
@f @x @f @y @f @x @f @y
F = + u+ + v
@x @u @y @u @x @v @y @v

and corresponding formulas for the derivatives of z = F :

@F @f @x @f @y @F @f @x @f @y
= + and = + .
@u @x @u @y @u @v @x @v @y @v

41
 This can be written in matrix notation:
0 @F 1 0 @x @y 1 0 @f 1
@x
B @u C B @u @u
CB C
@ A = @ A@ A
@F @x @y @f
@v @v @v @y

The matrix 0 1
@x @y
B @u @u C
B C
J ⌘ B C
@ A
@x @y
@v @v
is an important quantity. It is called the Jacobian matrix
and is used frequently. Please remember its definition.

Example:

Suppose z = f (x, y) = x3 + y 3 and x = r cos ✓ and y = r sin ✓.

@z @z
Calculate and .
@r @✓

42
IV CONTOUR MAPS.

It is difficult to plot a graph of a function of two variables such as

f2 (x, y) because we have to produce a three-dimensional diagram.
It is still harder to produce a graph of a function of three variables
such as f3 (x, y, z) because we need a four-dimensional diagram.
We can get a good graphical idea of the behaviour of functions
such as these by drawing contour plots i.e. plots of lines or sur-
faces on which the functions are constant. Examples of these are
isotherms and isobars on a weather map or contour lines (i.e. lines
of constant height ) on an ordnance survey map. The method and
the features to note are best seen by looking at a few examples.

Examples:

x2 y 2
1. Sketch some contours of the function f (x, y) = + .
4 9
x2 y 2
2. Sketch some contours of the function f (x, y) = .
4 9
3. Sketch some contours of the function f (x, y) = (x + y)(x + y 2
2
4).

The following rules of thumb are worth knowing:

1. If the function factorises then it easiest to draw the contour
corresponding to f (x, y) = 0 first.
2. Many of the remaining contours can be sketched by continuity.
3. Notice that if the contour lines form a closed loop then the
function has at least one maximum or minimum within the
loop.
4. Contour lines corresponding to two di↵erent values of the con-
stant never cross.
5. If two contour lines corresponding to one value of the constant
cross then the function has a saddle at the crossing point.

43
We can get the slope of a contour line at a particular point (x0 , y0 )
by a process called implicit di↵erentiation:
The equation of the contour line is f (x, y) = c and we have the
basic formula
f = fx x + fy y
for the change in f when we move from the point (x0 , y0 ) to the
@f
point (x0 + x, y0 + y). In this formula fx = is evaluated at
@x
(x0 , y0 ) and similarly for the second coefficient. We remember now
that on a contour line f (x, y) = c and if both points lie on the
contour line it follows that f = 0. Hence fx x + fy y = 0 and
dy fx
= .
dx fy
This gives the slope of the contour line at the point we started at
viz. (x0 , y0 ).

*********************************************************************

It also answers the question we put in item 6 on p16.

If you have an implicit relation between two variables of the form
f (x, y) then you can get the derivative of y with respect to x in
terms of two partial derivatives of f (x, y) from the formula
dy fx
= .
dx fy

*********************************************************************

If we need the equation of the tangent to the contour line at the

point then we write

fx (x x0 ) + fy (y y0 ) = 0.

The idea can be generalised to three or more dimensions. Consider

the surface f (x, y, z) = c. Then in the surface we have

df = 0 and df = fx x + fy y + fz z.

Consequently if z is also a constant so z = 0 then

fx x + fy y = 0

44
The final result is
@y fx
= .
@x fy
Further by a similar process we get
@z fy @x fz
= and = .
@y fz @z fx
These are implicit partial derivatives of one variable with respect
to another with the third variable kept constant.
The equation of the tangent plane at the point (x0 , y0 , z0 ) is

fx (x x0 ) + fy (y y0 ) + fz (z z0 ) = 0.

where the partial derivatives are evaluated at the point (x0 , y0 , z0 ).

