Further Pure 2

1
2
3
Questions from the Cambridge International AS & A Level Further Mathematics papers are reproduced
by permission of Cambridge Assessment International Education. Unless otherwise acknowledged, the
questions, example answers, and comments that appear in this book were written by the authors.
Cambridge Assessment International Education bears no responsibility for the example answers to
questions taken from its past question papers which are contained in this publication.
The Publishers would like to thank the following for permission to reproduce copyright material.
Photo credits
p.1 © imatthiasschulz - iStock via Thinkstock; p.16 © Alexander Zelnitskiy/123RF ©; P.60 tl,
tr, bl © Geograph.com br © David Burrows - Shutterstock p.85 © Digoarpi - iStock via Thinkstock;
p.117 © Olaf Schulz/stock.adobe.com; p.145 © World History Archive/Alamy Stock Photo.
t = top, b = bottom, l = left, r = right.
Every effort has been made to trace and acknowledge ownership of copyright. The publishers will be
glad to make suitable arrangements with any copyright holders whom it has not been possible to
contact.
Although every effort has been made to ensure that website addresses are correct at time of going to
press, Hodder Education cannot be held responsible for the content of any website mentioned in this
book. It is sometimes possible to find a relocated web page by typing in the address of the home page
for a website in the URL window of your browser.
Hachette UK’s policy is to use papers that are natural, renewable and recyclable products and made
from wood grown in sustainable forests. The logging and manufacturing processes are expected to
conform to the environmental regulations of the country of origin.
Orders: please contact Bookpoint Ltd, 130 Milton Park, Abingdon, Oxon OX14 4SE. Telephone:
(44) 01235 827720. Fax: (44) 01235 400401. Email education@bookpoint.co.uk Lines are open from
9 a.m. to 5 p.m., Monday to Saturday, with a 24-hour message answering service. You can also order
through our website: www.hoddereducation.com.
Much of the material in this book was published originally as part of the MEI Structured Mathematics
series. It has been carefully adapted for the Cambridge International AS & A Level Further Mathematics
syllabus. The original MEI author team for Pure Mathematics comprised Catherine Berry, Val Hanrahan,
Terry Heard, David Martin, Jean Matthews, Roger Porkess and Peter Secker.
© Roger Porkess, Jean-Paul Muscat and Rose Jewell 2018
First published 2018 by
Hodder Education,
An Hachette UK Company
Carmelite House
50 Victoria Embankment
London EC4Y 0DZ
www.hoddereducation.com
Impression number 10 9 8 7 6 5 4 3 2 1
Year 2022 2021 2020 2019 2018
All rights reserved. Apart from any use permitted under UK copyright law, no part of this publication
may be reproduced or transmitted in any form or by any means, electronic or mechanical, including
photocopying and recording, or held within any information storage and retrieval system, without
permission in writing from the publisher or under licence from the Copyright Licensing Agency Limited.
Further details of such licences (for reprographic reproduction) may be obtained from the Copyright
Licensing Agency Limited, www.cla.co.uk
Cover photo © Comaniciu Dan/Shutterstock
Illustrations by Aptara, Inc. and Integra Software Services
Typeset in Bembo std 11/13 Integra Software Services Pvt Ltd, Pondicherry, India
Printed in Italy
A catalogue record for this title is available from the British Library.
ISBN: 9781510421790
eISBN: 9781510422056
4
Contents
Introduction v
How to use this book vi
The Cambridge International AS & A Level Further
Mathematics 9231 syllabus viii
1 Hyperbolic functions 1
1.1 Hyperbolic functions 1
1.2 Inverse hyperbolic functions 8
2 Matrices 16
2.1 Important results relating to 2 × 2 matrices 16
2.2 Finding the inverse of a 3 × 3 matrix 23
2.3 Intersection of three planes 29
2.4 Eigenvalues and eigenvectors 38
3 Differentiation 60
3.1 Differentiating inverse trigonometric functions 61
3.2 Differentiating hyperbolic functions 62
3.3 Differentiating inverse hyperbolic functions 62
3.4 Finding the second derivative for relations given implicitly 64
3.5 Finding the second derivative for relations given parametrically 67
3.6 Polynomial approximations and Maclaurin series 70
3.7 Using the Maclaurin series for standard functions 79
4 Integration 85
4.1 Integration using the inverse sine function 85
4.2 Integrating hyperbolic functions 88
4.3 Integration using inverse hyperbolic functions 89
4.4 Integration using reduction formulae 92
4.5 Approximation of areas using rectangles 97
4.6 Using areas under curves to approximate series sums 99
4.7 The arc length of a curve 102
4.8 Surface area of a solid of revolution 109
iii
5
5 Complex numbers 117
5.1 The modulus and argument of a complex number 117
5.2 De Moivre’s theorem 121
5.3 The nth roots of a complex number 125
5.4 Finding multiple angle identities using de Moivre’s theorem 133
5.5 The form z = re iq 137
6 Differential equations 145

6.1 Integrating factors 147
6.2 Second order differential equations 154
6.3 Auxiliary equations with complex roots 167
6.4 Non-homogeneous differential equations 177
6.5 Use of substitutions to transform differential equations 185
Index 197
iv
6
Introduction
This is one of a series of four books supporting the Cambridge International
AS & A Level Further Mathematics 9231 syllabus for examination from 2020.
It is preceded by five books supporting Cambridge International AS & A Level
Mathematics 9709. This book follows on from Further Pure Mathematics 1.
The six chapters in this book cover the further pure mathematics required for
the Paper 2 examination. This part of the series also contains a book each for
further mechanics and further probability and statistics.
These books are based on the highly successful series for the Mathematics
in Education and Industry (MEI) syllabus in the UK but they have been
redesigned and revised for Cambridge International students; where
appropriate, new material has been written and the exercises contain many
past Cambridge International examination questions. An overview of
the units making up the Cambridge International syllabus is given in the
following pages.
Throughout the series, the emphasis is on understanding the mathematics as
well as routine calculations. The various exercises provide plenty of scope for
practising basic techniques; they also contain many typical examination-style
questions.
The original MEI author team would like to thank Jean-Paul Muscat and
Rose Jewell who have carried out the extensive task of presenting their work
in a suitable form for Cambridge International students and for their many
original contributions. They would also like to thank Cambridge Assessment
International Education for its detailed advice in preparing the books and for
permission to use many past examination questions.
Roger Porkess
Series editor
7
How to use this book
The structure of the book
This book has been endorsed by Cambridge Assessment International
Education. It is listed as an endorsed textbook for students taking the
Cambridge International AS & A Level Further Mathematics 9231 syllabus.
The Further Pure Mathematics 2 syllabus content is covered comprehensively
and is presented across six chapters, offering a structured route through the
course.
The book is written on the assumption that you have covered and
understood the work in the Cambridge International AS & A Level
Mathematics 9709 syllabus. The following icon is used to indicate material
that is not directly on the syllabus.
e There are places where the book goes beyond the requirements of
the syllabus to show how the ideas can be taken further or where
fundamental underpinning work is explored. Such work is marked as
extension.
Each chapter is broken down into several sections, with each section covering
a single topic. Topics are introduced through explanations, with key terms
picked out in red. These are reinforced with plentiful worked examples,
punctuated with commentary, to demonstrate methods and illustrate
application of the mathematics under discussion.
Regular exercises allow you to apply what you have learned. They offer a large
variety of practice and higher-order question types that map to the key concepts
of the Cambridge International syllabus. Look out for the following icons.
PS Problem-solving questions will help you to develop the ability
to analyse problems, recognise how to represent different situations
mathematically, identify and interpret relevant information, and select
appropriate methods.
M Modelling questions provide you with an introduction to the
important skill of mathematical modelling. In this, you take an everyday
or workplace situation, or one that arises in your other subjects, and
present it in a form that allows you to apply mathematics to it.
CP Communication and proof questions encourage you to become a
more fluent mathematician, giving you scope to communicate your work
with clear, logical arguments and to justify your results.
Exercises also include questions from real Cambridge Assessment
International Education past papers, so that you can become familiar with the
types of questions you are likely to meet in formal assessments.
Answers to exercise questions, excluding long explanations and proofs, are
available online at www.hoddereducation.com/cambridgeextras, so you can
check your work. It is important, however, that you have a go at answering
vi
8
the questions before looking up the answers if you are to understand the
mathematics fully.
In addition to the exercises, the following features are included to enhance
your learning.
ACTIVITY
Activities invite you to do some work for yourself, typically to introduce
you to ideas that are then going to be taken further. In some places,
activities are also used to follow up work that has just been covered.
? This symbol highlights points it will benefit you to discuss with

your teacher or fellow students, to encourage deeper exploration
and mathematical communication. If you are working on your own,
there are answers available online at www.hoddereducation.com/
cambridgeextras.
A variety of notes are included to offer advice or spark your interest:
Note
Notes expand on the topic under consideration and explore the deeper
lessons that emerge from what has just been done.
Historical note
Historical notes offer interesting background information about famous
mathematicians or results to engage you in this fascinating field.
Finally, each chapter ends with the key points covered, plus a list of the
learning outcomes that summarise what you have learned in a form that is
closely related to the syllabus.
Digital support
Comprehensive online support for this book, including further questions,
is available by subscription to MEI’s Integral® online teaching and learning
platform for AS & A Level Mathematics and Further Mathematics,
integralmaths.org. This online platform provides extensive, high-quality
resources, including printable materials, innovative interactive activities, and
formative and summative assessments. Our eTextbooks link seamlessly with
Integral, allowing you to move with ease between corresponding topics in
the eTextbooks and Integral.
MEI’s Integral® material has not been through the Cambridge International
endorsement process.
vii
9
The Cambridge International
AS & A Level Further
Mathematics 9231 syllabus
The syllabus content is assessed over four examination papers.
Paper 1: Further Pure Paper 3: Further Mechanics
Mathematics 1 • 1 hour 30 minutes
• 2 hours • 40% of the AS Level; 20% of the
• 60% of the AS Level; 30% of the A Level
A Level • Offered as part of AS;
• Compulsory for AS and A Level compulsory for A Level
Paper 2: Further Pure Paper 4: Further Probability &
Mathematics 2 Statistics
• 2 hours • 1 hour 30 minutes
• 30% of the A Level • 40% of the AS Level; 20% of the
• Compulsory for A Level; not a A Level
route to AS Level • Offered as part of AS;
compulsory for A Level
The following diagram illustrates the permitted combinations for AS Level

and A Level.
AS Level Further A Level Further
Mathematics Mathematics
Paper 1 and Paper 3

Further Pure Mathematics 1
and Further Mechanics
Paper 1, 2, 3 and 4
Further Pure Mathematics 1 and 2,
Further Mechanics and Further
Probability & Statistics
Paper 1 and Paper 4
Further Pure Mathematics 1
and Further Probability & Statistics
Prior knowledge
It is expected that learners will have studied the majority of the Cambridge
International AS & A Level Mathematics 9709 syllabus content before
studying Cambridge International AS & A Level Further Mathematics 9231.
viii
10
The prior knowledge required for each Further Mathematics component is
shown in the following table.
Component in AS & A Level Prior knowledge required from
Further Mathematics 9231 AS & A Level Mathematics 9709
9231 Paper 1: Further Pure 9709 Papers 1 and 3
Mathematics 1
9231 Paper 2: Further Pure 9709 Papers 1 and 3
Mathematics 2
9231 Paper 3: Further Mechanics 9709 Papers 1, 3 and 4
9231 Paper 4: Further Probability 9709 Papers 1, 3, 5 and 6
& Statistics
For Paper 2: Further Pure Mathematics 2, knowledge of Paper 1: Further
Pure Mathematics 1 subject content from this syllabus is assumed.
Command words
The table below includes command words used in the assessment for this
syllabus. The use of the command word will relate to the subject context.
Command word What it means
Calculate work out from given facts, figures or information
Deduce conclude from available information
Derive obtain something (expression/equation/value) from
another by a sequence of logical steps
Describe state the points of a topic/give characteristics and
main features
Determine establish with certainty
Evaluate judge or calculate the quality, importance, amount, or
value of something
Explain set out purposes or reasons/make the relationships
between things evident/provide why and/or how and
support with relevant evidence
Identify name/select/recognise
Interpret identify meaning or significance in relation to the
context
Justify support a case with evidence/argument
Prove confirm the truth of the given statement using a
chain of logical mathematical reasoning
Show (that) provide structured evidence that leads to a given
result
Sketch make a simple freehand drawing showing the key
features
State express in clear terms
Verify confirm a given statement/result is true ix
11
Key concepts
Key concepts are essential ideas that help students develop a deep
understanding of mathematics.
The key concepts are:
Problem solving
Mathematics is fundamentally problem solving and representing systems and
models in different ways. These include:
» Algebra: this is an essential tool which supports and expresses
mathematical reasoning and provides a means to generalise across a
number of contexts.
» Geometrical techniques: algebraic representations also describe a spatial
relationship, which gives us a new way to understand a situation.
» Calculus: this is a fundamental element which describes change in
dynamic situations and underlines the links between functions and graphs.
» Mechanical models: these explain and predict how particles and objects
move or remain stable under the influence of forces.
» Statistical methods: these are used to quantify and model aspects of the
world around us. Probability theory predicts how chance events might
proceed, and whether assumptions about chance are justified by evidence.
Communication
Mathematical proof and reasoning is expressed using algebra and notation so
that others can follow each line of reasoning and confirm its completeness
and accuracy. Mathematical notation is universal. Each solution is structured,
but proof and problem solving also invite creative and original thinking.
Mathematical modelling
Mathematical modelling can be applied to many different situations and
problems, leading to predictions and solutions. A variety of mathematical
content areas and techniques may be required to create the model. Once the
model has been created and applied, the results can be interpreted to give
predictions and information about the real world.
These key concepts are reinforced in the different question types included
in this book: Problem-solving, Communication and proof, and
Modelling.
12
1 Hyperbolic functions
1.1 Hyperbolic functions

As is well-
known, Physics
became a
science only
after the
invention of
differential
calculus.
Riemann
(1826–1866)
?
› How would you describe the curved shape in this suspension bridge?

The curved shape formed by the wires holding up a suspension bridge (as in
the photo above) is a catenary. A catenary is also formed when a chain is
hung between two posts.
y
6
−4 −3 −2 −1 0 1 234 x
▲ Figure 1.1
Figure 1.1 shows the same curve drawn with a different horizontal scale. If
you think it looks like a quadratic, you are in good company. Galileo made
1
Answers to exercises are available at www.hoddereducation.com/cambridgeextras
13
the same observation but on investigation
found that it was close to, but not exactly Note
1 the same as, a quadratic curve.
The equation of this curve inx its simplest
The sinh function is pronounced
in many different ways by different
x
+ e − and this people, but most commonly as
form is actually y e= ‘shine’, or ‘sine-aitch’, or ‘cinch’
2
function is called y = cosh x. (with a soft ‘c’).
xx −
−
A closely related curve is y = sinh x. Its equation is y e=e .
2
1 HYPERBOLIC FUNCTIONS
Hyperbolic functions and circular functions
Example 1.1
Simplify cosh2 x − sinh2 x.
Solution
2 2
2 2  ex + e−x   e xe − −x

cosh x − sinh x =   −  2 
2
2x −2 x 2x −2 x
= e + +2 e − e − +2 e
4 4
2x −2 x −2 x
e + +2 e − e 2 x + −2 e
=
4
= =4 1
4
?
› Compare the result in the example above to a similar result involving
the more familiar functions of sine and cosine.
You will have worked with the circular functions (often referred to as the
trigonometric functions, because of their uses in measuring triangles)
throughout your A Level Mathematics studies. They are properly called the
circular functions because they describe the coordinates of a point moving in
a circle – they parameterise the circle, since
x = cos q and y = sin q
give the circle x2 + y2 = 1.
y
1
(cosθ, sinθ)
1
sinθ Note
O cosθ 1 x This is, of course, where the identity
cos2 x + sin2 x ≡ 1 comes from.
2 ▲ Figure 1.2
14
If you use parametric equations x = cosh t, y = sinh t, you get one branch of
the rectangular hyperbola x2 − y2 = 1, rather than a circle. For this reason
the sinh and cosh functions and other related functions are known as the
hyperbolic functions.
1
y
(cosht, sinht) ?
› Why do the
parametric equations

O x x = cosh t, y = sinh t
only give one of the
branches of the
hyperbola? Which
branch do they give?
▲ Figure 1.3
Graphs and properties

The hyperbolic functions have similar properties to the circular functions, but
their graphs are not periodic.
y
6
e–x ex
5
cosh x
4
−4 −3 −2 −1 0 1 234 x
▲ Figure 1.4
Note
» Since cosh x = 1 (e xe+) − the
x graph of y = cosh x lies midway between the
2
graphs of y = ex and y = e−x .
» Notice that the function has a minimum point at (0, 1).
» The domain of the cosh function is x ∈ and its range is y  1.
3
15
ex
y Note
1 2
sinh x
» Similarly, the graph
of y = sinh x lies
1 midway between the
graphs of y = ex and
−3 −2 −1 0 1 2 3 x y = −e−x .
−1
» It passes through the
origin, where it has a
non-stationary point of
−2
inflection.
» The domain of the sinh
–e–x
function is x ∈ and
▲ Figure 1.5 its range is y ∈ .
The similarities with the circular functions are more obvious when it comes
to identities relating sinh x and cosh x. The reasons for the similarities
between hyperbolic and circular functions will become more apparent in
Chapter 5 on complex numbers.
The comparison with the circular functions also motivates a definition of the
ratio of sinh x to cosh x, to give an equivalent to the circular tan q function:
tanh x.
sinh x e
x
e − −x Note
tanh x = = x −x
cosh x e e+ Pronunciations of tanh x vary,
or (by dividing top and bottom by ex ) but usually, ‘tanch’ or ‘tan-
aitch’ or ‘than’ – with a soft ‘th’
−2 x
sinh x 1 e
tanh x = =− − as in ‘thistle’, not ‘this’.
cosh x 1 +e 2 x
This helps to visualise the function, since you can see that y → 1 as x → ∞
and y → −1 as x → −∞.
y
Note
1 y = tanh x The domain of the tanh
function is x ∈ and the
–2 0 2 x range is −1 < y < 1.
–1
▲ Figure 1.6
As you saw in Example 1.1:

cosh2 x − sinh2 x = 1
In fact, most of the circular trigonometric identities you already use have an
equivalent hyperbolic identity.
16
Trigonometric identity Hyperbolic identity
cos2 q + sin2 q ≡ 1
cos 2q ≡ cos2 q − sin2 q
cosh2 x − sinh2 x ≡ 1
cosh 2x ≡ cosh2 x + sinh2 x
1
sin 2q ≡ 2 sin q cos q sinh 2x ≡ 2 sinh x cosh x
θ
tan θ ≡ sin θ tanh x ≡
sinh x
cos cosh x
1 1
sec θ ≡ sech x ≡
cos θ cosh x

1 1
cosec θ ≡ cosech x ≡
sin θ sinh x
1 1
cot θ ≡ coth x ≡
tan θ tanh x
Note
You will notice that this table includes three further hyperbolic functions:
sech, cosech and coth. These are related to cosh, sinh and tanh in the same
way that the reciprocal trigonometric functions sec, cosec and cot are related
to cos, sin and tan.
Example 1.2
Solve the equation:
cosh x = 2 sinh x − 1
Solution
It is often easiest to convert the hyperbolic functions into their definitions in
terms of ex .
cosh x = 2 sinhx1−
x −x
e e+ = e−xe − x − 1
2
−x
0= −x
e 3e −2
(e ) − 3− 2e 0 =
x 2 x
Multiply by ex , to get a quadratic in ex .
(e ) − 2e 3−0=
x 2 x
(e 3−e 1)(0 ) + =
x x
e x3= Since ex can’t be negative.

x = ln 3
5
17
Example 1.3
Show that 1 +2sinh 2
θ = cosh 2θ .
1 Solution
1 ( x − −x )
Substitute sinh x = e e into the left-hand side
2
(2 (e e− ))
2
θ −θ
1 +2sinh 2
θ = 1+ 2 1
(e − 2e e + e )
θ θθ2 θ − −2
= 1+ 2× 1
4
= 1+ 1 (e θ − 2+ e θ )
2 −2
2
= 1 (e θ + e θ ) = cosh 2θ
2 −2
2
Exercise 1A 1 Using the definitions of the hyperbolic functions, prove that

CP sinh 2x = 2 cosh x sinh x.
2 Given that sinh x = 2, find the exact values of cosh x and tanh x.
3 (i) Rewrite the equation
cosh x + 2 sinh x = −1
in terms of ex , showing that it simplifies to:
3ex − e−x + 2 = 0
(ii) Multiply by ex to create a quadratic in ex .
(iii) Solve the quadratic equation to show that the only real root of this
equation is x = −ln 3.
4 Find all the real roots of each of the following equations:
(i) 10 cosh x − 2 sinh x = 11
(ii) cosh x − 5 sinh x = 5
(iii) 7 cosh x + 4 sinh x = 3
5 Given that
sinh x + sinh y = 25
12
cosh x − cosh y = 5
12
show that
2ex = 5 + 2e−y
and
3e−x = −5 + 3e y
Hence find the real values of x and y.
18
M 6 The diagram represents a cable hanging between two points A and B,
where AB is horizontal. The lowest point of the cable, O, is taken as the
origin of the coordinate system.
y
1
B A
O x

If the cable is flexible and has uniform density then the curve formed is
a catenary with equation:
( (xc )− 1)
y c=cosh
where c is a constant.
For a particular cable c = 20 m and AB = 16 m. Find the sag of the cable,
i.e. the distance of O below AB, and the angle that the tangent at A
makes with the horizontal.
CP 7 (i) Using the definition of cosh x in terms of the exponential function,
prove:
cosh 2 x = 21 (cosh 2x + 1)
Deduce that sinh x = 21 (cosh 2x1)− .

2
(ii)
CP 8 (i) Sketch the curve y = cosh x and the line y = x on the same axes.
Prove that cosh x > x for all values of x.
(ii) Prove that the point on the curve y = cosh x which is closest to the
((
line y = x has coordinates ln 1 + 2 , 2 ) )
and mark this point on
your graph.
PS 9 Find conditions on a, b and c which are necessary and sufficient to ensure
that the equation a cosh x + b sinh x = c has:
(i) two distinct real roots
(ii) exactly one real root
(iii) no real roots.
7
19
1.2 Inverse hyperbolic functions
1 Example 1.4
Solve the equation sinh x = 2.
Solution
sinh x = 2
x −x
e −e = 2
2
−
ex − e x = 4 Multiply through by ex .
e 2 x − 1= 4e x
This is a quadratic in ex .
e 2 x − 4e x − 1= 0
Using the quadratic formula.
e = ±4
x 20
2
= ±2 5 2 −5 is negative so ln(2 - 5)
is not real.
x = ln(2 ± 5)
=
The solution of the equation is x ln(2 + 5) .
The inverse function of the sinh function is denoted by sinh−1 , or sometimes

arsinh.
So the example above shows that sinh 2–1 ln(2
= 5) + .
The graph below shows the curves y = sinh x and y = sinh–1 x.
y y = sinh x
y=x
3
2
y = sinh–1 x
1
–4 –3 –2 –1 0 1 2 3 x
–1
–2
–3
–4
▲ Figure 1.7
As for any function and its inverse, the curves are reflections of each other
in the line y = x.
This allows you to find an expression for sinh–1 x, using a similar method to
Example 1.4.
−
y = sinh 1 x ⇒ =x sinh y
y −y
e − e
8 x =
2
y −y
2x = −e e
20
y −y
2 x = −e e
x 2+ is1 greater than
e 2 y − 2 xe 1y −0 =
± 4x + =
4+ x 2
x, so x −x + 2 1
is negative. As ey 1
e = 2x
y cannot be negative
x2 + 1
2 for any real y, the
(
y = ln x + x 2 + 1 ) negative square root
is discarded.
2
So sinh–1 x = ln x + x + .1( )
This result gives an easy method to solve an equation like sinh x = 2.

sinh x = ⇒
2 = x
−
sinh 21 ln=2 (+ 221+ln(2
= 5)) +
You can find the inverse of the tanh function in a similar way.
−1
y = tanh x ⇒ =x tanh y
y y = tanh–1 x
2y
−
x = 2y 1
e 3 y=x
e +1
xe 2 y + =x e 2 y − 1 2
1
e 2(1
y
− x ) 1= + x y = tanh x
e 2 y = +1 − x –3 –2 –1 0 1 2 3 x
1 x –1
2y = ln 1 − ()+x
1 x –2
1
y = ln 1 −
2 ()+x
1 x ▲ Figure 1.8
–3
x = ln (
1 −) .
1 +x
So tanh −1
2 1 x
Notice that both the sinh function and the tanh function are one-to-one, and
so the inverse functions are defined over the whole domain.
Example 1.5
Solve the equation cosh x = 2.
Solution
cosh x = 2
−
ex + e x = 2
2
x −x
e +e = 4 Multiply through by e x.
2x x
e +=
1 4e
This is a quadratic in e x.
2x x
e − 4e 1+0 =
Using the quadratic formula.
e x = ±4 12
2
= ±2 3 2 3− is positive so the equation
has two real roots.
x = ln(2 ± 3)
9
21
The two roots of the equation in Example 1.5 are shown on the graph below.
y
1
y = cosh x
2
ln(2 – √3) O ln(2 + √3) x
▲ Figure 1.9
You can see from the symmetry of the graph that one of the roots is the
negative of the other.
Note
−
It is probably not immediately obvious to you that ln(2 3) is the negative
of ln(2 + 3). However, it is quite easy to prove that this is the case by adding
them together.
ln(2 + 3) +ln(2 − 3) = (
+ 3)
ln (2 3)(2 − )
= ln (4 3− )
= ln1
=0
Since ln(2 + +
3) ln(2 − 3) 0, = ln(2 3) ln(2
− 3).= − +
Example 1.5 shows that the equation cosh x = k has two real roots if k > 1.
The inverse of the cosh function is denoted by cosh–1 (or sometimes arcosh).
For this to be a function, cosh–1 k must have only one value.
The graph of x = cosh y is shown in Figure 1.10.
y
Notice that this graph goes both above
and below the x-axis, so the negative part
has to be excluded to give the function
y = cosh–1 x.
O 1 x
▲ Figure 1.10
10
22
If the domain of the cosh function is restricted to y = cosh x
y
the non-negative real numbers, i.e. to x  0, then
the function is one-to-one, with the graph shown
by the red line in Figure 1.11. This restricted cosh
3
y=x
1
function has an inverse function – the cosh–1 2
y = cosh–1 x
function. 1
Partly because of the restriction in the domain of
cosh–1 x the derivation is more complicated: 0 4 3 2 1x
▲ Figure 1.11

-1
y = cosh x
x = cosh y
−y
2 x = +e y e
e 2 y − 2 xe 1y +0 =
2
2
y x ± 4 x −4
e =
2
2
= ±x x −1
( )(
y = ln x + x 2 − 1 or ln x − x 2 − 1 )
As in Example 1.5, the second root is the negative of the first (as you can see
in Figure 1.12 showing the y = cosh x curve).
y
y = cosh x
y1
–x 1 0 x1 x
▲ Figure 1.12
Since cosh–1 x > 0 by definition, the positive root is the required one.
Therefore:
cosh –1 x = ln x + ( x2 − 1 )
11
23
It is important to understand that there is a difference between solving the
equation cosh x = k and finding cosh–1 k. Compare this with solving an
1 equation like sin x = 0.5. When you use the inverse sine button on your
calculator, it gives you just one answer – the principal value. You then need to
work out any other roots in the required range. Similarly, if you use the square
root button on your calculator to solve the equation x2 = 2, you get the
positive square root only, but the full solution of the equation is x = ± . 2
ACTIVITY 1.1
Use
ln x + x 2 − 1 to find cosh-1 2.
( )
Howthe definition
is this relatedcosh–1 x =the equation cosh x = 2?
to solving
Example 1.6 2 2
Solve the equation 7cosh x + 2sinh x = 14
giving the answers in
logarithmic form.
Solution
Method 1
2 2
7cosh x + 2sinh x = 14
Substitute for sinh2 x
7cosh 2 x + 2 (cosh 2 x − 1=)14
2
9cosh x = 16
4
cosh x = ±
3 Notice there are two roots
of this equation.
But cosh x is not negative
()

() 
2
−1 4
So x = ± cosh 4 3 = ± ln 43 + 3
−1
 
(
= ± ln 1 4 + 7
3
)
Method 2
2 2
7cosh x + 2sinh x = 14
7 (1 +
sinh 2
x )+ 2 sinh2 x = 14
9 sinh2 x = 7
sinh 2 x = 7
9
7 Notice sinh x can take
sinh x = ± positive and negative values.
9
12
24
 7   7
2 
1
−1
= ln 7 +
 3  =  3  3  + 
x sinh 1
= ln 1
3
( 7 +4 )
   
=
Or x sinh
−1
− 7 ln − 7 + 4
 3  =  3 3 
It can be shown that this is the same
= ln 1 − +7 4 ( ) as the answer found by method 1. To

3 prove it, add the logarithms together.
2
Example 1.7 Solve the equation 4 cosh x − 19 sinhx9−0= giving your answers as exact
logarithms.
Solution
Substitute for cosh2 x in
4 (1 sinh
+ 2
x )− 19 sinhx − 9= 0 terms of sinh x
4 sinh2 x − 19 sinhx − 5= 0
(4 sinhx + 1)(sinh x − 5=)0
1
sinh x = − 4or sinh x = 5

() 
(( ))
2
= 1 −+− 1 +1
x ln ln 1 −1+17
 4 4  = 4
=5
or x ln (+ 2
5 1+ln= 5 )( + 26 )
Exercise 1B 1 Find the exact value of each of the following, giving your answers as
logarithms:
(i) cosh–1 3 (ii) sinh–1 1 (iii) tanh–1 0.5
(iv) sinh–1 (−2) (v)
−1
cosh 5
4
−1
(vi) tanh −
2
3 ()
2 Solve the following equations:
(i) sinh x = 3 (ii) cosh x = 3 (iii) tanh x = 0.2
3 Write the following as logarithmic functions:
(i) cosh–1 (2x) (ii) sinh–1 (x + 3) (iii) ()
tanh–1 2x
13
25
4 Solve the following equations. Give your answers in logarithmic form.
3 cosh x = 5 sinh x
1 (i)
(ii) 3 cosh x + 2 sinh x = 3
(iii) 2 sinh2 x − 7 cosh x + 8 = 0
(iv) cosh 2 x − 6 cosh x + 5 = 0
(v) sinh 2x = 5 sinh x
CP 5 Show that 2 tanh 1 x = tanh 1 2x 2
− −
1+ x
()
CP 6 Use the formula sinh ( A B + ) = sinh A B +sinh A B

cosh cosh to show
−1 −1 −1
that sinh x + sinh y = sinh x 1 + +y y x
2
( 2
+1 )
7
− − −
Prove that cosh 1 x + cosh 1 y = cosh 1 xy y+ ( 2
− 1 x2 − 1 )
KEY POINTS
e
x
+ e−x x
− −x
1 cosh x = sinh x = e e
2 2
y y
y = sinh x
y = cosh x x
1
x
e x
− e−x e 2x
−1
tanh x = x −x = 2x
e +e e +1
y
1 y = tanh x
–1
2 cosh 2 x − sinh 2 x ≡ 1
14
26
3 (
cosh –1 x = ln x + x 2 − 1 ) −
(
sinh 1 x = ln x x+ 2
+1 )
y
y = cosh−1 x y
y = sinh−1 x
1
x
1 x

tanh x = 12 ln (x11+−)x
–1
y
y = tan−1 x
−1 1 x
LEARNING OUTCOMES
Now that you have finished this chapter, you should be able to
■ understand the definitions of the following hyperbolic functions in
terms of the exponential function of
■ sinh x ■ tanh x ■ cosech x
■ cosh x ■ sech x ■ coth x
■ sketch the graphs of hyperbolic functions
■ prove and use the identities involving hyperbolic functions

2 2
■ cosh x − sinh x = 1
■ sinh 2x = 2sinh coshx x
■ cosh 2x = cosh x + sinh x = 2cosh x − = 1+
2 2 2 2
1 2 sinh x
■ understand and use the definitions of inverse hyperbolic functions
■ derive and use the logarithmic form of inverse hyperbolic functions.
15
27
2 Matrices
2 MATRICES
By relieving
the brain of all
unnecessary
work, a good
notation sets
it free to
concentrate on
more advanced
problems.
A.N. Whitehead,
(1861–1947)
Matrices are used extensively in computer games to apply rotations and

Note translations to images that will appear on the screen. The positions and
This chapter orientations of these images are defined by coordinates and vectors.
follows on from
the matrices work
you will have done 2.1 Important results relating
in Further Pure
Mathematics 1. to 2 × 2 matrices
However, the a b 
first few pages The determinant of the general 2 × 2 matrix M =
of this chapter c d  is given by
revisit some of
the work that ab
was done there det M = = −ad bc
cd
and the ideas
underpinning the If M represents a transformation, then the determinant of M represents the
rest of the chapter. scale factor associated with the transformation.
16
28
If a matrix has zero determinant, the area scale factor of the transformation
is zero, so that all points are mapped onto a shape with zero area. In fact, the
matrix maps all points in the plane to a straight line.
For square matrices P and Q, det (PQ
)= det×
det P Q.
2
a b  −1 1  db− 
The inverse of a 2 × 2 matrix M = M = .
c d  is −
ad bc − c a 
If the determinant is zero then the inverse matrix does not exist and the
matrix is said to be singular. If det M ≠ 0 the matrix is said to be
2.1
Important
non-singular.
If a matrix is singular, then it maps an infinite number of points in the plane
to the same point on the straight line. It is therefore not possible to find the
results relating
inverse of the transformation, because it would need to map a point on that
straight line to just one other point, not to an infinite number of them.
An important result about a matrix and its inverse is that MM−1 = M−1 M = I.
This is true for all square matrices, not just 2 × 2 matrices.
The inverse of the product of two non-singular n × n matrices M and N
is given in terms of the inverses M−1 and N−1 through the relationship:
to 2 × 2
(MN )−=1 N −M 1 1
.− This means that when working backwards, you must
matrices
reverse the second transformation before reversing the first transformation.
Note
The example shown here is for a pair of simultaneous equations in two
unknowns, which is not in the Cambridge International syllabus but provides
a good introduction to the problem of three simultaneous equations in three
unknowns, which is in the syllabus.
Using matrices to solve simultaneous equations

There are a number of methods to solve a pair of linear simultaneous
equations of the form
3x + 4y = 7
2x − y = 12
such as elimination, substitution or graphical methods.
Another method involves the use of inverse matrices. This method has the
advantage that it can more easily be extended to solving a set of n equations
in n variables.
Example 2.1 Use a matrix method to solve the simultaneous equations

3x + 4y = 7
2x − y = 12 ➜
17
29
Solution
2  3 4   x  7 
2 1− y  = 12 
Write the equations in
matrix form.
3 4  − −  1 4 
The inverse of the matrix   is − 1  1 4 = 1  .
 2 1−  11  −2 3  11  2 3− 
1  1 4   3 4   x 1 1 4  7 
2 MATRICES
11 2 3− 2 1− y  = 11 2 3− 12  As M −1M = I the

left hand side
x   55   5  x
1 simplifies to   .
Pre-multiply both  =   = −  y
sides of the matrix  y  11  −22   2 
equation by the
inverse matrix.
The solution is x = 5, y = −2.
Geometrical interpretation in two dimensions

Two equations in two unknowns can be represented in a plane by two
straight lines. The number of points of intersection of the lines determines
the number of solutions to the equations.
There are three different possibilities.
Case 1
Example 2.1 shows that two simultaneous equations can have a unique
solution. Graphically, this is represented by a single point of intersection, as
shown in Figure 2.1.
y
4
3
2 2x – y = 12
1 3x + 4y = 7
–3 –2 –1 O 1 2 3 4 5 6 7 8 9 10 11 x
–1
–2
–3
–4
–5
▲ Figure 2.1
This is the case where det M ≠ 0 and so the inverse matrix M−1 exists,
allowing the equations to be solved.
18
30
Case 2
If two lines are parallel they do not have a point of intersection. For example,
the lines 2
x + 2y = 10 The equations can be written in matrix
x + 2y = 4 form as 
1 2   x   10  .
1 2  y  = 4 
are parallel (see Figure 2.2).
y
2.1
Important
5
results relating
x + 2y = 4 x + 2y = 10
4 10 x
▲ Figure 2.2
 
The matrix M = 1 2  has determinant zero and hence the inverse
to 2 × 2
1 2 
matrices
matrix does not exist.
Case 3
More than one solution is possible in cases where the lines are coincident, i.e.
lie on top of each other. For example, the two lines
x + 2y = 10 The equations can be written in matrix
3x + 6y = 30  
form as  1 2  x  10 
.
are coincident (see Figure 2.3). You 3 6 y  = 30 
can see this because the equations
are multiples of each other.
y
3x + 6y = 30 x + 2y = 10
0 10 x
▲ Figure 2.3
 
In this case the matrix M is  1 2  and det M = 0.
3 6 
There are infinitely many solutions to these equations.
19
31
Exercise 2A
2x − 7 2 
2 1 The matrix   has determinant −4.
 9 8− x x + 1
Find the possible values of x.

2 (i) Write down the matrices A and B which represent:
A - a reflection in the y-axis
B - a reflection in the line y = −x
(ii) Show that the matrices A and B each have determinant of −1.
2 MATRICES
(iii) Draw diagrams for each of the transformations A and B to

demonstrate that the images of the vertices labelled anticlockwise
on the unit square OIPJ are reversed to a clockwise labelling.
3 The graph shows the unit square y
transformed by a shear. 5 P′
(i) Write down the matrix which 4 I′
represents this transformation. 3
(ii) Show that under this transformation the 2
J P
area of the image is always equal to the 1
J′
area of the object. I
O
–1 –2 1 23 x
 − 
4 For the matrix 3 5
1 5  :
(i) find the image of the point (2, −1)

(ii) find the inverse matrix
(iii) find the point which maps to the
image (3, 1)
3 − k k 
5 The matrix   is singular.
 22 − k
Find the possible values of k.

 −2 3 
6M =   and N = 4 7 
 −1 6  9 1−  .
(i) Find the determinants of M and N.

