© 2017 University of South Africa

All rights reserved

Printed and published by the

University of South Africa
Muckleneuk, Pretoria

MAT1613/1/2018–2020

70541337

Layout done by the Department of Mathematical Sciences

 iii MAT1613/GD001/2018–2020

Contents

Page
About the module v

Chapter 1 Applications of Derivatives 1

1.1 Related Rates 2

1.2 Graphs and Functions 7
1.3 Optimization 22
1.4 Rolle’s Theorem and The Mean Value Theorem 27

Chapter 2 Transcendental Functions 33

2.1 The Logarithmic and Exponential Functions 34

2.1.1 The Logarithmic Function 34
2.1.2 The Exponential Function 35
2.1.3 Functions of the Form f (x) = g(x)h(x) 36
2.2 Determining indeterminate forms with L’Hôpital’s Rule 37
±∞
2.2.1 Limits of the Form 00 and ±∞ 38
2.2.2 Limits of the Form ±∞ · 0 39
2.2.3 Limits of the Form ∞ − ∞ 40
2.2.4 Limits of the Forms 00 , 1∞ , ∞0 41
2.3 The Inverse Trigonometric Functions 43
2.3.1 Domains for the Inverse Trigonometric Functions 43
2.3.2 Derivatives of Inverse Trigonometric Functions;
Related Integrals 48
2.4 Hyperbolic Functions 49

Chapter 3 Applications of Integration (areas between curves, solids of revolution) 54

3.1 Areas between Curves 54

3.2 Volumes of Solids of Revolution 55
 iv

Chapter 4 Advanced Techniques of Integration 59

4.1 Basic Integration Formulas 59

4.2 Integration by Parts 60
4.3 Trigonometric Integrals (using substitution) 66
4.4 Trigonometric Substitution 69
4.5 Partial Fractions 75
4.5.1 Linear Factors in the Denominator 76
4.5.2 Repeated Linear Factors in the Denominator 79
4.5.3 Irreducible Quadratic Factors in the Denominator 80
4.5.4 Repeated Irreducible Quadratic Factors in the Denominator 81
4.6 z-Substitution 84

Chapter 5 The Improper Integral 89

5.1 Integrals of which the Integrand is unbounded over the Integration Interval 89
5.2 Integrals of which the Integration Interval is Unbounded 90
5.3 Integrals for which both (5.1(S)) and (5.2(S)) hold 91
5.4 The Limit Comparison Test 92

Chapter 6 Infinite Sequences and the Taylor Series 94

6.1 Limits of Infinite Sequences 94

6.2 Taylor Series 94
 v MAT1613/GD001/2018–2020

About the module

Calculus is the mathematics of motion and change. Where there is motion or growth, where variable forces
are at work producing acceleration, calculus is the right mathematics to apply. Differential calculus deals
with the problem of calculating rates of change. Integral calculus deals with the problem of determining
a function from information about its rate of change. It enables people to calculate the future location of
a body from its present position and includes the knowledge of the forces acting on it, to find the areas
of irregular regions in the plane, to measure the lengths of curves and to locate the centres of mass of
arbitrary solids.

Today, calculus and its extensions in mathematical analysis are far reaching indeed, and the physicists,
mathematicians and astronomers who first invented the subject would surely be amazed and delighted to
see what a profusion of problems it solves and what a wide range of fields now use it.

How to study for this module:

• Prescribed book

These notes were written in addition to the prescibed textbook.

• Referenced notes

References to these notes will be followed by an (S), for instance p. 40(S), means p. 40 of the notes.

The notes must always be used in conjunction with the prescribed book. The notes provides extra expla-
nations and examples to some of the prescribed study material.

• Content

In the notes we discuss the following two topics namely: Applications of the derivative (Chapters 1, 2 and
6) and Integration methods (applied to specific problems) (Chapters 3, 4 and 5).
 vi

• Exercises

At the end of each section in your textbook or in the notes some exercises on that section are given.
Each time when you finish a section, it is very important that you try to do as many of these exercises
as possible. Remember that very little theory is required of you in this module, i.e. you do not need to
study many theorems and their proofs. So, please, do as many exercises as possible!!

• Pocket calculators

Pocket calculators are not allowed during the examinations. Therefore we expect that you will not use
one when doing your assignments. To get used to working without a pocket calculator never use it when
working on this module. You need not do any of the exercises which indicate that they are calculator
based.

Prescribed Study Material for this module

The sections you must study in the prescribed textbook in conjunction with the notes in the study guide
are below.
Topic 1

1.1 Exponential and logarithmic functions and their derivatives and integrals.
1.2 The hyperbolic functions and their derivatives and integrals.
1.3 The inverse trigonometric functions, their definition sets, derivatives and integrals.

Topic 2

2.1 Limits involving infinity.

2.2 Indeterminate forms and L’Hôpital’s Rule.

Topic 3

3.1 The mean value theorem and Rolle’s theorem.

Topic 4

The sketching of graphs using the following:

4.1 Asymptotes.
4.2 The first derivative to find maximum and minimum values.
4.3 The sign pattern of the first derivative to determine where functions increase and decrease.
4.4 The second derivative and its sign pattern to find the concavity of a function.
4.5 The inflection point(s) of a function (if they exist).
4.6 The second derivative test.
 vii MAT1613/GD001/2018–2020

Topic 5

5.1 Optimization.
5.2 Related rates.
5.3 Rates of change in economics and sciences.

Topic 6

6.1 Area between curves.

6.2 Solids of revolution: Finding volumes by using slicing, disks and washers.

Topic 7

Integration
7.1 Review of basic formulas and techniques.
7.2 Integration by substitution.
7.3 Integration by parts.
7.4 Trigonometric substitution and integration using trigonometric identities.
7.5 Integration of rational functions using partial fractions.
7.6 Integration using z-substitution.
7.7 Identification and computing of improper integrals.
7.8 Using the limit comparison test to decide if an integral converges or diverges.

Topic 8

8.1 Finding the limit of a sequence.

8.2 Constructing Taylor polynomials for certain functions. (The remainder is not required.)

Outcomes
After studying this module, the learner should have acquired skills in

• solving related rates problems, graph sketching (using the derivative) and solving optimization prob-
lems.

• solving basic problems using the Mean value theorem and/or Rolle’s theorem.

• using L’Hôpital’s Rule to solve indeterminate forms.

• obtaining the inverse trigonometric functions, their derivatives and related integrals.

• finding the derivatives and integrals of hyperbolic functions.

• evaluating areas between curves.

• evaluating volumes of solids by revolving plane regions about the co–ordinate axes.
 viii

• applying more advanced techniques of integration.

• selecting a suitable method of integration to solve a problem.

• evaluating improper integrals.

• testing for convergence or divergence of improper integrals.

• determining the limit of a sequence if it exists.

• constructing the Taylor polynomial of a certain given function.

We have provided a mind map so that you can see how the topics that we will discuss in this module are
related. Since this module is a continuation of MAT1512 some topics are already partially dealt with in
MAT1512. In the MAT1613 module, we discuss some advanced methods and applications.

Mind map
Basic Integration methods (substitution)

Trigonometric Substitutions
Trigonometric Integrals

Applications of Integration
Integration by Parts

Partial Fractions

Solids of Revolution
Methods of Integration

Area between curves

z- Substitution

Improper Integrals
Specific Problems
Hyperbolic Functions

Other Transcendental
Module MAT1613

Functions
Inverses
Applications of the Derivative

L'Hopitals Rule

Taylor Formula
Limits

Graph Sketching

Mean Value Thm

Optimization
Related Rates

Rolles Thm
 1 MAT1613/GD001

Chapter 1

Applications of Derivatives

Introduction
This chapter shows how to draw conclusions from derivatives: how to calculate rates of change, how to
find a function’s extreme values, how to determine the shape of a function’s graph, and how to find a
function when we know its first derivative and its value at a single point.

The key to recovering functions from derivatives is the Mean Value Theorem, a theorem whose corollaries
provide the gateway to integral calculus.

Outcomes
After studying this chapter the learner should be able to:

• solve basic word problems concerning related rates, for example, change in volumes, areas, speeds,
etc.

• solve a range of basic mathematical problems using the Mean Value Theorem and/or Rolle’s Theo-
rem.

• use first and second derivatives to sketch the graphs of rational functions and trigonometric functions.

• solve basic word problems in optimization, involving, for example, volumes, areas, speed, etc.

In this chapter we will provide some examples on the following topics.You have to work on the relevant
topics in your textbook.

1.1 Related Rates.

1.2 Graphs of Functions.

1.3 Optimization.

1.4 Rolle’s Theorem and the Mean Value Theorem.

 2

1.1 RELATED RATES

Most quantities encountered in everyday life change with time. This is especially evident in scientific
investigations for instance a chemist may be interested in the rate at which a certain substance dissolves
in water. An electrical engineer may wish to know the rate of change of current in a part of an electrical
circuit. A biologist may be concerned with the rate at which bacteria in a culture decreases or increases.

Let us consider the following general situation which can be applied to all the preceding examples.

Suppose a variable w is a function of time such that at time t, w is given by w = g (t), where g is a
differentiable function.

We define the instantaneous rate of change of w with respect to t as

dw
= g 0 (t) .
dt

Sometimes we are given a problem with a rate of change that we cannot measure directly from the rate we
are given. For instance, we may have a function v and we need to find dv dv
dt , but we cannot find dt directly
if we have that v is a function of p where p is then a function of t.

In this case we have to use the Chain rule and obtain

dv dv dp
v 0 (t) = = .
dt dp dt

In applications it is not unusual to have two variables x and y that are differentiable functions of time
t, say x = f (t) and y = g (t). In addition, x and y may be related by means of some equation such as

x2 − y 3 + 2x + 7y 2 − 4 = 0, x = f (t) , y = g (t) .

Differentiating with respect to t and using the chain rule produces an equation involving the rates of
dy
change dx
dt and dt . The previous equation leads to

dx dy dx dy
2x − 3y 2 +2 + 14y = 0.
dt dt dt dt

dx dy
The derivatives dt and dt are called related rates, since they are related by means of the equation.
 3 MAT1613/GD001

Worked Examples

1. A spherical balloon is inflated with helium at a rate of 100π m3 s−1 . How fast is the balloon’s radius
increasing at the instant the radius is 5m? How fast is the surface area increasing?
Solution
Let V be the volume (in m3 ) of the balloon, r the radius (in m) of the balloon, t the time (in sec)
and S the surface area (in m2 ) of the balloon.
We are given that
dV
= 100π,
dt
and need to find
dr dS
and .
dt r=5 dt r=5

We have that
4
V = πr3 and S = 4πr2 .
3
Now
dV dV dr
= ·
dt dr dt
i.e.
dr
100π = 4πr2 ·
dt
dr 100π 25
so = 2
= 2. ...(∗)
dt 4πr r
Consequently
dr 25
= = 1.
dt r=5 52

Hence the balloon’s radius is increasing at the rate of 1ms−1 at the instant the radius is 5m. Also,

dS dS dr
= ·
dt dr dt
25
= 8πr · 2 (see (∗) above)
r
200π
= .
r

Thus
dS 200π
= = 40π,
dt r=5 5

and the surface area is increasing at the rate of 40π m2 s−1 when the radius is 5m.

2. The area of an equilateral triangle is increasing at a rate of 300 cm2 /min at the instant its edges are
√
each 3 cm long. At what rate are the edges changing at that instant?
Solution
 4

Let A denote the area (in cm2 ) of an equilateral triangle. Let ` denote the side length (in cm) of
such a triangle. Let t denote the time in minutes.

3l
l 2 l

_l _l _π
2 2 3

From the Pythagoras’s Theorem it follows that the height (in cm) of the triangle is
√
3
q
2 2
` − (`/2) = `.
2

Hence √ √
1 3 3 2
A= ·`· `= ` .
2 2 4
dA d`
It is given that √
= 300. We must find √
. Now
dt `= 3 dt `= 3
√ √
dA dA d` 3 d` 3 d`
= · = · 2` · = `· ,
dt d` dt 4 dt 2 dt

so
d` 2 dA
=√ · .
dt 3` dt
Hence
d` 2
√
= √ √ · 300 = 200.
dt `= 3 3· 3
√
Thus the edges are each increasing at a rate of 200 cm/min the instant the edges are each 3 long.

3. Water is running into a conical tank at a rate of 9m3 /min. How fast is the water level rising when
the water is 6 m deep? The dimensions of the tank (in metres) appear in the diagram.

10
x

y
 5 MAT1613/GD001

Solution
The volume V of a cone with radius x and height y is:
1
V = πx2 y.
3

5
A E

10 x
B D
y

To get a relationship between y and x we use a property of similar triangles, namely

BC BD
=
AC AE

i.e.
y x
=
10 5
i.e.
1
x = y.
2
Substitute this x into the equation for V :
1  y 2 1
V = π y = πy 3 .
3 2 12

Differentiate with respect to time:

dV π dy
= y2 .
dt 4 dt
dV
It is given that = 9 and y = 6.
dt
π dy
∴ 9 = (36) .
4 dt
dy 1
∴ = .
dt π

1
Therefore, the water level is rising at m/min when the water is 6m deep.
π
 6

4. According to Poiseuille’s law, the volume V of blood flowing through a vein in a unit of time at a
fixed pressure is related to the radius r of the vein by the formula

V = kr4 ,

where k is some fixed constant. Now cold water has the effect of contracting the blood vessels in the
hands, and the radius of a typical vein might decrease at the rate of 20% /minute. Use Poiseuille’s
law to calculate the percentage rate at which the volume of blood flowing through that vein decreases.
Solution
Let t denote the time in minutes. It is given that
dr/dt −20
= .
r 100
Now

V = kr4
dV dr
⇒ = 4kr3
dt dt
dV /dt 4kr3 · dr/dt
⇒ =
V V
4kr3 · dr/dt
=
kr4
dr/dt
= 4·
r
20 80
= −4 × =− .
100 100
Hence, the volume of blood flowing through the vein decreases at the rate of 80%/minute.

5. Ohm’s law for electrical circuits states that

V = IR

where V is the voltage, I is the current in amperes, and R is the resistance in ohms. Suppose that
1
V is increasing at a rate of 1 volt/s, while I is decreasing at a rate of amperes/s. Find the rate at
3
which R is
changing when V is 12 and I is 2. Is R increasing or decreasing?
Solution
Let t denote the time in seconds. It is given that
dV dI 1
= 1, =− and V = IR. ...(∗)
dt dt 3
Now
dV dR dI
= I +R .
dt dt
 dt 
dR 1 dV dI
∴ = −R
dt I dt dt
 
1 1V
= 1+ (from (∗)).
I 3I
 7 MAT1613/GD001

Hence, when V = 12 and I = 2, then

 
dR 1 1 12 3
= 1+ × = .
dt 2 3 2 2

3
Thus R is increasing by ohms/second at the instant in question.
2

1.2 GRAPHS AND FUNCTIONS

Before we do some problems on this topic, we first have to explain the method of using sign patterns.

Sign patterns
We use sign patterns to determine:

I. over which intervals a given function f is negative and over which intervals f is positive.

II. the intervals where the graph of a given function f rises and falls (we use the first derivative here).

III. the intervals where the graph of a given function f is concave up or concave down (we use the second
derivative here).

In all of the above cases, make sure that the given functions, first derivative and second derivative are fully
factorize that means that we have factorized the function into constants, powers of x, linear equations
and quadratic equations of the form x2 + k, k > 0.