45
V STATIONARY POINTS.

To guide us we first examine a function of one variable f (x) and

try to see when it has a maximum or minimum or point of inflex-
ion with horizontal tangent.
df
First we note that the point x0 is stationary if = 0 at x = x0 .
dx
Now we make a Taylor expansion of f (x) about the point x0 and
get
1 d2 f
f (x0 + h) = f (x0 ) + (x0 )h2 + · · ·
2! dx2
up to second order.
1. If the coefficient of h2 in the second term is positive then
f (x) has a minimum because a small change of h from 0 will
increase the function.
2. If the coefficient of h2 in the second term is negative then
f (x) has a maximum because a small change of h from 0 will
decrease the function.
3. If the coefficient of h2 in the second term is zero then there
are various possibilities including a point of inflexion with
horizontal tangent.
We now go through a similar analysis of a function f (x, y) of two
variables.
We first note that at a stationary point (x0 , y0 ) we have

@f @f
=0 and = 0.
@x @y
We can now write down the two dimensional Taylor expansion up
to second order
1 ⇣ ⌘
f (x0 + h, y0 + k) = f (x0 , y0 ) + fxx h2 + 2fxy hk + fyy k 2 + · · ·
2!
Our strategy is now to write the second order term as the sum or
di↵erence of two squares i.e. quantities with a consistent sign. To
this end we take out a factor fxx1 from this term and get
1 ⇣ ⌘
f (x0 +h, y0 +k) = f (x0 , y0 )+ fxx1 fxx
2 2
h + 2fxx fxy hk + fxx fyy k 2 +· · ·
2!

46
Now complete the square on the first two terms in the brackets to
get

f (x0 + h, y0 + k)
1 h⇣ ⌘ i
= f (x0 , y0 ) + fxx1 fxx2 2 2 2
h + 2fxx fxy hk + fxy k 2 2
fxy k + fxx fyy k 2 + · · ·
2!
1 h ⇣ ⌘ i
= f (x0 , y0 ) + fxx1 (fxx h + fxy k)2 + fxx fyy fxy 2
k2 + · · ·
2!
We can now analyse the result.
⇣ ⌘
2
1. Suppose fxx fyy fxy > 0 then both terms in square brack-
ets are always positive or zero. Suppose also that outside
fxx > 0. Then the second order term is by its nature positive
and the function can only increase as we move away from the
point (x0 , y0 ). i.e. the function has a minimum.
⇣ ⌘
2
2. Suppose fxx fyy fxy > 0 then both terms in square brack-
ets are always positive or zero. Suppose also that the factor
outside fxx < 0. Then the second order term is by its nature
negative and the function can only decrease as we move away
from the point (x0 , y0 ). i.e. the function has a maximum.
⇣ ⌘
2
3. Suppose fxx fyy fxy < 0 then one term in square brackets
is always positive or zero and the other is always negative or
zero. Then the second order term will change its sign depend-
ing on the direction along which we move from the stationary
point (x0 , y0 ) i.e. on the exact values of h and k - the function
will increase in some directions and decrease in others. It will
have a saddle.

This is what you need to remember:

1. ⇣Find a point⌘where fx = 0 and fy = 0. If at that point
2
fxx fyy fxy > 0 and fxx > 0 then the function has a
MINIMUM at the point.
2. ⇣Find a point⌘where fx = 0 and fy = 0. If at that point
2
fxx fyy fxy > 0 and fxx < 0 then the function has a
MAXIMUM at the point.
3. ⇣Find a point⌘where fx = 0 and fy = 0. If at that point
2
fxx fyy fxy < 0 then the function has a SADDLE at the
point.

47
 TOPIC E. COMPLEX NUMBERS.
I INTRODUCTION AND ALGEBRAIC OPERATIONS.

Complex numbers can be regarded as an extension of the number sys-

tem which provides convenient solutions to problems which cannot be
solved in terms of real numbers. They are a useful tool for solving
algebraic equations and di↵erential equations and are essential to un-
derstanding quantum mechanics. The simplest application is to solving
algebraic equations which cannot be solved in terms of real numbers.