(ii) Find the matrix MN and show that det(MN) = det M × det N.
 −   6 1−3 6 
7 Given that M = 1 3
2 1  and MN =5 2−6 12  , find the matrix N.
20
32
 3 9− 
8 The plane is transformed by the matrix M = .
(i)
− 2 6 
Draw a diagram to show the image of the unit square under the
2
transformation represented by M.
(ii) Describe the effect of the transformation and explain this with
reference to the determinant of M.
9 Use matrices to solve the following pairs of simultaneous equations:
(i)5x − 3y = 13 (ii) 4x − y = −16
2.1
Important
2x − y = 5 x − 3y = −15
10 Find the two values of k for which the equations
2x + ky = 3
results relating
kx + 8y = 6
do not have a unique solution.
How many solutions are there in each case?
a b 
11 M =   is a singular matrix.
c d 
to 2 × 2
(i) Show that M2 = (a + d )M.
matrices
(ii) Find a formula which expresses Mn in terms of M, where n is a
positive integer.
(iii)Prove the formula you found in part (ii) by induction.
a b 
12 The plane is transformed using the matrix   where ad − bc = 0.
c d 
Prove that the general point P(x, y) maps to P’ on the line cx − ay = 0.
13 Triangle T has vertices at (1, 0), (0, 2) and (−3, 0).
4 1 
It is transformed to triangle T’ by the matrix M =  .
1 1 
(i) Find the coordinates of the vertices of T’.
Show the triangles T and T’ on a single diagram.
(ii) Find the ratio of the area of T’ to the area of T.
Comment on your answer in relation to the matrix M.
(iii) Find M−1 and verify that this matrix maps the vertices of T’ to the
vertices of T.
 −1 2 
14 (i) The matrix =S   represents a transformation.
 −3 4 
(a) Show that the point (1, 1) is invariant under this transformation.
(b) Calculate S−1 .
(c) Verify that (1, 1) is also invariant under the transformation
represented by S−1 .
21
33
(ii) Part (i) can be generalised as follows:
2 If (x, y) is an invariant point under a transformation represented

by the non-singular matrix T, it is also invariant under the
transformation represented by T−1 .
 x   x
Starting with T
y  = y  prove this result algebraically.
15 The simultaneous equations
2x − y = 1
2 MATRICES
3x + ky = b
x   1.
are represented by the matrix M   =  
y  b 
(i) Write down the matrix M.
(ii) State the value of k for which M −1 does not exist and find M −1 in
terms of k when M −1 exists.
Use M −1 to solve the simultaneous equations when k = 5, b = 21.
(iii) What can you say about the solutions of the equations when
3
k=− ?
2
(iv) The two equations can be interpreted as representing two lines in
the x−y plane.
Describe the relationship between the two lines when:
(a) k = 5, b = 21
= − 3
(b) k ,b =1
2
(c) k = − 3, b = 3
2 2
3 2  1 3− 
16 Matrices M and N are given by M =   and N = .
0 1  1 4 
(i) Find M −1 and N −1 .
(ii) Find MN and (MN )−1 . Verify that (MN )−1 = N −1 M −1 .
(iii) The result (PQ )−1 = Q −1 P −1 is true for any two 2 × 2
non-singular matrices P and Q.
The first two lines of a proof of this general result are given below.
Beginning with these two lines, complete the general proof.
(PQ )−1 PQ = I
(PQ )−1 PQQ−1 = IQ−1
22
34
2.2 Finding the inverse of a 3 × 3
matrix 2
In this section you will find the determinant and inverse of 3 × 3 matrices
using a non-calculator method.
Finding the determinant of a 3 × 3 matrix without

using a calculator
2.2
Finding
To find the inverse of a 3 × 3 matrix start by finding its determinant.
 a1a2 3a 
 
If M is the 3 × 3 matrix  b1 b2 3b  then the determinant of M is defined by
the
inverse
 
 c 1c2 c3 
b2b3 b1b3 b1 b2
det M = a1 − a2 + a3
of a 3 × 3
c 2c3 c 1c3 c 1c2
which is sometimes referred to as the expansion of the determinant by
matrix
the first row.
 304 
For example, to find the determinant of the matrix A = −  2 1 0 
 
 121 
1 0 − −2 0 −
+4 21
Notice that you do not really
det A = 3 0 −2 0
21 11 12 need to calculate as it is
11
= 3(1 − 0) − 0(−2 − 0) + 4(−4 − 1) going to be multiplied by zero.
= 3 − 20 Keeping an eye open for helpful
zeros can reduce the number of
= −17 calculations needed.
b2b3
The 2 × 2 determinant is called the minor of the
c 2c3
element a 1. It is obtained by deleting the row and column containing a 1:
a1a a 2 3
b1b b 2 3
c 1c c 2 3
Other minors are defined in the same way, for example the minor of a 2 is
a1a a 2 3
b1b3
b1b b 2 3 =
c 1c3
c 1c c 2 3
23
35
You may have noticed that in the expansions of the determinant,
the signs on the minors alternate as shown: +−+
2 A minor, together with its correct sign, is known as a cofactor
and is denoted by the corresponding capital letter; for example,
−+−
+−+
the cofactor of a 3 is A 3. This means that the expansion by the first
row, say, can be written as
a 1 A 1 + a 2 A 2+ a 3 A 3.
2 MATRICES
Note
You could use the expansion of the determinant by the second row:
a 2a3 a1a3 a1a2

det M = − b1 + b2 − b3 ,
c 2c3 c 1c3 c 1c2
or the expansion of the determinant by the third row:
a 2a3 a1a3 a1a2

det M = + c 1 − c2 + c3 .
b2b3 b1b3 b1b2
It is fairly easy to show that all three expressions above for det M simplify to:
a1b2 3c + a 2b3 1c + a3b1 2c − a 3b2 1c 1a3−b2 2c1a3 b c −
Example 2.2
 372 − 
Show that the determinant of the matrix M =  0 2 1  is -23.
− − 
 413 
Solution
Expanding by the first row using the expression:
b2b3 b1 b3 b1 b2
det M = a1 − a2 + a3
c 2c3 c 1c3 c 1c2
gives:
21 − −7 0 1 −− 02
det M = −3 +−
( 2)
13 43 41
= 3 (6 – (−)1 ) – 7–(0 (−)4 2)−0 8−(− ( ) )
= 21 −28 16−
= −23
To find the determinant you can also expand by columns. So, for
example, expanding by the first column would give:
21 − 7 2− + −( )4 7 2−
3 0
−1 3 −1 3 21
which also gives the answer −23. Notice that expanding by the first
column is quicker here as it has a zero element.
24
36
Earlier you saw that the determinant of a 2 × 2 matrix represents the area
scale factor of the transformation represented by the matrix. In the case
of a 3 × 3 matrix the determinant represents the volume scale factor.
2 0 0 
2
For example, the matrix  0 2 0  has determinant 8; this matrix represents
an enlargement of  
0 0 2 
scale factor 2, centre the origin, so the volume scale factor of the
Recall that a minor,
together with its transformation is 2 × 2 × 2 = 8.
correct sign, is
2.2
To find the inverse of a 3 × 3 matrix using cofactors
Finding
known as a cofactor
and is denoted by
the corresponding  A1B C1 1

capital letter;  
The matrix  A2B C2 2 
is known as the adjugate or adjoint of M,
the
for example the
 
inverse
cofactor of a 3 is A 3.  A3B C3 3 
denoted adj M.
The adjugate of M is formed by
of a 3 × 3
» replacing each element of M by its cofactor;
» then transposing the matrix (i.e. changing rows into columns and columns
matrix
into rows).
The unique inverse of a 3 × 3 matrix can be calculated as follows:
 A1 A
23A

 
M
−1
= 1 adj M = 1  B1B2 3B  , det M ≠ 0
det M det M  
C 1C2 3C 
The steps involved in the method are shown in the following example.
Example 2.3
Find the inverse of the matrix M where
2 2 3 − 
 
M = 3 5−6 .
4 2 3 − 
Solution
Step 1: Find the determinant ∆ and check
∆≠0
Expanding by the first row
−5 6 −
∆= 2 −2 36 ( )3 3 5
+−
2 3− 4 3− 42
= ×(2−3× −) ( 2 33) −
( ×3 26 ) = − 6
➜
25
37
Therefore the inverse matrix exists.
2 Step 2: Evaluate the cofactors
−5 6 36 3 5−
A 1= =3 A 2=−− = 33 A 3 = = 26
2 3− 43 42
2 3− = 0 2 3− = 6
B 1=− B 2= B 3=−2 2 =4
2 3− 4 3− 42
2 MATRICES
−
C 1 = 2 3 = −3 C 2 = − − 2 3 = −21 C 3 = 2 2 = −16
−5 6 36 3 5−
1
Step 3: Form the matrix of cofactors and transpose it, then multiply by ∆
 3 33 26 T
  The capital T indicates the
-1 1
M = − − − 0−6 4 matrix is to be transposed.
6 

 3 21 16 
Matrix of cofactors.
 3 0 −3 
= 1  33 6 21−  1.
−6   Multiply by You can evaluate the
 26 4 16−  ∆
determinant Δ using these
cofactors to check your
 −3 0 3  earlier arithmetic is correct:
= 1  −33 6−21 
2nd Row:
6 
 −26 4−16  Δ = 3B 1 − 5B 2 + 6B 3
The final matrix could then be simplified = 0 − 30 + 24 = −6 ✓
and written as 3rd Row:
 1  ∆ = 4C 1 + 2C 2 − 3C 3
− 01  = −12 − 42 = −6 ✓
 2 2
M 1 =  − 11 −1 7 
−
 2 2
 13 2 8 
− − 
 3 3 3
 −3 0 3  2 2 3 − 
−1    
Check: M M 1 = −33 6 −21  3 5 −6
6  
 4 2 3

 −26 4 −16   − 
6 0 0  1 0 0 
   
= 1  0 6 0  =  0 1 0  as required.
6
0 0 6  0 0 1 
26
38
This cofactor method for finding the inverse of a 3 × 3 matrix is reasonably
straightforward but it is important to check your arithmetic as you go along,
as it is very easy to make mistakes. 2
As shown in the previous example, you can also multiply the inverse by the
original matrix and check that you obtain the 3 × 3 identity matrix.
Exercise 2B 1 Evaluate these determinants.

113 1 1−3
(i) (a) − (b)
2.2
102 101
Finding
314 324
1 5−4− 122−
the
(ii) (a) 233 (b) −5 3 1
inverse
−2 1 0 −4 3 0
212 150
of a 3 × 3
(iii) (a) 353 (b) 150
1 1−1 212−
matrix
What do you notice about the determinants?
CP 2 Find the inverses of the following matrices, if they exist.
1 2 4  326 
(i) 2 4 5  (ii)  5 3 11 
   
0 1 2   7 4 16 
 555 −   656 
(iii)  −9 3 5 −  (iv)  −5 2 4 − 
   
− −
 468  − −
 465 − 
 4 5−3 
3 Find the inverse of the matrix  3 3 4 −  and hence solve the
 
simultaneous equations: 5 4 6 − 
4x − 5y + 3z = 3
3x + 3y − 4z = 48
5x + 4y − 6z = 74
1 3 2 − 
4 Find the inverse of the matrix M = k  0 4  where k ≠ 0.
 
−
2 1 4 
For what value of k is the matrix M singular?
5 (i) Investigate the relationship between the matrices
 031  1 0 3  3 1 0 
A =  242  B = 2 2 4  C = 4 2 2 
     
 −1 3 5   5 1−3  3 5 1 − 
(ii) Find det A, det B and det C and comment on your answer.
27
39
22 x
Show that x = 1 is one root of the equation 1 1x = 0 and find the
2 6
other roots. x 14
 3 1−1 
7 Find the values of x for which the matrix  2 x 4  is singular.
 
x 1 3 
 k 21 
8 Given that the matrix M =  0 − k 2  has determinant greater than 5,
2 MATRICES
 
2k1 3
find the range of possible values for k.
CP 9 (i) P and Q are non-singular matrices. Prove that (PQ)−1 = Q−1 P−1 .
 0 3 1− 
(ii) Find the inverses of the matrices P =  −2 2 2  and
 
 −3 0 1 
2 1 2 
Q =  1 0 1 .
 
 4 3−2 
Using the result from part (i), find (PQ)−1 .
ka1ka ka2 3 a1a2 3a

10 (i) Prove that b1b b 2 3
= k b1b2 3b , where k is a constant.
c1 c2 c3 c 1c2 c3
(ii) Explain in terms of volumes why multiplying all the elements in

the first row by a constant k multiplies the value of the determinant
by k.
(iii) What would happen if you multiplied a different row by k?
123
CP 11 Given that 645 = 43, write down the values of the determinants:
751
10 2 3 4 10 21−
(i) 60 4 5 (ii) 24 20 35−
70 5 1 28 25 7 −
x4 1 12 y
x
x 43 y
2 20
(iii) 6x8 5 y (iv) 6x 4x y
7x10 y 7x 4 5 4y
2x
28
40
2.3 Intersection of three planes
In most cases, two planes intersect in a straight line. The exception occurs
2
when they are parallel. It is also possible that they are one and the same plane.
However in what follows, it is assumed that all the planes being considered
are distinct.
In this section you will look at the different possibilities for how three planes
can be arranged in three-dimensional space.
Intersection
2.3
You saw in Further Pure Mathematics 1, Chapter 7, that the cartesian equation
of a plane is of the form ax + by + cz + d = 0. Solving a system of three
equations in three unknowns x, y and z is equivalent to finding where the
three planes represented by those equations intersect.
There are five ways in which three distinct planes π 1, π 2 and π 2 can intersect
of
threeplanes
in 3D space.
If two of the planes are parallel, there are two possibilities for the third:
» it can be parallel to the other two (see Figure 2.4); or
» it can cut the other two (see Figure 2.5).
The planes
intersect in two
parallel lines.
There are no
intersection
The planes do
points that are
not intersect.
common to all
three planes.
▲ Figure 2.4 ▲ Figure 2.5
If none of the planes are parallel, there are three possibilities:

» The planes intersect in a single point (see Figure 2.6)
▲ Figure 2.6
29
41
» The planes form a sheaf (see Figure 2.7)
2 The three planes share a

common line so there are an
infinite number of points of
intersection of the three planes.
▲ Figure 2.7
2 MATRICES
» The planes form a triangular prism (see Figure 2.8)
Each pair of planes intersects

in a straight line; the three
lines are parallel. There is no
intersection point common to
all three planes.
▲ Figure 2.8
The diagrams above show that three planes either intersect in a unique
point, intersect in an infinite number of points or do not have a common
intersection point.
Finding the unique point of intersection

of three planes
3 × 3 matrices can be used to find the point of intersection of three planes
that intersect in a unique point or to determine that the planes have a
different arrangement.
Suppose the three planes have equations Notice that the vector a 1i +b 1j +
c 1k is perpendicular to the plane
a1x b+ y 1c z+d 1
= 1 a 1x + b 1y + c 1z = d 1.
a 2x b+y c2 z +d 2
= 2
Similarly, a 2i + b 2 j + c 2k and
a 3i + b 3 j + c 3k are perpendicular
a 3x b+y c3 z +d 3
= 3 to the two other planes.
x  d   a1b1 1c 
1
     
They can be written as M  y  =  d 2  , where M =  a 2b2 2c .
z     a b c 
d3  333
Example 2.4
Three planes have equations
π 1: x + 3y − 2z = 7 ①
π 2: 2x − 2y + z = 3 ②
π 3: 3x + y − z = k ③
30
42
(i) Explain how you know that none of the planes are parallel to any of the
other planes.
(ii) Investigate the intersection of the planes when: 2
(a) k = 10 (b) k = 12
You cannot Solution

multiply any of
(i) If planes are parallel the left-hand sides of their equations are the same
the equations by a
or one is a scalar multiple of the other. That is not true in this case, so
Intersection
2.3
number to obtain
one of the other none of the planes are parallel to any of the others.
equations.
 1 3 2 −   x   7
Writing the equation
(ii)  2 2−1   y  =  3
    
of the planes in 3 1 1 −   z  k
the matrix format
of
threeplanes
 x  d1  det M = 0 so the matrix is singular and hence the planes do not
  intersect at a unique point.
M  y  =  d2  .
   d 
 z 3
Eliminating the variable z produces two equations in x and y:
① – 2 × ③: −5x + y = 7 − 2k For the equations to be

consistent, the value of
To investigate ② + ③: 5x − y = 3 + k 5x − y must be the same
the intersection in each equation.
of the planes in
a case where the These equations are consistent if 2k − 7 = 3 + k
matrix is similar,
⇔ k = 10
try to solve
the equations and in this case both equations reduce to 5x − y = 13.
algebraically.
(a) The equations are consistent in the case k = 10 and so there are
infinitely many solutions. Since none of the planes are coincident,
they intersect in a straight line and form a sheaf.
(b) When k = 12 the two equations in x and y are inconsistent:
−5x + y = −17
5x − y cannot be
equal to both 5x − y = 15
17 and 15.
Therefore there are no solutions. Since the planes are not parallel
they must form a prism of planes.
Example 2.5
Solve the set of equations
2x +
y z+ = 3 7
x++ 2y=z 1
3x +z= 5 ➜
31
43
Solution  x   7
2 This may be rewritten in matrix form: M  y  =  1 , where
   
 z   5
2 1 3 
M = 1 2 1  .
 
3 0 1 
21 − 11 + 12
You can then show that det M = 2 1 3 = 4−
−12
1− 15
=≠ 0,
01 31 30
2 MATRICES
which means that there will be a single point of intersection between the
three planes.
2x +
y z+ = 3 7 A
x + 2+y=z 1 B
3x +z= 5 C
3B – A ⇒ x+
y 5= − 4 D
C–B⇒ 2x − =
2y 4 E
2D – E ⇒ 12y = − 12 ⇒ y = −1, x = − − −4 5 1 (1)and
=
z=−
5=32
The solution is the point 1( 1, −2, )
Example 2.6 Solve the set of equations

x y− + 2 z = 5
x−y + 2z = 9
2 x − 2+y4 z = 7
Solution
x y− + 2 z = 5
x y− + 2 z = 9
2 x − 2+y4 z = 7
The left-hand side of the first two equations are the same. The left-hand side
of the third equation is twice that of the other two. This means that since
the right-hand sides of the equations are all different, the three equations
represent planes which are parallel to one other, and there can be no point
of intersection.
There is no unique solution to the set of equations.

2x +
y z+ = 3 6
2x +
y z+ = 3 9
x−y − 2z = 0
32
44
Solution
2x +
y z+ = 3
2x +
y z+ = 3
6
9
A
B
2
x−
y − 2z = 0 C
A and B are the equations of two parallel planes. There is therefore no
intersection between these.
A and B are not consistent as 2 x +
y z+ 3 cannot be equal to 6 and 9 at the
Intersection
2.3
same time.
C is the equation of a third plane which intersects the other two planes
along parallel lines.
A+C ⇒ 3x +
z= 6
⇒ 3x +
z=
of
B+C 9
threeplanes
These are the equations of two parallel lines with no points of intersection.
There is no unique solution to this set of equations.

x+
y z3− 2= − 2
−3+x +y =z 6
−3+x11 4 y −
z= 6
Solution
x+y z3− 2= − 2 A
−3+x +y =z 6 B
−3+x11 4 y − z= 6 C
Eliminate z:
A +2B ⇒ − + 5x 5y10= ⇒ = y+ x 2
4B +
C ⇒ − 15x + 15y = 30 ⇒ = y+ x 2
A+2B and 4B+C both reduce to the equation y = x +2. You can only solve
for x and y in terms of a parameter λ.
If x = λ then y = λ + 2 and from equation B:
6 3− =x+y− + =6 +3 λ λ (
z=+ 2 )4 2 λ
There is a line of solutions which is given in terms of a single parameter λ
 x   1  0
 y  =  1 λ +  2
     
 z  2   4
33
45
Figure 2.9 summarises the decisions that need to be made and Examples 2.4
to 2.8 show how these decisions are carried out. Note that the trivial
2 situation where the matrix M consists entirely of zeroes has been ignored,
and also cases where two or three of the equations are the same (i.e. two or
three of the planes are coincident).
Three equations in three unknowns represent
3 equations the equation of three planes. These can be
in 3 unknowns
 x   d1 
 
expressed in the form M  y  =  d 2  .
   
2 MATRICES
 z   d3 
No Yes
detM = 0?
The three planes
intersect in a straight
line and form a sheaf.
M non-singular M singular
The three planes

form a prism,
or two planes
are parallel
unique solution and the third
No equations Yes
consistent? intersects, or all
three planes are
So the three parallel.
planes would
intersect at a infinitely many
no solution
unique point. solutions
▲ Figure 2.9
If the determinant of M is zero then there is no unique solution to the

equations i.e. no unique point of intersection of the planes represented by
the matrix. In this case one of the other configurations of the planes would
be relevant.
To summarise:
If M is non-singular, the planes intersect in a unique point.
If M is singular, the planes must be arranged in one of the other four possible
configurations:
» three parallel planes
» two parallel planes that are cut by the third to form two parallel lines
» a sheaf of planes that intersect in a common line
» a prism of planes in which each pair of planes meet in a straight line but
there are no common points of intersection between the three planes.
34
46
Exercise 2C 1 (i) Find the determinant and inverse of the
 2 3−1
matrix M =  3 0 4

.
2
 − 
113
(ii) Using your answer to part (i), show that the planes
2x − 3y z+ = 10
3x + 4 z = 25
Intersection
2.3
x y− + 3z = 20
intersect in the unique point (0, −1.25, 6.25).

2 Without carrying out any calculations, describe the arrangements of the
following sets of planes:
π1 : 2 x y+ z− =
of
(i) 4
threeplanes
π 2 : x y+ z− = 2
π 3 : 2 x y+ z− = 6
(ii) π 1: 2 x y+ z− = 4
π 2 : 6 x + 3 y − 3 z = 12
π 3 : 2 x y+ z− = 6
(iii) π1 : 2 x y+ z− = 4
π 2 : 10 x + 5y − 5 z = 15
π 3 : 2 x y+ z− = 6
3 (i) Express the equations of the three planes
π 1 : 5 x + 3y − 2 z = 6
π 2 : 6 x + 2 y + 3 z = 11
π 3 : 7 x y+ + 8z = 12
 x  d1 
 
in the form M  y  =  d 2  and show that det M = 0.
   d 
 z 3
(ii) Eliminate the variable y from the three equations and hence show
that the three planes form a prism of planes.
M 4 Determine the arrangement of the following planes in three dimensions.
(i) π1 : x + 2 y + 4 z = 7 (iii) π1 : 2 x y− = 1
π 2 : 3 x + 2 y + 5 z = 21 π 2 : 3 x2 z+
13=
π 3 : 4 x y+ + 2z = 14 π 3 : 3 y4z+23=
(ii) π 1: x y+ z+ = 4 (iv) π1 : 3 x2 y+ z+ = 2
π 2 : 2 x + 3y − 4 z = 3 π 2 : 5 x3 y+
4z−1 =
π 3 : 5 x + 8 y − 13 z = 8 π 3 : x y+ z+ = 4 5
35
47
(v) π 1: 2 x y+ z− = 5
π 2 : 8 x + 4 y − 4 z = 20
2 π : 2 xyz
−−+=−
3
5
CP 5 Solve, where possible, the equation
 1 3 2 −   x  −2 
 −3 1 m   y =  6 
 −    
3 11 4 −   z  k
2 MATRICES
in each of these cases:

(i) m= 2, k = 3 (ii) m= 1, k = 3 (iii) m =k =1, 6
In each case interpret the solution geometrically with reference to three
planes in three dimensions.
 k 7−..4 
PS 6 (i) Obtain an expression for the inverse of the matrix  2 2−..3 
 − − 
in terms of k. 132
State the value of k for which the planes
kx − 7 y + 4 z = 3
2x − 2y + 3z = 7
x − 3y − 2z = − 1
would not intersect in a unique point.
(ii) Describe the intersection of the planes
4 x − 7y + 4 z p=
2 x − 2 y + 3z = 1
x − 3y − 2 z = 2
giving your answer in terms of p.
(iii) Find the value of p for which the planes
5x − 7y + 4 z p=
2x − 2y + 3z = 1
x − 3y − 2z = 2
have at least one point of intersection and describe the

arrangement of the planes geometrically in three dimensions.
7 Solve the set of equations
3x +y z+ = 5
x + 2+y=z 1
x−
y z3− = 3
36
48
8 Find the values of k for which the set of equations
x +2 +y =kz
x +3 +y =z 3
0
2
kx +
y 8z+ 5= 6
has no unique solution.
Solve the set of equations for each of these values of k and interpret
them geometrically.
Intersection
2.3
9 Find the set of values of a for which the system of equations
ax +y z+ 2 =
0,
3x2−y 4, =
3x4−y6az
− = 14,
of
has a unique solution.
threeplanes
Cambridge International AS & A Level Further Mathematics
9231 Paper 13 Q2 October/November 2012
142 
10 Show that the matrix  3 0 2 −  has no inverse.
 
 3 3−4− 
Solve the system of equations
x + 4+y2z = 0,
3x 2− =z 4,
3x3−
y4z− = 5.
11 Find the value of a for which the system of equations
x −y z+ 2 = 4,
x+ ay−z=b 3 ,
x− y z+ 7= 13.
where a and b are constants, has no unique solution.
Taking a as the value just found,
(i) find the general solution in the case b = −5,
(ii) interpret the situation geometrically in the case b ≠ −5.
37
49
2.4 Eigenvalues and eigenvectors
2 2 × 2 matrices
In a reflection, every point on the mirror line maps to itself. The line may be
described as a line of invariant points, since every point on the line is itself
invariant.
A line of invariant points is a special case of an invariant line, where the
image of every point on the line is itself on the line but is not necessarily the
2 MATRICES
original point.
As an example, think of the lines through the origin in an enlargement, scale
factor 2, centre the origin. Each point on the line in Figure 2.10 maps to
another point on the line, but the origin maps to itself.
y
P ′1
P1
O x
P2
P ′2
▲ Figure 2.10
This idea is now developed in terms of matrices, but only for lines which pass
through the origin.
As an example, look at the effect of the transformation with matrix
4 2 
M =
1 3  .
 4 2   1  6 
Since
1 3 1 = 4  the transformation defined by pre-multiplying
   6
position vectors by matrix M maps the vector 1
1 to4  .
 k   6k 
Similarly, the image of   is   . Each point on the line y=xcan be
k 4k
 k
represented by the position vector of the form
k  and the points with
 6k  2
position vectors = x . This means that under the
4k  form the line y 3
transformation represented by the matrix M the image of the line y x = is
2  4 2   1  8
the line y = x. Similarly, since
3 1 3 2  = 7  , the image of the line
7
y = 2 xis the line y = x.
8
38
50
ACTIVITY 2.1
(i) Find the images of the following position vectors under the transformation
4 2 
given by M =
1 3  .
 1  2  0
(a) (b) (c)
0  1  1 
 − 1  −1  −2 
(d) (e) (f)
 2   1   1 
(ii) Use your answers to part (i) to find the equations of the images of the
following lines.
1
(a) y = 0 (b) y = x (c) x = 0
2
1
(d) y = −2 x (e) y = − x (f) y = − x
2
The information you have just gathered may be represented as in Figures 2.11
and 2.12, where the object lines and their images are shown in separate diagrams.
(e) (d)
(e) (d) (c) (c)
(f) (b)
(b)
(f)
(a)
O (a) O
However, you can show all the information on one diagram, as in Figure 2.13,
where the object lines are shown at the centre of the diagram and their image
lines are shown in the outer section of the diagram.
39
51
y
1
y= x
2
y = –x
▲ Figure 2.13
The shaded part of the diagram is not directly relevant but shows lines
connecting each object line to its image. You will notice that there are two
1
invariant lines, y = x and y = − xwhich map to themselves under this
2 1
transformation. The other lines appear to crowd towards y x = , moving
2
away from y = − x. This diagram prompts several questions.
» Are there other invariant lines?
1
» Why does y = x attract and y x= − repel?
2
» Do all transformations behave like this?
» How can such lines be found efficiently?
Terminology
To answer these questions you need suitable terminology.
If e is a non-zero vector such that Me e= λ , where M is a matrix and λ is
a scalar, then e is called an eigenvector of M. The scalar λ is known as an
eigenvalue.
 4 2   2  10   2
Therefore, since 5
1 3 1 = 5  = 1 
 4 2   − 1  − 2   −1
and 2
1 3  1  =  2  =  1 
 2  −1  
= 42
1  and 1  are eigenvectors of the matrix M 1 3  . The
corresponding eigenvalues are 5 and 2, respectively. It will become evident

later that these are the only two eigenvalues.
40
52
Properties of eigenvectors
Notice the following properties of eigenvectors.
 2  −1
1 All non-zero scalar multiples of
1  and  1  are also eigenvectors of
M, with respectively the same eigenvalues.
2 Under the transformation the eigenvector is enlarged by a scale factor
equal to its eigenvalue.
3 The direction of an eigenvector is unchanged by the transformation.

(If the eigenvalue is negative, then the sense of the eigenvector will be
reversed.)
When finding eigenvectors, you need to solve the equation Me e= λ .
Me = λ e
I is the identity
matrix: it is
⇔ Me − λ e = 0 unnecessary in
this line.
⇔ Me − λ Ie = 0
⇔ (M − λ I)e= 0 I is essential here.
Clearly, e = 0 is a solution, but you are seeking a non-zero solution for e.

For non-zero solutions, you require det (M I− λ )= 0. The equation
det (M − λ I )= 0 is known as the characteristic equation of M. The left-
hand side of the characteristic equation is a polynomial in λ; this polynomial
is known as the characteristic polynomial.
Note
The German word for ‘characteristic’ is eigen: eigenvectors are also known
as characteristic vectors; eigenvalues are also known as characteristic
values.
Finding eigenvectors
The following are the steps for finding eigenvectors, illustrated in the next
example.
( I− λ
1 Form the characteristic equation: det M )= 0.
2 Solve the characteristic equation to find the eigenvalues, λ.
3 For each eigenvalue λ find a corresponding eigenvector e by solving

(M − λ I )e= 0.
41
53
Example 2.9
4 2 
Find the eigenvectors of the matrix M =
1 3  .
Solution
( I− λ
1 Form the characteristic equation, det M )= 0.
4 2    4 − λ 2 
M − λI = λ 10
1 3  − 0 1  =  1 3 − λ 
4 −λ 2
( − λ I )= 0 ⇔
So that det M
−λ
=0
13
⇔4 ( − λ )(3− λ )− 2=0
⇔ λ 2 − 7+λ10
= .0
2 Solve the characteristic equation to find the eigenvalues, λ.
λ 2 − 7λ + 10
= 0 ⇔ (λ − 2 )(λ− 5)=
0
⇔ λ = 2 or λ = 5 .
3 For each eigenvalue λ, find a corresponding eigenvector e by solving
(M − λ I )e= 0.
 2 2   x   0  x
(
When λ = 2 : M − λ I )e = 0 ⇔ , where e =
1 1 y  =0  y 
This tells you that if y is any ⇔xy +=0

number, k say, then x is −k.
 −k   −1
⇔e = k .
 k  =  1 
 −1 2   x   0 
( − λ I )e= 0 ⇔
When λ = 5 : M  1 2 − y  = 0 
If y is any number, − x+ y=2 0
say, k, then x is 2k .
 2k   2
⇔ e = k
 k  = 1  .
 − 1  2
Thus the eigenvectors are   and   or any non-zero scalar multiples
1 1
of these vectors.
42
54
 −1  2
Expressing vectors in terms of the eigenvectors e
1 =  1  and e2 = 1 
1
explains why the line y = x attracts and the line y x= − repels under the
2
4 2 
transformation M =
1 3  . Since e 1 and e 2 are non-zero and non-parallel
you can express any position p as α e 1 + β e 2 . Then:
Mp =M e (α 1
+ βe2 )
= α Me1 + β Me 2
= 2 α e1 + 5β e 2
Showing that the image of p is attracted towards the eigenvector with the
numerically larger eigenvalue.
3 × 3 matrices
The definitions of eigenvalue and eigenvector apply to all square matrices.
The characteristic equation of matrix M is det( M −I λ= ) 0. When M is a
2 ×2 matrix the characteristic equation is quadratic, and may or may not
have real roots. When M is a 3 ×3 matrix of real elements the characteristic
equation is cubic, with real coefficients; this must have at least one real root.
This proves that every real 3 3× matrix has at least one real eigenvector,
and so every linear transformation of three-dimensional space has at least
one invariant line. You use the same procedure as before for finding the
eigenvalues and eigenvectors of a 3 3× matrix, although the work will
generally be lengthy.
Example 2.10  2 2−3 

Find the eigenvectors of the matrix M =  1 1 1 .
 1 3 1 − 
Solution
 2 2−3  1 0 0  2 − λ −2 3 
 −  = 
M − λI =  1 1 1  λ 0 1 0   11
−λ 1  ;
1 3 1 −  0 0 1   1 3 1 −− λ
➜
43
55
2−λ −2 3
det (M − λ I ) = 1 1− λ 1
1 3 −1− λ
(2 λ )(1 − λ )(−1− −λ ) 321+(− − −λ+ −1)+3(3 1

=− λ)
= (−2 λ )λ( 2
− 4+)+ ( 2 λ )
= (+2 λ )[− (−λ 2 ) + 1=]2+ −( + −λ )λ( λ

2 2
43 )
= − (+2 λ )(
λ − 1)(λ − 3)
( − λ I )= 0 ⇔
so that det M λ = − 2, λ1 = or λ = 3.
 4 2 −3  x   0 
When λ = − 2: M( I e− λ0 ) = ⇔  131  y  =  0 
    
 131  z   0 
 −+= 
 4 x2 y3z0 A 
⇔  x+ y z+
3= 0 B
 
 x y+ z3+ = 0 C 
A-3B: x − 11y0= ⇒ x y= 11
A-4B: −14 x −
z = ⇒0 z = − 14 y
 11 
The eigenvector corresponding to λ = − 2 is e1 = k  1 
− 
 14 
 1 2 −3  x   0
101  y  = 0
When λ = ,1: M( I e− λ0 ) = ⇔     
132 −  z   0
 x− y z2+ 3= 0 A 
 
⇔  x + z= 0 B 
 +−= 
 x y z3 2 0 C 
2B+C: 3x3+y0= ⇒ =− yx
B: x +z = ⇒0=− zx
 1
The eigenvector corresponding to λ = 1 is e 2 = k−  1 
 
 −1 
44
56
 −1−2 3  x   0
 1 2 −1  y  = 0
( I−eλ0 ) =
When λ = 3, : M ⇔     
 134 −  z   0
 − x− y+z2=3 0 A 
 
⇔  x − 2+y=z 0 B 
 x +y z3− 4= 0 C 
 
A+B: −4+y4=z 0 ⇒ y=z
A-3B: −4+x4=y0 ⇒ x y=
 1
The eigenvector corresponding to λ = 3 is e 3 = k  1 
 
 1
The ideas above also apply to larger square matrices, but if M is n n× , its
characteristic equation is of degree n, and solving polynomial equations of
higher degree is generally not straightforward – Evariste Gallois (1811–32)
proved that there is no general formula for solving polynomial equations
of degree 5 or higher. In practice, eigenvalues are not usually found by
solving characteristic equations! Numerical methods will usually be
applied to matrices using a computer, with consequent problems caused by
approximations and rounding errors.
Exercise 2D 1 Find the eigenvalues and corresponding eigenvectors of these 2 2 ×

CP matrices and check that the sum of the eigenvalues is the sum of the
diagonal elements.
5 3   72  1 2 
(i) (ii) (iii)
2 4  −12
− 4  1 1 
1 1−   1.1 0 −4 .   p 0
(iv) (v) 0.2 0 2 .
 (vi) , p≠
q
1 3  0 q 
CP 2 Find the eigenvalues and corresponding eigenvectors of these 3 3×

matrices and check that the sum of the eigenvalues is the sum of the
diagonal elements.
3 0 0  1 1 2 
(i)
  (ii)
 − 
0 2 1   4 2 3 
 0 0 1 −   4 2 3 
45
57
1 1 2   014 − 
(iii)  −  − − 
5 2 1  (iv)  10 7 20 
 1 1 2   −2 1 2 − 
0 0 2   1 3 3− − 
(v)   (vi) − − 
8 6 3
0 3 0   
 2 0 0   8 2 7− 
 cos 2θ sin 2θ 
3 The 2 2× matrix M = represents a reflection in the
sin 2θ − cos 2θ 
line y = x... > = x tanθ .
Find the eigenvalues and corresponding eigenvectors of M.
4 Vector s is an eigenvector of matrix A, with eigenvalue α, and also
an eigenvector of matrix B, with eigenvalue β. Prove that s is an
eigenvector of
(i) A +B (ii) AB
and find the corresponding eigenvalues.
Tip: s is an eigenvector of M ⇔ Ms = λ s, s ≠ 0.
M 5 Matrix M has eigenvalue λ with corresponding eigenvector s;k is a
non-zero scalar.
Prove that the matrix kM has eigenvalue kλ and that s is a
corresponding eigenvector.
 1  3 1 1− 
   
6 (i) Show that e = − 2  is an eigenvector of A =  −1 5 1 −  and
 1   1 1 3− 
determine the corresponding eigenvalue.
(ii) Find the other two eigenvalues of A and the corresponding
eigenvectors.
7 Matrix M is n n × . For n = 2 and n = 3 prove that if the sum of the
elements in each row of M is 1, then 1 is an eigenvalue of M . (This
property holds for all values of n.)
PS 8 Show that if λ is an eigenvalue of the square matrix M and the
corresponding eigenvector is s, then:
λ 2 is an eigenvalue of M 2
3
λ 3 is an eigenvalue of M
n
λ n is an eigenvalue of M
and even λ −1 is an eigenvalue of M −1 .
Show further that s is the corresponding eigenvector in all cases.
46
58
For
 322 
 4 −1  
(i) M= (ii) M = 1 4 1 
2 1   −2−4−1 
find the eigenvalues of:

(a) M (b) M 2
(c) M 5
(d) M −.1
9 Find the eigenvalues and eigenvectors of the matrix
 25 40 
M=−−
 12 19 
 3
Express the vector s = −
 2 as a linear combination of the eigenvectors
and deduce that
 3
n n
   5
n −2
MM =− 5
 2 =  1 +− 3
10 The 2×2matrix M has real eigenvalues λ 1 , λ 2 and associated

eigenvectors s 1s,2 , where λ 1 > λ 2 . By expressing any vector v in
terms of s 1 and s 2, describe the behaviour of Mn vas n increases when
(i) λ 1 1< (ii) λ 1 1= (iii) λ 1 1> .
M 11 The self-drive camper-van hire firm DIY has depots at Calgary and
Vancouver. The hire period begins on Saturday afternoon, and all vans are
returned (to either depot) the following Saturday morning. Each week:
» all DIY’s vans are hired out
» of the vans hired in Calgary, 50% are returned there, 50% to
Vancouver
» of the vans hired in Vancouver, 70% are returned there, 30% to
Calgary.
(i) One Saturday, the Calgary depot has c vans and the Vancouver
c
depot has v vans. Form matrix M so that the product M
v 
gives the number of vans in each depot the following Saturday.
(ii) At the start of the season each depot has 100 vans. Use matrix
multiplication to find out how many vans will be at each depot
two weeks later.
 c
(iii) Solve the equation Mx x = where x = and explain the
v 
connection with eigenvalues.
(iv) How many vans should DIY stock at Calgary andVancouver if they
want the number of vans available at those depots to remain constant?
47
59
Note
The process described above is an example of a Markov process. The
matrix governing it is a transition or stochastic matrix. Each column of the
transition matrix consists of non-negative elements with a sum of 1.
M 12 At time t, the rabbit and wolf populations (r and w, respectively) on a

certain island are described by the differential equations
 dr
 =− 5r3w
 dt
 ①
 dw = + rw
 dt
r  5 3− 
Throughout this question, p represents
w  and M represents1 1  .
(i) Show that the differential equations may be written as:
dp Mp
= ②
dt
(ii) Show that if p =p 1 )(t and p =p 2 ) (t satisfy ②, then
p =p a t1b( t) + p 2 ( ) also satisfies ②, where a and b are constants.
λ
(iii) Show that if p = e t k satisfies ②, where k is constant, then
Mk = λ k.
(iv) Find the eigenvalues and eigenvectors of M and hence solve
① given that there are 1000 rabbits and 50 wolves at t = 0.
Powers of square matrices.