I. Intervals where f is positive and intervals where f is negative:

Suppose
(x − 2) (x + 3) (x − 5)
f (x) = .
x (x + 4)

It should be clear that the sign of f (x) changes only at those points where the individual factors
change sign. The individual factors change sign at the points where they become zero. Hence, f (x)
changes sign at x = 2, x = −3, x = 5, x = 0 and x = −4. We may mark these points on a
number line as follows:
undefined undefined

−4 −3 0 2 5

See also that the function f (x) is undefined at x = −4 and x = 0. (division by zero).

We need to find the sign of f (x) over the intervals (−∞, −4), (−4, −3), (−3, 0), (0, 2) , (2, 5) and
(5, ∞).
 8

Say we choose the interval (5, ∞) . We have that y = x−5 is a straight line intercepting the x–axis at
5. Thus the sign changes at x = 5. When x > 5 (since y = x − 5), we have that y has positive values
and when x < 5 we have that y has negative values. For each linear factor in the above problem we
now look at the signs. To know where the function f (x) is positive or negative we multiply all the
signs in the different intervals.

The sign pattern will now look as follows:

undefined undefined
y =x−2 − − − − + +
y =x+3 − − + + + +
y =x−5 − − − − − +
y=x − − − + + +
y =x+4 − + + + + +
(x − 2) (x + 3) (x − 5)
f (x) = − + − + − +
x (x + 4)
−4 −3 0 2 5

A word of warning! Not all the factors that occur will change sign, for example:

(x2 + x + 1)(x − 2)
f (x) = .
(x − 3)(x + 1)2

f (x) changes sign only at x = 2 and at x = 3. There is no point at which x2 + x + 1 changes sign –
it is always positive. (This follows from the fact that its discriminant is negative.) Also, (x + 1)2 is
zero at x = −1, but it does not change sign there. (x + 1)2 is always non–negative.

Note also that at x = −1 and x = 3 the function is undefined.

The sign pattern is as follows

undefined undefined
y =x−2 − − + +
y =x−3 − − − +
2


x + x + 1 (x − 2)
f (x) = + + − +
(x − 3) (x + 1)2
−1 2 3
 9 MAT1613/GD001

II. Intervals where the graph of f rises and falls

The derivative test to find intervals where the function f is increasing and decreasing, is as follows:

Suppose that f is continuous on [a, b] and differentiable on (a, b) .

1. If f 0 (x) > 0 for all x in (a, b), then f is increasing on [a, b].

2. If f 0 (x) < 0 for all x in (a, b), then f is decreasing on [a, b].

x2 − 16
So let f (x) = .
x+5

Then we need the sign pattern for

2x (x + 5) − x2 − 16


0
f (x) =
(x + 5)2
x2 + 10x + 16
=
(x + 5)2
(x + 2) (x + 8)
=
(x + 5)2

All the rules in I (intervals where f is positive or negative) apply here again.

Thus x = −5 is the point where f (x) is undefined. At x = −2 and x = −8 the sign changes. Since
(x + 5)2 is always positive, we need not include this in the sign pattern for f 0 (x).

undefined
y =x+2 − − − +
y =x+8 − + + +
(x + 2) (x + 8)
f 0 (x) = + − − +
(x + 5)2
−8 −5 −2

From the first derivative test, we see that the graph f rises on (−∞, −8) and (−2, ∞) and falls on
(−8, −5) and (−5, −2).
 10

III. Intervals where the graph of f is concave up and concave down.

The test to determine concavity is as follows

Suppose that f 00 exists on an interval I.

If f 00 (x) > 0 on I, then the graph of f is concave up on I and if f 00 (x) < 0 on I,
then the graph of f is concave down on I.

So we need the sign pattern of f 00 (x).

Using the same problem as in II, that is

x2 − 16 (x + 2) (x + 8) x2 + 10x + 16
f (x) = , f 0 (x) = =
x+5 (x + 5)2 (x + 5)2

we obtain
18
f 00 (x) = (we use the quotient rule for differentiation)
(x + 5)3

We see again that f 00 (x) is not defined at x = −5. Further (x + 5)3 = (x + 5)2 (x + 5) the first
which is always positive. So we only need to look at y = x + 5.

The sign pattern for f 00 (x) is as follows:

undefined
x+5 − +
f 00 (x) − +
−5

Thus, following the second derivate test, we have that f is concave up on (−5, ∞) and concave down
on (−∞, −5).

Exercises
Determine the sign patterns of the following functions:

(3x + 2) (3 − x)
(i) f (x) = .
(4 + x) (x + 1)
−x2 + 3x − 4 (4 + x)


(ii) f (x) = .
(3 − x)3 x
(x + 3)2 x2
(iii) g 0 (x) = .
(x + 4)4
(x − 3) (x + 1)
(iv) h00 (x) = √ .
4 − x2
 11 MAT1613/GD001

Solutions
(i)
undefined undefined 0 0
y = 3x + 2 − − − + +
y = −x + 3 + + + + −
y =4+x − + + + +
y =x+1 − − + + +
(3x + 2) (3 − x)
f (x) = − + − + −
(4 + x) (x + 1)
−4 −1 − 32 3

(ii)
undefined undefined
y= −x2− 3x − 4
2
− − − −
(always negative b − 4ac < 0)
y =4+x − + + +
y =3−x + + + −
y=x − − + +
2


−x + 3x − 4 (4 + x)
f (x) = − + − +
(3 − x)3 x
−4 0 3

(iii) g 0 (x) is always positive:

undefined
g 0 (x) + + +
−4 −3

(iv) h00 (x) is only defined on (−2, 2). h00 (x) is undefined for x ≥ 2 and x ≤ −2

y =x−3 − −
y =x+1 − +
h00 (x) + −
−2 −1 2
 12

Symmetry
We define the concept because it will also help us to draw the graph of a function.

Symmetry
The graph of an equation is symmetric with respect to
the y–axis if the point (−x, y) is on the graph whenever
(x, y) is on the graph.

The graph of an equation is symmetric with respect

to the x–axis if the point (x, −y) is on the graph whenever
(x, y) is on the graph.

The graph of an equation is symmetric with respect to

the origin, whenever a point (x, y) is on the graph, (−x, −y)
is also on the graph.

Symmetry with respect to any other straight line can also be determined but we are not going to
discuss this here.

Remarks on the worked problems

In all of the following worked problems you have to determine the following:

(a) Determine the x– and y–intercepts of the graph of f .

(b) Determine the vertical and horizontal asymptotes of the graph of f .
(c) Use the first derivate of f (use the sign pattern of f (x)) to
(i) determine the intervals over which the graph of f rises, and the intervals over which it falls.
(ii) find the local extrema.
(d) Use the second derivative of f (use the sign pattern of f 00 (x)) to determine the intervals over
which the graph of f is concave up, and the intervals over which it is concave down.
(e) (i) Use the sign pattern of f 00 (x) and determine the inflection points.
(ii) Obtain the gradient of the inflection points, respectively.
(f) Use the above information to draw the graph of f .

Worked problems

1. Let f be the function defined by

4x
f (x) = .
x2 + 4
 13 MAT1613/GD001

Solution

1. (a) f (0) = 0. This is the x- (and y-) intercept.

(b)
4
4x x
lim = lim = 0.
x→∞ x2 + 4 x→∞ 1 + 42
x

Similarly,
4x
lim = 0.
x→−∞ x2 + 4
The horizontal asymptote is the line y = 0.
(c)

4x x
f (x) = =4· 2
+4 x2 x
+ 4
2
1 · x + 4 − x · 2x
⇒ f 0 (x) = 4 ·
(x2 + 4)2
4 4 − x2


=
(x2 + 4)2
4 (2 − x) (2 + x)
= .
(x2 + 4)2

2
We have that f 0 (x) is not undefined at any point. We have that x2 + 4 is always positive.
The sign pattern of f 0 (x) is:

y =2−x + + −
y =2+x − + +
f 0 (x) − + −
−2 2

Hence:

(i) the graph rises on (−2, 2) and falls on (−∞, −2) and (2, ∞);
(ii) there is a local minimum at (−2, f (−2)) = (−2, −1), and a local maximum at (2, f (2)) = (2, 1).

4 − x2
f 0 (x) = 4 ·
(x2 + 4)2

2
−2x · x2 + 4 − 4 − x2 · 2 x2 + 4 · 2x



00
⇒ f (x) = 4 ·
(x2 + 4)4
−2x x2 + 4 x2 + 4 + 2 4 − x2




= 4·
(x2 + 4)4
2


−8x 12 − x
=
(x2 + 4)3
 √
√

−8x 12 − x 12 + x
= .
(x2 + 4)3
 14

(d) The sign pattern of f 00 (x) is

√
y = 12 − x + + + −
√
y = 12 + x − + + +
y = −8x + + − −
f 00 (x)
− + − +
√ √
− 12 0 12
√
√

We have that the graph is concave up on − 12, 0 and 12, ∞ , and concave down on
√
√

−∞, − 12 and 0, 12 .
(e) (i) The inflection points are where the second derivative becomes zero.Thus there are inflection
√ √

√ √
√ √


points at − 12, f − 12 = − 12, − 3/2 ; (0, f (0)) = (0, 0) and 12, f 12 =
√ √

12, 3/2 .
(ii) The gradients at the inflection points are, respectively:
 √  1
0
f − 12 = − ,
8
0
f (0) = 1

and √  1
f0 12 = − .
8
(f)
y

1
_3
2 inflection point

_ _
12 2
x
0 2 12

_ _3
inflection point 2
_
1

Please see from the graph above that whenever (x, f (x)) is on the graph, (−x, f (−x)) is also
on the graph. We thus have symmetry with respect to the origin.
 15 MAT1613/GD001

2. Let f be the function defined by

x
f (x) = .
(x + 1)3

Solution

2. (a) The x-intercept is 0 (this is also the y-intercept).

(b) The vertical asymptote is the line x = −1, and the horizontal asymptote is the line y = 0.
(c)
x
f (x) =
(x + 1)3
1 · (x + 1)3 − x · 3 (x + 1)2
⇒ f 0 (x) =
(x + 1)6
(x + 1)2 [x + 1 − 3x]
=
(x + 1)6
1 − 2x
= .
(x + 1)4

We see that f 0 (x) is undefined at x = −1 and that (x + 1)4 is always positive. The sign pattern
for f 0 (x) is:
undefined
y = 1 − 2x + + −
f 0 (x) + + −
1
−1 2

(i) Hence, the graph of f rises on (−∞, −1) and on −1, 12 , and falls on 1



2, ∞ .
(ii) There is a local maximum at
    
1 1 1 4
,f = , .
2 2 2 27

(d)
1 − 2x
f 0 (x) =
(x + 1)4
−2 (x + 1)4 − (1 − 2x) · 4 (x + 1)3
⇒ f 00 (x) =
(x + 1)8
−2 (x + 1)3 [x + 1 + 2 (1 − 2x)]
=
(x + 1)8
−2 (3 − 3x) 6 (x − 1)
= 5 = .
(x + 1) (x + 1)5
 16

We see that the function is undefined at x = −1 and (x + 1)4 (x + 1) = (x + 1)5 is not always
positive
undefined
y =x−1 − − +
y =x+1 − + +
f 00 (x) + − +
−1 1
Hence, the graph of f is concave up on (−∞, −1) and on (1, ∞) , and concave down on (−1, 1).
(e) (i) The inflection point is where the second derivative is zero, i.e.
 
1
(1, f (1)) = 1, .
8
(ii) The gradient of the inflection point is
1
f 0 (1) = − .
16
(f)
y

local maximum at (1/2, 4/27)

inflection point at (1, 1/8)
_
1
x
0 1/2 1

3. Let f be the function defined by

x
f (x) = .
(x + 1)2
 17 MAT1613/GD001

Solution

3. (a) f (x) = 0 ⇔ x = 0. The x-(and y-) intercept is thus 0.

(b) The vertical asymptote is the line x = −1.

x 1/x
lim f (x) = lim = lim = 0.
x→∞ x→∞ x2 + 2x + 1 x→∞ 1 + 2/x + 1/x2

Similarly, lim f (x) = 0.

x→−∞
Hence, the horizontal asymptote is the line y = 0.
(c)

x
f (x) =
(x + 1)2
1 · (x + 1)2 − x · 2 (x + 1)
⇒ f 0 (x) =
(x + 1)4
(x + 1) [x + 1 − 2x]
=
(x + 1)4
1−x
=
(x + 1)3

We see that f 0 (x) is not defined for x = −1 and (x + 1)3 = (x + 1)2 (x + 1) is not always
positive. The sign pattern for f 0 (x) is:

undefined
y =1−x + + −
y =1+x − + +
f 0 (x) − + −
−1 1

Hence:

(i) the graph of f rises on (−1, 1) and

the graph of f falls on (−∞, −1) and (1, ∞).
(ii) there is a local maximum at (1, f (1)) = 1, 41 .

(d)

1−x
f 0 (x) =
(x + 1)3
−1 (x + 1)3 − (1 − x) 3 (x + 1)2
⇒ f 00 (x) =
(x + 1)6
− (x + 1)2 [(x + 1) + 3 (1 − x)]
=
(x + 1)6
2 (x − 2)
= .
(x + 1)4
 18

The function y = (x + 1)4 is always positive so we only need to investigate the linear function.
The sign diagram for f 00 (x) is:

undefined
y =x−2 − − +
f 00 (x) − − +
−1 2

Hence, the graph of f is concave up on (2, ∞) and the graph of f is concave down on (−∞, −1)
and (−1, 2).
(e) (i) There is an inflection point at (2, f (2)) = 2, 92 .

(ii) The gradient of the inflection point is

1
f 0 (2) = − .
27
(f)
y

_1
4 local maximum
_2 point of inflection
9

_ x
1 0 1 2

4. Let f be the function defined by

x2
f (x) = .
x2 + 3
 19 MAT1613/GD001

Solution

4. (a) f (x) = 0 ⇔ x = 0. The x- (and y-) intercept is 0.

x2 1
(b) lim = lim = 1.
2
x→±∞ x + 3 x→±∞ 1 + 32
x
Hence, the horizontal asymptote is y = 1.
(c)

x2
f (x) =
x2 + 3
2x x2 + 3 − x2 · 2x


⇒ f 0 (x) =
(x2 + 3)2
2x3 + 6x − 2x3 6x
= 2 = .
(x2 + 3) (x2 + 3)2

2
We see that x2 + 3 is always positive. The sign pattern of f 0 (x) is:

y=x − +
f 0 (x) − +
0

Hence:
(i) the graph rises on (0, ∞) and the graph falls on (−∞, 0).
(ii) there is a local minimum at (0, 0).
(d)
6x
f 0 (x) =
(x2+ 3)2

2
6 x2 + 3 − 6x · 2 x2 + 3 · 2x


⇒ f 00 (x) =
(x2 + 3)4
6 x2 + 3 x2 + 3 − 4x2

 
=
(x2 + 3)4
6 3 − 3x2


18 (1 − x) (1 + x)
= 3 = .
2
(x + 3) (x2 + 3)3

3
We see that x2 + 3 is always positive. The sign pattern of f 00 (x) is:

y =1−x + + −
y =1+x − + +
f 00 (x) − + −
−1 1

Hence, the graph is concave up on (−1, 1) and the graph is concave down on (−∞, −1) and
(1, ∞).
 20

(e) (i) There are inflection points at (−1, f (−1)) = −1, 14 and (1, f (1)) = 1, 14 .