Consider the algebraic equation x2 = 1 . This p has no real so-

lution so we introduce a new type of number i = 1 in terms of
which this and every other algebraic equation can be solved. There are
two solutions of this equation x = ±i . (Don’t worry about the fact
that we are going beyond the real number system. Everything can be
justified. You are not required to understand the justification but you
are required to be able to use complex numbers.)

i is called the imaginary unit.

Consider the quadratic equation x2 10x+40 = 0 . From the formula

for solving a quadratic we get
p
x=5± 15

which is not a real number because of the square of a negative number.

But in terms of the imaginary unit we can write
p
x = 5 ± i 15

This is a complex number i.e. a number of the form z = a+i b where

a and b are both real numbers. a is called the real part of z and b is
called the imaginary part ofz. We often write

a = Re z and b = Im z

. Remember these notations.

z = a + i b is called the standard form of a complex number and

you should always try to reduce a complex number to standard form.

48
The usual algebraic operations of addition, multiplication and division
can be applied to a pair of complex numbers z1 = a+i b and z2 = c+i d:
1. Addition.
z1 + z2 = (a + i b) + (c + i d) = (a + c) + i (b + d)
i.e. form a new complex number by adding the real parts of z1
and z2 to form the real part of the sum and the imaginary parts to
form the imaginary part of the sum. (Subtraction is just a special
case of addition.)
2. Multiplication
We use the usual rules of removing brackets in multiplication re-
membering that i2 = 1 and then group real bits together and
imaginary bits together. Thus
z1 z2 = (a + i b)(c + i d) = (ac bd) + i (ad + bc)
The first bracket is the real partof z1 z2 and the second bracket is
the imaginary part. There is no need to remember this formula
because it is easy to calculate the product using the usual rules of
multiplication.
3. Division. This is harder. We first define the conjugate of a
complex number
z2 = c i d
and then note from the process of multiplication that z2 z 2 =
c2 + d2 . This combination is called
p the squared modulus of z2 .
The modulus of z2 is |z2 | = c2 + d2 .

To get the ratio of z1 to z2 we mutiply the numerator and the

denominator of the fraction both by z 2 i.e. we multiply the whole
fraction by 1. This process leaves the value of the fraction un-
changed but makes the algebraic form better. Thus
z1 z1 z 2 z1 z 2
= = 2
z2 z2 z 2 z1 + z22
The numerator can be calculated by multiplication as above and
the denominator is real so it presents no problems and the stan-
dard form is
ac + bd bc ad
z1 z2 = 2 2
+i 2
c +d c + d2
Again there is no need to remember the formula but you must be
able to carry out the process.

49
II GEOMETRICAL INTERPRETATION.

The complex number z = x + i y is specified in terms of two real

numbers (x, y) and can thus be represented by a point in a plane
with coordinates (x, y) .
A point in a plane can be represented either by Cartesian coordinates
as above or by polar coordinates (r, ✓)

The relation betweenthe Cartesian and the polar coordinates is given

by either
x = r cos ✓ and y = r sin ✓
or by q
r= x2 + y 2 and ✓ = tan 1 (y/x)
You will recognise r as the modulus of z ( i.e. r = |z| ).
✓ is called the argument of z and we write ✓ = Arg z. Also we require
⇡ < ✓  ⇡.

r is the distance from the origin to the point z and ✓ is the angle
between Oz and the x-axis.

Finally we note that we can write

z = r cos ✓ + i r sin ✓ = r (cos ✓ + i sin ✓)

The factor in round brackets appears frequently and is a complex num-

ber of modulus 1 because

(cos ✓ + i sin ✓)(cos ✓ i sin ✓) = cos2 ✓ + sin ✓2 = 1.