A point R has position vector r. Pre-multiplying r by a matrix M gives r',
the position vector of R', which is the image of R under the transformation
represented by M. If you apply the same transformation to r' you get r'',
with position vector r '' = M (Mr )M= r 2 . Higher powers of M arise if you
continue to apply the same transformation. In this section, you will learn how
to use eigenvalues and eigenvectors to find powers of matrices.
The two statements The eigenvalues and
eigenvectors of
 4 2   −1  −2   −1
2 4 2 
1 3  1  =  2  =  1 
1 3  were found
 4 2   2   10   2 on page 43.
and 5
1 3 1  = 5  = 1
48
60
can be combined into the single statement:
 4 2   −1 2   −1 2   2 0 
1 3  1 1  =  1 1 0 5 
4 2   −1 2 
=
which you may write as MQ QD where M = =
1 3  , Q  1 1  and
2 0 
D =
0 5  .
In just the same way if any 2 2× matrix has eigenvectors e 1 , e 2,
corresponding to eigenvalues λ 1λ,2 , then Me 1 1=1 λ and
e Me 2 2=2 λ soe that
MQ =QD , where Q e e 1
= ( )
2 , the 2 2
× matrix which has eigenvectors
 λ1 0 
as columns, and D = , a matrix with the corresponding eigenvalues
0 λ 2 
on the leading diagonal and zeros elsewhere.
=
If Q is non-singular, then Q −1 exists, and pre-multiplying MQ QD by
−1 D= ; you then say that M has been reduced to diagonal
Q −1 gives Q MQ
form or that M has been diagonalised.
Although there are square matrices which cannot be reduced to diagonal
form, being able to reduce M to diagonal form D helps if you want to raise
−1
M to a power. Post-multiplying MQ =QD by Q−1 gives M QDQ =
so that M 4
= (
QDQ
−1
)(QDQ QDQ
)( QDQ)(
−1 −1 −1
) This simplification
makes extensive use of
= QD (Q) Q(D
) (Q) Q−1 D Q Q−1 DQ
−1 −1
the associate property
−1
of matrix multiplication
= QD4Q together with properties
of inverse identities.
Similarly: M nQD
= Qn −1 , which can be proved formally by induction.
 λ1 0  n  λ 1n 0  n
Since D = ,D =  and you can evaluate M readily,
0 λ 2   0 λ2 
n
by doing only two matrix multiplications whatever the value of n.

Expressing a matrix M in the form QDQ–1 allows you to find M raised to
any power without having to multiply it out time and again. This is shown in
the next example.
49
61
Example 2.11  
Find M n , where M =  4 2  .
1 3 
Solution
 − 1 2 
You have already seen that   and   are eigenvectors, with eigenvalues
1 1 
You could use
−1 2   2 0  any non-zero
=
2 and 5, respectively, so take Q   and =
D   . multiples of
11  0 5 
 −1  
  and  2 
 1 2 1 1 
−1
− 3 3 n
 2n0 
but these are the
Then Q =  and D =  .
05 n
 simplest.
 1 1
 3 3
−1 2
 −1 2   2 0   3
n
3
Therefore M n = QDn Q −1 =
 1 1 0 5 n  1 1 

3 3
 5 2+n2n 5 2 × −n n +1 
1 2 × .
= 3 +1
 5n2− n nn
5 2+ 
Again the work with 3 3× and other square matrices follows the same
pattern, although the calculations are more complicated.
The Cayley–Hamilton theorem

You have already seen that the characteristic equation for the matrix
4 2  4 −λ 2
=
M  ( − λ I )= 0 ⇔
 is det M =0
1 3  13 −λ
⇔4 ( − λ )(−
3 λ )− 1=0
⇔ 10 7 − +λ=λ 2
0.
 4 2   4 2   18 14 
Notice that M 2 =
1 3 1 3  = 7 11  and that
2
1 0   4 2   18 14   0 0 
10 I7− M +
M = 10   − 7 + =  
0 1   1 3   7 11   0 0 
2
so that 10 I7−M M + = ,0 the zero matrix.
50
62
This illustrates the Cayley−Hamilton theorem which states:
‘Every square matrix M satisfies its own characteristic equation.’
Note that the matrices I and O have to be inserted appropriately so that the
terms in the equation are all compatible.
You can readily prove this result for the general 2 2× matrix by direct
multiplication.
a b 
11
×
For the general 2 2 matrix, M =    . The characteristic equation
 a 2b2 
a1 − λ b1
det (M − λ I )= 0 is written as = 0 or, equivalently, as
a2 b2 − λ
(a1b2 a−b 21 )−(+a1b 2 )λ+λ= 2

0.
You will need to show that:
(a1b2 a−b 2 1 )I− +(a1b 2 )M+ =M 2
0
 a b a−b 0 
12 21
(a1b2 a−b 21 )I =   
 0 a1b2 a−b 21 
 a 2 + ab a b bb + 
(a1b+ 2 )M =  
1 121112
2 
 a1a2 a+b a2 b2 1b2 2 + 
 a1b 1  a1b1   a1a2 + b a 2b1 11 1 12

+b b 
2
M =  
a2b 2 a 2b2  = a1a2 a2 +2b2 a1 2b b + 2
⇒ (a1b2 a−b )  a b a−b − (a12 + ab1 2 ) + a12 + a 2b1 −(+

a1b1 1b2 b1 1a1 b2 )b+(b+ ) 
21
= 1 2 2 1

 2 
I − (+a1b 2 ) M +M 2  −(a1a2 a+b 2 2 ) +
( a1a2 a+b 2 2 ) (a1b2 2a− −
1b a b )b ( 122
+ + 2+) ( ab2 1b2 )
 
= 0 0 
0 0 
The general 2 2× matrix satisfies its own characteristic equation.
Example 2.12
4 2 
=
Given M   use the Cayley–Hamilton theorem to find M 8.
1 3 
Solution
2
As before, the characteristic equation is 10 7− +λ=λ 0 . By the Cayley–
Hamilton theorem:
➜
51
63
2
10 I7−M M + =0
⇒M 2
= 7M − 10 I
⇒M 4
= (7M − 10 I )2
= 49 M 2 − 140 M + 100I
= 49 7( M − 10 I )− 140M + 100 I
= 203 M − 390 I
M 8 = (203M − 390 I)
2
⇒
= 41209 M 2 − 158 340M + 152100 I
= 41209 (7 M − 10 I158
) − 340 152100
M+ I
= 130123 M − 259 990I
4 2   1 0   260502 260 246 

8
Thus M = 130123   − 259990  =  
1 3   0 1   130123 130 379 
2 ×
5 +n2 5 2n
2 × −n n +1 
1 n
(Check this against M =  n  with n = 8;
3  5 2− n 5n2n+
+1

see Example 2.12)
Example 2.13
Use the Cayley–Hamilton theorem to show that:
5 2  +2
M= Mn = 8 M n +1 − 9 M n .
3 3  ⇒
Solution
5 −λ 2
= λ 2 − 8λ .+The
9 characteristic equation is λ λ 2
− 8+ 9= 0.
3 3 − λ
2 2
By the Cayley–Hamilton theorem: M − 8M +
I =9 ⇒
0 M M= I−8 .9
Multiplying by M n gives the required result.
+2
Mn = 8 Mn + 1 − 9 Mn
This expression, giving M n+2 in terms of M n+1 and M n , is an example of a

recurrence relation.
?
› The Cayley–Hamilton theorem states that a matrix M satisfies its own
characteristic equation. The characteristic equation may be written as
det (M − λ I ) = 0. Replacing λ by M produces a determinant consisting
entirely of zeros. Is this sufficient proof of the theorem?
52
64
Historical note
The Cayley–Hamilton theorem was first announced by Arthur Cayley in
‘A Memoir on the Theory of Matrices’ in 1858, in which he proved the theorem
× matrices and checked it for 3 3 × matrices. Amazingly, he went
for 2 2
on to say, ‘I have not thought it necessary to undertake the labour of a
formal proof of the theorem in the general case of a matrix of any degree.’
Essentially the same property was contained in Sir William Hamilton’s
× matrices.
‘Lectures on Quaternions’ in 1853, with a proof covering the 4 4
The name ‘characteristic equation’, is attributed to Augustin Louis Cauchy
(1789–1857), and the first general proof of the theorem was supplied in 1878
by Georg Frobenius (1849–1917), complete with modifications to take account
of the problems caused by repeated eigenvalues.
Exercise 2E
1 =
Find matrices Q and D such that M QDQ −1 .
5 4   7 10−   0.5. 0 5 
(i) M= (ii) M= (iii) M=
3 6  3 4 −  0.3. 0 7 
 1.9 −1.5 −1
2 Express M = in the form QDQ and hence find M 4 .
0.6 0 
What can you say about Mn when n is very large?
CP 3 Find the following
5 10 4
6 6 −   3 −1  0.7. 0 3 
(i) (ii) (iii)
2 −1 −1 3  0.6. 0 4 
4 Find examples of 2 2× matrices to illustrate the following.
(i) M has repeated eigenvalues and cannot be diagonalised.
(ii) M has repeated eigenvalues and can be diagonalised.
(iii) M has 0 as an eigenvalue and cannot be diagonalised.
(iv) M has 0 as an eigenvalue and can be diagonalised.
 12 
5 Demonstrate that
−1 4  satisfies its own characteristic equation.
a c 
6 Prove the Cayley-Hamilton theorem for M = by calculating M 2
b d 
and substituting directly into the characteristic equation.
7 Find the eigenvalues and corresponding eigenvectors of the matrix
 4 1− 1 
A =  −1 0 3 −  .
 
 −
13 0 
Find a non-singular matrix P and a diagonal matrix D such
that A 5 = PDP −1.
9231 Paper 11 Q8 June 2011 53
65
8 The vector e is an eigenvector of the matrix A, with corresponding
eigenvalue l, and is also an eigenvector of the matrix B, with
corresponding eigenvalue m. Show that e is an eigenvector of the matrix
AB with corresponding eigenvalue lm. State the eigenvalues of the
matrix C, where
 − 1−1 3 
C =  0 12 ,
 
 0 02 
and find corresponding eigenvectors.
1
Show that  6  is an eigenvector of the matrix D, where
 
3
 1 1−1 
D =  −6−3 4 ,
 − − 
937
and state the corresponding eigenvalue.

Hence state an eigenvector of the matrix CD and give the
corresponding eigenvalue.
9 The matrix A, where
 100 
 ,
A = 10 7 −10 
 7 5 −8 
has eigenvalues 1 and 3. Find corresponding eigenvectors.
1
It is given that  6  is an eigenvector of A. Find the corresponding
 
eigenvalue. 3
Find a diagonal matrix D and matrices P and P−1 such that P−1 AP = D.
9231 Paper 11 Q6 November 2015
CP 10 The vector e is an eigenvector of the 3×3matrices A and B, with
corresponding eigenvalues λ and µ respectively. Justifying your answer,
state an eigenvalue of A +
B.
The matrix A, where
 6 1−6− 
A = 1 0 2 − ,
 −− 
3 1 3 
54
66
 1  1  2
has eigenvectors  1 ,  − 1 ,  0  . Find the corresponding eigenvalues.
     
1 1 1
The matrix B, where
 8 2−8− 
B = 2 0 4 − ,
 −− 
4 2 4 
 1  1  2
also has eigenvectors  1 ,  −1 ,  0  , for which −2,2 4, , respectively,
     
1 1 1
are corresponding eigenvalues. The matrix M is given by M =A+B−I 5,

where I is the 3 3× identity matrix.
State the eigenvalues of M.
5
Find matrices R and S and a diagonal matrix D such that M RDS = .
[You should show clearly all the elements of the matrices R, S and D.]
11 The square matrix A has λ as an eigenvalue with e as a corresponding
eigenvector. Show that if A is non-singular, then
(i) λ≠0,
−
(ii) the matrix A 1 has λ −1 as an eigenvalue with e as a corresponding
eigenvector.
The 3 ×
3 matrices A and B are given by
 −2 2 4 − 
−
A =  0 1− 5  and B =
A(+I 3 ) 1,
 
 0 0 3 
where I is the 3 ×
3 identity matrix. Find a non-singular matrix P, and a
diagonal matrix D, such that B =PDP −1 .
1  141 − 
   −  and state the
PS 12 Show that 1  is an eigenvector of M =  −1 6 1 
0  2 2 −4 
corresponding eigenvalue. By finding the other eigenvalues and their
eigenvectors express M in the form S SA −1 .
55
67
 2 2 −3 
CP  
13 You are given the matrix =
A 1 1 1 .
1 3 1 − 
(i) Show that 1 is an eigenvalue of A, and find the other two

eigenvalues.
(ii) Find an eigenvector corresponding to the eigenvalue 1.
(iii) Using the Cayley–Hamilton theorem, or otherwise, find integers
p, q and r such that
A 3 = pA 2 + q+Ar ,I
and show that
A 4 = 9 A 2 + 4−A
12 I.
M 14 The Fibonacci sequence 1, 1, 2, 3, … is defined by
f 1f2= = 1, fn +1 = +
ff −1
nn.
f + 
Let u n =  n 1  
 fn 
1 1 
(i) Show that u n 1 Mu n, where M =  .
+= 1 0 
1
(ii) Show that the eigenvalues of M are λ 1 = + 1 5 ,
2
( )
1 λ λ  
( )
λ 2 = 1 − 5 , with associated eigenvectors  1 2 , 1  .
2 1  
1  λ 1λ2   λ 1 0   1 −λ 2 
(iii) Deduce that M =    and
λ 1λ− 2 1 1  0 λ 2   −1 λ 1 
 2 2 
1  1 +5  1 −5
hence show that =f n   −   
5  2   2  
56
68
KEY POINTS
aaa 
 123 
1 The determinant of a 3 × 3 matrix M =  b1b2 3b  is given by
ccc 
 123 
b2b3 b1b3 b1b2

det M = a 1 − a2 + a3 .
c 2c 3 c 1c3 c 1c2
2 The determinant of a 3 × 3 matrix represents the volume scale factor
of the transformation represented by the matrix.
a a a 
123
3 
For a 3 × 3 matrix b1b2 3b  the minor of an element is formed
 
 c c c 
123
by crossing out the row and column containing that element and
finding the determinant of the resulting 2 × 2 matrix.
b1b3
e.g. minor of a 2 =
c 1c3
4 A minor, together with its correct sign, given by the matrix
+−+
−+− is known as a cofactor and is denoted by the
+−+
corresponding capital letter;
b1c1
e.g. the cofactor of a 2 = A2 = −
b3c3
 a1a2 a3 
5 The inverse of a 3 × 3 matrix M =  b1b2 b3 
 can
 c c c 
123
be found using the formula

 A1B1 1C 
1 1   , det M ≠ 0
adjM =  A2B2 2C 
detM detM  
 A3B3 3C
ABC 
1 1 1
 
The matrix  A2B C2 2  is the adjoint or adjugate matrix, denoted
ABC 
 3 3 3
by adjM, formed by replacing each element of M by its cofactor and then

transposing (i.e. changing rows into columns and columns into rows).
57
69
6 (MN)–1 = N–1 M–1
7 If the determinant of a matrix is non-zero, the matrix is said to be
non-singular. If the determinant of a matrix is zero, then the matrix is
singular.
8 If A is a non-singular matrix, AA −1 = A −A
1
I= .
9 If the determinant of a matrix representing a transformation is zero, all
the points are mapped to either a straight line (in two dimensions) or
to a plane (in three dimensions).
10 A set of simultaneous equations can be written as Mx r = . Their
= r −1 provided that det M ≠ 0 (ie M is not
solution is given by x M
singular).
11 Three distinct planes in three dimensions will be arranged in one of
five ways.
● They meet in a unique point of intersection.
● All three planes are parallel and therefore do not meet.
● Two of the planes are parallel, and these are cut by the third plane
to form two parallel lines.
●The planes form a sheaf of planes that intersect in a common line.
● The planes form a prism of planes in which each pair of planes
meet in a straight line but there are no common points of
intersection between the three lines.
12 Three distinct planes
a1x b+ y1c z+ d 1 = 1
a 2x b+y 2c z+d 2 = 2
a 3x b+ y 3c z+d 3 = 3
can be expressed in the form
 x   d1 
   
M  y  =  d2 
 z   d 
3
 a1b c 11

 
where M =  a 2b c 22 .
 a b c 
3 33
If M is non-singular, the unique point of intersection is given by

 d1 
 
M–1 d 2  .
 d 
3
58
70
Otherwise, the planes meet in one of the other four possible
arrangements. In case of a sheaf of planes, the equations have an infinite
number of possible solutions, and in the other three cases the equations
have no solutions.
13 An eigenvector of a square matrix M is a non-zero vector e such that
Me = λ e; the scalar λ is the corresponding eigenvalue.
14 The characteristic equation of M is det M( I − λ )= 0.
15 If Q is the matrix formed of the eigenvectors of M and D is the
diagonal matrix formed of the corresponding eigenvalues then:
MQ =QD ,
−1
D = Q MQ ,
Mn = QDnQ −1 .
A power of a square matrix can thus be found by multiplying the
matrix of eigenvectors by the power of the diagonal matrix formed of
the eigenvalues and by the inverse of the eigenvectors matrix.
16 The Cayley–Hamilton theorem states:
‘Every square matrix M satisfies its own characteristic equation.’
LEARNING OUTCOMES
■ formulate a problem involving the solution of 3 linear simultaneous
equations in 3 unknowns as a problem involving the solution of a
matrix equation
■ understand the cases that may arise concerning the consistency or
inconsistency of 3 linear simultaneous equations, relate them to the
singularity or otherwise of the corresponding matrix
■ solve consistent systems
■ interpret results geometrically in terms of lines and planes
■ understand the terms ‘characteristic equation’ ,‘eigenvalue’ and
‘eigenvector’, as applied to square matrices;
■ find eigenvalues and eigenvectors of 2 2× and 3 3 × matrices;
■ express a square matrix in the form QDQ −1 , where D is a diagonal
matrix of eigenvalues and Q is a matrix whose columns are
eigenvectors, and use this expression
■ use the fact that a square matrix satisfies its own characteristic equation.
59
71
3 Differentiation
If I feel
unhappy, I do
mathematics
to become
happy. If I feel
happy, I do
mathematics
to keep happy.
Alfred Renyi
(1921−1970)
The four photographs above were taken over a period of time. They build up
to the final picture which shows the complete situation.
Later in this chapter you will meet a similar idea building up a polynomial
series to represent a function. Instead of showing how a stadium developed
over time as new parts were added to it, you will be expressing a function
as a series of ever increasing accuracy by adding on successive terms of a
polynomial.
y
y=x x3 x5
1 y=x−+
6 120
x3
y=x−
6 y = sin x
x
0 –π2 π
▲ Figure 3.1
60
72
?
› Look at Figure 3.1. Can sin x be represented as a polynomial for all
values of x?
3.1 Differentiating inverse

trigonometric functions
In Pure Mathematics 3, you will have seen that the derivative of tan –1x is given
by 1 2 . The derivatives of sin –1x and cos –1x can be found in a similar way,
1+ x
using implicit differentiation.
y
y = sin–1 x
2
Notice that the gradient of the curve
is always positive for −1 < x < 1.
–1 0 1 x
–
2
▲ Figure 3.2
−
y = sin 1 x
sin y= x
dy = Differentiate implicitly
cos y 1
dx with respect to x.
dy = 1
dx cos y
=±− 1
2
1 sin y
=±− 1
1 x2
− π π
But y = sin 1 x has a range of − 2 ഛy ഛ 2 , which implies thatcos x0 ≥ 0, and
so you can ignore the ± symbol since it must be positive in this case.
The conclusion is that:
d (sin −1 x) = − 1
dx 1 x2
There are several things to notice with this result:
» it is positive, and only defined for − 1< < x 1
» it has a minimum at x 0=
» it tends to ∞ as x → ± 1
and these points are consistent with the graph of y = sin−1 x in Figure 3.2.
61
73
ACTIVITY 3.1
Use a similar method to show that:
(i) xd (cos −1 x) = − − 1 2
d 1 x
(ii) dx (tan −1 x) = + 1 2
d 1 x
3.2 Differentiating hyperbolic functions

Differentiating shows that Note
d (cosh x) = d e x + e x  = e xe −
− −x
= sinh x Notice that these
dx dx  2  2 results are similar, but
and not quite the same,
as the corresponding
d (sinh x) = d e − e  = e e +
x −x x −x
= cosh x results for the circular
dx dx  2  2 functions.
Example 3.1 Use the identity tanh x = sinh x to find the derivative of tanh x
cosh x
Solution
Using the quotient rule
d (
dx
tanh x) = ()
dd (sinh x ) = cosh
tanh x
ddxx cosh x
d sinh
()
xcoshx x= −cosh
sinhxcosh
dx cosh x(cosh x)2
xsinh x − sinh x x
x sinh
(cosh x)2
2 2 2 2
cosh x − sinh cosh
x x − 1sinh
= 2
= = 2 2
=x sech
= 2 x1 2 = sech 2 x
cosh x cosh
coshx x cosh x
3.3 Differentiating inverse hyperbolic

functions
It is possible to differentiate the logarithmic versions of cosh−1 x and sinh−1 x, but
it is easier to work with the hyperbolic functions themselves and their identities:
−
y = cosh 1 x
cosh y x=
dy
sinh y =1 Differentiating implicitly.
dx
dy
= 1
dx sinh y
1 Using the identity
=±
cosh 2 y − 1 cosh 2 y − sinh 2 y ≡ 1.
=± 1
62 2
x −1
74
Since the gradient of y = cosh−1 x is always positive you must take the
positive square root, and so:
d (
cosh x) = − 2
−1 1
dx x 1
ACTIVITY 3.2
Use y = sinh−1 x to show that
d (
sinh x ) = + 2
−1 1
dx x 1
You should recognise a similarity with the inverse trigonometric results in
Section 3.1.
Exercise 3A
1 State the domain and range of the sin−1 , cos−1 and tan−1 functions.
2 Differentiate the following functions with respect to x:
(i) sin−1 (3x) (ii) cos
−1
(21 )x
(iii) tan−1 (5x) ()
(iv) sin −31 2 x
3 tan 1( − x )
−1 2
(v) tan−1 (e x) (vi)
3 Use the chain rule or product rule (as appropriate) to differentiate the
following functions, with respect to x:
(i) sinh 4x (ii) cosh x 2
(iii)cosh 2 x (iv) cosh x sinh x
CP 4 Find the exact coordinate of the turning point of the graph y = e 2x sinh x.
5 Differentiate the following:
(i) sinh−1 3x (ii) cosh−1 4x
(iii) sinh−1 (2x 2) (iv) cosh−1 (2x + 1)
CP 6 (i) (
Use the chain rule to differentiate xlnx + 2
−1. ) 1
Show, by multiplying the numerator and denominator by(x ) − ,1
2 2
(ii)
1
that your answer simplifies to x 2 − .
1
(
(iii) Differentiatelnx + x 2 + to
1 )
1 show that the derivative of sinh−1 x
is x 2 .
+1
CP 7 Use implicit differentiation to show that:
(i)
d
dx ()− x
sinh 1 = +
a
1
x2 a2
(ii) d
dx () −1 x
cosh = −
a
1
x 2 a2
63
75
CP 8 (i) Prove that d (tanh − 1 x) = − 1
dx 1 x2 .
(ii) By using partial fractions and integrating, deduce the logarithmic
form of tanh -1 x.
CP 9 Use implicit differentiation to prove:
(i)
dx ()
d sin −1 x
a= − a 2
1
x 2
(ii)
dx ()
d tan −1 x
=
a + 22
ax
a
()=
10 Differentiate the function xf x sin (x ) .
−1 2
3.4 Finding the second derivative

for relations given implicitly
In Pure Mathematics 2 and 3, you learned how to differentiate a function
defined implicitly. The second derivative can be found in a similar way.
Example 3.2 d y
2
d
(i) Find an expression for for the relation
dx 2 in terms of y and y dx
2 2 2
defined by x + y = ,r where r is a constant.
d 2y x2 + y2
(ii) Show that 2 = − 3 .
dx y
Solution
(i) Differentiating both sides of the equation with respect to x gives:
dy =
2 x + 2y 0
dx
Differentiating again with respect to x, using the product rule gives:
dy   d y  d2y
2 +  2    + 2y 2 = 0
dx dx dx
() ()
2 2
dy dy
2 2 2+ 1+
d y =− dx dx
2
=−
dx 2y y
+ dy = , dy = − x
(ii) Since x2 2y 0 dx y
dx
2
1+ 2 x
2
⇒ d y2 = − y
2
dx y Notice that d y 1
dx 2 ≠ d 2x
x 2
+ y2 dy2
=−
y3
64
76
Example 3.3 Figure 3.3 shows the graph of the relation between x and y given by
x 3 + y 3 = 3xy .
y
2
A
−2 −1 0 1 2 x
−1
▲ Figure 3.3
(i) (
Verify that the point A 2,
2
3
4 3 ) is a stationary point on the curve.
d y
(ii) Use the value of
dx 2 to prove that A is a maximum point.
(iii) Explain why substituting x = 0 and y = 0 into the equations will not
dy d2y
find values for and
dx dx 2 at the origin O.
Solution
(i) Substituting x = 3
2, y = 3
4 into the original equation gives:
(3
)2 ( +) 3(4)( =) 3 23 4
3 3
3
2 4+ 3= × ( )83
=6
So A lies on the curve.

Differentiating both sides with respect to x gives:
dy =  d y +  = dy +
3x 2 + 3y 2 3 x y 3x 3y
dx  dx  dx
Substituting x = 3
2, y = 3
4 gives
dy = 3 dy + 3
3 (23 ) ( +) 3 4 3 (2 ) 3 (4 )
2 2
3
dx dx
(3 (4 ) ( −) 3 2 )
3
2
3 dy = 3 − 3 2 = 3 −
dx
3 (4 )
3 2( ) ( ) ( ) 3 4 3 4 0 3
=
It is not necessary So A is a stationary point on the curve.

to rearrange
this to find an (ii) Differentiating x3 2 + 3y 2 dy = 3x d y + 3y with respect to x gives:
dx dx
expression for
 dy 2 2   dy 
()
2
2 d y d y d y
d 2y 6x + 3 2 y +y 2  = 3
+x 2 + 
 d x dx   d x dx dx 
dx 2 . ➜
65
77
dy = gives:
Substituting x = 3
2, y = 3
4, 0
dx
d2y 3
3 2 ( 16 − 3 2) = − 6 (23 )
dx
2
d y = −2 32 <0
dx 2
( 16 − 2)
3 3
So A is a maximum point on the curve.

dy
(iii) Substituting x = 0 and y = 0 into the expression for
dx
dy = × dy + × from which y d cannot be
3 0× 3+ 0× 30 30
dx dx dx
evaluated.
dy = gives
Similarly, substituting x = 0, y = 0, 0
2 2 dx
d y d y
0 × 2 = so 0
dx dx 2 cannot be determined from this expression.
Looking at the graph shows that the curve passes through the origin
twice with different gradients and rate of change of gradient.
ACTIVITY 3.3
Use graphing software to draw the curve given by x y3 3xy+= 3 .
Examine the gradient of a point on the curve as the point moves towards
zero on the two parts of the curve that cross at the origin.
Example 3.4 Show that the function given by

x
e y = cos x satisfies the differential 2
1
d 2y dy +
equation 2 2y = 0
dx 2 + d x 0
–2 –1 1 2345
–1
▲ Figure 3.4
Solution
Differentiating ye x = cos x implicitly twice gives:
x dy + x = −
e e y sin x
dx
d2y
x x dy
+ e x dy + e x y = − cos x
e e
dx 2 + dx dx
x
From the original equation, − cos x = − e y
2
d y
x x dy
+ ex y = − ex y
e 2e
dx 2 + dx
66
78
2
d y
x x dy
+ 2e x y = 0
Giving e 2e You will see in chapter 6 how
dx 2 + dx to solve differential equations
of this type.
e x ≠ for
0 all values of x
d 2y dy +
So 2 2y = 0
dx 2 + d x
ACTIVITY 3.4
−x
Rewrite the original function in the form y = e cos x and differentiate
twice. Show by substitution that the function satisfies the differential
equation.
3.5 Finding the second derivative

for relations given parametrically
In Pure Mathematics 2 and 3, you learned how to differentiate a function defined
parametrically. The second derivative can be found by differentiating the
−
xy yx
result implicitly or by using the formula . A proof of this, and an
(x) 3
example of its use, can be found in Example 3.6.
Example 3.5 A curve is defined by the parametric y

6a
equations x at= y 2 , = 2at where a is a 5a
d 2y 4a
3a
constant. Find an expression for
dx 2 in 2a
terms of a and t. 1a
–a x
–1a 1a 2a 3a 4a 5a 6a 7a 8a 9a
–2a
–3a
–4a
–5a
–6a
▲ Figure 3.5
Solution
Differentiating each equation with respect to t gives:
dx = 2at, dy = 2a
dt dt
dy = dy × dt = 2 a = 1
Using the chain rule
dx d t dx 2at t
Differentiating with respect to x gives
2
d y = − −2 d t
2
( )1 t
dx dx
= − ×12 1 =− 1
3
t 2at 2at
67
79
ACTIVITY 3.5
Find the Cartesian equation of the curve with parametric equations
2
x at= y , = 2at . Use implicit differentiation to check the answers to the
previous example.
Example 3.6 d x d 2
(i) Using the notation x = and x =x2 show that, for a curve given
dt dt
dy = and
y y d 2
−
xy xy
parametrically, 2
=
dx x dx (x) 3
(ii) Use the formula to find the second derivative for the relation given by
x = sin t, y = sin 2t
Solution
(i) Using the chain rule
dy
dy = dy × dt = dt
= y
dx
dx dt dx dt x
d 2 y = d  d y  =
dx
2
d x d x 
d y
dx x ()
−
= xy xy2 × dt
(x) dx
−
= xy xy2 × 1
(x) x
−
= xy xy3
(x)
(ii) x =cos t , y =2cos 2t
x =sin−t , y 4= sin
− 2t
dy = =y 2cos 2t
dx x cos t
2
d y = xy xy −
dx 2 (x) 3
(− ) −( − sin t2cos
) 2 t
= cos t4sin 2 t
(cos
) t3
= − 4 cost sin 2t 2+sin cost 2 t

) t3
(cos
Exercise 3B d2y
1 Find for the relations given implicitly by the following:
dx 2
x + y = 16
4 4
(i) (ii) x2 y2+xy
−= 45
y
(iii) x3 + y 3 = 4 xy (iv) xe = +x 3
68
(v) sin y + cos y = 2x
80
2
d y
2 Find for the relations given parametrically by the following:
dx 2
(i) x = 3cos θ, y = 2sin θ (ii) x = t y4t , =1
(iii) x = cos 2θ , y = sin θ 2 1
(iv) x = t y t ,t = +
t
(v) x = e , y = sin t
3 (i) Find the stationary point on the curve given by x y2 3xy−= 2 .
2
d y
(ii) Use to determine the nature of the stationary point.
dx 2
4 The parametric equations of a curve are x = cosh t, y = sinh t
2
dy d y
(i) Find and in terms of t.
dx dx 2
(ii) Show that the curve has no stationary points.
(iii) Find the Cartesian equation of the curve.
5 The diagram shows part of the graph of sin xcos +21 y =
y
x
0 p 2p 3p
–p
(i) Explain why the curve is not defined for π < <x π. 2
2
dy d y
(ii) Find and 2
dx dx
(iii) Find the stationary points on the section of the curve that passes
through the origin.
d2y
(iv) Use to prove that one of the stationary points is a maximum
dx 2
point and one is a minimum.
1
6 The curve C has parametric equations x =t y 2 , = −(2), fort 0  t  2.

2
2
d y
Find
dx 2 in terms of t.
9231 Paper 11 Q12 (part question) June 2014
3
7 The curve C is defined parametrically by x = 2cos t and y = 2 sin3 t ,
CP for 0 < <t π. 12
Show that, at the point with parameter t,
d2y = 1 4
sec cosec
t .t
dx 2 6
69
81
CP 8 The parametric equations of a curve are x = + θ sin θ, y = − 1 cos θ
(i) Show that yd = tan 1 θ
dx 2 2 tan 12 θ so θ = + 2 tan 12 θ
2 tanθ = − sin
d y 1 tan2 12 θ 1 tan2 12 θ
(ii) Find
dx 2 in terms of q. 1 − tan2 12 θ
CP 9 A function is defined by and cosθ =
t
1 + tan2 12 θ
x ct os = .eBy differentiating
implicitly, show that
2
d x 2 tantdx − 2x = 0
dt2 − dt
CP 10 The curve C has equation x3 y+ =3 3xy , for x > 0 and y > 0. Find a
d =.
relationship between x and y when y 0
dx
Find the exact coordinates of the turning point of C, and determine the
nature of this turning point.
9231 paper 13 Q10 November 2012
3.6 Polynomial approximations and

Maclaurin series
Since polynomial functions are easy to evaluate, y
differentiate and integrate (among other things), 2 f(x) = e x
they are often useful approximations to more
complicated functions. Here you will see one way
of building these approximations, starting with the 1
p(x) = 1
example of the exponential function:f x( ) = .e
x
To build up a polynomial function p(x) which

−1 0 1 2 x
approximates f (x), start by making sure it has the
correct value at x = 0 (where the graph cuts the
y−axis). ▲ Figure 3.6
Since e 0 = 1, the first term of the polynomial
approximation is 1, and so xp ( ) = +1 ... y
(see Figure 3.6). 2

p(x) = x + 1
You can certainly find a better approximation
than this. The next step is to consider the
1
gradient of your approximation. The derivative
f(x) = e x
of e x is e x , and so at x 0= the gradient is 1.
The next term of p(x) will be a multiple of x −1 0 1 2 x
and, since its derivative is 1, this will be x itself.
So xp ( ) = +1 + (Figure
x ... 3.7). ▲ Figure 3.7
This is a better approximation, but, again, can be improved further. The next
step is to ensure that the second derivatives of e x and p(x) have the same
70
values when x = 0. The second derivative of xf ( ) = e x at x 0= is also 1. The
82
next term of p(x) will be a multiple 1 2
p(x) = 1 + x + x
2
of x². Since the second derivative of y
x² is 2, the quadratic term must be
1 2
2 x 2 to give a second derivative of
2
1. So xp( ) 1= + + x 12 x + ,...and
the graph of this function is a much f(x) = e x
1
better approximation to y = f (x)
(Figure 3.8).
−1 0 1 2 x
▲ Figure 3.8
ACTIVITY 3.6
(i) Extend this method one more step and show that the cubic
approximation to xf ( ) = ise
x
1 2 + 1 x3
p (x) = +1 + x x
2 6
(ii) Extend the method two further steps to find a degree 5 polynomial
approximation for f (x).
You will see Figure 3.9 that, for positive values of x around x = 0, the
accuracy of the approximation improves as you add more terms. This is also
true for negative x−values, although it is not so clear from the diagram.
y
y=ex
ACTIVITY 3.7
20 cubic
Write down the cubic
quadratic approximation to e x and
15
substitute (−x) in place
of x to generate a cubic
10 approximation to e−x .
Multiply these two cubics
5 linear together and comment
on your answer.
constant
−5 0 5 x
▲ Figure 3.9
Example 3.7
(i) Express e x as the sum of an infinite series.
(ii) Investigate whether the terms of the series converge for all values of x.
➜
71
83
Solution
(i)
x
e = +1 + x 1 x 2 + 1 x 3 + 1 x 4 + 1 x 5 + 1 x 6 + ...
2! 3! 4! 5! 6!
(ii) When x 2= ,
e 2 = +1 +
2×1 + × + ×2+2 × +1× + 2 3 1 2 4 1 25 1 2 6...
2! 3! 4! 5! 6!
In the 5th term, 4! is greater than 2 4 so this term is less than 1. After this,
the terms continue to get smaller, so it appears that the terms converge.
If you take any value of x, at some point the value of r ! will be greater
than x r , so after that, the terms are less than 1 and will continue to get
smaller. So it seems that the terms converge for all values of x.
?
› Is it always the case that if the terms of a series converge, the sum of
the series also converges?
In Example 3.7 you saw that the terms of the series for e x converge, whatever
the value of x. In fact, the sum of the terms also converges for all values of x.
The formal proof of the convergence is beyond the scope of this book.
ACTIVITY 3.8
This spreadsheet shows the start of a method for approximating e x.
A B
1 0.5 =SUM(B3:B10)
2
3 0 1
4 =A3+1 =$A$1*B3/A4
Copy these cells into your own spreadsheet. Copy and paste the formulae
in Row 4 down to Row 10. Check the formula does indeed behave as
explained above.
The number in cell A1 is the value of x, try changing it and watch the
effect on the other cells. In particular, explain why the value in cell
B1 is an approximation to e x , and state the order of the polynomial
approximation this example achieves.
Change your spreadsheet so that it gives a polynomial approximation up to
the term in x 10 . Use it to calculate the value, and the percentage error, of
10 2
the Maclaurin approximation up to the term in x , for the calculation e .
Note
Notice the $ signs in the formula in cell B4: these indicate an absolute
reference, meaning this part of the formula will always refer to cell A1,
whereas the other parts of the formula are relative references, meaning
they will change relative to where the formula is copied.
72
84
Maclaurin approximations and series
In general, you can find a polynomial p(x), of order n, for any function f (x),
for which its first n derivatives at x 0= exist.
To ensure that p(x) takes the same value as f (x) for each derivative, at x 0= ,
then you need:
p (0) = f (0 )
p′ (0=)′ f 0( )
p′′0( ) = ′′f 0( ) A third derivative f ″′(x) can be
(3) (3) written as f (3)(x), and similarly
p 0( ) = f 0 ( ) for higher derivatives.
...
p(n0) ( ) = f ((0)
n)
This is a list of n + 1 conditions, and you can see the general nth order
polynomial has n + 1 constants to determine:
p (x) = +a 0a x a1 x +a x2 2
+ 3
3
+ … + ar x r +… + anx n
Substituting in x = 0 immediately gives:
a f (0)
= 0
Doing this with the derivatives of p gives the various values of a r :
− −
p′ (x) = +a1 2a2x + 3a 3x 2 +… + ra r x r 1 + … + nanx n 1
so
a f (0)
1 =′
Then
− n −2
p′′ (=x) 2a 2 + 6a3x + … + r−(r ) 1 ar x r 2 + … + n−(n )a x1 n
2a 2 = ′′f (0)
a 2 = 1 f ′′(0)
2
and
− n −3
p(3)0 ( ) = 6a 3 +… + r (r − 1)(r − 2) ar x r 3 + … + n−(n−)(n a)1x2 n
6a 3 = f (3)
(0)
a 3 = 61 f ((0)
3)
In general, the nth derivative gives the condition

n !a n = f ((0)
n)
so
an = n1! f ((0)
n)
73
85
Putting all these back into the p (x) definition to find the approximation for
f (x) gives:
f ′′0( ) 2 f (03) ( ) 3 f
(r )
0 () r f
(n)
0() n
(
f x ) ≈ (
f0f0) + ′ ( ) x + x + x +… + x + … + x .
2! 3! r! n!
This is the Maclaurin approximation or Maclaurin expansion for f (x)
up to the term in x n.
The accuracy of the approximation is usually defined as follows:
=
error approximate − value
value exact
−
approximate value exact value
percentage error = × 100%
exact value
1
Example 3.8 Find the Maclaurin expansion for (1 − x)− , as far as x n .
Solution
Let f (x) = (1 − x)−1
f (x) = (1 − x)−1 f (0) = 1 It is useful

to tabulate
f ′(=
x) −( 1 x)
−2
f ′0(=1) the function
−3 f (x) and its
f ′′(x=) 21( − x) f ′′0(=2) derivatives,
−4 (3) then evaluate
f (3) (x) = 61( − x) f 0 (6 ) =
them at x = 0.
f
(4)
(x) = 24 (1 − x)−5 f (04) (24) =
(n)
f (x) = n !1( − x)
−(n+)1
f (0n) ( ) = n !
So
(1 − x)−1 ≈ +1 + x 2 x 2 + 6 x 3 + 24 x 4 + … + n ! x n
2! 3! 4! n!
2 34 n
= +1 + x+ x+ x+…x + x
ACTIVITY 3.9
Compare the result from Example 3.8 with the binomial expansion for
(1 − x)−1(see Chapter 7 in Pure Mathematics 2) and the sum to infinity of
2
the geometric series 1 + +x x x3
+ +…
If the function and all its derivatives exist at x 0= then, of course, this
expansion could be continued indefinitely, in which case you would get an
infinite order polynomial:
f ′′0( ) 2 f (03) ( ) 3 (r )
f 0( ) r
f (x) = f (0 )f +
0 ′ ( )x + x + x +… + x +…
2! 3! r!
74
86
This is the Maclaurin series or expansion for f (x). Care must be taken with
infinite polynomials like this, but if the sum of the series up to and including
the term in x n tends to a limit as n tends to infinity, and this limit is f (x), then
you can say that the expansion converges to f (x).
Note
A Maclaurin expansion for a function involves a finite number of terms
and is an approximation to the function. A Maclaurin series is the sum of an
infinite number of terms.
Validity of Maclaurin series

1
Example 3.8 showed that the nth Maclaurin expansion for (1)− x − is the
2
geometric series 1 + +x +x +… +x.3 If you letx nn tend to infinity this
would become the sum to infinity, but you will already know that this sum
only converges if x| | <1 . This is an example of a Maclaurin series which only
converges for a limited range of x values – these are described as the values
for which the series is valid.
Other Maclaurin series are valid for different ranges of x. As you saw in
x
Example 3.7, the Maclaurin series for e
e x ≡ +1 + x 1 x 2 + 1 x 3 + … + 1 x r +…
2! 3! r!
is valid for all values of x.
A Maclaurin series may be regarded as incomplete without a statement of the
values of x for which it is valid. However, if there is no such statement it may
be taken that it is valid for all x.
Example 3.9 Find the first three non-zero terms of the Maclaurin series for sin x.
Solution
Note
=
Let xf ( ) sin x.
All the odd
=
f (x) sin x f (0) 0= derivatives are
zero at x = 0. The
f ′( x=) cos x f ′(0)
= 1 even derivatives
alternate between
f ′′( )x= − sin x f ′′(0)= 0 1 and −1 at x = 0.
f (3)
( ) x = − cos x f (3)
( ) x = −1
( ) xsin=
f (4) x (0) 0 =
f (4)
(5) (5)
f ( ) xcos= x f (0) 1 =
… …
➜
75
87
f ′′0( ) 2 f (03) ( ) 3
Using f (x) = f (0 )f +
0 ′ ( )x +
2!
x +
3!
x +…
gives
sin x x= − 1 x 3 + 1 x 5 −…
3! 5!
Note
Note that for this series, x is measured in radians, as the rules for
differentiating sin xand cos xrequire x to be measured in radians.
Example 3.10 2
(i) Show that the derivative y ' of y = tan x can be written as y ' 1= + . y
(5)
(ii) Find similar expressions for derivatives as far as the fifth derivative y .
(iii) Hence find the first three non−zero terms of the Maclaurin series for
y = tan x.
Solution
2 2
(i) =
y ' sec x = +1 tan x = +1 y2
(ii) Implicit differentiation gives:
y '' 2= ' yy
2
y ''' 2= ' ' y2 y'' 2+ ' 2yy'' = (y) + yy
y (4) = 4 y' ''y 2 +' '' (y y yy+ ''')6=' '' y2 y''' yy+
y (5) = 6 (''y ''y' y'''+y2 ' ''' ) + (y y yy+ (4)
) =( )6 y′′ + ′ ′′′8y+y yy 2
2 (4)
(iii) Evaluating the derivatives at zero gives:

y (0) = 0
y '(0) 1=
y ''(0) = ×2 ×0 =1 0
y ''' 0 =2 ×1 +2 ×02=2
y (4)0 ( ) = ×6 ×1 +0 ×2 ×0 =2 0
y (5)0 ( ) = ×6 +0 ×8 2×1 +2 ×2 ×0 =
0 16
1 3 1 5
So tan x = ×1 + ×
x 2 3! x
+ 16
×
5! x
+ ...
= +x 1 x3 2 x 5...+
+ 15
3
76
88
Exercise 3C
1 (i) Show that the Maclaurin series for cos xis:
2 1 4 1 6 (−)1 r x 2 r
cos x = −1 1 x + x − x + … + +…
2! 4! 6! (2)r!
(ii) Use the first three terms of the series to calculate an approximate
value for cos (0.1).
(iii) Use your calculator to find cos (0.1) and find the percentage error
in your answer to (ii).
2 The Maclaurin series for e x is 1 + +x 1 x 2 + 1 x 3 +… + 1 x r + …
2! 3! r!
1
(i) Use this to calculate correct to five decimal places, stating
e
how many terms you need to use to be sure.
(ii) Use the ex function on your calculator to calculate the percentage
error in this approximation.
3 (i) Write down the cubic approximation to sin x.
2
(ii) Use this approximation to rewrite the equation sin x x= as a
polynomial equation.
(iii) By solving this polynomial equation, show that an approximate
root of the original equation is
x = 15 3−
4 (i) Explain why it is not possible to find Maclaurin expansions for
f (x) = ln x.
(ii) (a) Show that the Maclaurin series for ln (1)+ is
x
2
x 3
x 4
(−)1 n +1 x n
x x− + − +… + +…
2 3 4 n
(b) This series is valid for − <
1≤ x1 1 only; by drawing graphs of
y = ln (1 + and
x) several successive approximations show that
this is plausible.
5 The third Maclaurin approximation to a function isfx( ) = −1 +3 x 2x 5 3
.
2 2
( )
Write down the values of f 0′(, )f 0 ′′(and
) f 0 ( ).
3
Sketch the graph of y = f (x) near x 0= .