(ii) The gradients of the inflection points are f 0 (1) = 3

8 and f 0 (−1) = − 38 .
(f)
y

_1
4

_ x
1 0 1

Exercises

1. Suppose that the derivative of a certain function y = f (x) is

y 0 = (x − 1)2 (x − 2) .

Find the intervals over which the graph of f is concave up, and the intervals over which it is concave
down. Also, find the values of x at which the graph of f has inflection points.

2. (a) Determine the y-intercept of the curve of

f (x) = 2x3 + 3x2 − 36x + 10.

(b) Use the first derivative of f (x) to find the intervals over which the curve of f (x) is increasing
(rising) and the intervals over which it is decreasing (falling). Also find the local extrema.
(c) Use the second derivative of f (x) to determine the intervals over which the curve of f (x) is
concave up and the intervals over which it is concave down. Determine the points of inflection.
(d) Determine the slope of the tangent to the curve of f (x) at the point of inflection.

Answers

1.

f 0 (x) = (x − 1)2 (x − 2)
⇒ f 00 (x) = 2(x − 1)(x − 2) + (x − 1)2 · 1
= (x − 1)[2(x − 2) + (x − 1)]
= (x − 1)(3x − 5).
 21 MAT1613/GD001

The sign pattern for f 00 (x) is:

y =x−1 − + +
y = 3x − 5 − − +
f 00 (x) + − +
5
1 3

Hence, the graph of f is concave up on (−∞, 1) and ( 53 , ∞), and concave down on (1, 53 ). There are
inflection points where x = 1 and x = 53 .

2. (a) One can find the y-intercept by putting x equal to 0. Therefore the y-intercept is

f (0) = 10.

(b) If f (x) = 2x3 + 3x2 − 36x + 10, then

f 0 (x) = 6x2 + 6x − 36.

There are possible extremes where f 0 (x) = 0,

i.e.
6x2 + 6x − 36 = 0

i.e.
x2 + x − 6 = 0

i.e.
(x + 3) (x − 2) = 0

i.e.
x = −3 or x = 2.

The sign pattern for f 0 (x) is:

y =x+3 − + +
y =x−2 − − +
f 0 (x) + − +
−3 2

Therefore the curve is increasing over (−∞, −3) and (2, ∞) and decreasing over (−3, 2).
There is a maximum at (−3, f (−3)) = (−3, 91) and a minimum at (2, f (2)) = (2, −34).
(c) If
f 0 (x) = 6x2 + 6x − 36

then
f 00 (x) = 12x + 6 = 6 (2x + 1) .

Possible points of inflection appear where

f 00 (x) = 0
 22

i.e.
12x + 6 = 0

i.e.
1
x=− .
2
The sign pattern for f 00 (x) is:

y = 2x + 1 − +
f 00 (x) − +
− 21

Therefore the curve is concave down over −∞, − 12 and concave up over − 12 , ∞ . There is a

point of inflection at − 12 , f − 21 = − 12 , 57




2 .

(d) The point of inflection is at x = − 12 . The gradient of the inflection point is then
   2  
1
0 1 1
f − = 6 − +6 − − 36
2 2 2
6
= − 3 − 36
4
75
= − .
2
1.3 OPTIMIZATION

The theory (used in 1.2(S)) for finding the extrema of functions can be applied to certain practical problems.
We see here how the first and second derivative test are used to find the extreme values and to establish
whether they are local maximum or a local minimum values.

We state the second derivative test:

Second Derivative Test

Suppose a function f is differentiable on an open interval containing
a point c and f 0 (c) = 0. If f 00 (c) < 0 then, f has a local maximum at c
and if f 00 (c) > 0 then, f has a local minimum at c.

Say we have the following problem to solve: We have a rectangular piece of cardboard 16 metres wide
and 21 metres long. We want to make a box by cutting out a square from each corner and bending the
sides. Find the size of the corner which will produce a box having the maximum possible volume.

Guidance: we first have to find the equation for the volume V as a function of x. Thus if we let x be the
height of the box, we have

V = x (16 − 2x) (21 − 2x)

= 2 168x − 37x2 + 2x3 .


 23 MAT1613/GD001

_
x 16 2x

_
21 2 x
To find the critical points we first find the derivative
dV
= 4 3x2 − 37x + 84


dx
= 4 (3x − 28) (x − 3) .
dV 28
Possible critical points are where dx = 0, that is at x = 3 and 3.

28
3 is impossible because it is outside the domain of x.

The second derivative test will tell us where the maximum value is
d2 V
= 4 (6x − 37) .
dx2
Then
d2 V

= 4 (18 − 37) = −76 < 0
dx2 x=3
so that V has a relative maximum at x = 3.

We also have to check for the endpoints. We see that since 0 ≤ x ≤ 8 and since V = 0 if either x = 0 or
x = 8, the maximum value of V does not occur at the endpoints of the domain of x.

Thus, a 3-metre square should be cut from each corner to maximise the volume of the resulting box.

Worked Examples

1. Find the largest possible value of s = 2x + y if x and y are side-lengths in a right triangle whose
√
hypotenuse is 5 units long.
Solution
The Pythagorean Theorem tells us that

x2 + y 2 = 5

from which follows that p

y= 5 − x2 .

Hence, p
s = 2x + y = 2x + 5 − x2 .
√ √
Observe that 0 < x < 5. x cannot be 0 or 5, for then one of the sides in the triangle would have
√ √
zero length. If x > 5 then 5 − x2 < 0, so 5 − x2 is not a real number.
 24

Using the first derivative to find the possible critical points, we have
p
s = 2x + 5 − x2
ds −2x
⇒ = 2+ √
dx 2 5 − x2
√
2 5 − x2 − x
= √ .
5 − x2

ds ds √ √ √
√ and √ are undefined, and so 5 and − 5 could be critical values. However 5
dx x= 5 dx x=− 5
√
and − 5 lie outside the domain of s. Furthermore
ds
= 0
p dx
2
⇔2 5−x −x = 0
p
⇔ 2 5 − x2 = x
⇒ 4 5 − x2 = x2

⇒ 5 (2 − x) (2 + x) = 0.

The number 2 lies in the domain of s, so 2 is a critical point.

Using the second derivative we can investigate if this point is a local maximum or local minimum:
ds
− 1
= 2 − x 5 − x2 2
dx
d2 s
 

− 1 1
− 3
= − 5 − x2 − x · − 5 − x2 2 (−2x)
2
dx2 2
1 x 2
= −√ − 3 ,
5 − x2 (5 − x2 ) 2

and
d2 s 4
= −1 − = −5 < 0.
dx2 x=2 1

Consequently s has a maximum when x = 2. If x = 2, then y = 1 and the maximum value of s is 5.

2. Suppose that the perimeter of a rectangle is 8 cm. What is the smallest possible length of the
diagonal of such a rectangle?
Solution
Let r and s denote the side lengths (in cm) of a rectangle.
s

r l

It is given that
2r + 2s = 8,
 25 MAT1613/GD001

i.e.
r = 4 − s.

If ` denotes the length (in cm) of the diagonal, then from the Pythagoras’s Theorem it follows that
√ q √
` = r2 + s2 = (4 − s)2 + s2 = 2s2 − 8s + 16. ...(∗)

Since s is a side length, and since the perimeter is fixed at 8 cm, it follows that 0 < s < 4. Now

d` 4s − 8 2 (s − 2)
= √ =√ .
ds 2
2 2s − 8s + 16 2s2 − 8s + 16

For f (s) = 2s2 − 8s + 16, the discriminant, b2 − 4ac = 64 − 4 (2) (16) < 0, which implies that f (s)
is always positive. Thus ` > 0.
d`
The sign pattern for is then only dependent on y = s − 2.
ds

y =s−2 − − + +
0 2 4

√
(Remember that denotes only the positive root.) Hence ` is decreasing on (0, 2) and increasing
on (2, 4), so there is a local (and, in fact, absolute) minimum when s = 2. From (∗), we get that the
√
smallest possible length of the diagonal is 2 2cm.

3. Find two positive real numbers with sum 36 and where the product of the square of one number
multiplied with the other is a maximum.
Solution
Let x be the one number and y the other.
We have x + y = 36, so that y = 36 − x.
Let T = x2 y.
Thus, T (x) = x2 (36 − x) = 36x2 − x3 .
dT
Now = 72x − 3x2 .
dx
Then
dT
=0 when 72x − 3x2 = 0;
dx
that is where 3x (24 − x) = 0.
The values are then x = 0 or x = 24.
Using the second derivative test, we have

d2 T
= 72 − 6x.
dx2

So
d2 T

= 72 > 0
dx2 x=0
 26

and
d2 T

= −72 < 0 (maximum when x = 24).
dx2 x=24

Thus, we have that the numbers are 24 and 12.

4. A rectangle has its base on the x-axis and its upper two vertices on the parabola y = 12 − x2 . What
is the largest area the rectangle can have?
Solution
y

_ 2
_ 2 y = 12 x
12 l

_ _ x
2 3 l 0 l 2 3

Since opposite sides of the rectangle are parallel, it is clear that the rectangle must be symmetric
about the y-axis. Let the distance from the lower right–hand corner of the rectangle to the origin
be `. If A denotes the area of the rectangle then,
√
A = 2`(12 − `2 ) = 24` − 2`3 , where 0 < ` < 2 3.

Now
dA
= 24 − 6`2 = 6(4 − `2 ) = 6(2 − `)(2 + `).
d`
dA √
Hence, = 0 if ` = 2. (We cannot have ` = −2 because 0 < ` < 2 3.) A possible local extremum
d`
thus occurs when ` = 2. Now
d2 A
= −12` = −24, if ` = 2.
d`2
This shows that A has a local maximum, and in fact, a maximum value when ` = 2. If ` = 2 then

A = 24 × 2 − 2 × 8 = 32,

which is the largest area that the rectangle can have.

Exercises
√
1. A right angle, whose hypotenuse is 3 long, is revolved about one of its legs to generate a right
circular cone. Find the radius, height and volume of the cone of greatest volume that can be made
this way.
 27 MAT1613/GD001

2. What is the smallest perimeter possible for a rectangle whose area is 16 square metres?

3. An isosceles triangle has its vertex (the point where the two equal sides meet) at the origin, and its
base parallel to the x-axis with its vertices above the x-axis on the curve y = 27 − x2 . Find the
largest area the triangle can have.

Answers

1. The volume is given by the formula

1
V = π 3 − h2 h.


3

(This comes from V = 13 πr2 h and from Pythagoras’s Theorem that r2 = 3 − h2 .)

dV
We see that if dh = 0, h = 1.
2π 3
√
The largest volume is 3 m and the corresponding radius 2m.

2. If we let x and y be the side lengths, then xy = 16 and the perimeter P = 2x + 2y.
Thus we obtain the function P (x) = 2x + 32x−1 .
Differentiation gives dP
dx = 2 − 32x−2 and if dP
dx = 0, then x = 4.
The smallest possible perimeter occurs when x = y = 4, then P = 16 m.

A (x) = x 27 − x2 .

At x = 3 or x = −3, there are local extrema.

The largest area is 54 square units.

1.4 ROLLE’S THEOREM AND THE MEAN VALUE THEOREM

There is strong geometric evidence that between any two points, where a smooth curve crosses the x-axis,
there is a point on the curve where the tangent is horizontal.

The theorem stating this is called Rolle’s Theorem.

Rolle’s Theorem
Suppose y = f (x) is continuous at every point of the closed interval
[a, b] and differentiable at every point of its interior (a, b). If f (a) = f (b) = 0,
there is at least one number c between a and b, with f 0 (c) = 0.

The Mean Value Theorem is a generalised version of Rolle’s Theorem and it is stated here.
 28

The Mean Value Theorem

If y = f (x) is continuous at every point of the closed interval [a, b] and
differentiable at every point of its interior (a, b) , then there is at least one
number c between a and b at which
f (b) − f (a)
= f 0 (c) .
b−a

Worked Examples

1. Suppose f 0 (x) = 2 and f (0) = 5. Use the Mean Value Theorem to show that f (x) = 2x + 5 at every
value of x.
Solution
We consider the three cases x > 0, x < 0 and x = 0.
Suppose x > 0. The function f is assumed to be differentiable everywhere, so in particular it is
differentiable on (0, x) and continuous on [0, x]. According to the Mean Value Theorem there exists
c ∈ (0, x) such that
f (x) − f (0)
= f 0 (c)
x−0
i.e.
f (x) − 5
=2
x
or
f (x) = 2x + 5.

Similarly, if x < 0, we can apply the Mean Value Theorem to the function f and the interval [x, 0]
to obtain
f (x) = 2x + 5.

Finally, if x = 0, it clearly holds that

f (x) = 2x + 5.

2. Prove that the equation

x4 − 4x + c = 0

(where c is an arbitrary fixed constant) does not have two different roots that are less than 1.
Solution
Let
f (x) = x4 − 4x + c.

Suppose that the given equation does have two different roots that are less than 1, i.e. there exist
numbers α and β such that α < β < 1, f (α) = 0 and f (β) = 0. The function f , being a polynomial
 29 MAT1613/GD001

function, is differentiable and continuous everywhere. In particular, f is continuous on [α, β] and

differentiable on (α, β). According to the Mean Value Theorem there exists γ ∈ (α, β) such that

f (β) − f (α)
= f 0 (γ)
β−α

i.e.
0 = f 0 (γ).

However,
f 0 (x) = 4x3 − 4,

so

f 0 (x) = 0
⇔ 4x3 − 4 = 4 (x − 1) x2 + x + 1 = 0

⇔ x = 1.

Hence γ = 1. This is a contradiction since γ < β < 1. Thus the given equation does not have two
different roots that are less than 1.

3. Given the function

f (x) = xn + px + q

where n is a fixed positive even integer, and p and q are arbitrary constants, show that f cannot
have more than two zeros.
Solution
Suppose that f has at least three different zeros, say α, β and γ. Also suppose that α < β < γ. Now
f is continuous on [α, β] and differentiable on (α, β). By assumption f (α) = f (β) = 0. By Rolle’s
Theorem there exists δ 1 ∈ (α, β) such that f 0 (δ 1 ) = 0. By applying Rolle’s Theorem to the function
f and the interval [β, γ] we also obtain a δ 2 ∈ (β, γ) such that f 0 (δ 2 ) = 0. We have thus found two
different roots (viz δ 1 and δ 2 ) of the equation
nxn−1 + p = 0. ...(∗)

But from (∗) we obtain

 p 1
n−1
x= − ,
n
i.e. the equation (∗) has only one root because n − 1 is odd. This contradiction shows that f cannot
have more than 2 zeros.

4. Use the Mean Value Theorem to show that

| sin b − sin a| ≤ |b − a|

for any real numbers a and b.

Solution
 30

If a = b it holds that
| sin b − sin a| = 0 ≤ |b − a|.

Suppose a < b. The function f (x) = sin x is continuous over [a, b] and differentiable over (a, b).
According to the Mean Value Theorem there exists a c ∈ (a, b) such that
f (b) − f (a)
= f 0 (c).
b−a
Therefore
sin (b) − sin (a)
= cos(c)
b−a
i.e.
sin (b) − sin (a)
= | cos(c)| ≤ 1.
b−a
∴ | sin(b) − sin(a)| ≤ |b − a|

for all real a and b.