50
III EXPONENTIALS.

Remember the Taylor expansion of the exponential, the cos and the
sin:
1 1 1
e x = 1 + x + x2 + x3 + x4 + · · ·
2! 3! 4!
1 2 1 4
cos ✓ = 1 ✓ + ✓ + ···
2! 4!
1 3 1 5
sin ✓ = ✓ ✓ + ✓ + ···
3! 5!
Now consider
1 1 1
ei ✓ = 1 + i ✓ + (i ✓)2 + (i ✓)3 + (i ✓)4 + · · ·
2! 3! 4!
We now write this in standard form
✓ ◆ ✓ ◆
1 2 1 4 1 3 1 5
ei ✓ = 1 ✓ + ✓ + ··· + i ✓ ✓ + ✓ + ···
2! 4! 3! 5!
and comparing the expansions we see
ei ✓ = cos ✓ + i sin ✓
This is an extremely important formula and you MUST remember it.

The exponential obeys the normal laws of indices so we get

z1 r1
z1 z2 = r1 r2 exp(i (✓1 + ✓2 )) and = exp(i, (✓1 ✓2 ))
z2 r2
In other words multiplication and division are very easy in polar form.
Powers are also very easy
z n = rn exp(in✓)
This formula works when n is an integer but also when n is a fraction
but requires a bit of caution when n is a fraction.

The exponential above also obeys the normal rules of calculus so that
d i✓
e = i ei ✓
d✓
d2 i ✓
e = i2 ei ✓ = ei ✓
d✓2
d3 i ✓
3
e = i3 ei ✓ = i ei ✓
d✓
d4 i ✓
e = i4 e i ✓ = e i ✓
d✓4

51
TOPIC F. DIFFERENTIAL EQUATIONS.
I INTRODUCTION

There are two iconic di↵erential equations which you must be able to
recognise whenever they appear and you should also be able to write
down the solutions quickly and easily. These are the equation of ra-
dioactive decay or exponential growth and the equation of simple
harmonic motion. They exhibit most of the general features of ordi-
nary di↵erential equations and we will write down the solutions. (Later
we shall see how to derive the solutions of these and many others.)

(a) Radioactive decay and exponential growth.

These processes are described by the same equation but the sign of
a constant is di↵erent. The experimental observation from which
they arise is that the rate of change of the number of objects (ra-
dioactive atoms in a lump of matter or bacteria in a colony) is
proportional to the number of objects itself i.e.
dN
= N (t) for radioactive decay
dt
dN
= + N (t) for exponential growth
dt
The only di↵erence between these equations is the sign of the con-
stant on the rhs and is determined by the nature of the atoms
or bacteria. The solutions of these equations are

t
N (t) = A e for radioactive decay
t
N (t) = A e for exponential growth.
Note that the solutions involve one arbitrary constant A which
is not determined by the equation itself. It is determined by the
initial conditions i.e. by the number of radioactive atoms or bac-
terial cells initially present.

You should now check the solutions by substituting back into the
original di↵erential equations and then sketch them on a graph.

52
 (b) Simple Harmonic Motion

This equation describes the motion of a simple pendulum or the

motion of a mass at the end of a spring. It derives from the ob-
servation that the force or acceleration restoring the mass to its
centre of equilibrium is proportional to the displacement of the
mass from that centre. The equation is written in terms of the
displacement x(t) and is

d2 x
= !2 x
dt2
The solution of this equation is

x(t) = A cos ! t + B sin ! t

Note that there are now two arbitrary constants A and B to be

determined from the initial conditions i.e. the initial position and
velocity of the mass.
Note also that the motion is oscillatory as we expect for a pendu-
lum or bobbing mass.

You should now check the solution by substituting back into the
original di↵erential equation and then sketch it on a graph.

You should note the following general feature:

A first order di↵erential equation (i.e. an equation involving a first
derivative only) has one arbitrary constant and a second order di↵er-
ential equation (i.e. an equation involving a second derivative) has two
arbitrary constants. A third order equation has three arbitrary con-
stants and so on.

Many di↵erential equations do not have a solution in simple analyt-

ical form and we have to write down a solution in the form of a table
of values obtained numerically. But there are some equations which
we can solve analytically and we concentrate on these. We have to
distinguish between first order and second order equations.

53
II METHODS FOR FIRST ORDER EQUATIONS.

1. Separable Equations.

These are equations of the form

dy
= f (x) g(y)
dx
where the rhs splits up into two factors one of which depends
on x only and the other on y only. We rearrange the equation
so that everything that depends on y appears on one side of the
equation and everything that depends on x on the other side and
then integrate. We get
Z Z
dy
= f (x) dx
g(y)

If we can do the integration then we get a closed form solution.