6 Find the Maclaurin expansion for xf ( ) = tan x up to the term in x 3.
7 An approximate rule used by builders to find the length, c, of a circular
arc ABC is
−
c = 8b a , B
3
where a and b are as shown in the
diagram. b
(i) If O is the centre of the circle, A a

C
θ
show that b = 2rsin 2 and ()
a = 2rsin (θ). θ θ
O
77
89
(ii) Using the cubic approximation to sin x, show that b 8a r− = θ . 6
Hence verify the rule.
(iii) Find the percentage error caused by using this rule when θ = π .
3
8 A curve passes through the point (0,);2its gradient is given by the
dy
differential equation dx = −1 . Assume
xy that the equation of this curve
can be expressed as the Maclaurin series
2 3 4
y a=a+x0 a x 1 + 2 + a 3x a +x 4 +…
(i) Find a 0 .
(ii) Show that
a1 + 2a 2x + 3a 3x 2 + 4a4x 3 +… = −1−2 x a x 1a x2 −a x−2 − 3… 3
4
Equate coefficients to find the first seven terms of the Maclaurin

(iii)
series.
(iv) Draw graphs to compare the solution given by these seven terms
with a solution generated (step by step) on a computer.
9 Find the Maclaurin series for cosh x including the general term:
(i) by finding the values of successive derivatives at x = 0
(ii) by using the definition cosh x = 12 (e e + − . )
x x
(The Maclaurin series for cosh x and sinh x are valid for all values of
x, like the ones for cos x and sin x.)
10 (i) Write down the Maclaurin expansions of
(a) cosθ (b) sinθ θ +
(c) cos isin θ
giving your answers in ascending powers of q.
(ii) Substitute x i= θ in the Maclaurin series for e x and simplify the
terms.
(iii) Show that your answers to (i)(c) and (ii) are the same, and hence
θ
e i = cos θ + i sinθ
(See page 137 in Chapter 5 Complex numbers.)
(n) (n)
PS 11 In this question y n and a n are used to denotef (x) and f 0 ( )
respectively.
−1
(i) Let fx( ) = sin x .
Show that (1 − x y) 1 = and (1 − x 2y) xy2 − = .1 0
2 2
1
(ii) Find a 1 and a 2 .
Prove by induction that (1 − x y) (2n1xy )n y n +1 − =2
2
(iii) n +2
−+ n 0 , and
deduce that a n + 2 = n 2a n .
1 −
(iv) Find the Maclaurin expansion of sin x , giving the first three non−
zero terms and the general term.
78
90
3.7 Using Maclaurin series for
standard functions
You will often need to use the Maclaurin series for some common functions, which
are listed below, together with the values of x for which the expansion is valid.
2
e = +1 + x x
x
++
... x r + ...
Note 2! r!
Valid for all x
It is always good 3 2
ln(1 + =
x )− x x + x −+
...−( 1)
r +1 x r + ...
practice to state Valid for −1< < x 1
2 3 r
the values of x 3 5 21 r+
for which it is sin x x= x− + x −+ r x
...−( 1) (2 1)! + + ...
3! 5! r Valid for all x
valid when you
write down a 2 4 2r
Maclaurin series, x
cos x = −1 2! 4! + x −+ r x
...−( 1) (2 )! + ...
r Valid for all x
particularly in
cases where this n(n − 1) 2 + + n(n1) ( nr 1)
is not for all x.
(1 + x )n = +1 nx + x xr
2! r!
− 1) ( − + 1) Valid for x < 1, n ∈
+ ( r
+
!
ACTIVITY 3.10
(i) What happens when you differentiate these series, term by term?
(ii) What do you notice about the powers in the series for sin x and
cos x? How does this relate to the symmetry of the graphs of sin x
and cos x?
Example 3.11 These standard series can be used to find Maclaurin series for related functions.
−2 x 4
(i) Find the Maclaurin expansion for e up to the term in x .
(ii) For what values of x is the expansion valid?
Solution
2 3 4
(i) x
e = +1 + x x + x + +x ...
2! 3! 4!
Substituting x−2for x:
−2 x ( −2 )x 2 ( −2 )x 3 (−2 )x 4
e = +1 −
(2) x + + + + ...
2! 3! 4!
2 3 4
= −1 2 x + 4 x − 8 x + 16 x + ...
2 6 24
3 4
= −1 2 x + 2x −2 4 x + 2 x + ...
3 3
(ii) Since the expansion for e x is valid for all values of x, this expansion is
also valid for all values of x.
79
91
Sometimes a Maclaurin series can be found by adapting one or more
known Maclaurin series. An example you will have already verified is that of
differentiating the series for sin xto obtain the series for cos x. It is important
to question whether this is a justifiable method:
» Is it valid to integrate or differentiate an infinite series term by term?
» Can you form the product of two infinite series by multiplying terms?
» Is the series obtained identical to the series that would have been obtained
by evaluating the derivatives?
Answering these questions in detail is beyond the scope of this book, but
you may take it that the answer to all of them is,‘Yes, subject to certain
conditions.’ In the work in this book you may safely assume the conditions
are met, but strange things can happen with infinite series in other situations.
ACTIVITY 3.11
Try these suggested methods of deriving new series, and explain why they
work.
(i) The Maclaurin series for ln (1 + canx) be found by integrating
the terms of the binomial series for 1 . Why is the constant of
1 +x
integration zero?
x
(ii) The start of the Maclaurin series for
e can be found by
1+ x
multiplying together the first four terms of the series for e x and
(1 + x)−1, then discarding all terms in x 4 and higher powers.
(iii) The first few terms of the Maclaurin series for sec x can be found from
24
when y = − + .x
x
−1
the first three terms of the series for (1)+ y
2! 4!
Exercise 3D
1 (i) Write down the first four non−zero terms in the Maclaurin series for
cos u.
(ii) Substitute u = 2x in this series to obtain the first four non−zero
terms in the Maclaurin series for cos 2x.
2 Use known Maclaurin series to find the Maclaurin series for the
4
following functions as far as the term in x :
(i) sin 3x
(ii) ln(1 + 2)x
1x
(iii) e2
3 Use known Maclaurin series to find the Maclaurin series for the
4
following functions as far as the term in x :
(i) sin 2 x (ii) ln (1 + sin x)
−x sin x
(iii) e sin x (iv) e
80
92
1
4 (i) Find ∫ − 2
dx .
14 x 1
(ii) By expanding(1 − 4 x )2 − 2and integrating term by term, or
otherwise, find the series expansion for sin −21 x , when x < as12 far
as the term in x 7.
−1
(iii) Use your expansion to find an approximate value for sin 0.5 and
find the percentage error for this approximate value.
5 In this questions give all numerical answers to four decimal places.
(i) Put x 1= in the expansion
2 3 10
ln (1 + ≈x)− x x + x −…− x
2 3 10
and calculate an estimate of ln 2.
Approximately 1000 terms would be needed to obtain an estimate of
ln 2 accurate to 3 d.p. by this method.
(ii) Show that ln 2 = − ln 1 1−

six terms.
() 2
and hence estimate ln 2 by summing
(iii) Write down the series for ln (1 + − x) ln (1)− as

x far as the first
three non−zero terms and estimate ln 2 by summing these terms
using a suitable value of x.
6 (i) Write down the first four terms of the series for 1
1 − .x
(ii) By comparing with part (i), find an expression for the sum of the
2 3
series 1 +2 x + 3x + 4 x + …
CP 7 (i) Prove by induction that:
f (x) = e sin x ⇒f
x (n)
nx
()
(x) = 2 2e sin x n+ π
4
x 6
Use this result to obtain the Maclaurin series for e sin x as far as x .
(ii) Multiply the cubic Maclaurin approximation for e x by the cubic
approximation for sin xand comment on your answer.
x
(iii) Find a Maclaurin approximation for e cos x by multiplying the
cubic Maclaurin approximation for e x by the quartic Maclaurin
approximation for cos x, giving as many terms in your answer as
you think justifiable.
PS 8 (i)
2x
Given that xf ( ) e=sin 3 x , show that fx′′( )=4f′ −( ) 13fx ( ) x .
(3)
(ii) Differentiate this result twice to find expressions for f ( ) x and
f (4)
( ) x in terms of lower derivatives.
(iii) Hence find the Maclaurin expansion for xf ( ) up to the term in x 4.
81
93
 
M 9 A projectile is launched from O with initial velocity u  relative to
v
horizontal and vertical axes through O. The path of the projectile may
be modelled in various ways. The table below shows the position x( ,y )
of the projectile at time t after launch, as given by two different models.
Both models assume that g (gravitational acceleration) is constant.
Assumptions Position at time t
Model 1 No air resistance =
x ut
= gt
y vt − 1 2
2
Model 2 Air resistance is proportional
x u= − (1 e kt )
−
to the velocity (with k
proportionality constant k ). +
g kv gt
y =
k
2 (
1 −
e −− kt
) k
kt
Use the Maclaurin expansion for e − , where k is constant, to show that
the results given by Model 1 are a special case of the results from
Model 2, with k 0= .
PS 10 The diagram illustrates a Maclaurin expansion. Find it.
(You may assume that 0 < <r .) 1
r
1 (1 – r)
r2
r 1–r
1–r
r
1 1 r3 r(1 – r)
r2 r3 r3
r2 1–r
r2(1 – r)
4
r
1
1–r r2 r2
r4
r r4 r4 r4
r3 r3
82
94
KEY POINTS
d (sin −1 x) = − 1 d (cos −1 x) = − − 1
1 , x
dx 1 x d
2
1 x
2
2
d (sinh x) = cosh x d (cosh x) = sinh x
dx , xd
3 d (tanh x) = sech 2 x
dx
4 d (sinh −1 x) = + 1 d (cosh − 1 x) = 1
dx , x
1 x2 d x2 − 1
5
d2y
dx 2 can be found for a relation given implicitly by differentiating the
equation twice implicitly.
6
d2y
dx 2 can be found for a relation given
dy parametrically by
differentiating the equation for in terms of the parameter using
dx
the chain rule.
d 2 y = xy xy
−
7 The formula 2 can be used for parametric equations.
dx (x) 3
8 The general form of the Maclaurin series for f(x) is:
f ′′0( ) 2 f ((0)
3)
f (x) = f (0 )f +0 ′ ( ) x +
3
x + x +…
2! 3!
9 Series which are valid for all x:
2 3
x
e = +1 + x x + x +…
2! 3!
3 5 7
sin x x= x− + x x
−+…
3! 5! 7!
2 4 6
x
cos x = −1 2! 4! 6!
+ x x
−+…
10 Series valid for x 1 < :

1 x3 + x5 − x7 + …
tan − x x = −
3 5 7
11 Series valid for − 1<
x ഛ1:
2 3
ln (1 + x ) = −x x + x −…
2 3
12 Series where validity depends on n:
( − )
(1 + x )n = 1+ nx + n n 1 x 2 + …
2!
If n is a positive integer: the series terminates after n 1+ terms, and is
valid for all x.
If n is not a positive integer: the series is valid for x < ;1
also for x = if1 n 1 > ;
and for x = − 1if n 0 > .
83
95
LEARNING OUTCOMES
■ differentiate hyperbolic functions
■ differentiate inverse trigonometric functions
■ differentiate inverse hyperbolic functions
d 2y
■ obtain an expression for in cases where the relation between x
and y is defined d x 2
■ implicitly
■ parametrically
■ use the outcome from an initial implicit differentiation to find

successive derivatives.
■ derive and use the first few terms of Maclaurin’s series for a function
■ know that a Maclaurin series may converge only for a restricted set of
values of x
■ recognise and use the Maclaurin expansion of standard functions
x
■ e
■ ln (1 + x)
■ sin x
■ cos x
n
■ (1 + x)
84
96
4 Integration
The moving
power of
mathematics
invention is not
reasoning, but
imagination.
Augustus
de Morgan
(1806−1871)
?
 How could you estimate the number of birds in this picture?
Note
You should be confident in all the integration methods you have covered
previously.
4.1 Integration using the inverse sine

function
−
()
In Pure Mathematics 3, you used the substitution u = tan 1 x to show that
a
()
∫ a 2 +1 xd2 x a= 1 tan−1 xa + c .
In this section, the trigonometric substitution x = sin u or, equivalently,
u = sin 1 x, is used to find ∫ − 1 2 d x . Alternatively, you could also use
−
1 x
the substitution x a=u cos , or
Similarly x a=u sin or u = sin −1 x () a u ()
= cos −1 x . However, in practice
a
it is the inverse sine function that is
is used for the integral ∫ − 2 1 2 d x . more usually used.
a x
85
97
Example 4.1 1
Find ∫− 2
d x.
1 x
Solution
Method 1 (substitution)
Let x = sin u
⇒ dx = cos u
du
1
∫ cos ud u
1 −sin u 2
= ∫ 1 cos ud u
cos u
= ∫ d u u= c+
= sin −1 x c+
Method 2 (recognition)
Since x d (sin 1 x) = − 1 2
−
This is a standard result.
d 1 x
1
it follows that ∫− −
d x = sin 1 x c+ .
1 x2
Activity 4.1
Use the substitution x = 3sin u to show that ∫ 1
9−x
2
d xsin= 3
−1
(x c)+ .
The result in Activity 4.1 above can be generalised to give the formula which
can be quoted for use in integration.
∫ 1
a2 − x 2
d x = sin −1 x + c
a () You may quote this result where it is needed
when you are carrying out integration.
In the next example this work is taken a step further.
Example 4.2 1
Find ∫ −− 2
dx
54 xx
Solution
This is a quadratic
The denominator is the square root of −5−4 x x 2 . function with a
negative term in x2 .
86
98
Completing the square gives:
2 2
5 4− x x− 2
= −9 +2 ( x) = 32 − +
(x 2)
and so
1 1
∫ dx = ∫ dx
5 4− x x− 2 2
(x
3 −+ 2)
2
Let x + =2 3sin u.
So
dx = 3cos u
du
The integral is
1
∫ × 3cos ud u
3 − 3 sin u
2 2 2
= ∫ 3cos u d u = ∫ du
3cos u
= +u =c − ++
sin 1 x 2 c
3 () Check for yourself that
d sin −1  x + 2= 1
()
= sin −1 x + +2 c
3
dx  3  2 − +
3 (x) 2 2
.
The next example shows you how this sort of substitution can be used with
other types of integral.
Example 4.3
Find ∫ −4 x 2xd using the substitution x = 2 sinu.
Solution
In this case, you are given the
dx = 2cos u
Let x = 2 sinu, giving substitution. If the substitution
du is not given, compare the
So the integral becomes integral with the standard one
∫ 1
∫ 4 − x 2xd = ∫ − ( 2cos
4 2sin
2
u) d uu 4 x− 2
d x , which gives
= ∫ 2cos u × 2cos ud u = ∫ 4cos du u

2
()
sin −21 x c+ .
Using the double angle formula = 1 2 u − and

cos2 u2cos
u = 1 u+
so 2cos 2cos2
4 ∫cos d u u = 2 ∫cos
( 2 u + 1) du = sin 2u2u+ c+
2
= 2 sinucos
u + +2u=c 1 x 4 − +x 2 2sin 2 x c+
−1
2
87
99
Exercise 4A 1 Use the standard results to evaluate
∫ 1 2 dx
25 − x
2 Use the standard results to evaluate
∫ 0− 1 2 d s
1
3 s
3 Find
∫ − 1 2 dx
4 x
4 Evaluate the following definite integrals, leaving your answers in terms
of π:
1
(i) ∫
2 1 d x (ii) ∫ 4− 1 dx
0
−
4 x 2 0
14 x 2
5 Find the following indefinite integrals

1
(i) ∫ 2
dx (ii) ∫ 1 dx
16 (− −x 3) 3 3− 1−( x ) 2
1 1
(iii) ∫ −− 2
dx (iv) ∫− 2
dx
32 xx xx 2
6 Find the following indefinite integrals
(i) ∫ −9 x 2xd (ii) ∫ −4 9 d x x 2
1
(iii) ∫ dx
(1 − x )
32
2
4.2 Integrating hyperbolic functions

Using the derivatives for hyperbolic functions gives:
∫ sinh xd x = cosh x c+ , ∫ cosh xd x = sinh x c+ and ∫ sech dxtanh
2
x = x c+
Example 4.4
Find ∫ sinh 5xdx .
Solution
Using the substitution u = 5x gives:
∫ sinh 5xdx = 51 cosh 5x c+
88
100
Example 4.5
∫
Find sinh 2 x dx .
Solution
2
Using the identity cosh 2x = +1 2 sinh x, the integral becomes:
1 1
∫ cosh 2x − 1)d x = sinh 2x x− c+
(
2 4 2
4.3 Integration using inverse

hyperbolic functions
From the derivatives of cosh-1 and sinh-1 , given on page 63, it is clear by
integrating that
∫ 21 dx = cosh −1 x c+
x −1
and
1
∫ 2
−
dx = sinh 1 x c+
x +1
Activity 4.2
Use the chain rule and the derivatives for sinh-1 and cosh-1 given on page
63, to show that
d ()
(i) x d sinh −1=x+
a x2
1
a2 d ()
(ii) xd cosh −1=x−
a
1
22
xa
The results in Activity 4.2 lead to the following results:
∫ 2
1
x +a
2
d x = sinh −1 x + c or xlnx a +c
a ( 22
++ )
∫ 1
x −a
2 2
d x = cosh −1 x + c
a or
ln x x− a c
(
22
−+ )
The range of functions that you can integrate has now been extended.
Just as before, more complicated examples use techniques such as taking
out constant factors or completing the square. You can often choose an
appropriate substitution, even if the integral is not quite a standard one.
Note
When evaluating a definite integral, it is usually easier to work with the
logarithmic form.
89
101
Example 4.6 1
Find ∫ 2
9 x − 25
dx.
Solution
This is now a standard
1 = 1∫ 1
∫ 2
9 x − 25
d x
3 2 25
dx integral with a = 53 .
x − 9
To use the standard x
integral, the coefficient of 1 −1
= cosh  5  ++c
x2 must be 1, so you need 3 ()
 3
to take out a factor of 9,
which becomes 3 when it
leaves the square root. 3 ( ) +c
= 1 cosh −1 3x
5
Example 4.7
Find ∫ 1 dx .
(9 +4 )
3
2 2
x
Solution
Compare with the standard integral to determine the substitution
x = 3 sinh u giving:
2
dx = 3
cosh u
du 2
1 1 3 cosh ud u
∫ dx = ∫
(9 4+ x 2) (9 9sinh u) 2
3 3
2
2
+ 2
= 3 ∫ cosh u3 du
2 27cosh u
= 1 ∫ sech 2du u = 1 tanh u + c = sinh u + c
18 18 18cosh u
2x 2x
= 3 +=c 3 +c
2 2
18 1 + (23 )x +
18 9 4 x
9
= 2x +=
c x +c
18 9 4+ x 2
9 9 4+ x 2
90
102
Example 4.8
Use integration by parts to find ∫ cosh d x x−1
. Remember:
∫ u ddvx d x uv
= −v u ∫ d dx
dx
Solution
Write the integral as ∫ 1× cosh −1
x dx .
Use u = cosh −,1 dx v = 1.

dx
du = 1 , v x=
dx 2
x −1
x
∫ cosh −1
x dx x =
−
cosh 1 x − ∫ x −1 2
dx
= x cosh −1 x − x2 − 1
Note
The second integral is of the form (f )(x)( ) f ' x so can be integrated by
n
inspection. It can also be achieved by using the substitution u x = − . 1

2
Exercise 4B 1 Use the standard results to find the following indefinite integrals:
(i) ∫ 21 dx (ii) ∫
2
1 dx
x −4 x +4
2 Use the standard results to find the following definite integrals, giving
your answers in terms of logarithms.
6 6
1 (ii) ∫ 1 dx
(i) ∫4 x 2 − 9 d x 3 2
x +9
3 Find the following indefinite integrals:
(i) ∫ 12 dx (ii) ∫ 1 dx
9x − 1 2
9 x1 +
1 1
(iii) ∫ d x (iv) ∫ dx
4x + 9
2
4x − 9
2
4 Evaluate the following definite integrals, giving your answers as

logarithms:
2
(i) ∫
2 1 1
d x dx
1 2
25x − 16 (ii) ∫0
9x 2
+ 4
CP 5 (i) Prove that cosh 2 x = 1 (cosh 2x1)+ .
2
(ii) Use the substitution x = 2 sinhu and the result from (i) to find
∫ + x. 2
4 dx
CP 6 (i) Prove that sinh 2 x = 1 (cosh 2x1)− .

2
(ii) Use a suitable substitution and the result from (i) to find
∫ − x. 2
9 dx
91
103
CP 7 (i) Use the substitution x = a sinh u to prove that
1 = sinh −1 x + c .
∫ d x
a
x2 + a2
(ii) Use the substitution x = a cosh u to prove that
∫ 1 d x = cosh −1 x + c .
x2 − a2 a
4 x + 1 dx = +8 ln 3 .
∫
4
8 Show that
2
0
x +9
9 Find the following indefinite integrals:
(i) ∫ sinh −1
x dx (ii) ∫ tanh 2 dx x
−1
10 (i) Write 4x2 + 12x − 40 in completed square form.

Evaluate the definite integral
(ii)
∫ 10 2 1
20
dx
4 x + 12 x − 40
CP 11 The points P 1 (a cos q, a sin q ) and P 2 (a cosh f, a sinh f) lie on the circle
x2 + y2 = a2 and the rectangular hyperbola x2 − y2 = a2 , respectively
(see below).
P 1(acos θ, a sin θ)
P 2(acosh φ, a sinhφ)
O A O A
In both diagrams A = (a, 0).

Prove that area OAP 1 is proportional to q and that area OAP 2 is
proportional to f, with the same constant of proportionality.
4.4 Integration using reduction

formulae
1
Integrals such as x∫ 0
4 −x
e d x can be found by repeated use of integration by
parts. At each stage a very similar integral is needed, so it makes sense to find
1
∫ 0
x e d x in terms of x∫
n x−
0
1 n −x1−
e d x and this is what is meant by a reduction
formula.
92
104
Example 4.9 1
Given that I n = ∫ exdn x 2 x,
0
(i) show that I n

= 1e 2 − n I n − 1
2 2
1
42 x
(ii) hence evaluate I 4 = ∫ 0e d .
x x
Solution
,d v = e
n 2x
(i) Using integration by parts with u x =
dx
du = nx n − 1 , v = 1 e 2 x
dx 2
1
1 n x2  1
1 e dx x
I n =  x e  −
2 0
∫ 0
nx
n −1 2
2
This has the same form
2 ()
= 1 e2 − −02 n x
∫0
1 n −x1
ed x as In but with n replaced
by n –1.
= 1 e 2 − n I n −1
2 2
(ii) So I 4
= 1 e2 − 4 I 3
2 2
2 ( 2 2 )
= 1 e2 − 2 1 e2 − 3 I 2
= − 1 e + 3(1 e − 2I )
2 2
1
2 2 2
= e − 3(
2
1 e − 1I ) 2
0
2 2
1
= − 1 e 23+ ∫ e 2dx x =− 1 e 2 + 3 1 e 2 x 
1
2 2 0 2 2 2 0
− 1 e 2 + 3 e 2 − =3 1 (e 23− )
2 4 4 4
Example 4.10
1
2n I
It is given that I n = ∫ 1x
0
n
− x dx . Show that I n
=
2n3 +1 andn −hence
1
find I 3 = ∫ 0
x 3 1 − x dx .
Solution
, d v = −(1) x
1
n
Using integration by parts with u x = 2
dx
3
du = nx n −1 , v =− (1)− x
2
= − −2 (1) x
3
2
dx 3 3
2
➜
93
105
( ) ( )
1
= x n − 2 (1 − x)  −
1 1
− 2− (1) x x d
3 3
−1
In = ∫ 0
x n
1 − x xd
 3 0
2
∫ 0
nx n
3
2
1
n x
= +0 2 ∫
3 0
n −1
(1 − x ) 1 − x xd
Note
In =
2n
3 0
x
∫
1
n −1
( 1
1 − x xd − ∫ x n 1 − x xd
0
) You want to
recreate the
nI
I n = 2 ( n −1 − I n ) shape of the
3 original integral.
Collecting terms: This gives you an
idea about how
()
I n 1 +2 n = 2n In −1
3 3
to rearrange the
algebra here.
I n (3 2+ n) = 2nI n − 1
I n = + 2n I n − 1
(3 2 n)
2 3× 6  2× 2  2 4 2 × 1 
So I 3 = I2 =  I 1 =  
(3 +2 3× ) 9 (3 +
2 2× )  3  7 (3 +2 1× ) I 0
1
= 16 × − −2 (1 1
I 3 = 16 ∫ x)  = − 32 (0)1− = 32
1 3
(1 − x)x d 2 2
105 0 105  3 0 315 315
Example 4.11 d tan xsec

(i) Find x
d
( n
x)
1 π n
1 π
( + 1) ∫ sec n+2
n ∫ secn xdx
4 4
(ii) Hence show that n

1
π
0
xxd
= ()+ 2
0
28 .
∫
4
6
(iii) Show that sec xdx =
0 15
Solution
d (
(i) Using sec x) = tan xsecx ,
dx
dx
(
d tan sec n ) =
x x x n x + tan x( sec
sec 2sec n −1
n )(ta x nnx sec
x )
x ( 2 x − 1)
= sec n + 2 x n+ sec nsec
= n( + 1) sec x n− x sec
n+2 n
(ii) Integrating both sides:

1 π
1
π 1
x 0 = +(n 1) ∫ sec n+2

∫
4
n π
tan
 xsec xxd n − sec n xxd x
4 4
0 0
1
π n
1
π
So n( + 1) ∫ sec xxd n ∫ sec n xdx
n+2 4 4
0 = ()+ 2
0
1 π
(iii) π
∫ sec xdx = [ tan x]0 = 1
4 1
2 4
0
π π
( )2 2 +sec d∫
2
1 1
3 ∫sec
4 2
xx =
4 4
Substituting n = 2: 0
d 0
xx
π
sec 4dx x = +2 =2 4
1
∫ 4
0 3 3
94
106
π π
( )2 4 sec
1 1
5 ∫0 + d∫
6 4 4
sec dx x =
4 4
Substituting n = 4: 0
xx
1
π 4 4+ × = 4
∫ 4
0
6
sec dx x =
5
3 28
15
π
Example 4.12 1 − I−
∫
4
( tan
) dx x
n
(i) Show that = I n I = for n ജ
2 .
n − 1 n 2 where n 0
π
π
(ii) Hence show that ∫
4
0
(tan x )xd= −
4
4
2
3
.
Solution
π π
∫ (tan x )xd ∫ (tan x )tan 2dx x

4 n 4 n −2
(i) In = =
0 0
∫ (tan x )(sec 2 x − 1)d x

4 n− 2
0
π
∫ (tan x )sec 2dx x I −

4 n −2
= n −2 n
f (f xd) x( )x ′
0
π
This is of the form ∫ ()
=  1 tan n−1 x  − I n− 2
4
n − 1 0
= 1 − I− n 2
n−1
π
4 1 −
tan 4 x xdI I2
(ii) ∫()
0 == 4
4 −1
1− 1 −I
3 2 −1 0( )
π
1 − 1 − 
∫
4
1 dx
3 2 −1 0 
= − + [x] 04 = π
π
2 − 2
3 4 3
Example 4.13 1 n
3n
The integral I n is given by I n = ∫ (x )1 − x
0
3
dx . Show that =I n I
2 +3 n n−1
Solution
(i) Use integration by parts with u = (
−1 x ) and ddxv = x
3 n
du = ( − 3 )n− 1 × 2 and = 1 2
Giving n 1 x 3x vx
dx 2
➜
95
107
1 n 1 1 n −1
In = ∫ x (1 − x )
0
3
dx= 1 x (1 x ) −

2
2
− 3 n
0 ∫ 3n x4 ( − 3 )
0 2
1 x dx
n −1 This needs to be rewritten

=−
03
n x1
2 0 ∫( 1 −1 (− x 3
)) (1 − x 3
) dx to give a factor of x. The x3
factor needs rewriting in
terms of (1 − x3 )
This gives I n = 3n (I n −1
− In )
2
n = 3n I
I n 1 +3
Rearrange to give
3n
() 2 2 1 n−
So =
In I
2 +3 n n −1
Exercise 4C 1
π n
You are given I n
= ∫ 0sin d xx .
1 2
CP 1 π
∫ sin sin x x x d n −1
2
By writing I n = and using integration by parts, show

0
n 1
that =
In− I
n n −2 π
4
π 1
2
1
2
Use your answer to find I 4 = ∫ sin d xx and I 5 = ∫ sin d 5 xx

0 0
1 −
You are given I n = ∫ e (1 − x)x d .
x n
CP 2
0
(i) Show that In = −1 − . nI n 1

1
(ii) Find I3 = ∫e −x
(1 − x)x3 d .
0
1
3 Let I n n π
= ∫ 0sin
x
2
xdx , where n is a non-negative integer. Show that
n −1
I n = n (12 )π − n (n − 1) I n 2
− , for n ജ .
2
Find the exact value of I 4 .

9321 Paper 12 Q7 June 2015
1
4 Let I n = ∫ (1 +1x )
0 2 n
dx . Prove that, for every positive integer n,
−
2nI n +1 = 2 n + (2n − 1) I n .
Given that I 1 = π,14 find the exact value of I 3 .

π
9321 Paper 11 Q4 June 2013
π +
5 Show that ∫ x
e sin dx x = 1e .
0 2
π
xn
Given that I n = ∫ e sin xxd , show that, for n ജ ,2
0
π
n−2
I n = n(n − 1)∫e cos sin2 x
x
x xd nI− n ,
0
96
108
and deduce that n( 2
+ 1)I n = n (n − 1) I n − 2 .
1 2n
9321 Paper 11 Q11 (part question) June 2012
π sin x x
6 It is given that I n =∫0
4
d , where n 0≥ .
cos x
− (+n ) 1
Show that I n − I n + 1 = 2 + . 2
2n 1
π 6
( ) −120
1
Hence show that ∫ sin4

x 73
dx = ln 1 + 2 2.
0 cos x
9321 Paper 11 Q10 June 2014
4.5 Approximation of areas using

rectangles
The area under a curve cannot always be found by integration. You can find
an approximate value by using rectangles with heights equal to the y-values
at either the left or right boundaries of the rectangles. The calculations can
give you an estimate for the area; they can also sometimes give bounds within
which the area must lie.
Example 4.14
Figures 4.1 and 4.2 show approximations to ∫ (f )dx, xwhere x f()=+ 1 2 ,
1
0 1 x
using 5 rectangles of width 0.2.
y
y
1
1
0.5
0.5
x x
0 0.5 1 0 0.5 1
(i) Use the areas of the rectangles to estimate the value of the integral and
to obtain upper and lower bounds for it.
(ii) Explain how the bounds may be improved.
(iii) The total area of n rectangles of equal width that gives a lower bound
for the integral is denoted by L. Express L as the sum of a series with n
terms.
97
109
Solution
(i) A table of values can be used to calculate the total area for each. Notice
that the final value in the first column and the first value in the second
column are not required.
x y Left area Right area

0 1 0.2
0.2 0.961538 0.192308 0.192308
0.4 0.862069 0.172414 0.172414
0.6 0.735294 0.147059 0.147059
0.8 0.609756 0.121951 0.121951
1 0.5 0.1
Totals 0.833732 0.733732
The estimated value is between 0.733732 and 0.833732 so
approximately 0.78
The diagrams show that 0.733732 is definitely less than the area so
gives a lower bound.
Similarly, 0.833732 is definitely greater and so gives an upper bound.
(ii) There is a large difference between these values. This can be reduced by
using more, thinner rectangles as shown in Figures 4.3 and 4.4.
y
1
y
0.5
0.5
0 x 0 x
0.5 1 0.5 1
(iii) When n rectangles are used, the width of each is n 1 .

Height of first is nf 1( ) , the nextf 2(n) etc.
n
Total area = n1 nf ( ) ( ) + + n1 f (nn n)= 1 ∑f (nr )
1n 1n +f 2
r =1
n
= 1∑ 1
()
2
n r =1 r
1+ n
n
= n∑ 1
n 2
r =1
+ r2
98
110
Activity 4.3
Use a spreadsheet to approximate the same integral with 10 and 20 4
rectangles. Which are the best upper and lower bounds? To how many
significant figures could you give an answer for the integral? The value of
π
the integral is known to be . How accurate is your answer?
4
4.6 Using areas under curves to

approximate series sums
Example 4.15 ∞ Example 4.13
Investigate ∑ 12 .
1 r
Is the sum of an infinite number of terms finite or infinite?
Solution
The series sum is the total area of the rectangles.
y
2
1.5
−0.5
0 1 2 3 4 5 6 7 8 9 10 x
▲ Figure 4.5
A lower bound for the area can be found using the curve y = 1 .
(x) + 1 2
N  −1 N
( + )
1 dx =  x 1  = −
∑ r1
N
1 +1
2
> ∫ 1 (x + 1)2  −1 1 N +1
1
∞
As N → ∞, r ∑ 12 > lim 1 −
1
N →∞
( 1
+
N 1 )
= 1.
An upper bound for the area can be found using the curve y x = 1
2 .
y
2
1.5
−0.5
0 x
1 2 3 4 5 6 7 8 9 10
▲ Figure 4.6 ➜
99
111
The first term needs to be separated from the rest as the curve is not
defined at x = 0.
N N −1 N
∑ r12 = 112 + ∑ r12 < 112 +
N
1 dx = x  11
∫ 1 x2
1
−1 1 = + − + N 1
1 2
()
∞
As N → ∞, ∑ 1< 2
N →∞ N
1
lim 2 − = 2.
r1
So the sum of an infinite number of terms is between 1 and 2. (In 1735 Euler
π 2≈
proved the total is 6 1.645 )
?
 Does it improve the upper and lower bounds to remove more terms
from the beginning?
Exercise 4D 1
(i) Use 5 rectangles to find upper and lower bounds for ∫ 0− 1 2 d x.