5. Refer to the Mean Value Theorem above. It states that f must be continous at every point of the
closed interval [a, b] and also diffferentiable at every point of (a, b) for a c to exists.The converse of
the theorem does not hold, i.e. if the conditions do not hold there still may be a c which satisfies
the equation. Given a = −8, b = 1 and f (x) = x1/3 , show that there exists a value of c so that the
equation is satisfied. Which condition of the theorem does not hold?
Solution
If
f (x) = x1/3

then
1
f 0 (x) = x−2/3 .
3
Since 0 ∈ (−8, 1) and f 0 (0) does not exist, f (x) is not differentiable over (a, b) = (−8, 1). Therefore
the differentiability condition of the Mean Value Theorem does not hold.
Now,
f (b) − f (a)
f 0 (c) =
b−a
1 1
1 (1) 3 − (−8) 3
⇒ 2 =
3c 3 1 − (−8)
3
=
9
1
=
3
2
⇒ c3 = 1
⇒ c = ±1.

Therefore there exists a c ∈ (−8, 1), namely c = −1, such that

f (b) − f (a)
f 0 (c) = .
b−a
 31 MAT1613/GD001

Exercises

1. Use the Mean Value Theorem to prove that

tan β − β ≥ tan α − α

whenever α and β are any numbers with β ≥ α which lie in the interval − π2 , π2 .

2. Use Rolle’s Theorem to prove that the equation

2x − cos x = 0

has exactly one solution in the interval [−π, π].

3. Suppose that f is a differentiable odd function with domain the set of all real numbers. Prove that
for each positive real number r there exists a number c such that

−r < c < r

and
f (r) = r.f 0 (c).

Answers

1. Let α, β ∈ − π2 , π2 , and suppose that β ≥ α. If α = β then tan β − β = tan α − α. Suppose

then that β > α. Now the function f (x) = tan x is continuous on [α, β] and differentiable on (α, β).
According to the Mean Value Theorem there exists a γ with α < γ < β such that

f (β) − f (α)
= f 0 (γ),
β−α

i.e.
tan β − tan α
= sec2 (γ) . ...(∗)
β−α
Now for all values of x for which the function sec is defined, we have | sec x| ≥ 1, so for such x, we
have sec2 x ≥ 1. From (∗) it then follows that

tan β − tan α
≥ 1.
β−α

Since β − α > 0, we can multiply through by β − α to obtain that

tan β − tan α ≥ β − α,

and hence that

tan β − β ≥ tan α − α.
 32

2. Let
f (x) = 2x − cos x.

Then

f (−π) = −2π − (−1)

= 1 − 2π < 0,

while
f (π) = 2π − (−1) = 2π + 1 > 0.

Because f is continuous, it follows from the Intermediate Value Theorem that f (c) = 0 for at least
one number c in the interval (−π, π), i.e. the equation

2x − cos x = 0

has at least one solution in the interval (−π, π).

If there are more than one solution in the interval (−π, π), then there are numbers a and b in (−π, π)
for which a < b and f (a) = f (b) = 0. Now f is continuous on [a, b] and differentiable on (a, b), so it
follows from Rolle’s Theorem that there is a number d with a < d < b and f 0 (d) = 0, i.e.

2 + sin d = 0,

or
sin d = −2.

This is impossible. Consequently, the given equation has exactly one solution in the interval (−π, π).

3. Let r be a positive real number. By assumption f is differentiable on the interval (−r, r), and
since f is continuous everywhere (being differentiable), f is continuous on [−r, r]. According to the
Mean-Value Theorem there exists a number c ∈ (−r, r) such that

f (r) − f (−r)
f 0 (c) =
r − (−r)
f (r) + f (r)
= (−f (−r) = f (r) , since f is odd)
2r
f (r)
= .
r

Hence
f (r) = r · f 0 (c).

Summary
We have focussed in this chapter on a large range of applications. The learner should now be able to
solve related rates problems, using the Mean Value Theorem and Rolle’s Theorem and also optimization
problems. The learner should now also be able to sketch graphs of functions using derivatives.
 33 MAT1613/GD001

Chapter 2

Transcendental Functions

Introduction
Many of the functions in mathematics and science are inverses of one another. The functions ln x and ex
are probably the most famous function–inverse pair. The trigonometric functions when suitably restricted,
have useful inverses. Less widely known are the hyperbolic functions and their inverses. We also show
using L’Hôpital’s rule, how a limit of a fraction can be calculated when both the numerator and the
denominator approach zero (we call this an indeterminate form).

This chapter consists of the following sections:

2.1 The logarithmic and exponentical functions.

This has already been done in MAT1512. Because of the importance of these functions and their
properties we do revision on these parts.

2.2 Determining indeterminate forms with L’Hôpitals’ rule.

2.3 The inverse trigonometric functions.

2.4 Hyperbolic functions.

Outcomes
After studying this chapter the learner should be able to:

• use L’Hôpital’s rule to evaluate limits of indeterminate forms.

• use the rules of differentiation and integration to differentiate or integrate f, where f is an inverse
trigonometric function or a hyperbolic function.

• obtain integrals and derivatives whenever these functions also appear in products, sums, quotients,
or composites.
 34

2.1 THE LOGARITHMIC AND EXPONENTIAL FUNCTIONS

The inverse of the exponential function ex is called the natural logarithm function y = `n x. In this
section, we study the natural logarithm both as a differentiable function and as a device for simplifying
calculations.

2.1.1 THE LOGARITHMIC FUNCTION AND ITS PROPERTIES

The natural logarithmic function y = `n x is the inverse

function of the exponential function.

Properties of the `n function.

(a) The graph of `n x looks like this:

x
0 1

(b) `n is defined for all x > 0, in other words

Dom (`n) = { x| x > 0} .

(c) Im (`n) = R.

(d) `n x < 0 if x < 1,

`n 1 = 0,
`n x > 0 if x > 1.
d 1
(e) (`n x) = , for x > 0.
dx x
(f) `n(xy) = `n x + `n y, x, y > 0.
x
(g) `n = `n x − `n y, x, y > 0.
y
(h) `n xr = r `n x, x > 0.
logc a
(i) logb a = .
logc b
 35 MAT1613/GD001

(j) If x → ∞, then `n x → ∞.

(k) If x → 0+ , then `n x → −∞.

The following important mistake is often made when applying rule (h):
`n (xr ) = r `n x (correct)
but
(`n x)r 6= r `n x (incorrect).
Note that the power r must be a power of x but not of `n x.

2.1.2 THE EXPONENTIAL FUNCTION AND ITS PROPERTIES

The inverse function of the function `n x is known as the

exponential function and is indicated by exp x (or ex ).

Properties of the exp function

(a) The graph of ex is the mirror image of the graph of `n x in the line y = x. Therefore, the graph
looks like this:
y
x
y =e y =x

1 y = ln x

x
0 1

(b) exp is defined for all x ∈ R, in other words

Dom (exp) = R.

(c) Im (exp) = {y|y > 0}.

 36

d x
(d) (e ) = ex .
dx
(e) ex+y = ex · ey .

(f) ex−y = ex /ey .

(g) erx = (ex )r .

(h) If x → ∞, then ex → ∞.

(i) If x → −∞, then ex → 0.

2.1.3 FUNCTIONS OF THE FORM f (x) = g(x)h(x)

We apply logarithmic differentiation to functions when they have an indeterminate form:

f (x) = g(x)h(x) .

We apply the logarithmic function to both sides and then make use of the properties of the logarithmic
function.
Then
`n f (x) = `n g(x)h(x) .

Therefore
`n f (x) = h(x).`n g(x).

Now we have that

f (x) = eln f (x)

= eh(x) ln f (x)

This is very important when we use L’Hôpital’s rule.

Worked Example
Determine

√ √
x
f 0 (x) if f (x) = x , x > 0.

Solution
By applying the logarithmic function on both sides, we obtain
√ √x
`n f (x) = `n x
√ √
= x · `n x.

By differentiating, we obtain from the chain rule and the product rule that
1 √ 1 1 √ 1
f 0 (x) = x · √ · x−1/2 + `n x · x−1/2
f (x) x 2 2
1 1 √
= √ + √ `n x.
2 x 2 x
 37 MAT1613/GD001

Therefore
√
 
0 1 1
f (x) = f (x) √ + √ `n x
2 x 2 x
√ √x √

 
1
= x √ 1 + `n x
2 x
√ √x
x √

= √ 1 + `n x .
2 x

2.2 DETERMINING INDETERMINATE FORMS WITH L’HÔPITAL’S RULE

This section is very important and as you will use the methods discussed here regularly in future. Since
it is so important, we discuss it in detail.

Limits having one of the following forms

0 ±∞
, , ± ∞.0, ∞ − ∞, 00 , 1∞ , ∞0
0 ±∞

are called limits of indeterminate form. A very useful rule, which we can apply for determining limits
of indeterminate form, is L’Hôpital’s rule.
(a) L’H ôpitals rule for the form 00

Let lim indicate one of the limits lim , lim , lim , lim , lim ,
x→a x→a− x→a+ x→∞ x→−∞

and suppose lim f (x) = 0 and lim g(x) = 0. If lim [f 0 (x)/g 0 (x)] has a finite value L or if this limit is +∞
or −∞, then

f (x) f 0 (x)
lim = lim 0 .
g (x) g (x)

±∞
(b) L’Hôpitals rule for the form ±∞
Let lim indicate one of the limits in (a) and suppose lim f (x) = ±∞ and lim g(x) = ±∞. If lim [f 0 (x)/g 0 (x)]
has a finite value L or if this limit is +∞ or −∞ , then

f (x) f 0 (x)
lim = lim 0 .
g (x) g (x)

0 ±∞
We always try to modify the other forms into one of 0 or ±∞ .

It is very important to see that we have to find the derivative of the function f (x) separately from the
derivative of the function g (x). You also have to simplify first before applying the next step, for example
applying L’Hôpital’s rule again.
 38

In the following example, we first follow the correct method and then we do the same problem incorrectly.

Worked Example

−x2 + 1
lim
x→1 x − 1

The correct method (finding the derivative of the numerator and denominator separately) is:

−x2 + 1 −2x
lim = lim = −2.
x→1 x − 1 x→1 1

(We first determine the derivative of −x2 + 1 which is −2x and then separately the derivative of x − 1
which is 1.)

The incorrect method (using the quotient rule and not obtaining the derivatives separately) is:

(x − 1) (−2x) − −x2 + 1 (1)

−x2 + 1
lim = lim (∗)
x→1 x − 1 x→1 (x − 1)2
−2x2 + 2x + x2 − 1
= lim
x→1 (x − 1)2
−x2 + 2x − 1
lim
x→1 (x − 1)2
x2 − 2x + 1


= lim −
x→1 (x − 1)2
(x − 1)2
= lim −
x→1 (x − 1)2

= −1 which is incorrect.

Some examples (indicating how one handles the various limits) follow below:

2.2.1 LIMITS OF THE FORM 00 AND ±∞ ±∞

In these cases we can simply apply L’Hôpital’s rule directly.

Worked Examples

1. Determine

`n x
lim .
x→∞ x

Solution
∞
It is an indeterminate form of the type ∞. Thus, we apply L’Hôpital’s rule

`n x 1/x
lim = lim = 0.
x→∞ x x→∞ 1
 39 MAT1613/GD001

2. Determine

ex
lim , k > 0.
x→∞ xk

Solution
∞
It is an indeterminate form of the type ∞ and therefore

ex ex
lim = lim ,
x→∞ xk x→∞ kxk−1

∞
which is again of the form ∞. By repeated application of L’Hôpital’s rule, we obtain

ex ex
lim = lim (ex is an increasing function)
x→∞ xk x→∞ k!
= ∞.

3. Determine

x ex
lim .
x→0 1 − ex

Solution
0
It is of the type 0 and therefore

x ex x ex + ex
lim = lim
x→0 1 − ex x→0 −ex
= lim (−x − 1) (simplify)
x→0
= −1.

2.2.2 LIMITS OF THE FORM ±∞ · 0

If lim [f (x) · g(x)] is an indeterminate form of the type ±∞ · 0, we write it as
x→∞

f (x)
lim
x→∞ 1/g (x)

0 ∞
which is again of the form 0 or ∞.
 40

Worked Example
Determine

lim x2 `n x.
x→0+

Solution
1
Since x2 can be written as 1 we have
x2

`n x −∞
lim x2 `n x = lim (of the form )
x→0+ x→0+1/x2 ∞
1/x
= lim
x→0 −2/x3
+

1
= lim − x2 (simplify)
x→0+ 2
= 0.

2.2.3 LIMITS OF THE FORM ∞ − ∞

We try algebraic manipulation (for example by using a common denominator). Then we rewrite the form
∞ − ∞ into the form 00 or ∞
∞.

Worked Example
Determine
 
1 1
lim − .
x→0 x `n (1 + x)

Solution
Obtaining a common denominator we have
   
1 1 `n (1 + x) − x 0
lim − = lim (of the form )
x→0 x `n (1 + x) x→0 x `n (1 + x) 0
1
−1
= lim x 1+x
1+x + `n (1 + x)
x→0

1 − (1 + x)
= lim (simplify)
x + (1 + x) `n (1 + x)
x→0
−x 0
= lim (of the form )
x→0 x + (1 + x) `n (1 + x) 0
−1
= lim
x→0 1 + 1+x + `n (1 + x)
1+x
1
= − .
2
Remark
Note carefully that we simplify the expression after each step before applying the rule again.
 41 MAT1613/GD001

2.2.4 LIMITS OF THE FORMS 00 , 1∞ , ∞0

We shall once again try to reduce these limits to the form 00 or ±∞±∞ . To do this, we make use of the
logarithmic function. As the logarithmic function is continuous on R+ (the positive real numbers),

lim `n f (x) = `n[ lim f (x)].

x→a x→a

Worked Example

1. Determine

1 x
 
lim 1 + .
x→∞ x

(Please see 2.1.3(S).)

Solution
It is a limit of the form 1∞ , therefore we consider

1 x 1 x
     
`n lim 1 + = lim `n 1 +
x→∞ x x→∞ x
 
1
= lim x `n 1 + (of the form ∞ · 0)
x→∞ x
`n 1 + x1


0
= lim 1 (of the form )
x→∞
x
0
1
1+ x1
· − x12
= lim
x→∞ − x12
1
= lim
x→∞ 1 + 1
x
= 1.

Now
1 x
   
`n lim 1 + = 1,
x→∞ x

x
from which it follows that lim 1 + x1 = e1 = e.
x→∞

2. Determine
1
lim (1 + 3x) 2x .
x→∞
 42

Solution
g(x)
It is a limit of the form ∞0 therefore we consider the method eln(f (x)) . We have
 1

ln lim (1 + 3x) 2x
x→∞
1
= lim ln (1 + 3x) 2x
x→∞
1
= lim ln (1 + 3x)
2x
x→∞
ln (1 + 3x) ∞
= lim (of the form )
x→∞ 2x ∞
3
(1+3x)
= lim
x→∞ 2
3
2 c
= lim (Remember lim = 0, where c is a constant)
x→∞ (1 + 3x) x→∞ x
= 0

Consequently
1
lim (1 + 3x) 2x = e0 = 1.
x→∞

Remarks
We draw your attention to the fact that the following are not indeterminate forms:
A limit of the form

0
(a) ∞ has the value 0

(b) 0∞ has the value 0

(c) ∞ · ∞ has the value ∞

(d) +∞ + (+∞) has the value ∞

(e) +∞ − (−∞) has the value ∞

(f) −∞ + (−∞) has the value −∞

(g) −∞ − (+∞) has the value −∞.