Note that the radioactive decay/exponential growth equation is a
very easy type of separable equation.

Examples:
dy
(a) Solve = xy
dx
dy x
(b) Solve =
dx y
dy
(c) Solve = y cos x
dx
dy x
(d) Solve = 2
dx y

54
2. First order Linear Equations.

These are equations of the form

dy
+ f (x) y = g(x)
dx
We use an integrating factor to solve these equations. The
integrating factor is defined as
✓Z ◆
Q(x) = exp f (x) dx

(There is no need to put an arbitrary constant into Q(x).) We

multiply the equation through by Q(x) and note that it can be
written ✓ ◆
d
Q(x) y = Q(x) g(x)
dx
and now integrate to get finally
1 Z
y= Q(x) g(x) dx
Q(x)

Examples:
dy
(a) Solve + x y = exp( x2 /2) x
dx
dy 1
(b) Solve + y = 3x
dx x
dy
(c) Solve + (tan x) y = cos x
dx

55
3. Exact Equations.

These are equations of the form

dy
P (x, y) + Q(x, y) =0 (1)
dx
where P (x, y) and Q(x, y) are functions of both x and y which
satisfy the relations
@Q @P
= (2)
@x @y

There is a theorem which says that for functions of this type we

can find a function F (x, y) which is related to P and Q as follows
@F @F
=P and =Q (3)
@x @y
We can now write equation (1) in the form
P (x, y) dx + Q(x, y) dy = 0 (4)
and it then follows from equation (3) that
@F @F
dx + dy = 0 (5)
@x @y
We compare this with the basic formula of partial di↵erentiation
on p41 which is
@F @F
dF = dx + dy (6)
@x @y
and it then follows from equations (5) and (6) that
dF = 0 (7)
and the solution of the original equation (1) is
F (x, y) = c (8)
In order to apply this method you need to recognise when
an equation is exact. In other words you have to recognise
that the equation can be written in the form (1) with coef-
ficients which satisfy (2). You then have to solve equation
(3) to get F (x, y). The rest is easy.

56
Examples:

dy y
(a) Solve =
dx x
Solution: Rewrite the equation as
dy
y+x =0
dx
@P @Q
Then P = y and Q = x so that = 1 and = 1.
@y @x
These are equal so the equation is exact. We then have
@F @F
=y and =x
@x @y

By inspection we find F (x, y) = xy and the solution is

xy = c.

Later we shall see that there is a better way of solving equa-

tion (3) than inspection.

dy y
(b) Solve =
dx x
Solution: Rewrite the equation as
dy
y+x =0
dx
@P @Q
Then P = y and Q = x so that = 1 and = 1.
@y @x
These are not equal so the equation is not exact so we cannot
use this method. We can use separation of variables.

57
 ! !
y2 x2 dy
(c) Solve x(y + 1) + + + xy =0
2 2 dx
Solution: ! !
y2 x2
Then P = x(y + 1) + and Q = + xy so that
2 2
@P @Q
= x + y and = x + y.
@y @x
These are equal so the equation is exact.

We then have
! !
@F y2 @F x2
= x(y + 1) + and = + xy
@x 2 @y 2

We cannot easily find F (x, y) by inspection so we use the

following method:
@F
We integrate the formula for with respect to y keeping x
@y
constant to get

x2 y xy 2
F (x, y) = + + g(x)
2 2
where g(x) is constant with respect to the variable of integration y
and hence acts as the arbitrary constant of integration.
@F
We now substitute this into the formula for to get
@x
y2 y2
xy + + g 0 (x) = x(y + 1) +
2 2
x2
which gives g 0 (x) = x and g(x) =
2
Substituting into the formula for F (x, y) above we get the
final form of the solution:
x2 y xy 2 x2
+ + = c.
2 2 2

58
 dy
(d) Solve (y cos x + sin y sin x) + (sin x + x cos y) =0
dx
Solution:
Then P = (y cos x + sin y sin x) and Q = (sin x + x cos y)
@P @Q
so that = cos x + cos y and = cos x + cos y.
@y @x
These are equal so the equation is exact.