2
1
(ii) Find the exact value for the integral. 1 x
(iii) Use your answers to find upper and lower bounds for π .
(iv) If you have access to a computer, use a spreadsheet to repeat this
with 10 rectangles.
2 The diagram shows part of the graph of y = sin x .
0.8
0.6
0.4
0.2
0 x
0.5 1
1
Use the rectangles shown to find upper and lower bounds for∫ 0 sin xd x .
100
112
∫ ( 9 x x− x )d
2 3
3 George needs to evaluate . He uses a graph drawing
1.4
package to estimate area. The diagram shows part of the graph and the
rectangles used.
y
1.5
1.4
1.3
1.2
1.1
1
9.9
9.8
9.7
9.6
9.5
0 x
1.3 1.4 1.5 1.6 1.7 1.8 1.9 2 2.1 2.2
2
(i) Use calculus to find the exact value of ∫ ( 9 x x− x 3 )d .
1.4
(ii) Use three rectangles to estimate the area:
(a) using the left edges
(b) using the right edges.
(iii) Explain why these values can be used as estimates but may not as
reliable upper and lower bounds.
1
1 dx
CP 4 An estimate of ∫ 0+1 is found
x
by using n rectangles with the
n −1
right corner on the curve. Show that ∫ 0 +1 dx ≈ 11+ r
1
1 x ∑r=
n n
1
()
1  3 1−
n
Show that ∫ 1 dx n≈ 1 2 + r  . Use n = 4 to estimate the
CP 5
integral.
0 2 + x3 ∑
r =1
 n  ()
6 The diagram shows the graph of y = log 10 x.
y
2
0 x
2 4 6 8 10
101
113
(i) Show that the total area of the rectangles is log10!
Use the identitylog x = ln x to find ∫ log xd x.
10
(ii)
ln10 1
(iii) Hence find a lower bound for 10!
n
Show that r∑ 1>
n
1 d x and hence show that the sum to
CP 7 ∫ 0 x +1
r =1
infinity does not exist.

n
∑ r1 1< + ∫
n n
1 dx < 1 d x and hence find upper
8 Show that ∫ 1
(x + 1)
3
2
3
2 1
x
3
2
r =1
∞
1
and lower bounds for ∑ r .
3
2
r =1
4.7 The arc length of a curve

One practical way to find the length of a curve between two of its points (for
example to find the distance along the road between two towns, from a map)
is to mark several intermediate points along the arc, join them successively
with straight lines to form a polygonal line, and measure the total length of
this open polygon to give an approximation to the length of the arc. If you
start with a polygon P1 and construct a new polygon P2 by inserting extra
points along the arc, then P2 will fit better than P1, and the length of P2 will
be greater than the length of P1 (see Figure 4.7).
P1
P2
▲ Figure 4.7
In this way, you can form more and more polygons with successively greater
lengths. But since the shortest route between two points is the straight line
joining them, the length of any such polygon does not exceed the length of
the curve. So you would expect that the lengths of successive approximations
would be bounded above (by the length of the curve), and that by putting the
immediate points sufficiently close together you could get an approximation
as close as you like to the length.
For most curves that occur in practice, this approach works and leads to the
calculus method of finding arc length, which is given below. But, contrary
to intuition, there are some curves for which there is no upper bound to the
length of the inscribed polygon between two fixed points. One example is
Von Koch’s ‘snowflake’. Essentially what happens with these exceptional
102
114
curves is that they wiggle so much that no chord, however short, is a good
approximation. To rule out this possibility, only curves for which the arc
length PQ is nearly the same as the length of the chord PQ whenever the
two points P and Q on the curve are close together are considered here. To
be more precise, the assumption is made that:
arcPQ →
1 as P → Q.
chordPQ
The positive sense along a curve

If the coordinates of a point P on a curve are given in terms of a parameter
p then the sense in which P moves along the curve as p increases is called the
positive sense. The same applies if you are using x or y instead of p as the
independent variable. The positive sense on a curve depends on the particular
way in which the Cartesian equation or parametric equations are expressed.
For example, the equations:
(i) y = − x3 (ii) x =−3 y
(iii) x p y= , = − p3 (iv) x q y=q − =,
3
all give the same curve, using independent variables, x, y, q, p, respectively. In

(i) and (iii) the positive sense is from left to right across the page, but in (ii)
and (iv) the opposite sense is positive (see Figure 4.8).
(i) (ii) (iii) (iv)

y y y y
x x x x
▲ Figure 4.8
Activity 4.4
Draw diagrams to show the positive sense when the unit circle is
expressed as:
(i) x = cosθ, y = sinθ
(ii) x = sinθ, y = cosθ
(iii) =y± − 1 x2.
103
115
Arc length with Cartesian coordinates
4 To see how to find the length of an arc, look at part of a general curve, as
shown in Figure 4.9. There C is a fixed point on the curve, P is the point
with parameter p, and s is the arc length from C to P, where s is positive if
and only if the motion from C to P is in the positive sense along the curve.
Let P and Q have coordinates (x ,y ) and x( x +y δy , + δ, corresponding
) to
parametric values p and p + δ, respectively,
p and let arc PQ δ= . s
Q
δs δy
P
δx
C
s
▲ Figure 4.9
2
The chord length PQ is PQ = ( ) δ x
+( )
δy 2.
s = δ s × PQ
Therefore δδ δp
p PQ
2 2
δ s × δ x  + δ y 
= δ p  δ p 
PQ
δy
→ 0, δ, s , δ x
dy
As pδ tend to sd , dx , , respectively, and by the
δp δp δp dp dp dp
δ → . So taking limits as p 0 δ → gives the
assumption stated above, s 1
PQ
basic result:
2 2
ds =  dx  +  dy 
dp  dp  d p
From this s can be found by integrating with respect to p.

If the independent variable is x (i.e. the equation of the curve is given in the
form y = f (x) ), then you put p = x in the basic result. Then:
dx = dx = and dy = dy
1
dp dx dp dx
104
116
so that:
2
ds = dy 
1 +  
dx dx
Similarly, when the independent variable is y ,
2
ds =  dx  +
 dy  1
dy
All these are easy to remember from the right-angled ‘triangle’ (see Figure 4.10)
in which s(δ)
2
≈ (δ )x 2 + (δ)y 2 by Pythagoras’ theorem. The three results
2
follow in the limit from dividing by p(δ) (2 ), δorx y2 (δ ) , as appropriate, and
then taking the positive square root of each side. Notice that the positive
root is needed in each case, since, by definition, s increases with the
independent variable.
δs δy
δx
▲ Figure 4.10
Example 4.16
Find the length of the astroid x a= cos 3,θ y a= sin 3θ .
Solution
dx = − 3acos sin
2
θθ dy = 2
θθ
θ θ 3asin cos
d d
⇒ ds = 9 cos
a 2 4
sinθ 2
θ + 9a 2sin
4
θ
cos 2
θ
dθ
2 2
(
= a3 cos sinθ cos θsin
2
θ + 2
θ)
sinθ ➀
= a3 cos θ
The values of θ at the four cusps of the curve are as shown in Figure 4.11, so
the length of an arc in the first quadrant is:
π
2 π
=  3a sin θ 
2
3a
∫ 3acos θsinθ θd 2
2
0 = 2
0
and the length of the complete astroid is:

3a =
4× 6a
2
105
117
y
π
θ =
2
θ =π θ =0
O x
3π
θ =
2
▲ Figure 4.11
Note
If you try to find the whole length in a single integration you get:
2π 2π
a 2 
∫ 0
θ θ θd 3 =  sin θ
3acos sin 2 0 = 0
This is because cosq sinq < 0 in the second and fourth quadrants, and the
positive and negative contributions in the fourth quadrants have cancelled. When
taking the square root at ➀ (and in similar cases) it is essential to check that you
use an expression which is never negative throughout the range of integration.
Arc length with polar coordinates

The method of finding the length of a curve from its polar equation comes
from differentiating the relations x r = cos θ, y r= sinθ with respect to θ ,
remembering that r is a function of θ . Thus:
dx = dr cosθ − r sinθ dy = dr
θ θ and θ θ sinθ + r cosθ
d d d d
( ) ( ) cos θ − 2rdrdθ cosθsinθ sin+ r

2 2
so that dθx = dr 2 2 2
θ
d dθ
2
and  θ  = (d θ) sin θ + 2rdr θ cosθsin

 dy  r
2
2
θ cosr . θ 22
d  d d
Adding these and using cos 2 θ + sin 2 θ = 1 gives:
 d y 2
() ( ) +r
2 2
dx + = d θr 2
dθ  dθ  d
2
s dr
so that d θ r2,
d
= ()
+
dθ
from which s is found by integrating with respect to q.
106
118
?
 Figure 4.12 shows a curve through neighbouring points P and Q with
polar coordinates θ(r, ) and r ( + δ,rθ δθ+ ). Identify in the diagram
a right-angled ‘triangle’ with sides sδr δ, and rδθ, and explain how it
can be used as a reminder of this result.
r + δr
P
δθ
r
θ
O
▲ Figure 4.12
Example 4.17 from θ = to0 θ=.π

Find the length of the equiangular spiral r 3e= 3
θ
Solution
dr = e θ
3
dθ
⇒ sd =
2θ 2θ θ
e + 9e
3 3
= 10 e 3
dθ
π π
⇒s = ∫ 0
10e d θ = 3 10e 0  
θ
3
θ
3
0 3
= 3 10 e( 1) − ≈
π
▲ Figure 4.13
3
17.5
Exercise 4E
1 Find the length of the semi-cubical parabola y 2x = 3
from (0, 0) to (4, 8).
Historical note
This curve was the first for which the length was found by calculus methods,
by the Dutchman Heinrich van Heuraet, the Englishman William Neil and the
Frenchman Pierre de Fermat independently, all between 1658 and 1660.
107
119
p
2 Find the length of the curve x = +1 2 , y 1= + from p 0= to
1 p p2
p 1= , and draw a sketch of the curve to explain the answer.
3 Find the length of the catenary y c = cosh x from x 0= to x X = .
c
CP 4 Show that x a= sinh ,p y = 2asinh p are parametric equations
2
of the parabola y 2 = 4ax , and that the arc length from p 0= to p P =
(
is a P + 1 sinh 2P .
2 )
CP 5 Prove that the length of one complete arch of the cycloid
x a= (θ − sinθ), y a= (1 −cos ) θ is a8 . (This result was first given by
Christopher Wren in 1659.)
6 Find the length of the nephroid
x = 3acos θ − a cos 3θ, y = 3asin θ − a sin 3θ
CP 7 Prove that the length of the arc of the equiangular spiral r ae= kθ from
(r1, 1)θ to r ,( 2 2 θ ) is proportional to r( 2r 1− , )and find the constant of
proportionality in terms of k.
8 Find the length of the cardioid r a(1 = +cos ). θ
l 1 ecosθ
CP 9 For the conic = + show that, if e is small,
r
s l≈ (θ − e sinθ) , when s is measured from where θ = . 0
CP 10 (i) For the ellipse x a = cos θ, y b= sinθ prove that
ds = a 1 − e 2 cos 2 θ 2
= 2
− e 2
, where b a (1 ).
dθ
(ii) Prove that the perimeter of this ellipse is exactly the same as the
x
length of one complete wave of the curve =y ea cos .
b
(iii) Prove that if e is small then the perimeter of this ellipse is
()
approximately 2π a1 1 − e 2 .
4
11 A circle has polar equation r = a for 0  q < 2π, and a cardioid has polar
equation r = a (1 - cos q ), for 0  q < 2π, where a is a positive constant.
Draw sketches of the circle and the cardioid on the same diagram. [3]
Write down the polar coordinates of the points of intersection of the
circle and the cardioid. [2]
Show that the area of the region that is both inside the circle and inside
the cardioid is
5
2( )
π − 2 a2 . [6]
12 The curve C has polar equation r 2=sin (1θcos −), forθഛ ഛ0θ π.
Find the exact area of the sector from q = 0 to θ = π. [6]
9231 Paper 13 Q10 (part question) November 2013
108
120
4.8 Surface area of a solid of
revolution
First consider a simple case, the curved surface area of a right circular cone. If
the cone has base radius r and the slant height is l its curved surface, flattened
out, is a sector of a circle with radius l and arc length 2π r, as in Figure 4.14.
l
l
2πr
▲ Figure 4.14
πrradians, and so the curved
The angle at the centre of this sector is 2
surface area of the cone is
1 2 × 2πr = π .
l rl
2 l
Now consider the solid of revolution formed by rotating about the about the
x axis the line segment joining the points (x 1, y 1) and (x 2, y 2),
where the distance between these points is sδ . This solid, which is called a
frustum, is the difference between two cones. Let the slant heights of these
cones be l 1 and l 2 as shown in Figure 4.15. Then sδl =l − .2 1
δs
l1
y1 y2
x1 x2 x
l2
▲ Figure 4.15
109
121
The curved surface area of the frustum, Sδ , is given by
δS = πy 2l2 − πyl1 1 = π (yl2 2 − yl ).
11
l1 l2
But by similar triangles = , so that y l y2l1 1 =2 . Therefore:
y1 y2
δ S = π (2y2 l − y 2l1y1+2l 1y1 l − ) Inserting extra
terms which cancel.
= π (2y + y1l)(
2 l1
− )
= 2πy δ
s
1
where y = 2 (y1y+ 2 ) , the average radius of the frustum.
It is now easy to see how to find the curved surface area of a solid of
revolution: divide the arc AB which is rotated into elements of arc sδ , each
of which generates a surface with area approximately 2πy s δ. Then the total
curved surface area is:
B
B
S = lim ∑
2π y sδ = ∫ 2π y ds
δ s→ 0 A
A
The detailed evaluation of this integral depends on which independent

variable is being used. If you are working in terms of x, with x a= at A and
x b= at B, then:
b
S = ∫a
2π yds d x
dx
but if the independent variable is parameter p, with p = α at A and = βp at
B, then:
β
S = α∫ 2πyds d p
dp
The corresponding results when the solid is generated by rotating about the
y axis are:
d
S = ∫c 2π x sddy d y
where y c= at A and y d = at B, and
β
S = α∫ 2πxds d p
dp
110
122
Example 4.18 Find the curved surface area of the paraboloid formed by rotating the
parabola y = 4 ax about the x axis from x 0= to x a = .
2
2a
x
0 a
–2a
▲ Figure 4.16
Solution
2 2
2 dy dy 2a  dy  4 a a
y = 4 ax ⇒ 2y = 4a ⇒ =⇒
 dx 
== 2
dx dx y y x
so that:
ds = 1 + =a x a+
dx x x
and
+=
y sd = 4 ax x a 2 a x a+
dx x
The required area is:
a a
4π a ∫ = 4π a × 2 (x a+ ) 
3
x a+ dx 2
0  3 0
3
(
= 8 π a a(2a) −
3
2
3
2
)
= 8 πa22 1 −
3
()
3
2
111
123
Example 4.19 Find the surface area of the solid generated by rotating the equiangular spiral
θ
=
r 3e 3
from θ = to0 θ = aboutπ the line θ=.0
(a bx b bx− cos
e axsin )
You may use the result of ∫ e sin dbx x ax
= 22
.
a b+
Solution
As in Example 4.17, dθs = 10e
θ
3
d
θ
=
and y rsin θ = 3e sin θ so the required area is:
3
π θ θ π 2θ
∫0
× sin
2π 3e 3
θ × 10e d θ = 6 10π e∫ sin d θ θ
3
0
3
ax
ax e a bx b bx− cos )
= ( sin
e sin dbx x
Using the result above for ∫ 22
a b+
with a 2= and b 1 = , the area is:
3
π
 2 θ − cosθ 
6 10π e 3 sin

2θ
4 +
3  = 54 10π e 1 +
 13
( 2π
3
)
 91 0
≈ 376.4
Exercise 4F 1 Find the curved surface area of the solid generated by rotating the curve
about the x axis. Leave your answers in terms of π .
(i) The line 4y = 3x from x 4= to x 8. =
(ii) The circle x2 + y 2 = from
a2 x = − ato x a. =
(iii) The catenary y c= cosh x from x a= − to x a = .
c
(iv) = y , = 2ap from p 1= to p 2. =
The parabola x ap 2
(v) One arch of the cycloid x a= (θsin− ), θ y a= − (1 cos . θ )

3 3
(vi) The astroid x a= cos ,p y a = sin .p
(vii) One loop of the lemniscate
(
r 2 = a 2cos 2 θ i.e. from θ = 0 to θ = π .
4 )
(viii) The cardioid r a(1 = cos +
). θ
2
2 Show the arc of the circle ( x a+ cos α ) + =y 2a2 , for which ജ x 0,
subtends an angle α2 at the centre of the circle. Find the area of the
curved surface generated when this arc is rotated about:
(i) the x axis
(ii) the y axis.
112
124
M 3 Archimedes’ tombstone
The diagram shows a sphere
circumscribed by a cylinder with
a vertical axis which touches the
sphere at its horizontal equator.
Two horizontal planes cut both the
sphere and the cylinder. Prove that
the portions of the sphere and the
cylinder between these planes have
equal curved surface areas.
Historical note
Archimedes (287–212 BC) was so pleased to discover this that, at his
request, a representation of a sphere circumscribed by a cylinder was
carved on his tombstone in Sicily, where it was found and restored by the
Roman author Cicero about a century later.
4 Use the ‘Archimedes’ tombstone’ theorem to find

(i) the surface area of a sphere of radius a
(ii) the surface area of ‘the tropics’, i.e. the part of the Earth between the
circles of latitude 23.47°N (the Tropic of Cancer) and 23.47°S (the
Tropic of Capricorn). Take Earth to be a sphere of radius 6370km,
and give your answer in km2 to 3 significant figures.
1
A solid of revolution is generated by rotating the curve y x= about the
CP 5
x axis from x 1= to x k = , where k 1 > .
(i) ()
Prove that the volume of V of this solid is π 1 1− .
k
(ii) Prove that the curved surface area S is given by
k
2π 1 1 +d .
S = ∫1 x x4
x
Noting that 1 + 1 4 > ,1deduce that S > 2π ln k.

x
Note
This gives a paradox. If an infinitely long hollow vessel of this
shape were placed with its axis vertical then a volume π of paint
poured into it would completely fill it, but no amount of paint
however large would be enough to cover the surface!
113
125
6 A curve has parametric equations
2 1 3
x t=y t , =− t , for 0  t  1.
3
Find
(i) the arc length of C,
(ii) the surface area generated when C is rotated through 2π radians
about the x-axis.
9231 Paper 11 Q8 June 2014
7 The curve C has parametric equations
= 1 t 4 − ln t,
2
x t=y ,
4
for 1  t  2. Find the area of the surface generated when C is rotated
through 2π radians about the y-axis.
8 The curve C has equation y = 2 sec x, for ഛ x π 1 . Show that the

0 ഛ
4
arc length s of C is given by
1 π
S = ∫ 4
0
(2 sec2 x − 1) dx.
Find the exact value of s.
The surface area generated when C is rotated through 2π radians about
the x-axis is denoted by S.
Show that
1 π
S = 4 π ∫ 0 (2 sec3 x − sec x)dx,
4
(i)
(ii) d (sec xtan ) x2sec

= 3
x − sec .x
dx
Hence find the exact value of S.
9231 Paper 11 Q11 June 2013
114
126
KEY POINTS
1
∫( ) 2
1
a +x
2 d x =
a 1 tan−1 x + c
a ()
2 ∫ (sinh
) = xd x cosh x c+
3 ∫ (cosh xd)x = sinh x c+
4 ∫ (sech x)xd=
2
tanh x c+
 
5 ∫  2 2  dx = sin −1 xa + c
1
 a −x 
()
 
6 ∫  2 2  dx = sinh −1 xa + c
1
 a +x 
()
 
7 ∫  2 2  dx = cosh −1 xa + c
1
 x −a 
()
8 Trigonometrical and hyperbolic substitutions can be used to integrate
related functions.
9 Reduction formulae can be used to simplify definite integrals.
10 Rectangles can be used to approximate integrals
● to set bounds for integrals
● to derive inequalities
● to derive limits concerning sums.
11 Arc length can be found by integrating
ds = dy  ds =
2
 d x 2
● 1 +   or   + for1 Cartesian form
dx dx dy  dy 
2 2
●
ds =  d x  +  dy  for parametric form
dp  dp   dp 
2
ds dr
r 2 for polar form.
●
dθ
= ()
+
dθ
12 The volume of solid of revolution formed by rotating about the
b B
2
x-axis is ∫ aπy x d and its surface area is ∫ Aπy
2 ds .
13 The corresponding results for a solid formed by rotating about the
d D
2
y-axis is ∫ c πx y d and its surface area is ∫ Cπx2 sd .
115
127
LEARNING OUTCOMES
■ integrate hyperbolic functions
■ use inverse trigonometric functions to integrate functions of the
form 1
2 2
a −x
■ use inverse hyperbolic functions to integrate functions of the form
1 1
and x
a2 + x 2 2
− a2
■ use completing the square to rewrite integrals into the correct form
■ derive and use reduction formulae to evaluate definite integrals
■ understand how area can be approximated by the areas of rectangles
■ use rectangles to
■ estimate the area under a curve
■ derive bounds for an integral
■ derive inequalities or limits concerning sums
■ use integration to find arc length

■ with equations given in Cartesian form
■ with equations given parametrically
■ for curves given in polar form
■ use integration to find surface areas of revolution about one of the axes
■ with curves given in Cartesian form
■ with equations given parametrically.
116
128
5 Complex numbers
The shortest
path between
two truths in
the real domain
passes through
the complex
domain.
Jacques
Hadamard
(1865–1963)
Complex numbers may appear to be a mere mathematical curiosity but this

is far from the truth. They have many applications in the real world. For
example, electrical engineers use i to analyse oscillating currents. Physicists
have found that imaginary numbers provide the best language for describing
some real-world phenomena, such as the flow of fluid around a pipe or
solutions to differential equations modelling shock absorbers.
5.1 The modulus and argument of a

complex number
Figure 5.1 shows the point representing z = x + yi on an Argand diagram.
Im
x + yi
Note
The work in this chapter follows on
from the complex numbers covered
y in Chapter 11 of Pure Mathematics 2
and 3. However, the first few pages of
this chapter revisit some of the work
that was done there and the ideas
O x Re underpin the rest of the chapter.
▲ Figure 5.1
117
129
The distance of this point from the
origin is x 2 + .y 2 Using Pythagoras’ theorem.
This distance is called the modulus of z, and is denoted by |z|.

22
So, for the complex number z = x + yi, |z| = x y +.
Notice that since zz∗ = (x + iy)(x − iy) = x2 + y2 , then |z|2 = zz∗.
Figure 5.2 shows the complex number z on an Argand diagram. The length
r represents the modulus of the complex number and is denoted |z|. The
angle q is called the argument of the complex number.
Im
z When describing complex

numbers, it is usual to give
r the angle θ in radians.
θ
O Re
▲ Figure 5.2
The argument is measured anticlockwise from the positive real axis.

This angle is not uniquely defined since adding any multiple of 2π to q gives
the same direction. To avoid confusion, it is usual to choose that value of q for
which −π < q < π. This is called the principal argument of z and is denoted
by arg z. Every complex number except zero has a unique principal argument.
Always draw a diagram when finding the
argument of a complex number. This tells you The argument of
zero is undefined.
which quadrant the complex number lies in.
In Figure 5.3, you can see the relationship between the components of a
complex number and its modulus and argument.
y
Using trigonometry, you can see that sinθ = and so y = r sin q.
r
x
Similarly, cos θ= r so x = r cos q. Im
Therefore, the complex number z = x + y i can

be written
r
z = r cos q + r sin q i y
or z = r (cosq + i sin q)
θ
O x Re
This is called the modulus–argument
or polar form of a complex number and is
▲ Figure 5.3
denoted (r, q ).
?
› Give one similarity and one difference between modulus-argument
form and polar coordinates.
118
130
Multiplying and dividing complex numbers in
modulus–argument form You may need to add
To multiply complex numbers in modulus- or subtract 2π to give
the principal argument.
argument form, you multiply their moduli and
For example, if
add their arguments. 7π
arg (z)1 + arg (z)2 =
= z1z2 3
z1z2 π.
then arg (z)1z2=
arg (z)1z2= arg z1 + arg z 2 3
You can prove these results using the compound angle formulae.
To divide complex numbers in modulus-argument form, you divide their
moduli and subtract their arguments.
z1 z
= 1
z2 z2
z 
arg  1  = arg z1 − arg z 2
 z2 
z
You can prove this easily from the multiplication results by letting 1 = ,wso
z2
that z 1 = wz 2.
The effect of multiplication by a complex number in

an Argand diagram
Much of the geometrical power of complex numbers comes from the
result about multiplication of complex numbers in polar form:‘multiply the
moduli, add the arguments’.
ACTIVITY 5.1
(i) Write the numbers i and −2 in modulus-argument form.
(ii) Using the result about multiplication of complex numbers in
modulus–argument form investigate:
(a) multiplication of a complex number z by i
(b) multiplication of a complex number z by −2.
You will have found in Activity 5.1 that multiplication of complex numbers
in modulus-argument form gives rise to a simple geometrical interpretation
of multiplication.
Im
To obtain the line representing z 1 z 2 enlarge the
This combination z 1z 2
of an enlargement
line representing z 2 by the scale factor |z 1| and
rotate it through arg z 1 anticlockwise about O r 1r 2
followed by a
rotation is called a (see Figure 5.4).
spiral dilation.
In modulus-argument form, you can say that θ1 r2
multiplication by r (cosq + i sinq) corresponds to θ2 z2
O Re
enlargement with scale factor r with (anticlockwise)
rotation through q about the origin. ▲ Figure 5.4
119
131
Example 5.1 Explain the geometrical effect of multiplying a complex number z by:
(i) −3i
(ii) − 3 3 3i+
Solution
(i) −3i has modulus 3 and argument 2 −π.
Multiplying by −3i would enlarge the vector z by scale factor 3 and
rotate it through 2π radians clockwise (or 3 π radians anticlockwise).
2
π
(ii) − 3 3 3i+has modulus 6 and argument 5 .
6
Multiplying by − 3 3 3i+would enlarge the vector z by scale factor 6
π
and rotate it through 5 radians anticlockwise.
6
Exercise 5A 1 Write each of the following complex numbers in the form x + yi, giving
surds in your answer where appropriate.
(i) ( ( ) (i sin) 4− )
π
6 cos 4− +
π
( ()() )
(ii) π
3 cos 6− + − i sin 6
π
6 )
2 cos 3 + i sin 3 ) 7 cos 5(− +)−( i)sin 5
π π π π
2
(iii) ( 4 4
( (iv)
6
For each complex number, find the modulus and principal argument,
and hence write the complex number in modulus-argument form. Give
the argument in radians, as a multiple of π.
(i) 2 3 2i+ (ii) −2+ 3 2i
(iii) 2 3 2i− (iv) −2 −3 2i
3 Represent each of the following complex numbers on a separate Argand
diagram, and write it in the form x + yi, giving surds in your answer
where appropriate.
π π
(i) |z| = 3, arg z = 4 (ii) |z| = 5, arg z = 2
5π π
(iii) |z| = 4, arg z =− 6 (iv) |z| = 6, arg z = – 4
4 ( ( ) (i )sin 3− )and w 3=cos( 5

π
Given that z 5= cos 3 − +
π
6
π
+ i sin 5π ,
6 )
find the following complex numbers in modulus-argument form:
(i) wz (ii) w (iii) z (iv) 1z
z w
5 Explain the geometrical effect of multiplying a complex number z by:
−
(i) 5 5i (ii) − −
1 1 3i
4 4
120
132
M 6 Write down the complex number w such that the product wz represents
the following transformations of z:
(i) an enlargement by scale factor 2 and a rotation of 3π radians
anticlockwise
(ii) an enlargement by scale factor 1 and a rotation of 2 π radians
3 3
clockwise
5.2 De Moivre’s theorem

On page 119 you saw that when you multiply two complex numbers in
modulus-argument form you multiply their moduli and add their arguments.
ACTIVITY 5.2
+ 0.1) write down z2 , z3 , z4
For the complex number =z 2(cos0.1 i sin
and z5 .
Use your answers to write down an expression for zn .
For the general complex number z = r (cos q + i sin q) write down an
expression for zn .
The product of:

( θ 1 + isin θ 1 ) and z 2r2cos
z1 = r1 cos = ( θ 2 + isin θ 2 )
is: ( (θ 1θ+
z1z2 = r1 r2 cos 2 ) + isin (θ 1θ2+ ))
Using this result repeatedly with a single complex number z with modulus 1
allows you to concentrate on what happens to the argument.
If =z cos θ + isin θ
then z = cos (θ θ+ +) isin (θ θ+ =) cos 2θ isin
+ 2 θ
2
= 2 ( θ θ+ +) i sin(2θ θ+ =) cos 3θ isin

z 3 = z 2z cos + 3 θ
and so on.
n
The diagram below shows (cos θ + isin θ) for n = 1, 2, 3, …, 6.
Im
z3
z2 In this example 0 < <
θ π.
2
z4
z
z5
θ
Re
z6
▲ Figure 5.5
121
133
?
› How does the pattern on the Argand diagram continue when higher
powers of n are added?
This suggests the following result which is called de Moivre’s theorem.

If n is an integer then (cos q + i sin q)n = cosn q + i sin n q
The proof of this result is in three parts, in which n is 1 positive 2 zero
3 negative.
1 When n is a positive integer de Moivre’s theorem can be proved by

induction.
The theorem is obviously true if n = 1.
Assume the result is true for n = k, so
(cosθ + isin θ )k = cos kθ + i sinkθ
You want to prove that the result is true for n = k + 1 (if the assumption
is true).
k +1
(cosθ + isin θ ) = (cos kθ + isin k)(cos
θ θ isin +
) θ
= cos (kθ θ+ +) isin (kθ θ+ )
= cos ((k + 1)θ ) + i sin((k) + 1 θ )
If the result is true for n = k, then it is true for n = k + 1. Since it is true

for n = 1, it is true for all positive integer values of n.
e
2 By definition, z0 = 1 for all complex numbers z ≠ 0.
Therefore ( cosθ + isin θ ) = =1 cos0 i sin

0
+0 .
3 For negative n the proof starts with the case n = −1.

If a × b = 1 it follows As (cos θ + isin θ)(cos (−)θ+ isin (−)θ= ) cos (θ θ− +) isin (θ)θ− = 1 it
that b is the follows that
reciprocal of a. −1
(cosθ + isin θ ) = cos (−)θ+ i sin(−)θ †
If n is another negative integer, let n = −m where m is a positive integer.

−
Then (cosθ + isin θ ) (= cos θi sin )m
n
+ θ
 cos θisin m1−
+ θ) 
As m is a positive integer = (
de Moivre’s theorem
holds using part 1.  mθ + isin mθ −1
= cos
= cos (−)m+θ isin (−)mθ
Using † with mq in
place of q.
= cos nθ + isin nθ
Therefore de Moivre’s theorem holds for all integers n.
122
134
De Moivre’s theorem can also be used for simplifying powers of complex
numbers when the modulus is not 1, as in Figure 5.6. If z r=cos( isinθ + θ)
then
z n ==[r (cos θθ ++isin θθ)] ==r n (cos n θθ ++i sin n θθ)
n
Im
z3 z2
z4 z
r
θ
Re
z5
z6
▲ Figure 5.6
Example 5.2 Use de Moivre’s theorem to simplify each of the following.

π+ π 18
cos 12 i sin 12 ) (
−1+3i )
5
(i) ( (ii)
Solution
(i) (
π+ π 18 =
cos 12 i sin 12 )×π+
cos 18 12 ( ) ( ) i sin 18 12× π
= cos 3π + i sin 3π = − i
2 2
= cos 3π + i sin 3π = − i
2 2
(ii) First convert to modulus-argument form:
z = − 1+ ( )1
3i ⇒ z = − +
2
( )3 2
= 2
 3
−1
tan   = π so arg z = π2 . z is in the second quadrant.
 1  3 3
( ) ( 2 cos 10 3π + i sin 103π )

5
π + π =
So (−1+ 3i )
5
= 2 cos 2
5
i sin 2 5
3 3
 1 3 
= 32 − − i
 2 2 
= − 16
− 16 3i
Example 5.3
4(cos5 θ + isin5 θ ) 5
Simplify the expression .
4
 3(cos4 θ + isin4 θ ) 
➜
123
135
Solution
5 4(cos5 θ + isin5 θ ) 5
3(cos4 θ + isin4 θ ) 4
=
1024 (cos 25θisin25
+
81(cos16 θisin16
+ θ
θ
)
)
= 1024 (cos9 θi sin

+ 9 θ)
81
In Activity 5.3 you will prove the useful result:
(cos φφ −−i sin φφ)n ==cos nφφ −−i sin nφφ
Activity 5.3
Use de Moivre’s theorem with q = −f and the facts:
» cos (−)θ= cos θ
» sin (−)θ= − sin θ
to show that (cos φ − isin φ ) = cos nφ − i sinnφ .

n
Historical note
Abraham de Moivre (1667–1754) went to England from France as a Huguenot
refugee at the age of eighteen and spent the rest of his life in London. In papers from
1707 onwards he made use of ‘his’ theorem, though he never published it explicitly.
Exercise 5B 1 Use de Moivre’s theorem to simplify each of the following:

(i) in the form cos  + i sin 
(ii) in the form a + ib
) )
4 −8
cos π + i sin π cos π + i sin π
(a) ( 6 6
(b) ( 3 3
(cos (π− 12) (+ i)sin π− 12 )

10
(c)
2 Given that =w cos π + i sin π, write each of the following complex

4 4
numbers as a power of w:
(i) z1 = cos 3π + i sin 3π (ii) z 2 = cos π + i sin π
4 4 2 2
(iii) z 3 = cos π +i sin π
Illustrate w, z 1, z 2 and z 3 on an Argand diagram.
3 Use de Moivre’s theorem to simplify each of the following:
(cos 4π + i sin 4π)

3
(cos 3θ + i sin 3θ ) 4
(i) (ii)
(cos5 θ + i sin 5θ )3 (cos 6π + i sin 6π)
2
)(cos 6π + i sin 6π)

5 −4
cos π + i sin π
124
(iii) ( 3 3
136
4 Given that w cos= π + i sin π , write each of the following complex
6 6
numbers as a power of w.
(i) z cos π − i sin π (ii) z cos π − i sin π
1= 6 6 2= 2 2
Illustrate w, z 1 and z 2 on an Argand diagram.
5 By converting to modulus-argument form and using de Moivre’s
theorem, find the following in the form x + y i giving x and y as exact
expressions or correct to 3 significant figures.
1 − 3i )
4
(ii) −( 2+2i )
7
(i) (
(27 3i )
6
+
(iii)
6 Without using a calculator write (
− −3 i ) in the form x + y i where x
7
and y are exact values.