Now you can try to do the following yourself. We again give the answers so that you can make sure that
you understand this section thoroughly.

Exercises
Determine the limits:

xk
(1) lim x, k > 0.
x→∞ e

(2) lim (`n x)1/x .

x→∞
 43 MAT1613/GD001

(3) lim xx .
x→0+

x4 +x2
(4) lim x .
x→∞ e +1
 
x−1
(5) lim x `n x+1 .
x→∞
 
1 1
(6) lim x − ex −1 .
x→0

(7) lim (1 + kx)1/x .

x→0
1 1


(8) lim x − sin x .
x→0
x−sin x
(9) lim x .
x→0 e −1

Answers

(1) 0

(2) 1

(3) 1

(4) 0

(5) −2
1
(6) 2

(7) ek

(8) 0

(9) 0.

2.3 THE INVERSE TRIGONOMETRIC FUNCTIONS

This section is very important since here you will come across further inverse functions that you will use
frequently. Also see that the inverse trigonometric function for sin x can either be denoted by arcsin x or
sin−1 x (similarly for the other inverse trigonometric functions). Look carefully at the following important
facts in connection with these functions:

2.3.1 DOMAINS FOR THE INVERSE TRIGONOMETRIC FUNCTIONS

An inverse function f −1 of the function f can only be defined if f is one-to-one on its domain, that is for
each value y in the range, we have precisely one x value in the domain. The trigonometric functions are
not one-to-one. Thus their inverses do not exist. If we restrict the domains of the trigonometric functions,
then their inverses exist on these specific intervals.

When we restrict the domains of the trigonometic functions to specific intervals, their inverses exist on
these intervals. Let us look at the following example:
 44

We know that the sine function is strictly increasing on the closed interval − π2 , π2 and if we restrict sin x
 

to this interval, the inverse does exist. See the following figure:

y =x
_1
y = sin x

y = sin x

Now consider the graphs of the trigonometric functions:

1 1
_ _π _π 2_π
_π 2 2 π 2

0
_π π 0
2
_1 _1

y = sin x y = cos x

1 1
_ _π _π
2 2
_ _π _π
2 0 2 0
_
1 -1

y = tan x y = cot x
 45 MAT1613/GD001

1 1
_ _π
π π 2

0 0
_ _
1 1

y = sec x y = cosec x
By restricting the domains of the trigonometric functions as shown in the graphs, it is now possible to
define the inverse trigonometric functions.

See the following table:(Note that the range of all the inverse functions are NEVER in the third quadrant)

Inverse function Domain Range

y = sin−1 x −1 ≤ x ≤ 1 − π2 ≤ y ≤ π2
quadrants I & IV
y = cos−1 x −1 ≤ x ≤ 1 0≤y≤π
quadrants I & II
y = tan−1 x −∞ < x < +∞ − π2 < y < π2
quadrants I & IV
y = sec−1 x x ≥ 1 or x ≤ −1 0 ≤ y ≤ π y 6= π2
quadrants I & II
y = cosec−1 x x ≥ 1 or x ≤ −1 − π2 ≤ y ≤ π2 y 6= 0
quadrants I & IV
y = cot−1 x −∞ < x < ∞ 0<y<π
quadrants I & II
 46

The following diagrams give the graphs of the six inverse trigonometric functions:

y
y

π
_π
2
1 _π
2

_ x _ x
1 0 1 1 0 1
_
1
_ _π
2

y = sin−1 x y = cos−1 x

y y

_π _π
2 2

x _ x
_ 0 0
1 1 1 1
_ _π _ _π
2 2

y = tan−1 x y = cot−1 x

y y

π π

_π _π
2 2

x x
_ _ _ _ _ _
3 2 1 1 2 3 3 2 1 1 2 3
_ _π _ _π
2 2

π π

y = sec−1 x y = cosec−1 x
 47 MAT1613/GD001

As you can see from the above, it is extremely important always to consider the domains of the different
inverse trigonometric functions when doing proofs or calculations using these functions. Consider the
following examples:

Worked Examples

1. Calculate

tan sin−1 − 21



.

Solution
Let sin−1 − 12 = θ. Then


1
sin θ = − and − π2 ≤ θ ≤ π
2
.
2
Hence it is obvious that θ must be in the first or fourth quadrant. Since sin θ < 0, we know that
sin θ must be in the fourth quadrant. We can represent it as follows:

3
0
_
1
2

Hence   
−1 1 1
tan (sin − = tan θ = − √ .
2 3
(Note: We do not use a pocket calculator to find an approximate value!)

2. Prove that

d 1
cos−1 x = − √ .
dx 1 − x2

Solution
We have y = cos−1 x therefore x = cos y. Because the cosine is 1 − 1 and differentiable on the
open interval (0, π), the inverse of cosine is also differentiable on the open interval (−1, 1). We will
determine the derivative of y = cos−1 x :

cos y = x
d d
(cos y) = (x) .
dx dx
 48

dy
Using the chain rule, we have − sin y dx = 1.
If we now devide by sin y > 0 (where 0 < y < π), we have

dy −1 −1 −1
= =p =√ .
dx sin y 2
1 − cos y 1 − x2

The derivative of y = cos−1 x with regard to x is then

d −1
cos−1 x = √


.
dx 1 − x2

2.3.2 DERIVATIVES OF INVERSE TRIGONOMETRIC FUNCTIONS; RELATED IN-

TEGRALS
Just as important as the differentiation is the simplification of the derived function. Consider the following
examples, then attempt the given exercises.

Worked Example
Determine the derivative of
√
f (x) = x cosec−1 x1 + 1 − x2

and simplify your answer.

Solution  
−1 1 d d −1 1 d p 
0
f (x) = cosec · (x) + x · cosec + 1 − x2 .
x dx dx x dx
Using the chain rule    
d −1 1 −1 d 1
cosec = q · .
dx x 1 1
− 1 dx x
x x2

Hence
 
 
1 −1 1
f 0 (x) = cosec−1 · 1 + x   −
x2
q
x 1 1
−1
x x2
1
−1/2
+ 1 − x2 (−2x)
2
1 x x
= cosec−1 + q −√ .
x 1 1−x 2
· x2 1 − x2
x x2
√
We know that x2 = |x|, with |x| = x if x ≥ 0 and |x| = −x if x ≤ 0. Also |x|.|x| = x2 . Thus
1 x x
f 0 (x) = cosec−1 + 1 1
√ −√
x |x| · |x|
2
1−x ·x2 1 − x2
1 x x
= cosec−1 +√ −√
x 1−x 2 1 − x2
1
= cosec−1 .
x
 49 MAT1613/GD001

Exercises

1. Calculate the exact values of each of the following:

(a) sin cos−1 − 53 .

(b) cos tan−1 12 .

2. Determine the derivative of each of the following and simplify your answers:
 
(a) f (x) = (x2 + 1) tan−1 x+1
x−1
, x 6= −1.
√
(b) f (x) = x a2 − x2 + a2 sin−1 xa , a < 0.

Answers

1. (a) 4/5.
√
(b) 2/ 5.

2. (a) 2x tan−1 x−1

x+1 + 1.
2
(b) √−2x .
a2 −x2

2.4 HYPERBOLIC FUNCTIONS

Certain combinations of the exponential function appear so often in the application as well as the theory of
mathematics, that it is worthwhile to assign special names to them. You will notice that their properties
are reminiscent of those of the trigonometric functions which also explains the names of these functions.
The hyperbolic functions are defined as follows:
2.4.1 x −x
hyperbolic sine function: sinh x = e −e 2
x −x
hyperbolic cosine function: cosh x = e +e 2
x −e−x
hyperbolic tangent function: tanh x = eex +e −x
ex +e−x
hyperbolic cotangent function: coth x = ex −e−x
2
hyperbolic secant function: sech x = ex +e −x
2
hyperbolic cosecant function: cosech x = ex −e −x

The following identities on the hyperbolic functions can be proved easily.

2.4.2
Identities:
cosh2 x − sinh2 x = 1
sinh 2x = 2 sinh x cosh x
cosh 2x = cosh2 x + sinh2 x
cosh2 x = cosh 22x+1
sinh2 x = cosh 22x−1
tanh2 x = 1 − sech2 x
coth2 x = 1 + cosech2 x
 50

The derivatives of the hyperbolic function follow easily from the above definitions as well as the facts that
d x x d −x −x
dx (e ) = e and dx (e ) = (−1) e .
The following table is important:

2.4.3
Derivatives of the hyperbolic functions:
d
dx (sinh x) = cosh x
d
dx (cosh x) = sinh x
d 2
dx (tanh x) = sech x
d 2
dx (coth x) = − cosech x
d
dx (sech x) = − sechx tanh x
d
dx (cosech x) = −cosech x coth x

The integral formulas for the hyperbolic functions are:

2.4.4
Integrals of the hyperbolic functions:
R
sinh xdx = cosh x + c
R
cosh xdx = sinh x + c
sech2 x dx = tanh x + c
R

cosech2 xdx = − cothx + c

R
R
sech x tanh xdx = − sechx + c
R
cosechx coth xdx = − cosechx + c

Worked Examples

1. Prove that

d
(sinh x) =cosh x.
dx

Solution

d ex − e−x
 
d
(sinh x) =
dx dx 2
1 x
e − e−x (−1)



=
2
ex + ex
=
2
= cosh x.

2. Determine

d
{`n(coth x)} .
dx
 51 MAT1613/GD001

Solution
d 1
−cosech2 x


{`n (coth x)} =
dx coth x
− sinh x 1
= ·
cosh x sinh2 x
−1
= .
sinh x cosh x

3. Determine

tanh2 2x dx.
R

Solution
Z Z
2
1 − sech2 2x dx


tanh 2x dx =
1
= x− tanh 2x + c.
2

4. Determine
Z
sech x dx.

Solution
Z Z
dx
sech x dx =
cosh x
Z
cosh x
= 2 dx
Z cosh x
cosh x
= dx
1 + sinh2 x
= arctan (sinh x) + c

(by using the substitution u = sinh x).

5. Determine
Z
ex cosh x dx.

Solution
ex + e−x
Z Z  
x x
e cosh x dx = e dx
2
Z
1
= (e2x + 1)dx
2
1 2x 1
= e + x + c.
4 2
 52

6. Determine
Z `n 2
sinh
dx.
0 1 + cosh x

Solution
Z `n 2
sinh x
Consider dx.
0 1 + cosh x
Let u = 1 + cosh x. Then du = sinh x dx. Also, if x = 0, then

e0 + e−0 1+1
u = 1 + cosh 0 = 1 + =1+ = 2,
2 2

and if x = `n 2, then
1
e`n2 + e−`n2 2+ 2 9
u = 1 + cosh(`n 2) = 1 + =1+ = .
2 2 4

Hence
Z `n 2 Z 9/4
sinh x
dx = du/u
0 1 + cosh x 2
9/4
= `n u]2
9
= `n − `n 2
4
9
= `n .
8

Note how the definition of cosh x is used in Example 6 to calculate the values of cosh 0 and cosh (`n2).

Again note that `n 98 is an exact answer and a pocket calculator is not used to approximate it.

Now try the following:

Exercises

1. Rewrite the expressions in terms of exponentials and simplify the results

(a) 2 cosh (ln x)

(b) cosh 8x + sinh 8x
(c) ln (cosh x − sinh x) + ln (cosh x + sinh x)

2. Determine

d
cosh x2 − 1 .


(a) dx
d
sinh3 (x/2) .
 
(b) dx
d
 x/2 
(c) dx e tanh 2x .
 53 MAT1613/GD001

d
tanh x1


(d) dx

3. Determine

sinh2 x dx
R
(a)
sech4 x dx
R
(b)
R
(c) x sinh x dx.
R ln 2 −x
(d) 0 4e sinh xdx

Answers

1
1. (a) x + x

(b) e8x
(c) 0

2. (a) 2x sinh (x2 − 1).

(b) 3
2 sinh2 x
2 cosh x2 .
1
(c) ex/2 (2 sech2 2x + 2 tanh 2x).
(d) − x12 sech2 x1

1
3. (a) 4 sinh 2x − 12 x + c.
1
(b) tanh x − 3 tanh3 x + c.
(c) x cosh x− sinh x + c.
3
(d) ln 4 − 4

Summary
The learner will now be able to obtain limits of indeterminate forms when they appear in the forms
0 ∞ h(x)
0 , ∞ , 0.∞ and also when they occur in the functions f (x) = g (x) as 10 , 1∞ , ∞0 . L’Hôpital’s rule
can be used in all these forms by applying an adaptation of the original given form. The learner will also
be able to demonstrate skills in differentiation and integration of the inverse trigonometric function and
hyperbolic function.
 54

Chapter 3

Applications of Integration
(areas between curves, solids of revolu-
tion)

Introduction
This section shows how to find the areas of regions in the coordinate plane by integrating the functions
that define the boundaries of the regions. Solids generated by revolving plane regions about axes are solids
of revolution. Threaded spools are solids of revolution and so are billiard balls. These solids have volumes
we can find using geometry. But sometimes we want to find the volume of a blimp instead or to predict
the weight of a part that will be turned on a lathe. In these cases geometric formulas are of little help and
we have to return to calculus for the answers.

Outcomes
After studying this chapter, the learner should be able to:

• find the area between curves of a variety of functions.

• find the volume of a solid of revolution (by rotation about either the x-axis or the y-axis) using the
disk method or the washer method.

3.1 AREAS BETWEEN CURVES

Study this whole section carefully and work through Examples 1.1 – 1.6. When studying this section
you have to keep the following in mind: It is important to make sketches whenever possible, when you
have to determine the area underneath the curve, the area between curves or the volume of a solid of
revolution. You should always show all calculations, even, for example, the calculations to determine
points of intersection. If you do not show all your calculations, we cannot identify your problem areas and
then we would not be able to help you. Sometimes it could be easier to integrate with respect to y than
to integrate with respect to x.
 55 MAT1613/GD001

3.2 VOLUMES OF SOLIDS OF REVOLUTION

The volume of a solid of revolution about the x–axis (see that the formula is in terms of x):
The volume of a solid of revolution between the x–axis and the graph of a continuous function y = R (x),
a ≤ x ≤ b about the x–axis is

Rb Rb
V = a π (radius)2 dx = a π (R (x))2 dx

and the volume of a solid of revolution about the y–axis is (see that the formula is in terms
of y):

Rd Rd
V = c π (radius)2 dy = c π (R (y))2 dy.

Study this whole section thoroughly. Work through all the Examples in this section carefully. Use the
following guidelines to calculate the volumes of solids of revolution:

• Sketch the given curves.

• Calculating a volume by revolving the region between the two curves about an axis, make sure that
you know which of the two curves will be the greater over the given interval, otherwise it is possible
that you may get a negative answer.

• Make sure about which axis, i.e. the x-axis or the y-axis, the area revolves.

• When using the washer method, make sure that you know what R(x) = outer radius and r(x) = inner
radius represent.

• When the rotation axis is parallel to the x-axis the formula is also in terms of x and similarly, when
the rotation axis is parallel to the y-axis the formula is also in terms of y.