We then have
@F @F
= (y cos x + sin y sin x) and = (sin x + x cos y)
@x @y
@F
We integrate the formula for with respect to y keeping x
@y
constant to get

F (x, y) = y sin x + x sin y + g(x)

where g(x) is constant with respect to the variable of integration y

and hence acts as the arbitrary constant of integration.
@F
We now substitute this into the formula for to get
@x
y cos x + sin y + g 0 (x) = y cos x + sin y sin x

which gives g 0 (x) = sin x and g(x) = cos x

Substituting into the formula for F (x, y) above we get the

final form of the solution:

y sin x + x sin y + cos x = c.

59
4. Homogeneous Equations.

These are equations of the form

✓ ◆
dy y
=f (1)
dx x
A change of variables transforms these equations into separable
equations which can be solved by the method outlined in section
1.
y
The new variable we introduce is v = or alternatively we
x
write y = v x. We then find
dy dv
=v+x (2)
dx dx
We substitute from equation (2) into equation (1) to get

dv
v+x = f (v) (3)
dx
and rewrite this as
dv
x = f (v) v (4)
dx
which is separable. We rearrange it so that everything depending
on v appears on one side and everything depending on x on the
other and then integrate to get v in terms of x. Finally we sub-
y
stitue v = .
x
More concretely from equation (4) we get
Z Z
dv dx
= (5)
f (v) v x

60
You will see exactly how it is done in the following examples:

Examples:

dy
(a) Solve 2 x2 = x2 + y 2 .
dx
Solution:
Divide through by 2 x2 to get
✓ ◆2 !
dy 1 y
= 1+
dx 2 x

This clearly homogeneous. Now substitute y = vx and

the result is
dv 1⇣ ⌘
v+x = 1 + v2
dx 2
Now separate the variables and integrate to get
Z Z
2 dv dx
2
=
1 2v + v x
The integral on the rhs is easy and is equal to ln | x | +c. To
get the integral on the lhs we factorise the denominator and
find
Z
2 dv
lhs =
(1 v)2
2
=
1 v
Solve for v in terms of x to get
2
v=1
ln | x | +c
y
Write v = , substitute in and rearrange to get the answer
x
!
2
y=x 1
ln | x | +c

61
 ✓ ◆
dy y y
(b) Solve: = sec + .
dx x x
Solution:

Substitute y = v x to get
dy
v+x = sec v + v
dx
Thus Z Z
dx
cos v dv =
x
sin v = ln | x | +c
and in terms of y and x

y = x sin 1 (ln | x | +c)

62
III METHODS FOR SECOND ORDER EQUATIONS.

We will deal only with constant coefficient linear second order equa-
tions i.e. equations where each term contains y and its derivatives only
to power 1 or 0.

There are two types of these equations, homogeneous and inhomoge-

neous. In the following a, b and c are constant and a 6= 0.

Homogeneous:
d2 y dy
a 2
+ b + c y = 0. (1)
dx dx
Inhomogeneous:
d2 y dy
a + b + c y = f (x) where f (x) 6= 0. (2)
dx2 dx
These have to be dealt with separately.

1. Homogeneous Equations.

These have only the following functions in their solutions: sin, cos,
exp, polynomials and combinations of them. Also the linearity and
homogeneity of these equations implies that any solution remains
a solution if multiplied by an arbitrary constant and also the sum
of two solutions is also a solution.

In order to solve them we use the method of trial solutions i.e.

we substitute y = exp( x) into the lhs of the equation and
try to find out what values of work. There are only three
possibilities: is real, complex or imaginary and each will give
a di↵erent type of answer.

Now substitute y = exp( x) into equation (1) to get

2
(a + b + c) y = 0
In other words we have a solution if satisfies the quadratic equa-
tion
a 2+b +c=0 (3)

63
This is called the auxiliary equation and the roots are
1 ⇣ p ⌘
= b ± b2 4 ac
2a
There are four possibilities for the roots of equation (3).