7 Simplify the following expressions as far as possible.
5
(ii)  i(cos 3θisin θ ) 
4
(i) 3(cos
 2θ + i sin 2θ   + 3
 (cos7 θ + i sin7θ ) −3
 2i
(iii)
8 Show that
4
3(cos
 θ
2 − i sin2 2(cos
θ   θ + isin θ  5
2 8
4(cos
 θ +
3 i sin3 θ   
1 (cos θ − isin θ
)  
2
can be expressed in the form k(cos θ − isin θ) where k is a constant to be
found.
9 The three complex numbers in the diagram below each have modulus 1.
They form an equilateral triangle centred on the origin.
(i) θ isin
Write down each of z 1, z 2 and z 3 in the form cos + θ.
(ii) Use de Moivre’s theorem to show that Im
the cube of each of these complex z2
numbers is the real number 1.
(iii) Draw an Argand diagram to show z1
five complex numbers for which O 1 Re
z5 = 1, and write down these complex
z3
numbers in the form cos θ + isin θ .
5.3 The nth roots of a complex number

The nth roots of unity
You already know that all quadratic equations have two roots (which may be
a repeated root, or they may be complex roots). You have also solved cubic
equations to find the three roots, some of which may be complex.
As early as 1629 Albert Girard stated that every polynomial equation of
degree n has exactly n roots, including repeated roots. Some of these roots
125
137
may be complex numbers. This was first proved by the 18-year-old Carl
Freidrich Gauss 170 years later.
Therefore even the simple equation zn = 1 has n roots. One of these roots is z = 1
and, if n is even, then z = −1 is another. All the other roots are complex numbers.
In this section you will look at methods for finding the other roots of the
equation zn = 1, and the relationship between them.
Example 5.4 (i) Write down the two roots of the equation z 2 = 1 and show them on an
Argand diagram.
(ii) Use z 3 − 1 = (z − 1)(z 2 + z + 1) to find the three roots of z 3 = 1. Show
them on an Argand diagram.
(iii) Find the four roots of z 4 = 1 and show them on an Argand diagram.
Im
Solution
(i) Using properties of real numbers
z = ±1. These numbers are shown in
Figure 5.7.
O Re
(ii) The equation z3 = 1 can be rewritten

▲ Figure 5.7
as z3 − 1 = 0.
So z( − 1)(z + +z = 1)
2 This result is given in
0
the question.
One of the roots is z = 1.
1 3i
The equation z2 + z +1 has roots =z − ± .
2
The three roots
Using the
z = 1, z = − 1± 3i quadratic formula.
2
are shown on an Argand diagram in Figure 5.8.
Im
O Re
▲ Figure 5.8
126
138
(iii) z4 = 1 can be written in the form z
4
−=
1 (z 2
− 1)(z 2 + which
1) has
the four roots z = ±1, z = ±i.
These roots are shown on the Argand diagram in Figure 5.9.
Im
O Re
▲ Figure 5.9
In the previous example you may have noticed that the roots of the equations
z2 = 1, z3 = 1 and z4 = 1 all lie on a unit circle, centred on the origin, with
one root at the point 1.
In fact, every root of the equation zn = 1 must have unit modulus, as
otherwise the modulus of zn would not be 1. So every root is of the form
z = cos q + i sin q, and
1 ( cosθ + isin θ )n = 1
zn = ⇒
Using de Moivre’s
theorem. ⇒ cos nθ + i sinnθ = 1
⇒ nθ = 2kπ where k is any integer.
Since in modulus–
argument form 1 So, for example, in the case when n = 3 the result nq = 2kπ gives roots as
can be written follows:
(1, 0) or (1, 2π) or
(1, 4π) etc. when k 0= 3θ = ⇒
0 = θ 0 + 1
so =z cos0 i sin0 =
2π z cos 2π + i sin 2π = − +1 3
Note that these when k 1= 3θ = 2π ⇒ =
θ
3 so = 3 3 2 2
are the same 4π
roots of z3 = 1 as when k 2= 3θ = 4π ⇒ = θ so z cos 4π + i sin 4π = − −1
=
3
3 3 3 2 2
those obtained in
Example 5.4. For larger values of k the same roots are obtained, so k = 3 gives the root z = 1,
k = 4 gives the root z = − +1 3 , and so on.
2 2
So, when n = 3, the values of k that need to be considered are k = 0, 1 and 2.
Generally, as k takes values 0, 1, 2, 3, …, (n − 1) the corresponding values of q are:
(n − 1)π
n n, 6π ,…, 2
0, 2π , 4π
n n
giving n distinct values of z.
When k = n then q = 2π, which gives the same z as q = 0. Similarly, any
integer value of k larger than n will differ from one of 0, 1, 2, 3, …, (n − 1)
127
139
by a multiple of n, and so gives a value of q differing by a multiple of 2π from
one already listed; the same applies when k is any negative integer.
Therefore, the equation zn = 1 has exactly n roots. These are:
z = cos 2kπ i+sin 2 π k, , 1−(n)
k = 0, 1, 2, 3, …
n n
These n complex numbers are called the nth roots of unity. They include
z = 1 when k = 0 and, if n is even, z = −1 when k n= .2
It is usual to use w (the Greek letter omega) for the root with the smallest
positive argument:
ω = cos 2π i+sin 2π
n n
Then, by de Moivre’s theorem:
ω k = cos 2kπ i+sin 2 π k
n n
so the nth roots of unity can be written as:
Im
−
1, ω, ω, 2, … ω n 1
ω2
ω3
These complex numbers can be
represented on an Argand diagram ω
by the vertices of a regular n-sided
ω4
polygon inscribed in the unit circle,
with one vertex at the point 1. 1
O Re
Figure 5.10 shows the nth roots of
unity when n = 9. ω5
The sum of all of the nth roots of ω8

unity is a geometric series with
ω6
common ratio w: ω7
1+ω
+ω 2
+ … + ω n −1
ω n ▲ Figure 5.10
= −1
1−ω
Using the formula
=0 Since ω n
= 1. a (r1 − )
n
Sn = .
So the sum of the nth roots of unity 1− r
is always zero.
Activity 5.4
Verify that the sum of the nth roots of unity is equal to zero in the cases
where n = 2, n = and
3 n = 4.
128
140
Example 5.5 Solve the equation z = 1. Show the roots on an Argand diagram.
6
Solution
The sixth roots of unity are given by:
z = ω k = cos 2kπ + i sin 2kπ
6 6
where k = 0, 1, 2, 3, 4, 5.
This gives the following roots z:
k 0= z 1=
k 1= cos 2π + i sin 2π = +1
z = =ω 3i
6 6 2 2
k 2= z = ω 2 = cos 4π + i sin 4π =− +1 3i
6 6 2 2
k 3= z = ω 3 = cos 6π + i sin 6π =− 1
6 6
k 4= z = ω 4 = cos 8π + i sin 8π =− −1 3i
6 6 2 2
k 5= z = ω 5 = cos 10π + i sin 10π = −1 3i
6 6 2 2
The roots form a hexagon inscribed within a unit circle, as in Figure 5.11.
Im
1
−1 O 1 Re
−1
▲ Figure 5.11
The nth roots of any complex number

To find the nth roots of any given non-zero complex number a you have to
find z such that zn = a. The method to find the nth roots is the same as that
in the previous section on nth roots of unity, adjusted to take account of the
modulus s and argument f of a.
Suppose, for example, you wanted to find the fifth roots of the complex
number a = −1 + i.
You want to find the complex number z such that z5 = −1 + i.
Let =z r(cos θ + isin θ), then: Writing −1 + i in
modulus–argument form.
( 4 4)
r(cos θ + isin θ ) 5 = 2 cos 3π + i sin 3π
Using de Moivre’s
( θ + i sin 5θ ) =
5
⇔ r cos5 (
2 cos 3π + i sin 3π
4 4 ) theorem.
129
141
Two complex numbers in modulus-argument form are equal only if they have the
same moduli and their arguments are equal or differ by a multiple of 2π. Therefore:
1
r 2= and 5θ = 3π + 2kπ , where k is an integer.
10
4
Each of the roots will have the same modulus and so will lie on the circle
1
z = 2 .10
Larger values of k
3π + 2kπ generate the same
The argument of z is θ = 4 . set of arguments
5 so, for example,
As k takes the values 0, 1, 2, 3 and 4 the arguments
k = 5 gives 20
43π
3π
obtained are 3π ,, 11π
11π,, 19π 27π and 7π .
19π,, 27π
20
20 20 20 20 20 20 20 4 which is equivalent
The five roots are shown in Figure 5.12 and form to 3π .
20
the five vertices of a regular pentagon.
Im
−1 0 1 Re
−1
▲ Figure 5.12
In a general case, suppose =z r(cos θ + isin θ) and =a s(cos φ + isin φ).
Then:
n n n
z =⇔
a r (cos θ + isin θ) = s(cos φisin+ ) φ
⇔ r n (cos nθ + isin n)θ(cos
= isin
s )φ + φ
⇔ r n = s and nθ φ= + 2kπ, where k is an integer
Since r and s are positive real numbers the equation r n = s gives the unique
1 1
value r s = n
so all the roots lie on the circle z s = . n
φ + 2kπ . As k can take the values 0, 1, 2, …, n − 1,
The argument of z is =θ
n
this gives n distinct complex numbers z and, by the same argument as for the
roots of unity, there are no other roots.
Therefore the non-zero complex number s(cos φ + isin φ
) has precisely n
different nth roots, which are:


1
s  cos
n
()()
φ + 2kπ
n
+ i sin
φ + 2kπ 
n


where k = 0, 1, 2, …, n − 1.
130
142
You may also express these n roots as α ,αω
, αω
, ,
2
… αω n −1 where:
( )
1
φ φ
α = s n
cos and ω = cos 2π i+sin 2π
+ isin
n n n n
The sum of these nth roots of a is:
+ αω+
α αω 2
+ … + αω n−1 =
α (1 − ω n
)=0 Since ω
n
= 1.
1−ω
Example 5.6 Represent the five roots of the equation z5 = 32 on an Argand diagram. Hence,
− )5 = 32 on an Argand diagram.
represent the five roots of the equation (z 3i
Solution
You know mod 32 = 32 and arg(32) = 0 so the fifth roots of 32 are given by:
1
5
( (
32 cos 0 2π
+
5 )( + k
k + i sin 0 2π
5 ))
Each root has modulus 2 and the arguments of the roots are
0, 2π , 4π , −4π and −2π. The roots form a regular pentagon inscribed
5 5 5 5
in a circle, centre the origin and radius 2, as shown in by the blue points in
Figure 5.13.
The centre of the red
Im
pentagon is 3i. It is a
translation of the blue
pentagon.
4
−4 −3 −2 −1 0 1 2 34 Re
−1
−2
−3
−4
▲ Figure 5.13
The roots of (z − 3i )5 = 32 are therefore represented by the same pentagon

inscribed in a circle, centre 3i, radius 2 as shown by the red points in Figure 5.13.
131
143
Exercise 5C 1 Find the roots of the equation z5 = 1 giving your answers to 3 significant
figures.
Show these roots on an Argand diagram.
Describe the polygon formed by the points representing the roots.
2 Find the roots of the equation z8 = 1 and show these on an Argand
diagram.
Describe the polygon formed by the points representing the roots.
3 Find both square roots of −7 + 5i, giving your answers in the form
x + yi with x and y correct to 2 decimal places.
4 Find the fourth roots of −4, giving your answers in the form x+yi.
Show the roots on an Argand diagram.
5 In this question, give answers as exact values or to two decimal places
where appropriate.
(i) Find the cube roots of 1 − i.
(ii) Find the fourth roots of 2 + 3i.
(iii) Find the fifth roots of −3 + 4i.
6 Explain geometrically why the set of tenth roots of unity is the same as
the set of fifth roots of unity together with their negatives.
7 One fourth root of the complex number w is 2 + 3i. Find w and its other
fourth roots and represent all fourth roots on an Argand diagram.
8 Solve the equation z 3 − 4 3 4i+0= giving your solutions in the form
r(cos q + i sin q ), where r 0 and −π q π.
9 A regular heptagon on an Argand diagram has centre −1 + 3i and one
vertex at 2 + 3i.
Write down the equation whose solutions are represented by the vertices
of this heptagon.
10 (i) Find the fourth roots of −9i in the form r(cos q + i sin q ) where
r > 0 and 0 < q < 2π.
Illustrate the roots on an Argand diagram.
(ii) Let the points representing these roots, taken in order of q
increasing, be P, Q, R and S. The midpoints of the sides of the
quadrilateral PQRS represent the fourth roots of a complex
number w. Find the modulus and argument of the complex number
w and mark the point representing w on your Argand diagram.
11 If w is a complex cube root of unity, w ≠1, prove that:
(i) (1 + w)(1 + w2 ) = 1
(ii) 1 + w and 1 + w2 are complex cube roots of −1
(iii) (a b+ a )( + ωb )a( + ω b ) = +a b
2 33
(iv) (a b+ c+a )( + ωb + ω 2c )(
a + ω 2b c+aωb )c abc
= + 3+3 3− 3
132
144
12 Letω=cos 1 + 1
π isin π . Show that ω 5 + =1 0 and deduce that
5 5
ω4 − ω3 + ω2 − ω = − 1.
= 2 cos 1π and ω 3ω − = 2
4 3
Show further that −ω ω 2 cos π.
5 5
+ 3 π and cos 1 π cos 3 π.
Hence find the values of cos 1 π cos
5 5 5 5
Find a quadratic equation having roots cos 1 π and cos 3 π and deduce
5 5
the exact value of cos 1 π.
5
Note 5.4 Finding multiple angle identities

You need to be
familiar with
using de Moivre’s theorem
trigonometric When they were first introduced, complex numbers were not generally
identities such as
accepted by mathematicians. However, during the eighteenth century, their
cos 2θ + sin 2θ ≡ 1
and the double usefulness in producing results involving only real numbers was recognised.
angle formulae. The results could also be obtained without using complex numbers, but often
only with considerably more effort. One of these results is finding expressions
for the sine or cosine of multiple angles using de Moivre’s theorem, as shown
in the following example.
Example 5.7 Express cos 5q in terms of cos q.
Solution
Where c and s are
By de Moivre’s theorem: used as abbreviations
for cosq and sinq
cos5 θ + i sin 5θ = (cosθ + i sinθ )
5
respectively.
= +c 5 4 32 23 45
5ic s − 10c s − 10i c s + 5cs s + i
Equating the real parts gives:
cos5 θ = −c 5 10c 32s + 5cs 4
= −c 5 10c1 ( − c ) +( 5c1 − c )
3 2 22
Using sin 2cos
2
θ +1 θ ≡ .
= −c 5 10c 3 + 10c 5 + −5c10 c 3 + 5c 5

Therefore cos5θ = 16cos θ − 20cos θ + 5cos θ .
5 3
Activity 5.5
By equating the imaginary parts in Example 5.7, find sin5q in terms of sinq.
133
145
Example 5.7 expressed cos 5q in terms of powers of cos q. Sometimes it is
useful to do the reverse, for example if you wanted to integrate cos5 q.
The next example shows how this can be expressed in the form
a cos 5q + b cos 3q + c cos q, which is much easier to integrate.
To do this, you need expressions for cos nq and sin nq in terms of zn and z−n .
You can deduce these expressions from de Moivre’s theorem as follows:
If:
=
z cos θ + isin θ
then: As cos (−)θ= cosθ

z n = cos nθ + isin nθ and sin (−)θ= − sinθ.
and:
−n
z = cos (−)nθ + isin (−)nθ = cos nθ − i sinnθ
−n
Adding these two expressions gives +z nz = 2cos nθ
and so:
n −n
z + z
cos nθ =
2
n −n
Subtracting the two expressions gives z z− = 2i sinnθ
and so:
n −n
z − z
sin nθ =
2i
Example 5.8 Express cos5 q in terms of multiple angles.
Solution nn −
A rearrangement of cos nθ =
z z+
Let =z cos θ + isin θ. 2
with n = 1.
−1
Then 2cos θ = +z z .
⇒ (2cos θ )5 = +(z z )
−1 5 Expanding the
right-hand
− − −5 side using
⇒ 32cos 5θ = z 5 + 5z 3 + 10 z10+ z 1 + 5z z3 + the binomial
= (z 5 + z −5 ) +( 5 z 3z+ −3
) +(10 z z+
−1
) expansion.
= 2cos5 θ + 10cos 3θ20cos

+ θ nn −
Using cos nθ = z z+
θ + 5cos 3θ10cos
+ θ
cos θ = cos5 2
5
⇒
16 three times, with n = 1,
n = 3 and n = 5.
Activity 5.6
Use a similar method to that used in Example 5.8 to express sin5q in
terms of multiple angles.
134
146
Exercise 5D 3
1 (i) Using the result (cosθ + isin θ) cos= 3 i sinθ 3+ θ , compare real
and imaginary parts to show that:
3
cos 3θ = 4 cos θ − 3cos θ
and:
sin 3θ = 3sin θ − 4 sin3 θ
(ii) Hence express tan 3q in terms of tanq.
2 Let =z cos θ + isin θ.
(i) Write down expressions for z3 and z−3 .
33 −
z z+
(ii) Use your expressions from (i) to show that cos 3θ = and
3
− − 2
z z 3.
sin 3θ =
2i
3 Let =z cos θ + isin θ.
(i) Write down an expression for z−1 .
−1
(ii) (a) Use your answer to (i) to show that 2cos θ = +z z .
(b) Using the result in part (a), express cos4 q in terms of multiple
angles.
−
(iii) (a) Use your answer to (i) to show that 2i sin θ = −z z 1 .
(b) Using the result in part (i), express sin5 q in terms of multiple
angles.
sin 6θ
4 Find cos6q and sinθin terms of cos q.
5 Find an expression for sin6 q in terms of multiple angles.
6
Hence evaluate ∫ θsin
θ
d.
6 Express cos4 q sin3 q in terms of multiple angles and hence find
∫ cos 4
θ θ θsin d3 .
7 By first using de Moivre’s theorem, evaluate:
π
∫ 0
6 5
cos dθ θ
PS 8 (i) Use de Moivre’s theorem to show that:

5 3
cos5θ = 16cos θ − 20cos θ + 5cos θ
(ii) Given cos 5q = 0 but cos q ≠ 0, use your answer from (i) to find two
possible values for cos2 q. Give your answers in surd form.
(iii) Use (ii) to show that:
1
 + 5 2
cos18 ° =5
 8 
and find, in a similar form, an expression for sin18°.
135
147
n −n
z + z
CP 9 Use cos nθ = to express:
2
cosθ + cos 3θ + cos5 θ + … + cos (2 n1 − )θ
as a geometric series in terms of z.
Show that the sum of the geometric series can be expressed in the form
sin (2)nθ
.
2 sinθ
1
CP 10 (i) Given that z = cos q + i sin q, write down zn and z
n in the form
a + ib.
n 1 n 1
Simplify z + n and z − . n
z z
()()
2 4
1 1
(ii) Expand z z+ z z− and hence find the constants p, q, r
and s such that:
4 2
sin θcos θ = +p q cos 2θ + r cos4 θ + s cos6 θ
(iii) Using a suitable substitution and your answer to part (ii), show that:
2
4π +
∫1 x 4
4 − x 2
d x =
3
3
11 Letz=cos θ + isin θ. Show that

1+=
z
2 (
2cos 1 θ cos 1 θ + i sin 1θ .
2 2 )
By considering (1 + z)n , where n is a positive integer, deduce the sum of
the series
n n n
  sin θ +   sin 2θ + … +   sin n.θ
1 2 n

12 Use de Moivre’s theorem to show that
cos 5θ ≡ cos θ16( sin 4θ − 12 sin2θ + 1.)
By considering the equation cos 5θ = 0, show that the exact value of
(10 π) is 3 −58
sin21 .

13 Prove by mathematical induction that, for every positive integer n,

(cos θ + isin θ) n = cos nθ + isin nθ .
Express sin5θin the form p sin 5θ + q sin 3θ + r sin θ , where p ,q r, and
are rational numbers to be determined.
136
148
5.5 The form z = reiq
In Chapter 3, you saw that the series expansions of sin q, cos q and e x can be
written as:
r 2 r1+
θ 3 + θ 5 − θ 7 + … (−)1 θ
sinθ θ= −
3! 5! 7! (2r1+! )
θ 2
θ 4
θ 6
(−)1 r θ 2 r It can be shown
cosθ = −1 2! 4! 6!
+ − +… that this series
(2)r!
expansion is also
2 3 4 r
true for complex
x
e = +1 + x x + x + x +… + +x… powers.
2! 3! 4! r!
Replacing x by iq in the expansion for e x gives:
(iθ) (2 ) ( ) i(θ )3( ) iθ
θ θ4 5 6
e iθ
= +1 +
i θ + + i
+ + i +…
2! 3! 4! 5! 6!
22 33 44 5566
=1 +i + θ i θ + i θ + i θ + i θ + i θ +…
2! 3! 4! 5! 6!
θ 2 − iθ 3 + θ 4 + iθ 5 − θ+…
6 Collecting together
= +1 −
i θ real and imaginary
2! 3! 4! 5! 6!
terms.
 θ 2
θ 4
θ 6   θ 3θ5 
= −1 2! 4! 6!
+ − + … + i θ3!− + −…
5! 
   
Therefore: Using the series expansions

iθ
e = cosθ + i sinθ for cos q and sin q.
and so:
=
z r(cos θ + isin θ)
This is called the exponential
can be rewritten as: form of a complex number with
iθ
z re= modulus r and argument q.
This format is simply a more compact way of writing familiar expressions, as

the modulus-argument form z = r(cos q + i sin q) can now be abbreviated to
z = reiq .
This form allows you to derive de Moivre’s theorem very easily for all
rational n by using the laws of indices:
(cos θ + isin θ )n = (e iθ) n
= e inθ
= cos(n)θisin(
+ ) nθ
137
149
Example 5.9
(i) Write (
z 6= cos π
6
+ i sin π )
6 in the form re iq .
(ii) Write = −z + 1 3i in the form re iq .
Solution
(i) z 6= cos π( 6 )
+ i sin π
6 has modulus 6 and argument π 6 .
iπ
=
Therefore z 6e 6
.
2π
(ii) z = − 1+ 3i has modulus 2 and argument 3.
2iπ
=
Therefore z 2e 3
.
?
+ ) e θ = r iθbe adapted for a
› How would the result: =z r (cos θisin
complex number of the form: r(cos θ − isin θ) where − <πθ < π?
In the discussion point above, you should have noticed that since:
cos (−)θ= cosθand ( )sin −θ= − sinθ
then:
r (cosθ − isin θ) = r (cos (−)θ+ isin (−)θ )
Therefore:
− iθ
r (cosθ − isin θ) = re
Activity 5.7
For a complex number =z +x y i , show that:
(i) e = e (cos y + isin )y
z x
z + 2π ni
(ii) e = ez
(iii) e
iπ
= −1
The results in Activity 5.7 are useful when simplifying results involving
exponential functions with complex exponents. Part (iii) is often written in
iπ
the form e + 1= 0which is a remarkable result that links the five numbers
0, 1, i, e and π.
The results from Activity 5.7 also give rise to two very interesting
mathematical results that are useful when working with complex numbers:
cosθ = 1 (e + e )
iθ − iθ
2 Notice that these expressions for cos q and
sin q are very similar to the definitions of
sinθ = (e − e )
1 iθ − iθ
the hyperbolic functions cosh q and sinh q.
2i
To prove these results, you can use:
θ
138 e i = cosθ + i sinθ ①
150
and:
− iθ
= cosθ − i sinθ
e
Finding ① + ② gives:
②
5
θ −θ
e i + e i = 2cos θ
so:
cos θ = 1 (e + e )
iθ − iθ
2
5.5 The form

Similarly, finding ① – ② gives:
θ −θ
e i − e i = 2i sinθ
so:
sinθ = 1 (e − e )
z = re
iθ − iθ
2i
iq
You need to learn the proofs of these results.
?
› These results are essentially the same as the ones given on page 134.
Note Explain why this is the case.
You need to be
familiar with Summing series using de Moivre’s theorem
geometric
sequences and This section shows how complex numbers can be used to evaluate certain
series, including sums of real quantities. It may be possible to do these summations without
finding the sum using complex numbers, for example by induction if you already know the
to n terms and answer, but this is a lot more difficult.
the sum to infinity
of a geometric Sometimes it is worth setting out to do more than is actually required, as
series. shown in the next example.
Example 5.10 (i) Prove that:

iθ
iθ θ
1 +e = 2cos 2 e 2
and:
iθ
iθ θ
1 −e = − 2i sin 2 e 2
(ii) Show that the sum of the series:

n  n  n 
1 +1   cos θ +   cos 2θ +   cos 3θ + … + cos nθ
  2   3
is:
θ θ
2 cos 2 cos n2
n n
139
151
Solution
iθ
5 (i) The factor e 2 on the right-hand side suggests writing each term on the
left-hand side as a multiple of e 2 .
iθ
θ i − iθ
1 e= 2 ×e 2
iθ iθ
iθ
e =e 2 ×e 2
iθ
 iθ − iθ   iθ iθ 
So: +
1e =  e × e  + e × e 2 
2 2 2
   
5 COMPLEX NUMBERS
 iθ − iθ 
iθ iθ
= e e 2 2 +e 
2
Taking out a factor of e . 2
 
iθ Using the earlier result
=e 2 × 2cos θ2
cosθ = 1 (e e + )
iθ − iθ
.
2
Similarly,
iθ
 iθ − iθ   iθ iθ 
1 −e =  e × e  − e × e 2 
2 2 2
   
 iθ − iθ 
iθ
= e e 2 2 −e 
2
 
Using the earlier result
iθ
sinθ = 1 (e e − )
iθ − iθ
=e 2 × − 2i sinθ2 2i
.
(ii) At first sight this series seems to suggest the binomial expansion
(1 +cos θ )n . The binomial coefficients 1, n n2 , …,1 are correct, but ( )( )
there are multiple angles, cosr q, instead of1 ,powers of cosines, cosr q. This
suggests that de Moivre’s theorem can be used.
The method involves introducing a corresponding sine series too.
Let: C = +1 (1n )cosθ + (n2 )cos 2θ + (n3 )cos 3θ + … + cos nθ

and: S = (n1 ) sinθ + (n2) sin 2θ + (n3) sin 3θ +… + sin nθ
Then
C + iS = +1 (1n )(cosθ + isin θ) + (n2 )(cos 2θi +sin2 ) θ + … + (cos nisin
θ + nθ
C + iS = +1 (1n )(cosθ + isin θ) + (n2 )(cos 2θi +sin2 ) θ + … + (cos nisin

θ +) nθ
= +1 (n1)e θ + (n2) ei i2 θ
+… + e inθ
Using de Moivre’s
theorem and the fact
= +1 (1n )e θ + (n2)(e)θ
i i2
+… + (e i)θ
n
irθ θ
that e = e .
i r
()
This is now recognisable as a binomial expansion, so that:
C S+ i = +(1 e )
iθ n
140
152
To find C you need to find the real part of (
1 +e iθ n
) and here the
results from part (i) are useful. Using the result:
1 +e iθ θ
= 2cos 2 e
iθ
2
5
gives:
n
 iθ  iθ n
θ θ
C S+ i = +(1 C
e S+ i) = +(12cos
iθ n
e 2 )e =  2cos 2 e 2 
iθ n
2
  
inθ inθ
θ θ
5.5 The form

n n
= 2 cos 2 e= 2 cos 2 e 2 nn 2
= 2ncos n2θ cos (θn + θ nθ nθ

= 2n2ncos ( )θ
2 2 + i sin n2
2i sincos )
z = re
Taking the real part:
θ θ
C = 2ncos n2 cos n2
iq
ACTIVITY 5.8
For Example 5.10, state the corresponding result obtained by equating the
imaginary parts.
Exercise 5E 1 Write the following complex numbers in the form z re =

i
θ where
− π< < θ π.
( )
(i) 4 cos π + i sin π
3 3
(ii) ( ()()
3 cos 5π −
6
+ i sin 5π
−
6 )
(iii) −5i
(iv) − 3 −3i
(v) 3 i−
2 Write the following complex numbers in the form x + yi:
3π i − 3π i 23π i
iπ 4 4
(i) 5e (ii) 2e (iii) 2e (iv) 5e4
3π π
3 Two complex numbers are given by =z 2e and =w 3e 3 i . 4 i
θ
Find zw and z giving your answers in the form z re= i , where
w
r > 0 and − π< θ ഛπ.
4 (i) Write the complex number w = 32i in exponential form.
(ii) Find the five fifth roots of w, giving your answers in exponential
form.
141
153
CP 5 Let:
C = +1 cos θ + cos 2θ +… + cos (n − 1)θ
5 and:
S = sinθ + sin 2θ + … + sin (n − 1)θ
(i) Find C + iS and show that this forms a geometric series with
common ratio eiq
inθ
1
(ii) Show that the sum of the series in part (i) is − e
iθ
1 −e
(iii) By multiplying the numerator and denominator of this sum by
5 COMPLEX NUMBERS
−θ
1− e i , show that:
1 −cos θ + cos (n − 1)θ − cos nθ
C = − θ
2 2cos
and find S.
θ
CP 6 (i) Show that 1+ e i2 = 2cos θ(cos
θ isin+) θ .
(ii) The series C and S are defined as follows.
n  n 
C 1 1   cos 2θ   cos4 θ + … + cos 2nθ
= + +
  2 
n  n 
S =   sin 2θ +   sin 4θ + … + sin 2nθ
 1 2 
By considering C + iS, show that:

n n
C = 2 cos cosθ nθ
and find a corresponding expression for S.
CP 7 (i) Use de Moivre’s theorem to find the constants a, b, c in the identity
cos5θ ≡ a cos 5 θ + b cos 3 θ + c cos θ.
(ii) Let:
 (2n2−π ) 
= +
C cosθ cos θ + ()()
2π
n
+ cos θ + 4π
n
+… + cos θ

+
n


and:
 (2n2−π ) 
S = sinθ + sin θ + 2π
n()()
+ sin θ + 4π
n
+… + sin θ +
 n


where n is an integer greater than 1.
Show that C + iS forms a geometric series and hence show that C = 0,
S = 0.
∗
Use the result e i = cosθ + i sinθ to prove that e z = (e z) ∗ for all
θ
CP 8
complex numbers =z +a ib.
CP 9 Let C = ∫ e cos 2
3x
dx x and S = ∫ e 3sin x
2 dx x .
(i) Find C and S by using integration by parts twice.
(3)+2i x
+ iS = e + A where A is a constant.
(ii) (a) Show that C +
3 2i
(b) Hence verify your answers for C and S from part (i).
142
154
CP 10 The infinite series C and S are defined as follows:
θ − cos 2θ + cos 3θ − cos4 θ + …
=
C cos
=
S sin
2 4 8 16
θ − sin 2θ + sin 3θ − sin 4θ + …
5
2 4 8 16
iθ
2e +1
Show that C+S i = + and hence find expressions for C and S in
5 4cos θ
terms of cos q and sin q.
5.5 The form

KEY POINTS
1 The modulus r of z = +x iy is z = x 2y2+ . This is the distance
of the point z from the origin on the Argand diagram.
z = re
2 The argument of z is the angle q, measured in radians, between the
iq
line connecting the origin and the point z and the positive real axis.
3 The principal argument of z , arg z, is the angle q, measured in radians,
for which − π< θ ഛπ, between the line connecting the origin and
the point z and the positive real axis.
4 For a complex number z zz, * = z2.
5 The modulus–argument form of z is = z r (cos θisin +) θ , where
r = z and θ= arg z. This is often written as (r ), θ
6 Multiplication in polar form: multiply the moduli and add the
arguments
z1z2 = r1 r2 [cos (θ 1 + θ 2 ) + isin (θ1θ2+ )]
7 Division in polar form: divide the moduli and subtract the arguments
z1 r
= 1 [cos (θ1 − θ 2 ) + isin (θ1θ − 2)]
z2 r2
8 Geometrically to obtain the vector z z
1 2 enlarge the vector z 2 by the
scale factor z 1 and rotate it through () arg z1 anticlockwise about O.
9 De Moivre’s theorem: (cos θ + isin θ) cos n n = θ + θ
isin n , where n is
rational.
10 If = z cos θ + isin θ then
n + z −n n − −n
θ = z and sin n θ = z z
cos n
2 2i
iθ + − iθ iθ − − iθ
e e
iθ
11 e = cos θ + i sinθ , cos θ = and sin θ = e
e
2 2i
12 The nth roots of unity can be written as
1, ω, ω, ,2 … ω n−, 1where =ω cos 2π i+sin 2π

n n
13 The sum of all the nth roots of unity is zero
1 + +ω ω 2 + … + ω n −1 = 0
143
155
14 The non-zero complex number r (cos θ + isin θ) has precisely n
5 different nth roots, which are

1
n
( ()()
r cos
θ + 2kπ +
n
i sin
θ + 2kπ
n
)
where k = 0, 1, 2, …
, −n. 1
These roots can also be written as α ,αω
, αω
, , 2 … αωn −, 1where
1
(
θ
α = r n cos + isin
θ
) and ω = cos 2π sin
+ 2π
5 COMPLEX NUMBERS
n n n n
15 The exponential form of a complex number is
z r= (cos θ + i sin) θ = reiθ
For a complex number z x =y + i this can be written as
e z = e x(cos y + i sin) y
16 For a complex number in exponential form ez n+2π i = ez
17 For the complex number = z r (cos θ + i sin) θ
cos θ = 1 (e iθ + e− iθ ) sin θ = 1 (e iθ − e− iθ )
2 2i
LEARNING OUTCOMES
■ understand the geometrical effect of multiplication and division of
complex numbers
■ prove de Moivre’s theorem for a positive integer number
■ use de Moivre’s theorem
■ to express trigonometrical ratios of multiple angles in terms of

powers of trigonometrical ratios of the original angle
■ to express powers of sinq and cosq in terms of multiple angles
■ in the summation of series
■ in finding the nth roots of unity.
144
156
6 Differential equations
For the things

of this world
cannot be made
known without
a knowledge of
mathematics.
Roger Bacon
(1219–1294)
?
› The picture shows the Tacoma Narrows Bridge in the U.S. state of
Washington, which collapsed on 7 November 1940. Find out what
caused the collapse of the bridge.
Modelling
Modelling is the process of representing situations in the real world in
mathematical form. It is an important skill. Sometimes rates of change are
involved and in those cases the models involve differential equations.
Forming differential equations

If you are given sufficient information about the rate of change of a quantity,
such as the caffeine level in the body or the height of water in a harbour, you
can form a differential equation to model the situation.
145
157
It is important to look carefully at the problem before writing down an
equivalent mathematical statement. You have to decide whether you need a
6 model for a rate of change with respect to time or with respect to another
variable such as distance or height. You need to be familiar with the language
used in these different cases.
If the altitude, h, of an aircraft is being considered, the phrase rate of change
of altitude might be used. This actually means the rate of change of
d
altitude with respect to time. You could write it as h where t stands for
dt
time.
However, you might be more interested in how the altitude of the aircraft
changes with the horizontal distance it has travelled. In this case you would
talk about the rate of change of altitude with respect to horizontal
d
distance, and you could write it as h where x stands for the horizontal
dx
distance travelled.
Notation
2
d d d
» Any equation which involves a derivative such as q , y , or x 2 , is called a
dt dt dt
differential equation.
This is just like dq
» A differential equation which involves a first derivative such as is called
the convention dt
of naming a first order differential equation.
polynomials e.g. 2
d
» One which involves a second order derivative such as x 2 is called a
cubics which
dt
contain x 3 terms, second order differential equation. A second order differential
but might also
contain x 2 and x equation may also involve first derivatives – it is the highest derivative that
terms too. matters.
» A third order differential equation involves a third order derivative
3
d
(e.g. y 3 ), and so on.
dx
You should be aware of two shorthand notations.
» Differentiation with respect to time is often indicated by writing a dot
above the variable.
For example xd may be written as x You would say this as ‘x dot’.
2
d y d t
and
dt 2 may be written as y. You would say this as ‘y double dot’.
» Differentiation with respect to x may be denoted by the use of the symbol ´.
dy
For example y′ means You would say this as ‘y dash’.
d x
2
d
and f ′′ means f 2 . You would say this as ‘f double dash’.
dx
You have already met first order differential equations with separable
variables; this chapter extends that work to other first order equations and to
those of higher order.
146
158
6.1 Integrating factors
When a first order differential equation cannot be written in the form
dy
6
= f (gx),yit(cannot
) be solved using the method of separation of variables.
dx
If, however, it is a linear equation, you can multiply it by a special function
called an integrating factor which converts it to a form which can be
integrated.
Linear equations Strictly, it is linear in y.

A linear differential equation is one in which the dependent variable (y in
these examples) only appears to the power of 1. So the differential equation
dy = 2
x– xy is linear because the only terms that involve y are yd and −xy.
dx dx
d y  2 , etc.
There are no terms in y 2, y 3, sin (y), 
 dx 
Any linear first order differential equation may be written in the form:
dy
+ P (x)y = Q (x)
dx
where P and Q are functions of x only.
For example, the equation:
dy
= x 2 − xy
dx
can be rewritten as:
dy 2
+ xy x=
dx
This is in the form:
dy +
Py = Q
dx
where the functions P and Q are given by =Px =x
and Q 2 .
ACTIVITY 6.1
Which of the following differential equations
dy
(i) can be written in the form y ( which
= f (gx)(i.e. ) can be solved
dx
by separating variables)?
dy
(ii) can be written in the form + P (x)y = Q ((i.e.
x) which are
dx
linear)? Identify P and Q if so.
dy dy
(a) = x 2 + x 2y (b) = −x 2xy
dx dx
dy = 2 + + dy
(c) x x xy y + (d) = +x y 2
dx dx
147
159
Example 6.1 Find the general solution of the differential equation:
6 cos (x)
dy −
dx
sin (x)y x= 2
Solution
The equation looks forbidding until you notice that the left-hand side is a
perfect derivative.
Since x d (cos x) –= sin x , it follows (using the product rule) that:
d
d (y cos x)cos dy
= x – ysin
x
dx dx
So you can rewrite the differential equation as:
d (y cos x) = x 2
dx
You may now integrate both sides to obtain:
3
2
y cos x = ∫ x xdx = +c
3
Dividing both sides by cos x, the general solution is:
3
y = x + c
3cos x cos x
In the example above the left-hand side was already a perfect derivative. That is
not the case in the next example but it is a simple matter to convert it into one.
Example 6.2 Find the general solution of the differential equation:

dy + 2 = 4 for
yx 2 x ≠ 0.
dx x
Solution
First note that the equation is linear, because it can be written in the form:
dy There are no
+ P y =Q
dx terms in y 2, 1
y,
4
where =P x2 and = x Q 2 y, etc.
If you now multiply through by x 2, the equation becomes
dy +
x2 2 xy4 =
dx
The left-hand side of this equation is now the expression you obtain when
you differentiate x 2 y with respect to x, using the product rule:
d x 2y 2 dy
d x
( ) = x
d x
+2 xy
So the differential equation can be rewritten as:

d x 2y 4
dx
()=
148
160
Now integrating both sides with respect to x gives:
x 2y = 4∫ d 4x = x c+
The general solution is:

6
y x=4 + c
x2
In the previous two examples the differential equations could be rewritten in

the form:
d ()=
R y function of x
dx
where R was some function of x. In Example 6.1 R cos (x) in
= , and
2
Example 6.2 =Rx . Once the differential equation was written in this form,
it was a simple task to solve it. All you had to do was to integrate the function
of x on the right-hand side.
However, in Example 6.2 you needed to multiply each term in the equation
by a factor x 2 to bring the left-hand side into the required form. This factor
of x 2 is an example of an integrating factor; multiplying by it made the
left-hand side a perfect derivative.
To see how to calculate the integrating factor from a differential equation in
the standard form, think about the general case:
dy
+ Py = Q
dx
You need a function R to multiply everything by:
dy
R + RPy = RQ
dx
so that the LHS can be written as x d (R).y
d
This means you need:
d (R)y= R dy + RPy
dx dx
Differentiating the LHS (using the product rule and chain rule),
remembering that R and y are functions of x) gives:
dy + dR = dy +
R y R RPy
dx dx dx
Comparing the two sides, and realising that y 0≠ , you should be able to see
that this is only true if:
dR = RP
dx
This is a first order differential equation, in R and x, but the variables are
separable:
1 dR = P
R dx
ln R = ∫ xPd
149
161
so
∫ Pd x
6 R e=
This means that any first order linear differential equation written in the
∫ x
standard form can be multiplied by an integrating factor R e= Pd to
d
convert it to the compact form: x (R)y= RQ .
d
You can then solve the equation by integrating the right-hand side, which is
a function of x only.
Example 6.3
Solve the differential equation:
dy +
x 2y x= 4 for x ≠ 0
dx
Solution
Dividing through by x gives the standard form:
dy + 2 y = 4
2
dx x x
4
and with the standard notation =Px2 and = x Q
2 .
The integrating factor is:
The constant of integration c
∫ 2 dx +
2
becomes A e=
c
R e= x = e 2ln x c = Ae ln x = Ax 2
You will multiply the standard form equation by this. But the constant of
integration will always become a multiplier, and since it cannot be zero
(otherwise R would be zero) you can immediately divide by the constant
again, eliminating it. In practice this means you can safely ignore this
constant, so:
2
R =x
One of the few times you can safely ignore a constant of integration is when
calculating the integrating factor!
Multiplying the standard form of the differential equation by R gives:

2 dy
x + 2 yx = 4
dx
The left-hand side is now the derivative of a product, and can be written as
d ()
Ry. In this case R = x 2 and so:
dx
d (x 2)y = 4
dx
x 2y = 4 x c+
Dividing by x 2 gives the general solution:
y x= +4 c
2
x
150
162
This is, of course, the same solution as you obtained in Example 6.2, but this
time you didn’t need to ‘spot’ a convenient form for the LHS, but used an
explicit method to find an integrating factor.
Check you can follow the method in the following example, which also
6
requests a particular solution to satisfy a condition.
Example 6.4 Find the particular solution of:

dy +
x2 xy x= 2
dx
that satisfies the condition y = 1 when x = 2.
Solution
dy + =y 2
Write the equation in the standard form:
dx x x 3
∫ 1 dx
Find the integrating factor R: =R e x = e ln x = x
dy + =
Multiply through by R: x y x 22
dx
d (xy) = 2
dx x2
Integrate with respect to x: xy = ∫ 22 d x
x
xy = − +2 c
x
y = − +22 c
x x
which is the general solution.
The condition states that y = 1 when x = 2, so =1− + 22 c
2 2
c = 3
Therefore, the particular solution is: y x= x−3 2 2
Exercise 6A 1 Find the integrating factor for each of the following differential equations:
dy + 2 = dy +
(i) xyx (ii) x x( ) =
y sin
dx dx
dy − = 2 dy
(iii) 4 x yx 2
(iv) x + =xy 2
dx dx
dy + dy + −2 x
(v) 7y = 1 (vi) cos (x) x () = e
y sin
dx dx
151
163
2 Consider the differential equation
dy +
6
2
x xy = ,1with condition y = 0 when x = 1.
dx
(i) Rewrite the differential equation into the form
dy
+ P (x)y = Q (.x)
dx
∫ x
(ii) Find the integrating factor by calculating e Pd.
(iii) Multiply your answer from part (i) by the integrating factor from
part (ii).
d ()=
(iv) Rewrite your answer from part (iii) in the form RyRQ.
dx
(v) Integrate both sides with respect to x.
(vi) Rearrange your answer from part (v) to give the general solution.
(vii) Substitute the condition into the general solution to find the value
of the constant.
(viii) Write down the particular solution.
3 Find the particular solution of each of the following differential equations:
dy +
(i) x 2y x= 2 y = 0 when x = 1
dx
(ii) d y + xy = 4 x y = 2 when x = 0
dx
(iii) xy6 +
dy = y = 3 when x = 1
0
dx
(iv) dx − 2tx t= x = 1 when t = 0
dt
(v) dy − 3 y x= y = 0 when x = 1
dx x
(vi) dv + 3t 2v t = 2 v = −1 when t = 0
dt
4 An object falling vertically experiences air resistance so that the velocity
satisfies the differential equation:
dv = 10 0.4
− v
dt
(i) Use the integrating factor method to find the general solution of
this differential equation.
(ii) Find the particular solution if, initially, v 0 = .
(iii) Solve the original differential equation, with the condition, using
the method of separation of variables.
(iv) Compare your solutions and comment on which method you
would prefer to use, assuming that both were available (i.e. it is a
first order, linear, differential equation).
PS 5 A parachutist has a terminal speed of 30 m s−1 . The magnitude of the air
resistance acting when the parachute is open is modelled by F=kmv
newtons, where k is a constant, v is the speed and m is the mass of the
parachutist.
(i) By considering the forces acting at terminal velocity, find the value
for k, taking the acceleration due to gravity to be 10 m s−2 .
152
164
(ii) Formulate a first order differential equation for the velocity, v, at a
given time t, after the parachute opens.
(iii) Use the integrating factor method to find the general solution of
the differential equation.
6
(iv) Find the particular solution if the parachutist is moving at 60 m s−1
when the parachute opens.
6 A solution is sought to the differential equation:
dy + =y
cos x ( x > 0)
dx x
(i) Find the general solution for y in terms of x.
(ii) Given that y 0 = , when x 2 = π, find the particular solution.
(iii) Write down the function which approximates the solution as x gets
very large.
(iv) State the behaviour of y as x → and 0 sketch the shape of graph of
y against x, focusing on the behaviour at large and small x.
M 7 The radioactive isotope uranium-238 decays into thorium-234, which in
turn decays into protactinium-234. This can be summarised as
k1 k2
238
92 U → Th → 23491 Pa where k 1 and k 2 are reaction constants (i.e. the
234
90
constants of proportionality by which the rates of decay occur).
The amounts of U-238, Th-234 and Pa-234 at time t are denoted by
x,y , and z, respectively.
You may assume that the rate of decay of an isotope is proportional
to the amount present, and that the constant of proportionality is the
relevant k-value.
An experiment begins with an amount a of U-238, but no Th-234 or
Pa-234. The amount y of Th-234 present at time t satisfies the differential
equation:
dy
+ k2y k=a 1 e − k t
1
dt
(i) Find the integrating factor for the differential equation, and hence
its general solution.
(ii) Find the particular solution that satisfies the initial conditions.
(iii) Write down a differential equation for the variable x.
(iv) Solve the model you suggested in part (iii) and explain how its
particular solution has been incorporated into the suggested
differential equation for y.
(v) Find the particular solution of the differential equation for y, in the
case where k1 = k2 = .k
(vi) Write down a differential equation for the variable z (still assuming
k1 = k2 = .k)By substituting in the solution you found in part (v)
for y, solve this differential equation to find a particular solution for
the variable z at any time t.
153
165
8 The differential equation
(1 − x) dy + +2 y = −(1 x)2 is to be solved for x 1< .
6 dx 1 x
(i) Solve the differential equation.
(ii) Find particular solutions for
(a) y 0 = when x 0 =
(b) y 1 = when x 0 =
in each case stating the behaviour of the solution as x → − 1.
(iii) Find a particular solution which tends to a finite limit as x 1 → − .
Sketch the graph of this solution.
9 The function y = f(x) satisfies the differential equation:
dy +
x 2y = 1 + x 2
dx
(i) Using the integrating factor method, or otherwise, find the general
solution of this differential equation.
(ii) Find the particular solution which satisfies the condition that y 1 =
when x 1= . How does y behave when x becomes very small?
(iii) Write down the first three non-zero terms in the expansion of
3
(1 + x )
2 2
in ascending powers of x.
(iv) Using the expansion in part (iii) and the general solution found in
part (i), write down the power series expansion of the general
solution y for small values of x up to and including the term in
x 2. Hence find the particular solution of the differential equation
which crosses the y-axis from the region x 0 > into the region
x 0< .
6.2 Second order differential

equations
So far in this course you have mostly solved only first order differential
equations. In this chapter you will extend these techniques to second (and
higher) order differential equations. Second order equations are often needed
in mechanics, particularly to model situations which involve acceleration.
A reasonable model of a parachutist (with parachute open) is provided by
treating the parachutist as a particle of mass m, subject to two forces: weight,
mg downwards, and resistance, kv, against the motion (i.e. upwards – where v
is the velocity, and k is a constant).
154
166
Applying Newton’s second law with the downward direction as positive
gives: mg − kv = ma
Using some standard notation
kv
6
s = distance fallen Positive direction
v = sd = velocity
dt
2 mg
a = d v d s
= 2 = acceleration
dt dt ▲ Figure 6.1
the equation can be written in several different ways:
2
d v = −g k v d = −g k v d s gk v
1 2v v 3
dt m ds m dt 2 = − m
The first two equations are first order differential equations and can be solved
using the method of separation of variables.
However, the third equation is of a type that is new to you. It is a second
order differential equation, and this type of differential equation is the subject
of this chapter.
Notation and vocabulary

As an example, the differential equation
d 3 y − d 2 y + dy +
3 7 2 2 4 y = 3sin x
dx dx dx
is described as:
» third order (because the highest derivative is a third derivative)
» linear (because where y and its derivatives appear they are to the power 1)
» having constant coefficients (because the coefficients of the terms
involving y are constants – in this case 1, −7, 2 and 4)
» non-homogeneous (because the right-hand side, the part not containing
y, is not zero). In cases where the right hand side is zero the differential
equation is called homogeneous.
In this chapter you will meet second order, linear differential equations,
with constant coefficients. In general they can be written as follows:
d 2 y + dy + =
a 2 b cy f (x)
dx d x
with a, b and c constant. This may in fact be written without the a constant,
by dividing through by a, but the form given is helpful as it is very similar to
the form of quadratic equations which you will be familiar with. The rest of
this chapter concentrates on the method of solving this type of differential
equation and its applications.
155
167
The auxiliary equation method
6 Before solving second and higher order differential equations you will find
it helpful to look at the form of first order, linear, homogeneous differential
equations with constant coefficients. In general they are of the form:
First order. dy +
ky = 0 Homogeneous because the right-hand side is zero.
dx
Constant coefficients
You can solve this by separating the variables: because k is constant.
dy = −
ky
dx
1 dy =
∫ y
∫ −k xd
ln y = − kx
+ c
− kx
+ c
y = ±e
= Ae −kx
So, in general, the solution of any first order equation of this type will contain
an exponential function and one unknown constant.
Knowing the form of the solution allows you to find it without doing any
integration, as shown in the next example.
Example 6.5
Solve the differential equation:
dy + =
5 y 0
dx
Solution
Since this is first order, linear, homogeneous and with constant coefficients,
you know that the solution will be of the form
λx
=
y Ae
Note where l and A are constants.
In Example 6.5
If this is to be a solution, it must satisfy the original differential equation.
there cannot
be any other dy = λ λ x
Ae Differentiating y with respect to x.
solutions dx
λ λ
because this 5λeA x + Ae x = 0 Substituting y
one already 5λ + = 10 l x(since
dy
Dividing by Ae and into the
contains the it is not zero) gives an d x
one necessary λ =−1 equation just involving l.
5 differential equation.
arbitrary
constant to give So the general solution of the original differential equation is:
it the generality −1 x
=
y Ae 5
it needs.
156
168
This method of assuming the form of the solution is extremely powerful and
is called the auxiliary equation method. The equation 5 1l 0+ = is the
auxiliary equation. 6
ACTIVITY 6.2
All of the following differential equations can be solved by at least one of
the methods:
● separation of variables
● integrating factor
● auxiliary equation.
(i) Which of the equations are linear? Which ones have constant
coefficients?
(ii) For each equation, state which method (or methods) can be used and
use it (or them) to solve the equation.
dy dy − = 2
(a) – 17 y = 0 (b) y 0
dx dx
dy + dy − =
(c) xy = 0 (d) 3y0
dx dx
dy − 2 =
(e) y y 1
dx
Second order homogenous differential equations

The ideas from the work above can be extended to cover second order
equations.
Look at this differential equation. It is a second order, linear, homogeneous
differential equation with constant coefficients.
d 2y dy +
5 6y = 0
dx 2 − d x
Suppose that you assume a solution of the form y Ae=
x
λ (just as before),
where A and l are constants. Then
dy = λ λ x
A e
dx
and
d 2y = λ 2 λ x
A e
dx 2
Substituting these into the differential equation gives:
λ λ λ
Aλ 2e x − 5 Aeλ x + 6 A e x =0
Dividing through by Ae λ xyou obtain the auxiliary equation. It is a quadratic
equation in λ:
λ 2 − 5λ + =
60
157
169
Notice that the form of the auxiliary equation is very close to the form of
the original differential equation, and with experience you will often write it
6 down straight away, without the intermediate working shown above.
Factorising the auxiliary equation gives:
(λ − 3)(λ − =2) 0
You can see there are two different values for l which satisfy the auxiliary
equation: l = 3 and l = 2. So y =Ae 3 x and y Be = 2 x (with A and B
constants) are two solutions of the differential equation.
Each of the two expressions Ae 3 xand Be 2 xis a complementary function
3x 2x
of the differential equation. The sum of these expressions: A e + Be ,
is usually called the complementary function, and is also a solution of the
original differential equation. Since it has two arbitrary constants it is, in fact,
the general solution of the original second order equation.
ACTIVITY 6.3
3x
Verify that the complementary function y A= e + Be 2 x is a solution to
2
d y dy + = .
the original differential equation, 5 6 y0
dx 2 − d x
This method can be used on any linear differential equation with constant
coefficients, whatever its order. If l 1, l 2, l 3, … are the roots of the auxiliary
equation, then, assuming there are no repeated roots, the general solution of
the homogeneous differential equation is
=
yA e
λ 1x
+ Be λ x + Ceλ x +…
2 3
There are three important points to notice about the auxiliary equation
method:
» this method does not involve any integration;
» the number of terms in the complementary function is equal to the order
of the differential equation;
» the number of unknown constants in the solution is equal to the order of
the differential equation.
In general, a second order, linear, homogenous differential equation with
constant coefficients
2
d y dy + =
a b cy 0
dx 2 + dx
has auxiliary equation
aλ 2 + +
bλ= c 0
There are three types of solution to this equation and they each lead to
different types of solution:
» When b 2 > 4 ac , there are two real, distinct roots.
e.g. x 2 + 3x − 10 0= has roots x = 2, x = − 5
158
170
» When b 2 = 4 ac there is one real, repeated root.
e.g. x 2 + 4 x + = has root =x − 2 (twice)
40
» When b 2 < 4 ac both roots are complex; they are conjugates of each other.
2
6
e.g. x − 2 x + = 2 has
0 roots x = +1 i, x = −1 i
In this section you will look at the first two cases, and you will also see how
you can find values for the arbitrary constants to give a particular solution.
Auxiliary equation with real distinct roots, b > 4ac

2
d2 y dy + =
a b cy 0
dx 2 + d x
You have already seen an example of this situation at the start of this
section. In general, if b 2 > 4ac there will be two distinct roots of the auxiliary
equation, l 1 and l 2, which lead to the complementary function:
λ 1x
=
yA e + B eλ x
2
This complementary function is the general solution in this case, but to find
a particular solution you need to eliminate the two unknown constants. This
is only possible with two extra pieces of information, which may come in
two different ways:
» Initial conditions:
If the two conditions are given for the same value of the independent
variable (often x), you say that you have two initial conditions.
d y = when x = 0. Since both
For example: y = 0 when x = 0, and 1
dx
conditions are for x = 0, these are initial conditions, and the problem is
called an initial value problem.
» Boundary conditions:
If the two conditions are given for different values of the independent
variable, you say that you have two boundary conditions.
For example: y = 0 when x = 0, and y = 1 when x = 1. Since these use
different x-values they are boundary conditions, and the problem is called
a boundary value problem. Often the solutions to boundary value
problems are restricted to the region between the boundary points in the
conditions.
The following examples demonstrate one of each type of problem.
Example 6.6 d y
2
dy + =
(i) Find the particular solution of 4 3y0
dx 2 + dx
dy = when x = 0.
subject to the conditions y = 0 and 1
dx
(ii) Sketch the graph of the solution. ➜
159
171
Solution
6 (i) Auxiliary equation: λ 2 + 4λ + =
30
Factorise: (λ + 3)(λ + =
1) 0
Solving this gives two distinct roots, λ 1 and λ 3.
1 =− 2 =−
Complementary function:
−x
=
yA e + Be −3 x
and this is also the general solution.
To find the values of A and B you need to use the conditions y = 0 and
dy = when x = 0.
1 You need to differentiate your
dx
dy = − − x − −3 x
general solution in order to
Ae 3Be dy
dx use the condition.
dx
When x = 0, y = 0 ⇒ +A=B 0
dy =
When x = 0, 1 ⇒−A − B3 =1
dx
Solving these two equations simultaneously gives:
A = 21 and B =− 12
So the particular solution is:
− −
y = 12 e x − 12 e 3 x
(ii)
You know that the curve You also know that for positive x,
passes through the origin, e−x and e−3x are both positive, and
with positive gradient there. e−x > e−3x . So the curve is above
the x-axis for positive x.
y
You also know that for

large values of x, both e−x
and e−3x tend to zero.
▲ Figure 6.2
Example 6.7 The differential equation

2
d y dy +
4 3y = 0
dx 2 + dx
is used to model a situation in which the value of x lies between 0 and 2.
160
172
(i) Find the particular solution given the boundary conditions y = 0 at
x = 0, and y = 1 when x = 2.
(ii) Sketch a graph of the solution for ഛ0 ഛx 2. 6
Solution
(i) The general solution is the same as in the previous example:
−x
=
yA e + Be −3 x
Using the boundary conditions:
y = 0, x = 0 gives 0 = +A B
−2
y = 1, x = 2 gives 1 = Ae + B e −6
Solving these simultaneously gives:
Substituting for B
− −
1 = Ae 2 − Ae 6 from the first equation
2 into the second.
A =− e −4
1e
and then
2
e
B = − − −4
1e
The particular solution then is:
2
y =− e −x
− e −)3 x
−4 (e
1e
(ii) y Note
3 Note that you have two points
defined as the boundary
conditions, but the same
2
reasoning as before about the
variable terms.
1
−0.5 0 0.5 1 1.5 2 2.5 x
▲ Figure 6.3
?
› The last two examples started off exactly the same. However, when
it came to finding the particular solutions, they required different
approaches because of the different types of conditions involved.
Which is the easier to work with?
161
173
Auxiliary equation with repeated roots, b = 4ac
2
6 a
2
d y
b
dy + =
dx 2 + dx
cy 0
2
If b = 4ac then there is only one root from the auxiliary equation, and it is
λ = − b . (Check this with the quadratic formula).
2a
x
This leads to a complementary function of y Ae= λ . But, since the original
differential equation is second order, the general solution needs to have two
arbitrary constants. So, you need to find another solution to the equation in
order to provide this extra constant.
At this point, if someone suggested another possible solution of the form
λx
=
y Bxe (note the extra x in front), you could test to see if it satisfies the
original differential equation:
Using the product rule
dy = λ x + λ λ
Be Bxe x
dx
and differentiating again
d 2y = λ λ x + λ λ x + λ 2 λ x
2 B e B e Bx e
dx
= 2Beλ λ x + Bλx2 e λ x
x
=
If y Bxe λ is a solution, you should get 0 when you substitute these
expressions into the left-hand side of the original equation. You get:
a (B2 eλ + Bλx2 eλ x ) +(b B eλ x + Bλx eλ x ) +(c )Bx eλ x
λx
= Be λ2x ( aλ + + bλc ))
b x a (λ2 + +
b
But, in this case, λ = − , which rearranges to give a 2 λ = − , bmeaning that
2a
the first two terms in the above bracket sum to zero. (Notice this only works
in this case because of the value of λ).
The last bracket (the coefficient of x) is a λ 2 + +
bλc , which is the left-hand
side of the auxiliary equation, and therefore equal to zero too.
The whole expression is therefore zero, which means the expression
λx
=
y Bxe does satisfy the original differential equation, and so is a
complementary function – and it is different from your first solution
=
y Ae x
λ.
So, you can combine them to form the complementary function (which is
also the general solution), with two arbitrary constants:
λx
=
yA e + Bxeλ x
or
λx
y = (A +
Bx e )
See question 11 in Exercise 6B for a way to derive this result.
162
174
Example 6.8 (i) Find the particular solution of the equation
2
d z 6 dz + 9z = 0
d t 2 + dt
6
subject to the conditions z = 0 and zd = when
5 t = 0.
dt
(ii) Sketch the graph of the particular solution.
Solution
(i) Auxiliary equation:
λ 2 + 6λ + =
90
(λ + 3)2 = 0
λ = − 3 (repeated)
So, the general solution is:

−3 t
z = (A +
Bt e )
Differentiating this in order to use the condition gives:
dz = B −3t − (A Bt
3 + )e
−3 t
e
dt
And using the conditions:
0 = (A +B × × 0)1
A 0=
and
5 = −B 3A
B 5=
So, the particular solution is: Note
−
z = 5t e 3 t For all graphs of the form
(ii) z
= −kxwith
y xe
k > 0, y → as0 x → ∞. You
are multiplying an increasing
function, x, by a decreasing
function e −kx, and in this case the
decreasing one wins.
O t
▲ Figure 6.4
To summarise: if the auxiliary equation has a repeated real root λ , then the
complementary function is:
λx
y = (A +
Bx e )
163
175
Exercise 6B 1 Use the auxiliary equation method to find the general solution of the
6 following differential equations:

(i)
dy −
dx
3y = 0 (ii)
dy + =
dx
7y0 (iii) dx + =
dt
x 0
(iv) dp − 0.02 p = 0 (v) 5 dz − = z 0

dt dt
2 Use the auxiliary equation method to find the particular solutions of the
following differential equations:
dy +
(i) 2 y = 0 y = 3 when x = 0
dx
dy −
(ii) 2 5y = 0 y = 1 when x = 0
dx
(iii) 3 dx + =
x 0 x = 2 when t = 1
dt
(iv) dP = kP P = P 0 when t = 0
dt
(v) dm = − km m = m 0 when t = 0
dt
3 Find the general solution of the following differential equations:
2 2
d y d y dy
(i) 16 y = 0 (ii) 5y0
dx 2 − dx 2 + −d=x
2 2
(iii) 2 d x 4 dx + =
x 0 (iv) 7 y 2 dy = 0
d
dt 2 + dt dx 2 + dx
d2 y dy +
(v) 9 12 4y = 0
dx 2 − dx
2
4 You are given the differential equation
d y dy + = .
3 2y0
dx 2 − dx
(i) Write down the auxiliary equation in terms of λ.
(ii) Solve the auxiliary equation to find its two real roots.
(iii) Write down the complementary function in the form y(x) = …
(iv) Find the derivative of the complementary function.
dy = when x = 0. Use these
(v) You are also told that y = 1 and 0
dx
facts, the complementary function and its derivative, to eliminate the
arbitrary constants and find the particular solution that satisfies these
constraints.
164
176
2
d y dy +
5 Given 8 16 y = 0
dx 2 − d x
(i) Write down the auxiliary equation, and solve it. 6
(ii) Write down the complementary function (this is also the general
solution).
(iii) Find the particular solution which satisfies the initial conditions
dy = − when x = 0.
y = 1 and 1
dx
PS 6 Find the particular solutions of the following differential equations that
satisfy the conditions given. In each case sketch the graph of the solution
(you may wish to use graphing software to help visualise this). Assuming
each one models a real system, where x represents a variable changing
over time, use the graph to describe the motion of each system in words.
(i) d 2 x 5 dx = 0 x = 0, xd = ,4when t = 0
dt 2 + dt dt
2
(ii) 4 d x 4 dx + = x 0 x = 1, t = 0 and x = 1, t = 1
dt 2 + d t
M 7 In an electrical circuit, the charge q coulombs on a capacitor is modelled
by the differential equation:
d 2q d q +
0.2 1.25q = 0
dt 2 + d t
dq
Initially q = 2 and (the current in amperes) is 4.
dt
(i) Find an equation for the charge as a function of time.
(ii) Sketch the graphs of charge and current against time. Describe
how the charge and current change.
(iii) What is the charge on the capacitor and the current in the circuit
after a long period of time?
M 8 The temperature of a chemical undergoing a reaction is modelled by the
differential equation
2
2 d T dT = 0
dt 2 + dt
where T is the temperature in °C and t is the time in minutes.
For a particular experiment, the temperature is initially 50 °C, and it is
45 °C one minute later.
(i) Find an expression for the temperature T at any time.
(ii) What will the temperature be after two minutes?
(iii) Sketch a graph of T against t.
(iv) What is the steady state temperature?
165
177
9 The metal fins on a motorcycle engine help to cool it. A model for the
2
change of temperature along a fin is given by Td
6 4T0
dx 2 − = , where
T °C is the temperature and x m is the distance from the hot end.
(i) Find the general solution of this differential equation.
When the engine has been running for some time the temperatures at
the two ends of the fin are 100 °C and 80 °C. The fin is 5 cm long.
Hot
end
(ii) Find the particular solution.

(iii) Hence determine the temperature 3 cm from the hot end.
M 10 A car of mass 800 kg is travelling along a road at a constant velocity
of 100 km h –1. A catastrophic engine and brake failure renders the car
unable to brake in any way other than natural slow-down from the
friction from the air and road. The friction is modelled as a backwards
force equal to 40 000 times the velocity (in km h –1).
(i) Show that the equation of motion of the car as it slows (x in
2
kilometres, t in hours), simplifies to xd 50 dx = .0
dt 2 + dt
(ii) Solve this differential equation, using appropriate initial conditions,
to find a particular solution for x in terms of t as the distance
(in km) travelled after the brake failure.
(iii) Find the time taken for the car to travel 1 km further, and the
speed of the car at this moment.
(iv) State at least one problem with the solution given by this model.
CP 11 In this question you will prove the result for the general solution of a
second order differential equation for which the auxiliary equation has
repeated roots. 2
d y
(i) Find the general solution of the differential equation 0
dx 2 = .
(ii) Assume that the differential equation
d 2y dy +
4 4y = 0
dx 2 + dx
2x
has a solution of the form y = f (x)e − . Differentiate this twice and
substitute it into the original differential equation, and hence find
the form for f(x). 2
d y + dy + =2.
(iii) Repeat for the differential equation 2k ky 0
dx 2 dx
166
178
6.3 Auxiliary equations with
complex roots 6
d2 y dy + =
a b cy 0
dx 2 + d x
has auxiliary equation a λ2 + bλ + c = 0

b b 2
− 4 ac
and so λ = − ±
2a
If b < 4 ac then the two roots will be complex (and conjugates of each
2
other). For example, the differential equation

2
d y
4y = 0
dx 2 +
has auxiliary equation
λ2 + =
40
The roots of this quadratic are λ 2i and λ 2i (where i = − ).1
= 2= −
1
The general solution of this differential equation is therefore.
=
yA e
2i x
+ Be −2i x
The terms in the solution contain complex exponentials, and it is important
to recognise that these can always be written in terms of sine and cosine
functions. These are derived by using the relationships you met in Chapter 5:
θ
e i = cosθ + i sinθ
− iθ
e = cos θ − i sinθ
So you can rewrite the general solution as:

=
yA (cos 2x + i sin 2x)B+ (cos 2xisin
− 2 x)
= (A +
B ) cos 2x + i (A B − ) sin2 x
Note
Notice that complex numbers are needed to find this form for the general
solution, and yet this abstract result actually works as a model for oscillating
systems in the real world.
Notice now that the coefficients of cos 2x and sin 2x are constants, which
you can call P and Q, such that P = A + B and Q = −i( A. This
B )means the
general solution can be written simply as:
y P= cos 2x Q
+ sin 2x
It is important to note that this solution has an oscillating nature, and in fact
all differential equations with complex roots from their auxiliary equation
will give oscillating solutions.
167
179
Simple harmonic motion
6 Many real-life situations can be modelled by a differential equation of the form
d 2x = −ω 2 x
dt 2
For example:
» A mass fixed to the end of a spring is
pulled down vertically and then released.
The displacement x of the mass from its
equilibrium position, at time t, can be
Equilibrium position
modelled as
2
d x =− λ x
x
dt 2 ml 0
where l is the modulus of elasticity of
the spring, m is the mass of the weight
and l 0 is the natural length of the spring. ▲ Figure 6.5
» A mass suspended from a fixed point is moved
horizontally slightly from its equilibrium position
and then released, forming a simple pendulum. The
The right-hand side θ
of this equation
angle q which the string makes with the vertical at
should really time t, can be modelled as
2 l
involve sin θ but d θ = − gθ
the small angle dt
2
l
approximation of
sin θ≈θis very where l is the length of the string and g is the
accurate and is acceleration due to gravity. T
used in this model.
Both these equations have the same form:
2
d x = −ω 2 x
2
dt mg
2 k 2 g
In the first case ω = , in the second ω = . In the ▲ Figure 6.6
m l
second case the variable is q rather than x.
In both cases the motion is called simple harmonic motion (SHM); it
provides a model for many oscillations.
The general solution of the differential equation, as shown above, is
x P= cos ωt Q+ sin ω t
The constants of integration, P and Q, are unknown at this stage, but if you
know suitable initial or boundary conditions you can calculate their values.
e
Note
The focus of this part of the Cambridge International AS & A Level Further
Mathematics syllabus is on systems of three equations in three unknowns.
This introduction is based on sets of two equations in two unknowns; it
should be seen as providing an underpinning understanding but not a
sufficient coverage of the whole topic, which comes later in the chapter.
168
180
ACTIVITY 6.4
Given that the general solution x P= cos ω t Q
+ sin ω t can be written in 6
the form x a= sin(ω tε+ , )use the compound angle formulae to prove
2 2 P
that a = P Q+ and tan ε = .
Q
Expressing the general solution x P= cos ω t Q + t sin ω in the form

x a= sin(ω t + ε )tells you a lot about the solution.
» Since the sine function varies between +1 and −1, the solution varies
between −a and a.
» Since the sine function is periodic with period 2π, the solution is periodic
with period 2 π .
ω
ACTIVITY 6.5
If you have access to graphing software, use it to investigate the graph
x a= sin(ω t + ε )for different values of a, w and ε.
x
In the activity above you should have
x = a sin ωt
noticed that the effect of ε is to translate
the sine curve to the left by an amount O t= ε t
ε ω
ω
, as shown in the diagram below. The
quantity ε is called the phase shift.
x x = a sin(ωt + ε)
a sin ε
O t
▲ Figure 6.7
Example 6.9 d 2
(i) Find the general solution of the differential equation x 4 x0
dt 2 + = .
(ii) Find the particular solution in the case for which x = 2 and x
d = −2
dt
when t = 0.
(iii) Find the period and amplitude of the oscillations, and sketch the graph
of the particular solution.
Solution
(i) Auxiliary equation: λ2 + =
40
λ = ± 2i
General solution: =
xA + t sin 2
cos 2t B
➜
169
181
(ii) When t = 0, x = 2 ⇒ =2 A
6 dx = − 2 Asin 2 2t cos
dt
+ B2 t
x
When t = 0, d =− 2 ⇒ − 2= 2 B ⇒ =B − 1
dt
Particular solution: x = 2cos 2tsin − 2 t
To sketch the
(iii) The period of the oscillations is 2 π = 2 π = π graph and find
ω 2 the amplitude, it
− =
2cos 2tsin 2 t a sin(2 t + ε ) is helpful to write
= a sin 2t cos ε + a cos 2t sin ε the solution in the
form a sin(2
t ) + ε.
Comparing coefficients of sin 2t: −1=cosε
a
Comparing coefficients of cos 2t: =a
2 sinε Using the
compound angle
So a = 221+ =2 5. formula.
The amplitude of the oscillations is 5.

tan ε = − 2 ⇒ =ε − 1.107
The particular solution can be written in the form

x = 5 sin(2t1.107)
− .
x
√5
0
3 2 5 3 t
2 2 2
▲ Figure 6.8
Note
The graph is obtained from the graph of y = sin t by
» translation through 1.107 to the right;
» stretch scale factor 0.5 in the x-direction;
» stretch scale factor 5 in the y-direction.
In Example 6.9 the roots of the auxiliary equation were purely imaginary.
Often the roots also have a real part.
170
182
For example, suppose that the roots of the auxiliary equation are 2 3i± .
This gives the complementary function
=
yA e
(2)+3i x
+ B e(23i
)− x
6
Notice that the Notice that e 2x is a
real part of the =e 2x
(Ae 3 ix
+ Be −3i x
) common factor.
roots ends up in
the exponential But, as before, you can write A e
3i x
+ Be −3i x as P cos 3x Q
+ x sin 3 , so
outside the y = e (P cos 3x Q
2x
+ sin 3x).
bracket, and the
imaginary part This is the standard form of the general solution.
ends up as the
coefficient inside Note
the trigonometric
terms. The four forms of complementary function, for the four cases of solutions of
the auxiliary equation are:
= e λ x + Beλ x
» Real distinct roots, λ 1 and λ 2: y A 1 2
» Real, repeated root, λ: yA ( e ) λx

= +Bx
αx
» Complex roots, α β±: i y = e ( cos
P + x sin β)
βx Q
» Pure imaginary roots, i ±ω y P= cos ω x Q
+ x sin ω
Example 6.10 (i) Find the particular solution of the differential equation
d2y dy +
2 5y = 0
dx 2 + d x
dy = when x = 0.
which satisfies the initial conditions y = 0 and 1
dx
(ii) Sketch the graph of this particular solution.
Solution
(i) The auxiliary equation is:
λ 2 + 2λ + =
50
Completing the square gives:
Alternatively you
(λ + 1)2 − 1+ 5= 0 could use the
λ +1
=±− 4 quadratic formula.
λ = − 1± 2i
The general solution is y = e − x (P cos 2x Q

+ x sin 2 ).
To find the particular solution:
When x = 0, y = 0 ⇒ =P 0
➜
171
183
So the general solution is now:
6 =
yQ
−x
e sin 2 x
⇒ dy = − Qe −sin
x −x
2 x + 2Qe cos 2 x
dx
dy
When x = 0, dx = 1 ⇒ =1 2 Q
⇒ =Q 1
2
So the particular solution for these initial conditions is:
−x
y = 1 e sin 2 x
2
(ii) The graph of this particular solution is shown below. Notice that the
oscillating solution moves between the curves =y ± 1 e−,x which are
2
indicated by the blue dotted lines.
y
0.3
0.2
0.1
0 1 2 3 4 5 6 x
–0.1
–0.2
▲ Figure 6.9
It is important to understand the relationship between the solution and

its graph. In Example 6.10:
» the e –x factor tells you that the solution will decay as x increases (because
of the negative sign)
» the sin 2x factor tells you that there is an oscillation.
This form of solution can be described as exponentially decaying
oscillations or damped oscillations, and it arises frequently when
modelling real oscillating systems.
Damped oscillations
Simple harmonic motion has constant amplitude and goes on forever. For
many real oscillating systems, SHM is not a very good model: usually the
amplitude of the oscillations gradually decreases, and the motion dies away.
172
184
Example 6.11 A damped oscillating system is modelled by the differential equation
d 2 x k dx + 25x = 0
d t 2 + dt 6
where k is a constant that can be varied.
0 x d = .1
When t = 0, x = and
dt
Solve the differential equation for each of the following values of k, and
sketch the graph of the solution in each case.
(i) k = 26
(ii) k=6
(iii) k = 10
Solution
(i) Auxiliary equation: λ 2 + 26λ + =
25 0
(λ + 1)(λ + 25) 0=
λ = − 1 or 25
−
− −
General solution: =
xA e t + Be 25 t
When t = 0, x = 0 ⇒ +A=B 02
dx = − Ae −t − 25Be −25t
dt
When t = 0, dx = 1 ⇒ − A − 25B = 1
dt
1 , B =− 1
Solving these equations simultaneously gives: A = 24 24
1 e − t − 1 e −25
Particular solution is x = 24 .t
24
x
O t
▲ Figure 6.10
(ii) Auxiliary equation: λ 2 + 6λ + =
25 0
λ =−6
± −
36 100 = − 3± 4i
2
−3 t
General solution: x = e ( sin
A4 tB + cos4 t)
When t = 0, x = 0 ⇒ =B 0
= e −sin
xA 3t
4 t
dx = − 3 Ae sin
−3 t
+ A −3 t
4 4 et cos4 t
dt
x
When t = 0, ddt = 1 ⇒ 4 A = ⇒ 1 = A 1
4
➜
173
185
−3 t
Particular solution is x = 41 e sin 4 t.
6 x
O t
▲ Figure 6.11
(iii) Auxiliary equation: λ 2 + 10λ + =
25 0
(λ + 5)2 = 0
λ = −5
−
General solution: Bt )e 5t
x = (A +
When t = 0, x = 0 ⇒ =A 0
− 5t
=
x Bte
dx = Be −5t − 5Bt e
−5 t
dt
x
When t = 0, d = 1 ⇒ =B 1
dt
5t
Particular solution is x te= − .
x
O t
▲ Figure 6.12
The three parts of Example 6.11 illustrate the three different types of
damping. The type of damping that occurs depends on the nature of the roots
of the auxiliary equation.
» Overdamping: the discriminant of the auxiliary equation is positive and
the roots are negative; the system decays without oscillating.
y y
t t
▲ Figure 6.13
174
186
» Underdamping: the discriminant of the auxiliary equation is negative and
oscillations occur.
y 6
t
▲ Figure 6.14
» Critical damping: the discriminant of the auxiliary equation is zero.
y >y
t t
▲ Figure 6.15
Critical damping is the borderline between overdamping and underdamping.
It is not obvious in a physical situation when damping is critical, since the
pattern of motion for critical damping can be very similar to that in the
overdamped case.
Exercise 6C 1 For the differential equation

2
d x 9x = 0
dt2 +
(i) Write down the general solution of the differential equation.
(ii) Given the initial conditions x = 0 and x

d = when
1 t = 0, find the
dt
particular solution.
(iii) Write down the period and amplitude of the oscillations, and
sketch a graph of the solution.
2 For the differential equation
2
4dx x 0
dt 2 + =
(i) Write down the general solution of the differential equation.
(ii) Given the initial conditions x = 4 and x d = when

0 t = 0, find the
dt
particular solution.
(iii) Write down the period and amplitude of the oscillations, and
sketch a graph of the solution.
3 Find the general solution of each of the following differential equations:
2 2
d y dy + d y dy + =
(i) 4 5y = 0 (ii) 2 5y0
dx 2 − d x dx 2 − d x
2 2
d x 2 dx + 4 x = 0
(iii) (iv) 4 d x 4 dx + =
5x0
dt 2 + dt dt 2 + dt
175
187
2
4 Given 9 d x 4x = 0
dt 2 +
6 (i) Find the general solution.
(ii) Find the particular solution which satisfies the conditions
x = 4 and xd = when2 t = 0. Express the solution in the form
dt
x a= sin( t + ε. )
(iii) Sketch the graph of the particular solution.
PS 5 Find the particular solutions of the following differential equations that
satisfy the conditions given. In each case sketch the graph of the solution
(you may wish to use graphing software to help visualise this). Assuming
each one models a real system, where x represents a variable changing
over time, use the graph to describe the motion of each system in words.
2
d x 8x = 0
(i) x = 1, dx = 1 when t0 =
dt 2 + dt
d 2 x dx + =x
(ii) 0 x = 1, dx = 0 when t0 =
dt 2 − dt dt
2
d x 2 dx + 2 x = 0
(iii) x = 0, dx = 2 when 0t =
dt 2 + dt dt
2
(iv) 4 d x 8 dx + 5x = 0 x = 2, dx = 0 when 0t =
dt 2 − dt dt
2
(v) d x 2 dx + 5x = 0 π
x = 0, t
0 =
and 3, x t= =
dt 2 + dt 4
M 6 The motion of a spring-mass oscillator is modelled by the differential equation
2
d x 64 x = 0
dt 2 +
where x is the extension of the spring at time t.
(i) Find the general solution of the differential equation.
(ii) Initially x = 0.1 and x d = .0Find the particular solution
dt
corresponding to these initial conditions.
(iii) Write down the period and amplitude of the motion.
(iv) Sketch a graph of the solution.
7 The angular displacement from its equilibrium position of a swing door
is modelled by the differential equation
d 2θ 4 dθ + 5θ = 0 .
dt 2 + dt
The door starts from rest at an angle of 4π from its equilibrium position.
(i) Find the general solution of the differential equation.
(ii) Find the particular solution for the given initial conditions.
(iii) Sketch a graph of the particular solution, and hence describe the
motion of the door.
(iv) What does your model predict as t becomes large?
176
188
M 8 A damped spring-mass oscillator consists of a spring and an object with
mass m kg. The object is pulled down 10 cm from its equilibrium position
and released. The motion of the system is modelled by the differential
equation
6
2
m x 2 + k dx + 20 x = 0
d
dt dt
where x is the displacement from the equilibrium position at time t.
(i) In the case where m = 0.25, find the value of k if the system is to be
critically damped.
(ii) If the value of k does not change, describe the motion of the system if:
(a) the mass of the object is increased to 0.3 kg
(b) the mass of the object is decreased to 0.2 kg
CP 9 Prove that, if a, b and c are positive constants, then all possible solutions of
2
d
ax b dx + =
cx 0
dt 2 + d t
approach zero as t → ∞.
10 Find the general solution of the differential equation
d4 y
16 y = .0
dx 4 −
6.4 Non-homogeneous differential

equations
So far you have seen that equations of the form
d 2y dy + =
a b cy 0
dx 2 + dx
(linear, homogeneous, second order differential equations, with constant
coefficients) have general solutions of the form
u (x) + v( )
=xB
yA
where u(x) and v(x) may involve exponential and/or trigonometric functions
(and possibly a factor x). In modelling real situations, equations like this often
arise in which the right-hand side is non-zero. These equations are called
non-homogeneous linear equations:
2
d y dy + =
a b cy f (x)
dx 2 + dx
Examples of situations where such equations arise include the motion of
a parachutist and, like the bridge in the picture at the start of this chapter,
a structure being forced to vibrate.
177
189
In order to learn how to deal with this type of second order differential
Linear, first order. equation, it is useful to look at similar first order equations, e.g.
6 dy +
2y = 3x − 1
Non-homogeneous because the
dx right-hand side is non-zero.
Constant coefficients.
You should remember how to solve this using an integrating factor method,
but it is instructive to compare the solutions to several similar looking
equations.
ACTIVITY 6.6
Verify that the following solutions satisfy their respective first order linear
differential equations:
dy +
(i) 2y = 0 y =
Ae
−2 x
dx
dy +
(ii) 2y = 3x − 1 y = e −2 x + −3 x 5
A
dx 2 4
dy +
(iii) 2y = e 3 x y =
Ae
−2 x
+ 15 e 3 x
dx
dy +
(iv) 2y = sin x y = e −2 x − 1 cos x + 2 sin x
A
dx 5 5
Compare the different solutions: what is the same and what is different?
You should notice that all the solutions have strong similarities, and in
particular they consist of two distinct parts.
» The first part Ae −2 x contains an arbitrary constant, and is the general
solution for the homogeneous version (part (i) of the activity). It appears
to depend only on the left-hand side of the differential equation. This part
of the solution is the complementary function.
» The second part only exists in the non-homogeneous cases, contains no
arbitrary constants, appears to depend only on the right-hand side of the
differential equation – and is of a similar form to the original right-hand
side (e.g. linear in part (ii), exponential in part (iii), trigonometric in part
(iv)). This part of the solution is called the particular integral.
Since you already know how to find the complementary function, all that
remains in order for you to use this method of constructing a solution
is a way to find a particular integral. The following first order example
demonstrates a way to approach this, using a trial function.
Note
» The complementary function solves the homogeneous case, and contains
an arbitrary constant.
» The particular integral solves the full non-homogeneous case, and does
not contain an arbitrary constant.
178
190
Example 6.12 Solve the differential equation
dy +
dx
2y = e
3x This is one of the examples
from the previous activity.
6
Solution
The auxiliary equation is λ + =2 0
⇒ λ = −2
The complementary function is y Ae= −
2x
.
3x
To find the particular integral use a trial function of y ae= , since the
right-hand side is of the form ke 3 x where k is a constant.
y a= e 3 x Differentiate.
dy = 3x
3ae Substitute into the differential equation.
dx
x x x
3ae 3 + 2ae 3 = e 3
3x
The general solution is constructed
Comparing coefficients of e gives: by adding the complementary
function and the particular integral.
5a = 1
so a 1= 5 Notice that you still have one
−2 x arbitrary constant, as you would
=
y Ae + 1 e3x expect for a first order equation.
5
This example could also have been solved using first order methods, such as an
integrating factor, but this method of complementary functions and particular
integrals is more powerful since it can be used for higher order equations. The
next example shows how this method is used for a second order differential
equation, and also demonstrates a different form of particular integral.
Example 6.13 Find the general solution of the differential equation

d 2 y − dy −
2 2 3y = 6x − 2
dx d x
Solution The right-hand side is a

linear function, so your
The auxiliary equation is λ − −2λ2λ− −=3 =30 0
22 2
λ
trial function should be a
(λ(λ− −3)(3)(
λ λ+ +
=
1)=1)0 0 general linear function.
λ λ= =3 3
oror1−1−
3x x
So the complementary function is y A= e + Be − .
To find a particular integral, use a trial function y ax= b.
+
2
dy = d y = Differentiating twice.
a and 2 0
dx dx
0 2− −a 3 (ax b+ =) 6x − 2 Substituting into the original
differential equation.
➜ 179
191
Comparing coefficients of x gives
6 −3=a 6
a = −2
and comparing constant terms
−2−a 3b = − 2
4 3− b = − 2
−3=b − 6
b =2
The particular integral is y = − 2+x 2.
So the general solution is:
yA= e 3 x + Be − x − 2x + 2
Notice that the general solution has the properties you need:
» it satisfies the full differential equation (you may wish to check this);
» it has two arbitrary constants, which is consistent with it being the general
solution to a second order differential equation.
Particular integrals
In order to find a suitable particular integral, the trial function needs to
match the right-hand side of the differential equation.
The table is a guide to what trial function to use for the differential equation
2
d y dy + =
a b cy f (x)
dx 2 + dx
Right-hand side Trial function

linear function ax b+
n −1
polynomial of order n anx na +x n −1
+… + +a1x0 a
trigonometric function involving cos px px b px+ sin
a cos
and/or sin px
px
exponential function involving e ae px
sum of different functions sum of matching functions
You should note that this method finds a simple particular integral, but
that there are infinitely many possible particular integrals, which can be
constructed by adding on any term from the complementary function.
The following example demonstrates this trial function approach for finding
a particular integral where the right hand side is a trigonometric function.
180
192
Example 6.14 Find a particular integral of the differential equation
d 2 z − 2 dz − 3z = 6cos 3t
dt 2 dt
6
Solution
Use the trial function: z a= cos 3t b+ sin 3t
Notice that both the cos 3t and sin 3t functions are used, since they
differentiate to become each other.
Differentiating the trial function twice:
dz = − 3asin 3 t3+cosb 3 t
dt
d 2 z = − 9acos 3 t9−sinb3 t
dt 2
Substituting these into the differential equation gives
(−9acos 3 t9−sinb3 t) − −2( 3 sin
a 3 3t cos + b3 ) 3 cos t −3 sin (a t3b 6cos
t +3 )= t
(−9acos 3 t9−sinb3 t) − −2( 3 sin t −(3−12(a 6a )cos
+ b3 ) 3 cos
a 3 3t cos − tbb+ sint 63+)6cos
3 t( 12 −sin3
=+6cos
3b6acos
t3t 3 = t
( −12 a − 6b)cos 3 t( +12− 6 ) bsin3
a+ 6 cos 3 t = t
Equating coefficients of cos 3t gives −12

− a=6b6
Equating coefficients of sin 3t gives −12
+ b=6a0
Solving for a and b gives a =2− and
5
= −b 1
5
The particular integral is − 25 cos 3t − 15 sin 3t
Special cases
In some differential equations the function on the right-hand side has
the same form as one of the complementary functions. For example, the
complementary function of the differential equation
d 2 y − dy +
2 5 6 y = 4e 3 x
dx dx
is A e 2 x + Be 3 x , and e3 xoccurs on the right-hand side. In this situation it

is no good using the trial function ae 3 x, since upon substituting y ae= 3 x ,
2
dy = 3x d y
3ae and 2 = 9ae 3 x into the differential equation, you obtain
dx dx
9ae 3 x − 5(3ae 3)x 6(+e ) a4e3 x = 3x
3x
⇒ =0 4e
and so clearly this trial function does not work.
181
193
3x
=
Instead y axe is used as a trial function.
6 This gives
dy = 3 x +
dx
2
ae e 3x
3ax
d y =
and 2 6ae 3 x + 9ax
e 3x
dx
Substituting these in the differential equation gives:
3x 3x 3x 3x 3x 3x
(6 ae + 9ax
e ) 5( e− a + 3ax
e ) 6 e+4e ax =
⇒ ae 3 x = 4e 3 x
⇒ =a 4
The particular integral is x4 e 3 x .