Now we will do an example to show you once more how to work with the rotations about the different
axes:

Worked Examples

1. Determine the volume of the solid of revolution which results when the closed region enclosed by
y = x2 and the line y = x is rotated about the x-axis;

Solution

The graphs of y = x2 and y = x intercept where x2 − x = 0, i.e. where x = 0 and x = 1. So the

points are (0, 0) and (1, 1).
 56

We first sketch the region R as in the following figure:

y
2
y =x
y =x

We form the solid by rotating the region R about the x–axis. Note that the line y = x is always above
the parabola on the interval [0, 1] so we form a washer to approximate the volume of revolution, the
outer radius is y = R (x) = x and the inner radius y = r (x) = x2 . Thus the required volume is

Z 1
V = π [R (x)]2 − [r (x)]2 dx
0
Z 1
= π x2 − x4 dx
0
  1
1 3 1 5
= π x − x
3 5 0
2π
= .
15

2. Determine the volume of the solid of revolution which results when the closed region enclosed by
y = x2 and the line y = x is rotated about the y-axis.

Solution

Since we are revolving R about the y-axis we use horizontal washers to approximate the solid of
revolution. Please see the following figure:

y
2
y =x y =x

√
Note that the parabola x = y is to the right of the line x = y on the interval [0, 1]. The washer
 57 MAT1613/GD001

√
that we are going to work with has outer radius R (y) = y and inner radius r (y) = y. Then
Z 1
V = π [R (y)]2 − [r (y)]2 dy
0
1
√
Z
= π ( y)2 − y 2 dy
0
Z 1
= π y − y 2 dy
0
1
y2 y3

= π −
2 3 0
π
= .
6

3. Determine the volume of the solid of revolution which results when the closed region enclosed by
y = x2 and the line y = x is rotated about the line y = 2.

We always first determine the points of intersection of the two curves. Then

x2 = x

i.e.
x2 − x = 0.

Thus
x(x − 1) = 0.

Hence, the points of intersection are at x = 0 and x = 1. If x = 0, then y = 0 and if x = 1, then

y = 1.

Solution
The outer radius is R (x) = 2 − x2 and the inner radius r (x) = 2 − x is shown in the next figure

y
2
y =x
4
y =x
3
R r
2 y =2

x
1
 58

The volume is thus

Z 1
V = π [R (x)]2 − (r (x))2 dx
0
Z 1h
2 i
= π 2 − x2 − (2 − x)2 dx
0
Z 1
x4 − 5x2 + 4x dx


= π
0
1
x5 5x3 4x2

= π − +
5 3 2 0
8π
= .
15

Summary
After a learner has worked through all these examples, he/she will be able to evaluate areas between curves
and evaluate volumes of solids by revolving plane regions about different types of axes, for instance (about
the x-axis, y-axis, line y = 2 etc.).
 59 MAT1613/GD001

Chapter 4

Advanced Techniques of Integration

Introduction
We know in theory how integrals are evaluated using antiderivatives. The more sophisticated our models
become, however, the more involved our integrals become. We need to know how to change these more
involved integrals into forms we can work with. The first goal of this chapter is to change unfamiliar
integrals into integrals we can recognize.

Outcomes
After studying this chapter the learner should be able to:

• use substitution to obtain some basic integrals (integration by substitution is included in MAT1512).
Make sure you understand and can apply it.

• use the following integration techniques:

– integration by parts
– integration by trigonometric substitution
– integration by partial fractions
– integration by z–substitution.

• identify the integration method that should be used for a specific problem.

4.1 BASIC INTEGRATION FORMULAS

When you want to solve an integral, first determine whether the integrand is in a standard form. Then you
can directly apply the converse of differentiation, i.e. you ask of which function is the function under the
R
integral sign the derivative for example cos xdx = sin x + c because the function cos x is the derivative
of the function sin x.

In many cases it is also possible to change the integrand to a standard forms using algebraic manipulation
or trigonometric identities.
 60

Now try to do as many problems as possible in your textbook. The only way in which you can master the
basic integration formulas and also memorise them is by regularly doing as many exercises as possible.

4.2 INTEGRATION BY PARTS

This is one of the main methods of integration and thus you must study this section thoroughly.

Do you still remember the formula for the differentiation of a product? If u and v are two differentiable
functions then
d (uv) dv du
=u +v
dx dx dx
(the first × derivative of the second + the second × the derivative of the first). We write it in simple
terms as
d (uv) = udv + vdu

Now integrate both sides of this formula to obtain the formula for integration by parts:
Z Z Z
d (uv) = u dv + v du

so that Z Z
uv = u dv + v du.

When we rewrite this formula we obtain:

Formula for integration by parts

Z Z
u dv = uv − v du.

Before discussing some examples, keep the following remarks in mind when you do integration by parts.

Remarks

R
1. If the integrand consists of a product of two functions that are not related, for example x sin x dx
(it is obvious that substitution cannot be used), then we use the technique of integration by parts.

2. Remember the special cases for which we use integration by parts: `n, tan−1 , sin−1 , cos−1 , and
sec−1 . Note carefully that in every one of these cases, we must choose u and dv as follows: u = f (x)
where f (x) is one of the above-mentioned functions and dv = dx.

3. Furthermore, we also write sec3 x = sec x sec2 x and cosec 3 x = cosec x cosec 2 x and apply integration
by parts.
 61 MAT1613/GD001

4. The choice of u and dv is of major importance. We must make sure that we can in fact integrate
the function that is chosen as dv. If possible, we must also choose u in such a way that it becomes
simpler when we differentiate.
When we have a power of x (or the variable which we work with) in the integrand and we use the
method of integration by parts, we usually choose the power of x so that the power decreases (i.e.
p−1
it becomes simpler) in other word if u = xp ( p the power) then du = px dx.
The most important exclusion is when the functions under the integrand are such that the one is a
power of x and the other, one of the examples of remark 2 above i.e `n, tan−1 , etc., i.e. a function
of which we know the derivative but not the integral.

Now work through the following two examples

Worked Examples

(a) Determine
Z
I= x2 sin x dx.

Solution:
In this case, we let u = x2 and dv = sin x dx. Then du = 2x dx and v = − cos x dx so that
Z
2
I = −x cos x + 2 x cos x dx.

R
For the second integral that we have here i.e. I2 = x cos x dx we again use u = x and
dv = cos x dx, so that the derivative of x becomes a constant = 1.
(b) Determine
Z
I= x sin−1 x dx.

Solution:
In this case we set u = sin−1 x and dv = x dx because the derivative of sin−1 x is known but
the integral not.
Then du = √ 1 dx and v = 21 x2 . We use the formula for integration by parts and obtain
1−x2

x2
Z
1 1
I = x2 sin−1 x − √ dx.
2 2 1 − x2
R x2
For the integral I2 = √1−x 2
dx we use trigonometric substitution with x = sin θ and also the
2 1−cos 2θ
identity sin θ = 2 .

5. In the case of repeated application of the formula, we may not change the order of u and dv. We
will point this out again in the examples.
 62

6. The constant is added only to the last integral. It can be shown that it makes no difference.

7. When you are solving problems, you would do well to remember the following:
Z
cos ax
sin ax dx = − +c
a

and
Dx (sin ax) = a cos ax.

8. When you use a repeated application of the method of integration by parts, you must simplify the
integrand as far as possible after each step, i.e. you must take out all the constants and algebraic
signs from the integrand. The following example illustrates this:
Determine
Z
I= e2x cos 3x dx.

Solution
Put u = cos 3x and dv = e2x dx. Then du = −3 sin 3x dx and v = e2x /2. We use the formula for
integration by parts: Z Z
I = u dv = uv − v du

and then
e2x
Z 2x
e
I = cos 3x. − (−3 sin 3x)dx
2 2
e2x cos 3x 3
Z
= + e2x sin 3x dx
2 2
3
Take out − and write it in front of the integral sign.
2
e2x cos 3x 3 e2x e2x
Z 
= + sin 3x − 3 cos 3x dx
2 2 2 2
(Remember the bracket)
e2x cos 3x 3 2x
Z
9
= + e sin 3x − e2x cos 3xdx
2 4 4
e2x cos 3x 3 2x 9
= + e sin 3x − I.
2 4 4

Now take similar terms to the left-hand side to obtain:

Z
9 13 13 1 3
I+ I= I= e2x cos 3x dx = e2x cos 3x + e2x sin 3x.
4 4 4 2 4

Consequently
e2x
Z
I= e2x cos 3x dx = (2 cos 3x + 3 sin 3x) + c.
13
 63 MAT1613/GD001

9. After (repeated) application(s) of the formula, one of the following can occur:

(a) The integrand becomes simpler and the integral can be determined directly.
(b) The integrand reappears in its original form.

(The first three of the following examples are of case (a) and the following two are of case (b).)

Further Worked Examples

1. Determine
Z
I= x2 cos x dx.

(See a similar problem in remark (4) above).

Solution
Set u = x2 and dv = cos x dx. Then du = 2xdx and v = sin x. Hence
Z Z
2 2
I = x cos x dx = x sin x − 2 x sin x dx (*)
R
Now consider x sin x dx. We may not change the order of u and dv if we want to apply the
formula again. (Check, for yourself, what happens if you should in fact change the order, in
other words by setting u = sin x and dv = x dx.) Therefore, we set u = x and dv = sin x dx.
Then, again using the formula for integration by parts
Z Z
x sin x dx = x (− cos x) + cos x dx
= −x cos x + sin x.

Therefore it follows from (∗) that

Z
I = x2 cos x dx

= x2 sin x − 2(−x cos x + sin x) + c

= x2 sin x + 2x cos x − 2 sin x + c.

2. Determine

√
Z
I= x 1 + x dx.

Solution
Set u = x and dv = (1 + x)1/2 dx. Then
√ (1 + x)3/2 (1 + x)3/2
Z Z
I = x 1 + xdx = x. − .1 dx
3/2 3/2
2 2 (1 + x)5/2
= x(1 + x)3/2 − +c
3 3 5/2
2 4
= x(1 + x)3/2 − (1 + x)5/2 + c.
3 15
 64

3. Determine
Z π
2
I= x3 cos 2x dx.
0

Solution
R R
Use the formula udv = uv − vdu for integration by parts with

u = x3 and dv = cos 2x dx.

Thus
sin 2x
du = 3x2 dx and v= .
2
It gives
Z π  π Z π
3 sin 2x sin 2x
2 2 2
3
x cos 2x dx = x − 3x2 . .dx
0 2 0 0 2
Z π
3 2
= 0− x2 sin 2x dx.
2 0

By applying integration by parts twice more we see that

Z π "   π Z π   #
2
3 3 2 cos 2x 2 2 cos 2x
x cos 2x dx = − x . − − 2x. − dx
0 2 2 0 0 2
  Z π
3  π 2 1 3 2
= − − x cos 2x dx
2 2 2 2 0
"  π Z π #
3 2 3 sin 2x 2 2 sin 2x
= − π − x. − dx
16 2 2 0 0 2
Z π
3 2 3 2
= − π +0+ sin 2x dx.
16 4 0
 π
3 2 3 cos 2x 2
= − π + −
16 4 2 0
3 2 3
= − π − (−1 − 1)
16 8
3 2 3
= − π + .
16 4

4. Determine
Z
I= sin2 θ dθ.

Solution

Let Z Z
I= sin2 θ dθ = sin θ. sin θ dθ
 65 MAT1613/GD001

Let u = sin θ and dv = sin θ dθ. Then du = cos θ dθ and v = − cos θ. Hence
Z Z
2
I = sin θ dθ = sin θ(− cos θ) − (− cos θ dθ). cos θ dθ
Z
= − sin θ cos θ + cos2 θ dθ
Z
= − sin θ cos θ + (1 − sin2 θ)dθ
Z Z
= − sin θ cos θ + 1 dθ − sin2 θ dθ
Z
= − sin θ cos θ + θ − sin2 θ dθ
= − sin θ cos θ + θ − I.

By collecting similar terms on the left-hand side, we find that

Z
2I = 2 sin2 θ dθ = − sin θ cos θ + θ,

and therefore Z
1 θ
I= sin2 θ dθ = − sin θ cos θ + + c.
2 2

• 5. Determine

2 tan−1 x
Z
I= dx.
x2

(Since tan−1 is one of the special cases that can only be solved by integration by parts, this
example is included here.)
Solution
R R
Use the formula udv = uv − vdu with

1
u = tan−1 x and dv = dx.
x2
Then
dx 1
du = and v=− .
1 + x2 x
Thus
2 tan−1 x tan−1 x
Z  Z 
1 1
= 2 − − − · dx
x2 x x 1 + x2
tan−1 x
 Z 
dx
= 2 − + .
x x (1 + x2 )

Now consider
1 A Bx + C
= +
x (1 + x2 ) x 1 + x2
A 1 + x2 + x (Bx + C)


= .
x (1 + x2 )
 66

Hence
1 = x2 (A + B) + xC + A.

By now comparing the coefficients of equal powers of x we see that:

xo : 1 = A
x1 : 0 = C
x2 : 0 = A + B which implies B = −1.

Thus
2 tan−1 x tan−1 x
Z  Z   
1 x
I = dx = 2 − + − dx
x2 x x 1 + x2
2 tan−1 x
 
2
= − + 2`n|x| − `n(x + 1) + c
x
2 tan−1 x x2
= − + `n 2 + c.
x x +1

4.3 TRIGONOMETRIC INTEGRALS (USING SUBSTITUTION)

When we have an integral of the form cos4 x sin x dx, we immediately see that if we write this in the
R

form [cos x]4 sin x dx, that sin x is the derivative of cos x. We use the substitution formula:
R

Substitution formula

f (g (x)) · g 0 (x) dx =
R R
f (u) du.

where f and g are related.

In this case, we substitute u = g (x) = cos x so that du

dx = g 0 (x) and du = − sin x dx.

Then we obtain the integral

Z
1
− u4 du = − u5
5
1
= − (cos x)5 + c.
5
In order to do more difficult examples we use reduction methods:

For example, we use sin2 x = 1 − cos2 x to obtain again an integral where we can use the substitution
formula above.
 67 MAT1613/GD001

Worked example
Determine
Z
cos3 x sin4 x dx.

Solution
Now
Z Z
cos3 x sin4 x dx = cos x cos2 x sin4 x dx
Z
= cos x(1 − sin2 x) sin4 x dx
Z
= cos x(sin4 x − sin6 x)dx.

Now set
u = sin x.

Then
du = cos x dx,

and so it follows that

Z Z
3 4
cos x sin x dx = (u4 − u6 )du

u5 u7
= − +c
5 7
sin5 x sin7 x
= − + c.
5 7
Remark
From the example, we see that integrals of the form sinn x dx, cosn x dx, cosn x sinm x dx and
R R R

sinn x cosm x dx, with n odd, can be determined easily by first splitting the integrand and then writing
R

cos x in terms of sin x, or vice versa. Integrals of the form

Z Z
secn x tanm x dx and cosecn x cotm x dx

can be determined similarly using sec2 x = 1 + tan2 x or cosec2 x = 1 + cot2 x.

See the following worked examples.

Worked Examples

1. Determine
Z
I= tan2 x sec4 x dx.
 68

Solution
Z
I = tan2 x sec2 x sec2 x dx
Z
= tan2 x(tan2 x + 1) sec2 x dx
Z
= (tan4 x + tan2 x) sec2 x dx.

Now set
u = tan x,

then
du = sec2 x dx,

and
Z Z
2 4
I = tan x sec x dx = (u4 + u2 )du

u5 u3
= + +c
5 3
1 1
= tan5 x + tan3 x + c.
5 3

2. Determine
Z
I= cot3 x cosec5 x dx.

Solution
Z Z
I = cot3 xcosec5 x dx = (cot2 x cosec4 x)(cot x cosec x)dx
Z
= (cosec2 x − 1)cosec4 x cot x cosecxdx
Z
= (cosec6 x − cosec4 x) cot x cosecx dx.