(a) Di↵erent Real Roots.

If b2 4 ac > 0
1 ⇣ p ⌘
equation (3) has two real roots 1 = b + b2 4 ac
2a
1 ⇣ p ⌘
and 2 = b b2 4 ac and 1 6= 2 . Then the gen-
2a
eral solution of equation (1) is
1x 2x
y = Ae +Be (4)

This equation has two arbitrary constants so it is the most

general solution of (1).

This gives exponential growth or decay depending on

the signs of 1 and 2 .

(b) Equal Real Roots.

If b2 4 ac = 0
equation (3) has only one repeated root 2 = 1 = b/2 a.
Then solution (4) has really only one arbitrary constant A+B
and we have to modify it slightly. The general solution with
two arbitrary constants is
1x
y = (A + B x) e (5)

You can check this by looking at a simple example:

d2 y dy
2
+2 + y = 0,
dx dx

64
(c) Complex Roots.

If b2 4 ac < 0
then equation (3) has two complex roots 1 = µ + i ⌫ and
b 1 p
2 = µ i ⌫ with µ = and ⌫ = 4 ac b2 .
2a 2a
Then the general solution of equation (1) is

y = A eµx ei ⌫x + B eµx e i ⌫x
(6)

This equation has two arbitrary constants so it is the most

general solution of (1) but it is inconvenient because of the
exponentials with imaginary argument.
A much better form can be found by writing the two expo-
nentils with imaginary argument in terms of sies and cosines
to get
y = eµx (C sin ⌫x + D cos ⌫x)
This is the equation of damped or growing simple har-
monic motion depending on the sign of µ and you
should recognise it when you see it.

(d) Imaginary Roots.

If b = 0 and ac > 0
then equation (3) has two
r imaginary roots 1 = i⌫ and
c
2 = i ⌫ with ⌫ = .
a
Then the general solution of equation (1) is

y = A ei ⌫x + B e i ⌫x
(7)

This equation has two arbitrary constants so it is the most

general solution of (1) but it is inconvenient because of the
exponentials with imaginary argument.
A much better form can be found by writing the two expo-
nentils with imaginary argument in terms of sines and cosines
to get
y = C sin ⌫x + D cos ⌫x
This is the equation of simple harmonic motion and
you should recognise it when you see it.

65
2. Inhomogeneous Equations.

The equation is
d2 y dy
a 2
+ b + c y = f (x) where f (x) 6= 0. (8)
dx dx
We can deal with cases where f (x) is a sine, a cosine, an exponen-
tial or a polynomial or combinations of these. We use the method
of Trial Solutions. Other functions are difficult.

The general solution is the sum of two parts:

y = yCF + yP I .

The first is called the Complimentary Function (CF) and is the

solution of the corresponding homogeneous equation obtained in
the previous section.
The second is called the Particular Integral (PI) and can be any
solution of the inhomogeneous equation that you can find. The
PI is discussed below.

The most general function that we can deal with is of the form:

f (x) = e↵x (Pn (x) cos x + Qm (x) sin x) (9)

where Pn (x) and Qm (x) are polynomials in x of degree n and m

respectively.

The most general trial function that always works is of the form:

yP I (x) = e↵x (Rk (x) cos x + Sl (x) sin x) (10)

where Rk (x) and Sl (x) are polynomials in x of degree k and l

respectively. The degrees of these two polynomials are in most
cases equal to the larger of m and n in equation (9). But there
are cases where you need values of k and l up to 2 larger.

The strategy is to substitute yP I into the lhs of equation (8) and

choose the coefficients of all the powers of x in the polynomials so
that the lhs=rhs. This the most general way of putting it and it
looks quite complicated but most cases are really very simple and
we now look at examples which illustrate all the basic points.