This illustrates a general rule. If the function on the right-hand side of the
differential equation has exactly the same form as one of the complementary
functions, you multiply the usual trial function by the independent variable
to give a new trial function. In order to recognise these special cases when
they arise, it is worth getting into the habit of finding the complementary
function before the particular integral.
Exercise 6D 1 Find a particular integral for each of the following differential equations.
(i) d 2y dy + = − + (ii) d 2
x 4x t
4 y 2x 3 2
dx 2 − d x d t 2 + = +
d 2y dy + = d
2
x + +dx= 2 x3e −2 t
(iii) 2 y cos 3x (iv) 2
dx 2 + dx dt dt
2 Find the general solutions of the following differential equations.
2 2
d y + 2x d x − 2 dx − = −2 t
(i) 2 4y = e (ii) 3x5e
dx dt 2 dt
2 2
d y dy + d y dy + = +
(iii) 2 5y = cos x (iv) 2 5y3x2
dx 2 + d x dx 2 + d x
3 Find the general solutions of the following differential equations.
2 2
d x x sin t d y dy − = x
(i) (ii) 3 4 ye
dt 2 + = dx 2 + dx
2 2
(iii) d x x
− 4 d + 3x = e 3t
(iv) d x − 6 dx + = 9 x4e 3t
2 2
dt dt dt dt
PS 4 Find the particular solution of each of the following differential
equations with the given initial conditions.
2
d y dy + d y = − when x 0
(i) 5 6 y = 36 x y 0= and 10 =
dx 2 − d x d x
2
d x + 9 x = 20e t dx = when t0 =
(ii) 2 x 0= and dt
1
dt
d 2y dy + − dy = when x 0
(iii) 2 5y = 4e x y 0= and dx
0 =
dx 2 + dx
182
194
(iv) d 2 x dx − 2 x = 20 sin 2t x 2= and
dx = when
0 t0 =
dt 2 + dt dt
(v) d 2
x 3 dx + 2x = −1 e −t x 0= and
dx = when
1 t0 =
6
dt 2 − dt dt
d2y dy = when x 0
(vi) 4 y = 12 sin 2x y 0= and dx
1 =
dx 2 +
2
d y dy + dy = when x 0
(vii) 4 5y = 8sin x y 1= and dx
0 =
dx 2 + d x
2
d y + dy + 2x dy = when x 0
(viii) 4 4 y = 8e + 4 x y =
0 and 1 =
dx 2
d x d x
M 5 A biological population of size P at time t is growing in an environment
which can support a maximum population which is subject to seasonal
variation. The growth of the population is described by the first order
linear differential equation:
dP + = P 100 50 + sin t
dt
Find:
(i) the complementary function of this differential equation
(ii) the particular integral
(iii) the complete solution given that initially P = 20
(iv) the mean size of the population after a long time has elapsed
(v) the amplitude of the oscillations of the population
6 The pointer on a set of kitchen scales oscillates before settling down at
its final reading. If x is the reading at time t then the oscillation of the
pointer is modelled by the differential equation:
d 2 x 3 dx + 10 x = 0.5
dt 2 + dt
(i) Find the general solution of the equation for x.
(ii) Given that x = 0.1 and x
d = when
0 t = 0, find the particular
dt
solution.
(iii) At what reading will the pointer settle?
(iv) What length of time will elapse before the amplitude of the
oscillations of the pointer is less than 20% of the final value of x?
7 In an electrical circuit, the charge q coulombs stored in a capacitor at
time t is given by the differential equation:
d 2q dq +
100 10 000q = 1000 sin100t
dt 2 + d t
where t is the time in seconds.
(i) Find the complementary function of the differential equation.
(ii) Decide if the circuit is overdamped, critically damped or
underdamped.
(iii) Find the particular integral and hence the general solution.
183
195
dq
(iv) Initially both q and , the current in the circuit, are zero. Find the
dt
6 particular solution.
(v) If you have graphing software use it to sketch the solution and
describe how the charge in the circuit changes with time.
8 A simple model for the motion of a delicate set of laboratory scales,
when an object of mass m is placed gently on the scales, is given by the
differential equation
2
m dx 7 dx + 5x = 10m
dt 2 + dt
where x is the displacement from the initial position at time t.
(i) Find the particular solution that models the motion of the object
in each of the cases:
(a) m = 2 (b) m = 2.45 (c) m = 4
(ii) Sketch a graph of the solution for each value of m. Describe the
motion in each case.
CP 9 (i) Solve the equation dx − = kx e pt , where k p≠ ,
dt
(a) using the integrating factor method
(b) by finding the complementary function and particular integral
(ii) Repeat part (i) for the equation dx − = kt
kx e .
dt
PS 10 If you have graphing software to hand when doing this question it would
be helpful. In normal running, a machine vibrates slightly and this is
monitored by a marker on it. The marker moves along a straight line. Its
displacement in mm from an origin is denoted by x.
The motion of the marker is subject to the differential equation
d 2 x + 9 x = 0.
2
dt
At the start of the motion, when t = 0, the marker is stationary and x = 2.
(i) Find the particular solution.
Draw a graph of x against t.
One day the machine’s operator attaches another mechanism to it with
the result that a forced oscillation is applied to the machine and the
differential equation governing x becomes
d 2 x + 9x = 5cos 2t .
dt 2
The same initial conditions apply.
(ii) Verify that the solution is x = cos 3t cos
+ 2 t.
Draw a graph of x against t for ഛ
0 ഛ t 2π.
Describe what has happened to the motion of the marker.
Another day the operator changes the frequency of the forced oscillation
so that the differential equation becomes
d 2 x 9x = 5cos 3t .
dt 2 +
184 Again the same initial conditions apply.
196
(iii) Solve the equation.
Draw a sketch graph of x against t.
Predict what happens to the machine when this forced oscillation 6
is applied.
11 Find the general solution of the differential equation
d 2 x + 4 dx + 4 x = −7 2 . t 2
dt 2 dt
12 Obtain the general solution of the differential equation

2
d x 6 dx + 13x = 75cos 2t ,
dt 2 + dt
dx 0
Given that x 5= and dt = when t 0 = , find x in terms of t.
Show that, for large positive values of t and for any initial conditions,
x ≈ 5 cos(2 t − θ ),
where the constant q is such that tan θ = 4 .

3
6.5 Use of substitutions to transform

differential equations.
In this chapter, you have seen how to solve certain types of differential
equations, but what do you do when you meet one with a different pattern? In
this final section, the range of equations that you can solve is extended by the
y = dy
use of the substitutions v = , x = and e t
z
dx . Once you have worked
x
with these, you will find it straightforward to work with other substitutions.
y dy =  y 
The substitutionv= x : equations of the typedx f x 
y
If you introduce the new variable v = you can show, through use of the
x
chain rule, that:
dy = + d (vx d
y vx= ⇒ vxv d . dx
) =v +x v
dx
dx dx
dy y
So the equation x = x can be written as:
d
f()
v x+ v d = f (v)
dx
⇒ x vd = f (v)v−
dx
185
197
This equation has separable variables x and v:
6 dv
f (v) − v
= dx
x
The solution is obtained through integration as:
1 1 dx
∫ f (v) − v
dv = ∫ x
Example 6.15 Solve the differential equation

dy = y 2 + 2xy
dx x2
Solution
Rearrange the equation into the form:
( )+ 2( xy)
2
dy = y
dx x
=
You can then use the substitution y vx to give:
v x+ v d = +v 2 2v
dx
Rearranging gives: x vd = +v 2v
dx
separating the variables: dv = dx Using partial fractions:
v (v + 1) x
1 = +A+ B
1 1 dx v (v )+ 1 v v 1
∫ v (v +1 )
dv = ∫ x ⇒ A v( )+ 1+ ≡ Bv 1
⇒ A B+ A= =B =−
∫ (1v v− )
1 dv = 1 dx 0, 1, 1
+1 ∫ x ∴ 1 = −1 1
v (v )+ 1 vv +1
ln v − ln (v + 1=) ln (x)c+
 v  Using the property of
ln   = ln x c+ logarithms:
( +
v 1 ) A
ln A −ln=
ln B
B
v = Ax and A e= c
v +1
y =
Rearranging gives: Ax
y x+
2
or Ax
y = −
1 Ax
dy = y 2 + 2xy Ax
2
The general solution to the equation 2 is y = 1 − Ax .
dx x
186
198
The substitution x = et : equations of the type
ax 2 d
dx
2
y +
2
dy
bx dx + =
cy( ) f x 6
The differential equation
2
2 d y dy + =
ax 2
+ bx cy f (x)
dx d x
can be reduced to one with constant coefficients by using the substitution
x e= t
In this case:
dy = dy dy = ×t dy dt = × dy dy
x x e et e −=t
dx dt dx dt dx dt dt
d 2 y = d 2 y − dy
2
x 2 2
dx dt dt
d 2 y = ×2t  −t dy 
x2 eed −te 
dx 2 dt  dt 
 t dy 2 
−t d y
=×2−
t
eee − t− + e
 dt dt 2 
= − d+y d 2y
dt dt 2
The equation thus becomes:

d 2 y d y  dy + =
f (e )
t
a 2 −  + b cy
 dt dt  dt
d2y + − dy
i.e. a 2 (b a ) + =
cy f e( t)
dt dt
Example 6.16 Find the solution of the equation

2
2 d y dy + 2 dy = when x 1 = .
x 2
− 10 x 24 y = 6 x for which y =
1 , 1
dx dx dx
Solution
The equation reduces to the linear equation with constant coefficients
d 2 y − dy + 2t
2 11 24 y = 6e 2
d 2y dy
2 d y = −
dt dt Using x dx 2 dt 2 dt
The auxiliary equation for this equation is x dy = and
dy
x = .e t dx dt
λ − 11λ + 24 0=
2
(λ − 3)(λ − =8) 0
➜
187
199
The complementary function is thus
6 = e 3 t + B e 8t
yA
Trial function: y= a t e 2
2
⇒ dy = 2ae, 2t d y2 = 4ae 2t
dt dt
d 2 y − dy +
11 24y = (4 a22
− 24ae+ a ) 2t
dt 2 dt
= 6ae 2t ≡ 6e 2t
⇒ a 1. = The particular integral is = y e 2t
2t
The particular integral is y e=
The general solution is thus:
yA= e 3 t + B e 8t + e 2 t
t
If you now substitute x e= , to give:
=
y Ax 3
+ Bx 8 + x 2
dy = when x 1 = to arrive at the
Finally use the initial conditions y 1= , 1
dx
particular solution:
3
− +
x 8
5 x 2
y x=
5
y 1= when = x 1 ⇒ 1 = +A+B⇒ +1= AB 0

dy =
x1⇒ 1 =3 +
8 A+ ⇒ 5B1 = −
dx 1 when =
2B
⇒A= 1 B = − 1
5 5
dy
The substitution z = : equations where x or y do
dx
not both appear explicitly
2
d y = −  dy  or d y = dy + are to be
2 2
Equations of the type x 2 (1) y
d x 2
d x  d x 2
dx
solved here.
2
In this case, you can use the substitution z =
d y
which leads to d z = y2 .
d
dx dx dx
This reduces the second order equation into a first order equation between
z and x. If this equation can be integrated and z is found in terms of x, a
further integration then gives y in terms of x.
188
200
Example 6.17 Solve the equation
 dy   d 2 y 
   2  = ,xwhere = y 2 and
d x   d x 
dy = at x 1 = .
dx
1 6
Solution
The equation reduces to
d = dy = d 2 y = dz
zz x z and
dx dx dx 2 dx
Separate the variables:

zdz x=x d
and integrate to give:
1 z2 = 1 x2 + A
2 2
Apply the condition that z 1= when x 1 = to give A 0 = .
∴ dy =
z x= or x
dx
You now have to integrate to find y in terms of x:
y = 1 x2 + B
2
and using the condition y 2= when x 1 = gives B 3 = , so that the
2
solution is:
y = 1 (x 2 + 3)
2
Equations not containing x explicitly

dy
In this case, you can again use the substitution =z , but now rewrite as:
dx
d 2 y = dz = dz × dy = d
2 zz
dx dx dy dx dy
Using this substitution will reduce the second order equation into a first
order equation between z and y. Then if you can integrate this equation and
write z in terms of y (e.g. z = f (y)), a further integration will give you x in
terms of y.
dy = ( ) 1 dy x=
f y
dx
∫
⇒ ( )f y
Note
This type of equation is often seen in dynamics problems where y represents
dy
the displacement s, x the time t and x the velocity v.
d
189
201
Example 6.18 Solve the equation
6 d 2 y = dy + , given that y 0 = and

dx
2 2 (1) y
dx
dy = when x 0 = .
dx
1
Solution
dy
Since x is not involved explicitly, you can use the substitution z =
dx
which transforms the differential equation into the first order equation
6 DIFFERENTIAL EQUATIONS
between z and y.
z dz = 2 z1( )+ y dy = d 2 y
dy z, 2 = z dz
dx dx dy
Divide through by z
dz = 2 (1 )+ y
dy
Integrate with respect to y.
You know that z= 1 when = x 0
z = 2y y+ + 2 A and that y= 0 when . x 0= It thus
1= A follows that z= 1 when = y 0.
⇒ 2
z = 2y y+ + = +1 (y) 1 2
dy = +
(y) 1 2
dx
Separating the variables and integrating:

1
∫ (y)+ 1 2 dy
= ∫ dx
− (y 1+)1 = +x C
− ( 1) = −x 1 y 0= when = x 0 ⇒ = − C 1
y+1
Rearranging gives:
y = x
1− x
y + 1= − 1 ⇒ y = − 1 −1
1 x 1 x
1 −1 (− x )
⇒ y= = −x
1− x 1 x
190
202
Exercise 6E y
Use the substitution v = in questions 1–5.
x
1 Solve the equation
dy = x y+ , given that y 2 = when x 1 = .
dx x
6
dy = x 2 + y 2
2 Solve the equation 2 , given that y 4= when x 1 = .
dx 2x
d = y x−
3 Solve the equation y , given that y 0= when x 1 = .
dx y x+
()
2
dy = +y y
Solve the equation 2
6.5 Use of
4
dx x x
dy = +
x 2y2+
substitutions
5 Solve the equation x y
dx
PS 6 (i) Use the substitution z x=y+ to transform the equation
d y = 1 + 2 x 2+ y
d x 1 −2 x 2− y
into an equation involving z and x only.
transform differential equations.

to
(ii) Solve the equation.
7 Use the substitution x e= tot show that the differential equation
2
d y − dy +
x 2
2 2 x 2y x= 3
dx dx
reduces to a linear equation with constant coefficients. Solve the
dy
equation given that y 0= and = when
0 x1 =.
d x
8 Use the substitution x e= to show that the differential equation
t
2d 2y dy + =
x x y 0
dx 2 + d x
becomes
2
d y
y 0
dt 2 + =
Hence, or otherwise, find y in terms of x given that y 0= when x 1 = ,
dy = when x e = π .
and 2
dx
2
2d y − dy + 3
x 2 5 x 9y x= +6
dx d x
10 Use the substitution x e= tot show that
2 2 3 3 2
dy = dy 2 d y
= d y − d y 3 d y
= d y − d y + dy
x ,x 2 2 , x 3 3 3 2 2
dx dt dx dt d t dx dt dt dt
Hence or otherwise solve the equation
3d 3y + 2 d 2y + dy = 2
x 3 3x 2 x x
dx dx d x
191
203
2
d y +  dy  =
2
6 dx 2
2x
 dx 
0,
dy = when x 0. =
where y = 0, 1
dx
2
(1 + x )ddxy2 + 2x ddxy + 3x −2 = 0 ; x 0>
2

2
d y − 2 =
3y 0,
dx 2
dy = when x 0 = .
where y = 2, 4
dx
3
d 2 y +  d y  =
y 0
dx 2  dx 
15 Show that the substitution y z=x cos , where z is a function of x,
reduces the differential equation
2d 2y + dy + =
cos x 2 2cos xsinx 2y x x cos 3
dx dx
to the differential equation
d 2z z x .
dx 2 + =
dy = when x 0 = .
Hence find y given that y 0= and 1
dx
16 The variables x and y are related by the differential equation
2 2
d y 2 dy dy 
y 2
2 + 2y
+ 2y − 5y 3 = 8e −.x
dx dx  dx 
Given that =v y , 3 show that

2
d v v −x
2 d − 15v = 24e .
dx 2 + dx
Hence find the general solution for y in terms of x.
9231 Paper 11 Q7 June 2011
CP 17 Sometimes it is possible to solve a nonlinear equation by making a
change of the dependent variable that converts it into a linear equation.
The most important class of such equations is of the form
dy +
p (x) y = q (x) y n
dx
such equations are called Bernoulli equations.
192
204
(i) Show that by using the substitution v y= 1 n
−Bernoulli’s equation
can be reduced to the linear form
dv + −
dx
(1 n)xp( v) = −(1 n)xq( ) 6
(ii) Use this method to solve the differential equation
dy + 1 = 2
y xy x ; > 0
dx x
M 18 A particle of mass 0.5 kg is released from rest and falls a distance s
vertically. A resisting force of magnitude 0v.1 N acts vertically upwards on
6.5 Use of
the particle during its descent, where v m s –1 is the velocity of the particle
at time t s after release. The value of g is taken to be 10 m s –2.
substitutions
(i) Show that the acceleration of the particle is (10 0.2
− v) m s –2
(ii) Find the velocity of the particle after 5 s.
(iii) Find the distance fallen in that time.

to
KEY POINTS
Integrating factors
1 Any first order linear differential equation can be written in the form
dy +
P y = Q where P and Q are functions of x only.
dx
dy +
2 To solve the equation Py = Q you multiply throughout by
dx
P dx
the integrating factor R e = ∫ . The solution is then given by
Ry = ∫ RQdx.
Second order differential equations
3 A second order linear differential equation with constant coefficients
d 2 y + dy + =
can be written in the form a 2 b cy f (x) , where a, b and c
dx dx
are constants.
4 The equation is homogeneous if f (x) = 0. Otherwise it is non-
homogeneous.
5 When you are given a differential equation of the form above, you
can immediately write down the auxiliary equation λaλb c +0 + =
2
.
6 Each root of the auxiliary equation determines the form of one of the
complementary functions.
7 If the auxiliary equation has two real, distinct roots λ 1 and λ 2, then the
complementary function of the differential equation is:
yA = eλ x + B eλ x
1 2
193
205
8 If the auxiliary equation has a repeated root , then the
6 complementary function of the differential equation is:

α
y = e (x A Bx
+ )
9 If the auxiliary function has complex roots λ α β= ± i, then the
complementary function of the differential equation is:
α
y = e x (A sin β x B
+ cos β x)
2
10 Motion for which the differential equation is
d x + ω 2 x = is0 called
dt 2
simple harmonic motion (SHM). The solution of this differential
equation is of the form:
xA = sin ω t B+ cos ω t or x a= sin(ω tε+ )

22
The period of this motion is 2π and the amplitude is A B +.
ω 2
11 Motion for which the differential equation is
d x + α dx + ω = 2
x 0,
2
dt d t
where  > 0, is called damped harmonic motion. The following table
shows the features of damped harmonic motion.
 2 − 4w 2 Type of damping
 =0 no damping
 2− 4w 2 > 0 overdamping
 2− 4w 2 = 0 critical damping
 2− 4w 2 < 0 underdamping
12 The general solution of the non-homogeneous linear differential

equation with constant coefficients
2
d y + dy + =
a 2 b cy f (x)
dx dx
is the sum of the complementary function and a particular integral.
13 The number of unknown constants is the same as the order of the
equation.
14 A particular integral is any function that satisfies the full equation; it
does not contain any arbitrary constants.
15 To find the particular integral, use the trial function shown in the
following table:
Function Trial function
linear function ax + b
n −1
polynomial of order n anx na +x n −1
++
...+ a1x0 a
trigonometric function involving px b px+ cos
a sin
sin px and /or cos px
px px
exponential function involving e ae
sum of different functions sum of matching functions
194
206
16 If the trial function for a particular integral is the same as one of the
complementary functions, you multiply the trial function by x.
6
Use of substitutions to transform differential equations
17 Use the substitution y vx = to transform an equation of the form
dy = y
into an equation in v and x which is separable.
dx
f ()
x
t
18 Use the substitution x e= to transform an equation of the form
6.5 Use of
2
2 d y
ax 2
+ bx dy + =cy f (x) into a linear equation with constant
dx d x
substitutions
coefficients in y and t.
dy = d 2 y = dz
19 Use the substitution z; to transform a second
dx dx 2 dx
order equation not containing y explicitly into a first order equation
between z and x.
dy = d2y = d

20 Use the substitution z; z z to transform a second
to
dx dx 2 dy
order equation not containing x explicitly into a first order equation
between z and y.
LEARNING OUTCOMES
■ recognise differential equations which can be solved using an
integrating factor
■ find an integrating factor and use it to solve a first order linear differential
equation, to find both the general and particular solutions
d 2 y + dy + =
■ solve differential equations of the form a dx 2 b dx cy 0 using the
auxiliary equation method
■ understand and use the relationship between different cases of the
solution and the nature of the roots of the auxiliary equation
d 2 y + dy + =
■ solve differential equations of the forma dx 2 b dx cy x f ( ) by
solving the homogeneous case and adding a particular integral to the
complementary function
■ find particular integrals for cases where f (x) is a polynomial,
trigonometric or exponential function, including cases where the
form of the complementary function affects the form required for the
particular integral
195
207
2
solve the equation for simple harmonic motion, x2 = − ω 2 x, and be
d
6 ■
■
able to relate the solution to the motion dt
model damped oscillations using second order differential equations
■ interpret the solutions of equations modelling damped oscillations in
words and graphically
■ use a given substitution to reduce a differential equation to a first order
or second order linear equation with constant coefficients or to a first
order equation with separable variables.
196
208
Index
(1 + x ) , Maclaurin series 79, 83 with real distinct roots 159 non-homogeneous 177–82
2 × 2 matrices with repeated roots 162–3 notation 146, 155
Index
determinant 16–17 complex numbers second order 154–63
eigenvalues and eigenvectors 38–43 applications 117 simple harmonic motion 168–72
inverse 17 de Moivre’s theorem 121–4 substitution x = et 187–8, 195
3 × 3 matrices exponential form (z = reiθ ) 137–9, substitution y = vx 185–6, 195
determinant 23–5, 57 144 dy
substitution z = 188–90, 195
eigenvalues and eigenvectors 43–5 modulus-argument form 117–18, dx
finding the intersection of three 143 differentiation
planes 30–4, 58–9 multiplying and dividing 119–20 hyperbolic functions 62, 83
inverse 23–7, 57 nth roots of any complex inverse hyperbolic functions 62–3,
number 129 83
adjugate (adjoint) matrices 25–6, 57 nth roots of unity 125–9 inverse trigonometric
Archimedes’ tombstone 113 summing series using de Moivre’s functions 61–2, 83
arc length of a curve 102–3, 115 theorem 139–41 second derivative for relations
with Cartesian coordinates 104–5 constant coefficients, differential given implicitly 64–7, 83
with polar coordinates 106–7 equations 155 second derivative for relations
area under a curve cosech x 5 given parametrically 67–8, 83
approximating series sums 99–100 cosh x 2, 14 dividing complex numbers 119
approximation using derivative of 62
rectangles 97–8 graph and properties 3 eigenvalues and eigenvectors 59
Argand diagrams 117–18 integration 89 for 2 × 2 matrices 38–43
effect of multiplying and dividing cosh-1 x (arcosh) 9–13 for 3 × 3 matrices 43–5
complex numbers 119 derivative of 62–3 finding powers of matrices 48–9
nth roots of any complex use in integration 89–91 for larger matrices 45
number 130–1 cos x, Maclaurin series 79 properties 41
nth roots of unity 128–1 cos-1 x, derivative of 62 e, polynomial approximation 70–2, 79
powers of a complex number coth x 5 expansion of the determinant by the
123, 125 critical damping 175 first column 23–4, 25
argument of a complex number 118 exponential form of a complex
auxiliary equation method, differential damped harmonic motion 172–5, number (z = reiθ ) 137–9, 144
equations 156–9, 193–4 194 exponentially decaying (damped)
complex roots 167 de Moivre, Abraham 124 oscillations 172–5
real distinct roots 159–61 de Moivre’s theorem 121–4, 143
repeated roots 162–3 finding multiple angle first order differential equations 146
identities 133–4 see also differential equations
boundary value problems 159, 160–1 summing series 139–41 Frobenius, Georg 53
determinants frustum, surface area of 109–10
calculus see differentiation; integration of a 2 × 2 matrix 16–17
catenarys 1–2 of a 3 × 3 matrix 23–5, 57 Hamilton, William 53
Cauchy, Augustin Louis 53 for square matrices 17 homogeneous differential
Cayley, Arthur 53 zero value 17, 58 equations 155
Cayley–Hamilton theorem 50–3 diagonal form of a matrix 49 hyperbolic functions 2–3, 14
characteristic equation of a matrix 41 differential equations 193–5 derivatives of 62, 83
Cayley–Hamilton theorem 50–3 auxiliary equation method 156–63, graphs and properties 3–6
characteristic polynomial of a 167 integration 88–9, 115
matrix 41 boundary value problems 159, 160–1 inverse 8–13
circular (trigonometric) functions 2 damped harmonic motion 172–5 hyperbolic identities 5
cofactors 24, 25, 26, 57 formation of 145–6
communication x initial value problems 159–60 implicit functions, second derivative
complementary functions of integrating factors 147–51 of 64–7
differential equations 158, 178 linear 147–8, 155 initial value problems 159–60
197
209
integrating factors 147–51, 193 forming differential second order differential
integration equations 145–6 equations 146, 154–5, 193–5
arc length of a curve 102–7 modulus-argument form (polar form) auxiliary equation method 157–63,
hyperbolic functions 88–9, 115 of a complex number 117–18, 167
surface area of a solid of 143 damped harmonic motion 172–5
revolution 109–12 multiplying and dividing 119–20 non-homogeneous 177–82
Index
using inverse hyperbolic modulus of a complex number 118 notation and vocabulary 155
functions 89–91, 115 multiple angle identities 133–4 simple harmonic motion 168–72
using reduction formulae 92–6 multiplying complex numbers 119–20 series sums
using the inverse sine function effect in an Argand diagram 119 approximation of 99–100
85–7, 115 using de Moivre’s theorem 139–41
see also differential equations non-homogeneous differential sheaf of planes 30
intersection of three planes 29–34, equations 155, 177–8, 194–5 simple harmonic motion 168–72, 194
58–9 particular integrals 178–82 simultaneous equations
invariant lines 38–40 trial functions 179, 180–2 geometrical interpretation 18–19
inverse of a 2 x 2 matrix 17 nth roots of any complex matrix method 17–19
inverse of a 3 x 3 matrix 23–7, 57 number 129, 144 singular and non-singular matrices 17
inverse hyperbolic functions 8–13, 15 nth roots of unity 125–9, 143 sinh x 2, 14
derivatives of 62–3, 83 derivative of 62
use in integration 89–91, 115 order of a differential equation 146 graph and properties 4
inverse of a product of matrices 17 oscillations integration 88–9
inverse trigonometric functions, damped harmonic motion 172–5, sinh-1 x (arsinh) 8–9
derivatives of 61–2, 83 194 derivative of 63
simple harmonic motion 168–72, use in integration 89–90
linear differential equations 147–8, 194 sin x, Maclaurin series 79
155 overdamping 174 sin-1 x
lines of invariant points 38 derivative of 61
ln(1 + x), Maclaurin series 79, 83 paraboloid, surface area of 111 use in integration 85–7, 115
parametric equations 2–3 solids of revolution
Maclaurin approximations (Maclaurin parametric functions, second surface area 109–12, 115
expansions) 73–4 derivative of 67–8 volume 115
Maclaurin series 74–6, 83 particular integrals 178–82, 194 spiral dilation 119
for standard functions 79–80 phase shift 169 spreadsheets, polynomial
Markov processes 48 planes, intersection of three approximation of ex 72
matrices planes 29–34, 58–9 summing series
applications 16 polar form of a complex number see approximation of 99–100
Cayley–Hamilton theorem 50–3 modulus-argument form of a using de Moivre’s theorem 139–41
characteristic equation and complex number surface area, solids of revolution 109–12
polynomial of 41 polynomial approximations
determinant of a 2 x 2 matrix of ex 70–2 tan-1 x
16–17 Maclaurin approximations 73–4 derivative of 62
determinant of a 3 × 3 Maclaurin series 74–6, 79–80, 83 Maclaurin series 83
matrix 23–5 positive sense on a curve 103 tanh x 4, 14
eigenvalues and eigenvectors powers of complex numbers, de derivative of 62
38–45, 59 Moivre’s theorem 121–4 tanh-1 x (artanh) 9
finding the intersection of three powers of matrices 48–9 third order differential equations 155
planes 30–4, 58–9 principal argument of a complex transformations, invariant lines 38–40
inverse of a 2 x 2 matrix 17 number 118 transition (stochastic) matrices 48
inverse of a 3 x 3 matrix 23–7 problem solving x trial functions 179, 180–2, 194–5
inverse of a product of matrices 17 proof x trigonometric (circular) functions 2
powers of 48–9 trigonometric identities 5
singular and non-singular 17 rates of change, formation of multiple angle identities 133–4
solving simultaneous differential equations 145–6
equations 17–19 recurrence relations 52 underdamping 175
transition (stochastic) 48 reduction formulae 92–6 unity, nth roots of 125–9
minors 23–4, 57
198 modelling x, 145 sech x 5 validity of Maclaurin series 75
210
211
212
目录
Cover 1
Title Page 3
Copyright 4
Contents 5
Introduction 7
How to use this book 8
The Cambridge International AS & A Level Mathematics 9709 syllabus 10
1 Hyperbolic functions 13
1.1 Hyperbolic functions 13
1.2 Inverse hyperbolic functions 20
2 Matrices 28
2.1 Important results relating to 2 × 2 matrices 28
2.2 Finding the inverse of a 3 × 3 matrix 35
2.3 Intersection of three planes 41
2.4 Eigenvalues and eigenvectors 50
3 Differentiation 72
3.1 Differentiating inverse trigonometric functions 73
3.2 Differentiating hyperbolic functions 74
3.3 Differentiating inverse hyperbolic functions 74
3.4 Finding the second derivative for relations given implicitly 76
3.5 Finding the second derivative for relations given parametrically 79
3.6 Polynomial approximations and Maclaurin series 82
3.7 Using the Maclaurin series for standard functions 91
4 Integration 97
4.1 Integration using the inverse sine function 97
4.2 Integrating hyperbolic functions 100
4.3 Integration using inverse hyperbolic functions 101
4.4 Integration using reduction formulae 104
4.5 Approximation of areas using rectangles 109
4.6 Using areas under curves to approximate series sums 111
4.7 The arc length of a curve 114
4.8 Surface area of a solid of revolution 121
5 Complex numbers 129
5.1 The modulus and argument of a complex number 129
5.2 De Moivre’s theorem 133
5.3 The nth roots of a complex number 137
5.4 Finding multiple angle identities using de Moivre’s theorem 145
5.5 The form z = re iθ 149
6 Differential equations 157
6.1 Integrating factors 159
6.2 Second order differential equations 166
6.3 Auxiliary equations with complex roots 179
6.4 Non-homogeneous differential equations 189
6.5 Use of substitutions to transform differential equations 197
Index 209
213

Further Pure 2

Uploaded by

Copyright:

Available Formats

Further Pure 2

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Further Pure 2

Uploaded by

Copyright:

Available Formats

What are the main topics covered in the document?

What are the main topics covered in the document?

What are some examples of integration techniques mentioned?

What are some examples of integration techniques mentioned?

1

6 Differential equations 145

? This symbol highlights points it will benefit you to discuss with

The following diagram illustrates the permitted combinations for AS Level

Paper 1 and Paper 3

1.1 Hyperbolic functions

1.1 Hyperbolic functions

Hyperbolic functions and circular functions

1.1 Hyperbolic functions

Graphs and properties

As you saw in Example 1.1:

1.1 Hyperbolic functions

e x3= Since ex can’t be negative.

Exercise 1A 1 Using the definitions of the hyperbolic functions, prove that

1.1 Hyperbolic functions

Deduce that sinh x = 21 (cosh 2x1)− .

The inverse function of the sinh function is denoted by sinh−1 , or sometimes

1.2 Inverse hyperbolic functions

ln(2 – √3) O ln(2 + √3) x

1.2 Inverse hyperbolic functions

1.2 Inverse hyperbolic functions

CP 6 Use the formula sinh ( A B + ) = sinh A B +sinh A B

1.2 Inverse hyperbolic functions

■ cosh x ■ sech x ■ coth x

■ sketch the graphs of hyperbolic functions

■ prove and use the identities involving hyperbolic functions

Matrices are used extensively in computer games to apply rotations and

Using matrices to solve simultaneous equations

Example 2.1 Use a matrix method to solve the simultaneous equations

11 2 3− 2 1− y  = 11 2 3− 12  As M −1M = I the

The solution is x = 5, y = −2.

Geometrical interpretation in two dimensions

Find the possible values of x.

(iii) Draw diagrams for each of the transformations A and B to

(i) find the image of the point (2, −1)

Find the possible values of k.

(i) Find the determinants of M and N.

2 If (x, y) is an invariant point under a transformation represented

Finding the determinant of a 3 × 3 matrix without

which is sometimes referred to as the expansion of the determinant by

a 2a3 a1a3 a1a2

a 2a3 a1a3 a1a2

2 Step 2: Evaluate the cofactors

Exercise 2B 1 Evaluate these determinants.

ka1ka ka2 3 a1a2 3a

(ii) Explain in terms of volumes why multiplying all the elements in

▲ Figure 2.4 ▲ Figure 2.5

If none of the planes are parallel, there are three possibilities:

2 The three planes share a

» The planes form a triangular prism (see Figure 2.8)

Each pair of planes intersects

Finding the unique point of intersection

You cannot Solution

① – 2 × ③: −5x + y = 7 − 2k For the equations to be

Example 2.6 Solve the set of equations

Example 2.7 Solve the set of equations