Therefore set
u = cosec x,

then
du = − cosec x cot x dx,

and
Z Z
3 5
I = cot x cosec x dx = − (u6 − u4 )du

u7 u5
= − + +c
7 5
cosec7 x cosec5 x
= − + + c.
7 5
 69 MAT1613/GD001

Now try to do the following exercises yourself:

Exercises
Determine the following integrals:
Z
1. sin5 x dx
Z
2. tan3 x sec3 x dx

√
Z
3. tan x sec4 x dx
Z
4. sec6 x dx.

Answers:

2 cos3 x cos5 x
1. − cos x + 3 − 5 +c
Z Z
5

2
Hint: Write sin x dx = 1 − cos2 x sin x dx

Z
1 − 2 cos2 x + cos4 x sin x dx



=

1
2. 5 sec5 x − 13 sec3 x + c
Z Z
tan3 x sec3 x sec4 x − sec2 x tan x sec x dx


Hint: Write dx =

2
3. 7 tan7/2 x + 23 tan3/2 x + c
Z Z
1 1
(tan x) 2 sec4 x dx (tan x) 2 tan2 x + 1 sec2 x dx


Hint: Write =

4. tan x + 32 tan3 x + 15 tan5 x + c.

Z Z

2
Hint: Write sec6 x dx = tan2 x + 1 sec2 x dx

4.4 TRIGONOMETRIC SUBSTITUTION

This substitution is easy to recognise and you will encounter it frequently. Thus, make sure that you
understand this section well.

Trigonometric substitutions enable us to replace the binomials a2 + x2 , a2 − x2 and x2 − a2 by single

variables.

We once again discuss this method so that you may ensure that you understand it well.
 70

By respectively making use of the three identities

• 1 − sin2 θ = cos2 θ

• tan2 θ + 1 = sec2 θ

• sec2 θ − 1 = tan2 θ

we obtain the following three substitutions for a > 0 :

For use and obtain

√ p
a2 − x2 x = a sin θ a 1 − sin2 θ = |a cos θ|

√ √
a2 + x2 x = a tan θ a 1 + tan2 θ = |a sec θ|

√ √
x2 − a2 x = a sec θ a sec2 θ − 1 = |a tan θ|

When we make substitutions we always want them to be reversible so that we can change back to the
original variables after we are done. For example, if x = a tan θ, we want to be able to set θ = tan−1 xa
after integration have taken place. As we know from section 2.3.1(S), the functions in these substitutions
have inverses only for selected values of θ.
For reversibility
x = a tan θ requires θ = tan−1 xa with − π2 < θ < π2
x = a sin θ requires θ = sin−1 xa with − π
(2 ≤ θ ≤ 2
π

0 ≤ θ < π2 if xa ≥ 1 i.e. x ≥ a
x = a sec θ requires θ = sec−1 xa with π x
2 < θ ≤ π if a ≤ −1 i.e. x ≤ −a
It is important at each substitution to state clearly what the limits of θ are. The integral only exists where
sin−1 , tan−1 , or sec−1 is defined and where it is one-to-one. (See examples 1 to 3 that follows).

Worked Examples

1. Determine

3/2
16 − 9x2
Z
I= dx.
x6

Solution
Set
4 π π
x= sin θ, − ≤ θ ≤ .
3 2 2
(Remember we must restrict θ so that we have a one-to-one function whose inverse exists.)
Then
4
dx = cos θ dθ,
3
 71 MAT1613/GD001

and p
16 − 9x2 = |4 cos θ| = 4 cos θ,

since
π π
cos θ ≥ 0 for − ≤θ≤ .
2 2
Therefore
64 cos3 θ 34 cos θ
Z
I = 4096 6
dθ
729 sin θ
cos4 θ
Z
243
= 6 dθ
16 sin θ
Z
243
= cot4 θ cosec2 θ dθ
16
By making use of the first substitution rule, we therefore set

u = cot θ.

Then
du = − cosec2 θ dθ

and therefore

3
16 − 9x2 2
Z Z
243
I = dx = u4 (−du)
x6 16
243 cot5 θ
= − + c.
16 5
In order to write the integral back in terms of x we look at the following representation of the angle
θ in terms of x.

From
4 π π
x= sin θ, − ≤ θ ≤ ,
3 2 2
we have
3x π π
sin θ = , − ≤θ≤
4 2 2
so

4
3x

θ
_ 2
16 9x

Therefore

3
5
16 − 9x2 243 1 16 − 9x2 2
Z 2
dx = · +c
x6 16 5 243x5

5
1 16 − 9x2 2
= − + c.
80 x5
 72

2. Determine

Z
dx
√ when x > 35 .
25x2 − 9

Solution
√
We rewrite 25x2 − 9 :
s 
p 9
25x2 − 9 = 2
25 x −
25
s  2
2
3
= 5 x −
5

We now have a form x2 − a2 .

We now use substitution with

3 3  π
x= sec θ, dx = sec θ tan θdθ, θ ∈ 0,
5 5 2

[Note that θ ∈ 0, π2 since x > 35 ].

Also
 2
2 3 9 9
x − = sec2 θ −
5 25 25
9
sec2 θ − 1


=
25
9
= tan2 θ
25

so that
s  2
3 3
x2 − = tan θ .
5 5

If x = 35 sec θ, 0 < θ < π2 , then θ = sec−1 5x

3 and we can read the values of the other trigonometric
functions of θ from the right triangle:

5x
2 _
25x 9
θ
3
 73 MAT1613/GD001

We now use the substitutions above to obtain

Z Z
dx dx
√ = q
25x2 − 9 5 x2 − 3 2


5
3
sec θ tan θdθ
Z
5
= |tan θ| = tan θ
5 35 tan θ


Z
1
= sec θdθ
5
1
= ln |sec θ + tan θ| + c
5
√
1 5x 25x2 − 9
= ln + +c
5 3 3

Remark
If x < − 35 , we use the same substitutions i.e. x = 35 sec θ, dx = 35 sec θ tan θdθ with θ ∈ π


2,π (see
p. 72(S)). Also |tan θ| = − tan θ. The reference triangle for θ ∈ π2 , π is then

_
5x = 5x
2 _
25x 9
θ
_3

Then the integral becomes

Z Z
dx 1
√ = − sec θdθ
25x2 − 9 5
1
= − ln |sec θ + tan θ| + c
5
√
1 5x 25x2 − 9
= − ln − + c.
5 3 3

3. Determine
Z
dx
I= .
(4x2 − 24x + 27)3/2

We do not recognise any of the standard forms for trigonometric substitution in this integral, but
by completing the square, we obtain

4x2 − 24x + 27 = 4(x − 3)2 − 9.

Therefore, we set
3 nπ o
x−3= sec θ, θ ∈ (0, π) − ,
2 2
 74

then
3
dx = sec θ tan θ dθ
2
and
p
4x2 − 24x + 27 = 3 |tan θ| .

Therefore
3
sec θ tan θ
Z Z
dx 2
I = = dθ
(4x2 − 24x + 27)
3
2 27 tan2 θ |tan θ|
|cos θ|
Z
1
= dθ.
18 sin2 θ

sin θ sin θ
π
Note that |tan θ| = cos θ = |cos θ| for θ ∈ [0, π] − 2 . Therefore
(
1
R cos θ π
2 θ dθ, if θ ∈ (0, 2 ),
Z
dx 18
I= 3 = 1
R sin
cos θ
(4x2 − 24x + 27) 2 − 18 sin2 θ
dθ, if θ ∈ ( π2 , π).

Now set
u = sin θ,

then
du = cos θ dθ

and
Z Z
dx
1 du
I = 3 = ±
(4x2 − 24x + 27) 2 18 u2
1 1
= ∓
18 u
1
= ∓ cosecθ + c.
18

Now
2 (x − 3)
sec θ =
3
which can be represented as follows:
For θ ∈ 0, π2 :


_
2x 6
2 _
4x 24 x + 27

θ
3
 75 MAT1613/GD001

π

For θ ∈ 2,π :
_ _ _
2x 6 = (2x 6 )
_
4x2 24 x + 27
θ
_
3

From this, it follows that

( 2x−6
if θ ∈ 0, π2 , i.e. x > 23 + 3,


√
cosecθ = 4x2 −24x+27
−(2x−6)
if θ ∈ π2 , π , i.e. x < − 32 + 3,

√
4x2 −24x+27

Therefore
±2 (x − 3)
Z
dx 1
I = 3 =∓ √ +c
(4x2 − 24x + 27) 2 18 4x2 − 24x + 27
1 x−3
= − √ + c.
9 4x − 24x + 27
2

Summary
We have shown how to do trigonometric substitution.

4.5 PARTIAL FRACTIONS

P (x)
A rational function has the form Q(x) where P (x) and Q(x) are polynomial functions and Q(x) 6= 0. Such
a fraction is proper if the degree of P (x) is smaller than the degree of Q(x) and improper if the degree of
P (x) is greater than or equal to the degree of Q(x):
x+1
is a proper rational function
3x2 + 5x − 1
(the highest power of x in the numerator is smaller than the highest power of x in the denominator) and

x3 + 5
is an improper rational function.
(x + 1)3

(the highest power of x in the numerator is equal to the highest power of x in the denominator).

An improper fraction is reduced to a polynomial plus a proper fraction by long division, for example

x2 1
=x−1+ .
x+1 x+1
It is very important that the rational function that needs to be integrated must be in the form of a
polynomial plus a proper rational function.
x+7 2 1
The process by which a fraction such as x2 +8x+15 can be decomposed into the fractions x+3 − x+5 is called
splitting into partial fractions.
 76

Remarks

• Note that the examples on partial fractions always have positive integers as powers of x (or any
variable that we work with) in both the numerator and the denominator.

• We have two kinds of factors in Q (x) , namely linear factors (i.e. of the form ax + b), for example
2x + 3, x + 1 etc., and irreducible factors (i.e. of the form ax2 + bx + c) that cannot be decomposed
into further linear factors (i.e. those where b2 − 4ac < 0), for example 2x2 + 3, 3x2 + 2x + 1, etc.

Let’s consider a number of specific cases of linear factors, repeated linear factors, irreducible quadratic
factors and repeated irreducible factors in the denominator.

4.5.1 LINEAR FACTORS IN THE DENOMINATOR

P (x)
Consider a proper rational function Q(x) where Q(x) can be resolved into linear factors

(a1 x + b1 )(a2 x + b2 )(an x + bn )

Then
P (x) P (x)
=
Q (x) (a1 x + b1 ) (a2 x + b2 ) · · · (an x + bn )
A1 A2 An
= + ··· + .
a1 x + b1 a2 x + b2 an x + bn

Worked Example
Split

x + 35
x2 − 25

into partial fractions.

Solution
x + 35 x + 35
=
x2 − 25 (x − 5) (x + 5)
A B
= + .
x+5 x−5
To determine A and B, we add the fractions once more and find that
x + 35 A B
= +
(x − 5) (x + 5) x+5 x−5
A (x − 5) + B (x + 5)
=
(x + 5) (x − 5)

Therefore
x + 35 = A(x − 5) + B(x + 5). ...(1)

The unknown constants can be determined using many different methods. Here we show two, namely,
fixed values or equal powers of x.
 77 MAT1613/GD001

(i) Choose fixed values for x (usually the most convenient ones for which some terms become zero).
In (1) let x = 5, then
40 = A.0 + B.10,

and therefore
B = 4.

Also let x = −5, then

30 = A.(−10) + B.0,

and therefore
A = −3.

(ii) Comparing the coefficients of equal powers of x.

From (1), it follows that

x + 35 = Ax − 5A + Bx + 5B
= (A + B) x + (−5A + 5B) .

By now comparing the coefficients, we see that

x : 1=A+B ...(a)
Constant (x0 ) : 35 = −5A + 5B ...(b)

(a) × 5 + (b) gives that 40 = 10B and therefore, B = 4. By setting B = 4 in (a), we obtain that
A = −3.
From the above, we therefore see that
x + 35 −3 4
2
= + .
x − 25 x+5 x−5
At this stage method (i) is simpler than method (ii), but later both methods will have to be used to
determine certain constants. Next we use the method of splitting into the partial fractions to solve
integrals where the integrand is a rational function:

Worked Examples

1. Determine

23 − 2x
Z
I= dx.
2x2 + 9x − 5

Solution
23 − 2x 23 − 2x
=
2x2 + 9x − 5 (2x − 1) (x + 5)
A B
= +
2x − 1 x + 5
A (x + 5) + B (2x − 1)
=
(2x − 1) (x + 5)
 78

and therefore
23 − 2x = A(x + 5) + B(2x − 1).

By setting x = −5, we obtain that B = −3 and by setting x = 21 , we obtain that A = 4. Therefore

Z  
4 3
I = − dx
2x − 1 x + 5
1
= 4. `n |2x − 1| − 3`n |x + 5| + c
2
(2x − 1)2
= `n + c.
|x + 5|3
2. Determine

x2 + 10x + 6
Z
I= dx.
x2 + 2x − 8

Solution
We must first apply long division. After division
x2 + 10x + 6 8x + 14
=1+ 2 .
x2 + 2x − 8 x + 2x − 8
Furthermore
8x + 14 8x + 14
=
x2 + 2x − 8 (x − 2) (x + 4)
A B
= +
x−2 x+4
A (x + 4) + B (x − 2)
= ,
(x − 2) (x + 4)
and therefore
8x + 14 = A(x + 4) + B(x − 2).

By setting x = −4, we obtain that B = 3 and by setting x = 2 we obtain that A = 5.

Therefore
Z  
5 3
I = 1+ + dx
x−2 x+4
= x + 5`n |x − 2| + 3 `n |x + 4| + c.

The integrals that occur most generally after splitting into partial fractions are

Z
dx 1
= `n |ax + b| + c
ax + b a

1 (ax + b)n+1
Z
(ax + b)n dx = + c, n 6= 1.
a n+1
 79 MAT1613/GD001

4.5.2 REPEATED LINEAR FACTORS IN THE DENOMINATOR

Consider
3 2
2 − (x − 1) .
(x − 1)
Following addition

3 2 3 − 2 (x − 1)
2 − x−1 =
(x − 1) (x − 1)2
−2x + 5
= .
(x − 1)2
−2x+5
If the reverse process must be performed, namely to split (x−1)2
into partial fractions, provision must be
made for two fractions, in other words
−2x + 5 A B
2 = 2 + x − 1.
(x − 1) (x − 1)
P (x)
Similarly, for each factor (ax + b)r that occurs in the denominator of a proper rational function Q(x) ,
provision must be made by means of r terms in the decomposition, namely:
A1 A2 Ar
+ + ··· + .
ax + b (ax + b) 2 (ax + b)r

Worked Examples

1. Determine
Z
3x + 1
I= dx.
(x + 1)2

Solution
3x + 1 A B
2 = 2 + x + 1.
(x + 1) (x + 1)
Therefore
3x + 1 = A + B(x + 1).