66
Examples:
d2 y dy
(a) Solve: 4 + 3 y = x2 + 1.
dx2 dx
Solution:
The auxiliary equation is 2 4 + 3 = 0 with roots 1 = 1
and 2 = 3.
This gives a CF yCF = A exp(x) + B exp(3 x)
We take the PI to be of the form yP I = d x2 + e x + f and
substitute into the lhs and then choose the coefficients d, e
and f so that the two sides match.
We get 3 dx2 + (3 e 8 d)x + (3 f + 2 d 4e) = 3 x2 + 1
which gives the following relations:
3d = 3 3e 8d = 0 3f + 2d 4e = 1
or d=1 e = 8/3 f = 29/9

Thus the general solution is

8 29
y = yCF + yP I = A exp(x) + B exp(3 x) + x2 + x +
3 9
NOTE: If the rhs is a polynomial of order n then it is
enough to use a polynomial of the same order in the
trial function. This is always the case and makes life
much simpler.

d2 y dy
(b) Solve: 4 + 3 y = cos 2 x.
dx2 dx
Solution:
The CF is the same as the previous example.
We take the PI to be of the form yP I = d cos 2 x + e sin 2 x
and substitute into the lhs and then choose the coefficients d
and e so that the two sides match.
We get d 8e = 1 e + 8d = 0
which gives d = 1/65 e = 4/65

Thus the general solution is

1 4
y = yCF +yP I = A exp(x)+B exp(3 x) cos 2 x sin 2 x
65 65
NOTE: If the rhs is a single sine or cosine then you
must have both a sine and a cosine in the trial solution

67
 with the same argument as the trig function on the
rhs. But beware there is an exception which you will
see in the next example.
d2 y
(c) Solve: + 16 y = sin 4 x.
dx2
Solution:
The lhs appears in the equation of shm with ! = 4.
Thus the CF is yCF = A cos 4 x + B sin 4 x.
We now note that the CF contains a term with a trig function
which is exactly the same as the trig function on the right.
It is no longer enough to take a trial function of the form we
had in the previous example. It is necessary to multiply by an
extra polynomial and we try the lowest order. In other words
we try yP I = a x cos 4 x + b x sin 4 x and substitute
into the lhs to get

8 b cos 4 x 8 a sin 4 x = sin 4 x

which implies b=0 and a= 1/8. Hence the general

solution is
1
y = yCF + yP I = A cos 4 x + B sin 4 x x cos 4 x
8
NOTE:
1. If the inhomogeneous term f (x) coincides with
one of the terms in the CF then you always have
to introduce the multiplying polynomial - in this
case just x.
2. This example introduces the very important phys-
ical idea of resonance - if the forcing term f (x) os-
cillates with the same frequency as a natural fre-
quency of the system then the amplitude of the
oscillations grows to very large values. This idea
lies behind the whole telecommunications indus-
try and has important applications in many other
fields of Physics and Chemistry.

68
 d2 y dy
(d) Solve 4 + 3 y = e4 x .
dx2 dx
Solution:
The CF is the same as in example (a).
We take the PI to be of the form yP I = d e4 x and substi-
tute into the lhs and then choose the coefficient d so that the
two sides match.
We get 42 d 4 ⇥ 4 d + 3 d = 1 i.e. d = 1/3.

Thus the general solution is

1 4x
y = yCF + yP I = A ex + B e3 x + e
3
NOTE: If the rhs is a single exponential then you
must have an exponential in the trial solution with
the same argument. But beware there is an exception
which you will see in the next example.
d2 y dy
(e) Solve 2
4 + 3 y = e3 x .
dx dx
Solution:
The CF is the same as in example (a).
The inhomogeneous term on the rhs is the same as a term in
the CF. Once again it is necessary to multiply by an extra
polynomial and we try the lowest order. In other words we
try yP I = a x e 3 x and substitute into the lhs to get

(9 x + 9)e3 x a 4(3 x + 1)e3 x a + 3 x e3 x a = e3 x

which implies a = 1/5

Thus the general solution is

1
y = yCF + yP I = A exp(x) + B exp(3 x) + x e3 x
5
There are many other cases but they can all be dealt with by
small variations on the methods given. You will see some of them
in the problem sheets and also in the second part of the course
where many of these equations are examined from a more physical
poiont of view.

69