By setting x = −1 we obtain that A = −2 and by setting x = 0 we obtain that A + B = 1, from

which it follows that B = 3. (We can set x = 0 or use any other convenient value.) Therefore
Z  
−2 3
I = + dx
(x + 1)2 x + 1
Z  
−2 3
= −2 (x + 1) + dx
x+1
−2 (x + 1)−1
= + 3 `n |x + 1| + c
−1
2
= + 3 `n |x + 1| + c.
x+1
 80

2. Determine

x4 − x3 − x − 1
Z
I= dx.
x3 − x2

Solution
The integrand is improper. After division
x4 − x3 − x − 1 x+1
3 2
=x− 3 .
x −x x − x2

Furthermore
x+1 x+1
=
x3 − x2 x2 (x − 1)
A B C
= 2
+ +
x x x−1
A (x − 1) + Bx (x − 1) + Cx2
= .
x2 (x − 1)

Therefore
x + 1 = A(x − 1) + Bx(x − 1) + Cx2 .

By setting x = 1 we obtain that C = 2, by setting x = 0 we obtain that A = −1 and then by setting

x equal to any convenient value, say x = −1, we obtain that B = −2. Therefore
Z   
−1 2 2
I = x− − + dx
x2 x x−1
x2 1
= − + 2 `n |x| − 2 `n |x − 1| + c
2 x
2
 2
x 1 x
= − + `n + c.
2 x x−1

4.5.3 IRREDUCIBLE QUADRATIC FACTORS IN THE DENOMINATOR

For each irreducible factor

ax2 + bx + c where b2 − 4ac < 0

P (x) Ax+B
which occurs in the denominator of Q(x) , a proper fraction ax2 +bx+c
may occur in the decomposition.

Worked Example
Determine

x−1
Z
I= dx.
(x + 1) (x2 + 1)
 81 MAT1613/GD001

Solution
x−1 A Bx + C
= + 2 .
(x + 1) (x2 + 1) x+1 x +1
A x2 + 1 + (Bx + C) (x + 1)


= .
(x + 1) (x2 + 1)
Therefore

x − 1 = A(x2 + 1) + (Bx + C)(x + 1) ...(a)

2
= (A + B)x + (B + C)x + A + C. ...(b)

First set x = −1 in (a). Then it follows that A = −1. Now compare the coefficients of equal powers in
(b):
x2 : 0 = A + B, which implies that B = 1,

and
x : 1 = B + C, which implies that C = 0.

Therefore
Z  
−1 x
I = + dx
x + 1 x2 + 1
Z
1 du
= −`n |x + 1| + , where u = x2 + 1,
2 u
1
= −`n |x + 1| + `n x2 + 1 + c
2
√
2
x +1
= `n + c.
x+1

4.5.4 REPEATED IRREDUCIBLE QUADRATIC FACTORS IN THE DENOMINA-

TOR
As in the case of repeated linear factors in the denominator, for each factor of the form

(ax2 + bx + c)r with b2 − 4ac < 0,

which occurs in the denominator of a proper rational function, provision must be made by means of r
terms in the decomposition, namely:
A1 x + B 1 A2 x + B2 Ar x + Br
+ + ··· + .
ax2 + bx + c (ax2 + bx + c)2 (ax2 + bx + c)r

Worked Example
Determine

5x3 − 3x2 + 7x − 3
Z
I= dx.
(x2 + 1)2
 82

Solution
5x3 − 3x2 + 7x − 3 Ax + B Cx + D
2 = +
(x2 + 1) 2
x +1 (x2 + 1)2
(Ax + B) x2 + 1 + Cx + D


= .
(x2 + 1)2

Therefore
5x3 − 3x2 + 7x − 3 = Ax3 + Bx2 + (A + C)x + (B + D).

By now comparing the coefficients of equal powers of x, we obtain that

x3 : 5 = A,
2
x : −3 = B,
x : 7 = A + C, from which it follows that C = 2.
x0 : −3 = B + D, from which it follows that D = 0.

Therefore
Z  
5x − 3 2x
I = + dx
x2 + 1 (x2 + 1)2
Z Z Z
dx 5 2x 2x
= −3 + dx + dx.
2
x +1 2 2
x +1 (x2 + 1)2

By applying the substitution u = x2 + 1 in the last two integrals on the right, we therefore see that

5x3 − 3x2 + 7x − 3
Z
5 1
2 dx = −3 arctan x + `n(x2 + 1) − 2 + c.
(x2 + 1) 2 x +1

Summary of the Method of Splitting into Partial Fractions

The following is a very short summary of the method of splitting into partial fractions.

1. If the degree of P (x) is not smaller than that of Q(x), use long division to obtain the desired form.

2. Express Q(x) as a product of linear and/or irreducible quadratic factors.

3. (a) For each factor of the form (ax + b) which occurs only once in Q(x), there is a term of the form

A
.
ax + b

(b) For each factor of the form (ax + b), which occurs r times in Q(x), there are r terms of the form

A1 A2 Ar
+ +···+ .
ax + b (ax + b)2 (ax + b)r

(c) For each factor of the form (ax2 + bx + c) with b2 − 4ac < 0, which occurs only once in Q(x),
there is a term of the form
Ax + B
.
ax2 + bx + c
 83 MAT1613/GD001

(d) For each factor of the form (ax2 + bx + c) with b2 − 4ac < 0, which occurs r times in Q(x),
there are r terms of the form

A1 x + B 1 A2 x + B 2 Ar x + Br
+ +···+ .
2
ax + bx + c (ax2 + bx + c) 2 (ax2 + bx + c)r

Now try to do the following exercises yourself:

Exercises

1. Determine
Z
11x + 17
dx.
2x2 + 7x − 4

Solution
5
`n|2x − 1| + 3 `n|x + 4| + c.
2

2. Determine

3x2 − 10
Z
dx.
x2 − 4x + 4

Solution
2
3x − + 12 `n|x − 2| + c.
x−2

3. Determine

x2 − x − 21
Z
dx.
2x3 − x2 + 8x − 4

Solution
3 1 x 5
`n(x2 + 4) + tan−1 − `n|2x − 1| + c.
2 2 2 2

4. Determine

x3 − 4x − 1
Z
dx.
x (x − 1)3

Solution
3 2
`n|x| − + + c.
x − 1 (x − 1)2
 84

5. Determine

5x3 − 3x2 + 2x − 1
Z
dx.
x4 + x2

Solution
1 3
2`n|x| + + `n(x2 + 1) − 2 tan−1 x + c.
x 2
4.6 z-SUBSTITUTION
This method (using z = tan x2 ) reduces the problem of integrating a rational expression in sin x and cos x
to a problem of integrating a rational function of z. This is only used when simpler methods have failed.
Let z = tan x2 , then x = 2 tan−1 z and dx = 1+z
2
2 dz. From the following figure:

2
1+ z
z
_x
2
1

we have
x z x 1
sin =√ and cos = √
2 1+z 2 2 1 + z2
Using the identities

sin 2x = 2 sin x cos x and

cos 2x = cos2 x − sin2 x

we have
x x z 1 2z
sin x = 2 sin cos = 2 √ ·√ =
2 2 1+z 2 1+z 2 1 + z2
and
x x 1 z2
cos x = cos2 − sin2 = 2
−
2 2 1+z 1 + z2
1 − z2
=
1 + z2
We thus have the following three important formulas:

2
dx = 1+z 2
dz

2z
sin x = 1+z 2

1−z 2
cos x = 1+z 2
.

You must know these formulas well. We will now do one example. Note the method for changing the
limits of integration in the example:
 85 MAT1613/GD001

Worked Example
Calculate
Z π/2
dx
I= .
0 sin x + cos x

Solution
Set z = tan x2 , −π < x < π. If x = 0, then z = tan 0 = 0, and if x = π/2, then z = tan π4 = 1.

Therefore
π 2
Z Z 1 dz
2 dx 1+z 2
I = = 2z 2
0 sin x + cos x 0 1+z 2
+ 1−z
1+z 2
Z 1
2
= dz
0 −z 2 + 2z + 1
Z 1
2dz
= ,
0 2 − (z − 1)2
by completion of the square.

Now set
√ π π
z−1= 2 sin θ, with − ≤θ≤ ,
2 2
then
√
dz = 2 cos θ dθ.
√  
Furthermore, if z = 0, then 2 sin θ = −1, in other words θ = arcsin − √12 = − π4 , and if z = 1, then
sin θ = 0, in other words θ = 0.
Therefore
Z π/2 Z 0 √
dx 2 cos θ
I = =2 dθ
0 sin x + cos x −π/4 2 cos2 θ
√ Z 0
= 2 sec θdθ
−π/4
√
= 2`n |sec θ + tan θ|]0−π/4
√  √ 
= 2 `n (1 + 0) − `n 2−1
√ √ 
= − 2`n 2−1 .

Exercises:
Determine the following integrals by using z–substitution:
Z
dx
1.
1 + sin x
Z π
2 dx
2.
0 1 + sin x + cos x
 86

Z π
3 dx
3.
0 3 + 2 cos x
Z
dx
4.
sin x − cos x
Z π
2 dx
5.
π 1 − cos x
3

Z 2π
3 cos xdx
6.
π sin x cos x + sin x
2

Z
cos xdx
7.
1 − cos x

Answers:
−2
1. +c
tan x2 + 1
2. ln (2)
 
3. √2
5
tan−1 √1
15
 √ 
√1
(tan x )+1− 2
4. ln tan x2 +1+√2 +c
2 ( 2)
√
5. −1 + 3
1
√
6. 2 ln 3 − 21

− cot x2 − x + c


7.

Summary (of the Techniques of Integration)

At this point we have concluded the integration techniques. Whenever you have to determine a given
integral, you must first decide which method to use. Always keep in mind that there may be more than
one correct method to determine an integral. Sometimes the one method could involve simpler calculations
R 3/2 (x−1)dx
than the other. Try, for example, to evaluate 1 √ 2x−x2
by (i) the substitution u = 2x − x2 and (ii)
√
2− 3
completing the square and trigonometric substitution. (The answer is 2 .)

Remember, there are only two general methods of integration:

I. Integration by substitution and

II. Integration by parts.

All other methods constitute algebraic manipulation followed by the application of one of the above-
mentioned two methods.
 87 MAT1613/GD001

1. When you now have a given integral, first try to spot a suitable substitution that will transform
the integrand into a simpler form or standard form. Here you need to take particular note of the
following:

1.1 Always determine first whether the integrand does not perhaps consist of the product of two
functions one of which is the derivative of the other. (Then you can apply the method of
substitution.)
1.2 Integrals in which trigonometric substitutions are employed are usually easily recognisable.

2. Integration by parts: This method, as well as the circumstances in which it should be used was
discussed in paragraph 4.2(S).

3. You also need to keep in mind a number of special methods:

3.1 For trigonometric integrals you may try:

(i) Using the splitting method or by using identities, you can rewrite the integrand. (See the
worked examples in 4.3(S)).
(ii) By using the substitution z = tan x2 , where a rational function of sin x and cos x occurs,
R cos t
you obtain a rational function of z. (See p. 4.6(S)). 1−cos t dt is such an example.
√
(iii) By using identities, you can rewrite a function to get rid of a square root (e.g. 1 + cos 2θ =
√ √
2 cos2 θ = 2 cos θ since cos2 θ = 1+cos2
2Θ
).
(iv) As a last resort, you can rewrite all trigonometric functions in terms of cos and sin.
3.2 For integrals where rational functions play a part, you should keep the following in mind:
(i) Always use division if the integrand is an improper fraction as in the worked example 2 in
4.5.2(S).
(ii) If possible, factorise the denominator and then decompose the integrand into rational frac-
tions.
(iii) When you are confronted with an irreducible quadratic function (or the square root of a
quadratic function), try completion of the square.

Summary
The learner should now be able to identify which method to use for a specific problem and he/she must
also be able to apply all the integration methods.
 88

Chapter 5

The Improper Integral

Introduction
Up to now, we have required our definite integrals to have two properties. First, that the domain of
integration, from a to b must be finite and second that the range of the integrand must be finite on this
domain. In practice, however, we frequently encounter problems that fail to meet one or both of these
conditions. These integrals are called improper integrals. Thus in this chapter we are concerned with three
types of integrals, namely where the integrand is unbounded, the interval of integration is unbounded, or
both. We determine if such an integral converges or diverges and if possible, we evaluate the integral.
It is very important that you must use limit(s) whenever you have an improper integral

Outcomes
After studying this chapter the learner should be able to:

• evaluate improper integrals (i.e integrals with infinite limits of integration and integrals that become
infinite at a point within the interval of integration).

• use the limit comparison test to determine whether an integral converges or diverges.

We are concerned with three types of integrals:

5.1 integrals of which the integrand is unbounded over the integration interval

5.2 integrals of which the integration interval is unbounded

5.3 integrals for which the integrand is unbounded and the integration interval is unbounded.

We will discuss here the three types of improper integrals. Please study the relevant sections in your
prescribed textbook.

5.1 INTEGRALS OF WHICH THE INTEGRAND IS UNBOUNDED OVER THE IN-

TEGRATION
INTERVAL
 89 MAT1613/GD001

R3
For example, 0 (x − 2)−1 dx is unbounded in x = 2.
Rb
Consider the integral a f (x)dx. Suppose f (x) becomes infinitely large at the point c ∈ [a, b]. Then we
will investigate the following:

Z t Z b
lim f (x)dx and lim f (x)dx.
t→c− a u→c+ u

Rb
If both these limits exist and they are finite, we say that a f (x)dx converges and

Z b Z t Z b
f (x)dx = lim f (x)dx + lim f (x)dx.
a t→c− a u→c+ u

If we write the given example into this notation, we obtain

Z 3 Z t Z 3
−1 −1
(x − 2) dx = lim (x − 2) dx + lim (x − 2)−1 dx.
0 t→2− 0 u→2+ u

Worked Example
Determine whether
Z 3
dx
0 (x − 1)2/3

converges and, if so, calculate the integral.

Solution
1
f (x) = becomes infinitely large at the point x = 1. Therefore we consider the following two
(x−1)2/3
integrals: Z a Z 3
dx dx
2/3
and 2/3
0 (x − 1) b (x − 1)

and we investigate the following:

Z a
dx h
1/3
ia
lim = lim 3 (x − 1)
a→1− 0 (x − 1)2/3 a→1− 0
h i
= lim 3 (a − 1)1/3 − 3 (0 − 1)1/3
a→1−
= 3,

as well as
Z 3
dx h i3
lim = lim 3 (x − 1)1/3
b→1+ b (x − 1)2/3 b→1+ b
h i
= lim 3 (3 − 1)1/3 − 3 (b − 1)1/3
b→1+
√3
= 3 2.
 90

Both limits exist, therefore the integral converges and has the value
 √3

3+3 2 .

5.2 INTEGRALS OF WHICH THE INTEGRATION INTERVAL IS UNBOUNDED

In this case we again use limits to determine whether the integrals are convergent or divergent and,
provided the integral is convergent, we can also determine the value of the integral. The three cases which
we may find, are the following:

Z ∞ Z b
• f (x)dx = lim f (x)dx.
a b→∞ a

Z a Z a
• f (x)dx = lim f (x)dx.
−∞ b→−∞ b

Z ∞ Z 0 Rb
• f (x)dx = lim f (x)dx + lim f (x)dx.
−∞ a→−∞ a b→∞ 0

Worked Example
Ascertain whether
Z ∞
dx
2 (x − 1)2

converges and, if so, calculate the integral.

Solution
Z ∞ Z a
dx dx
= lim
2 (x − 1)2 a→∞ 2 (x − 1)2

−1 a
 
= lim
a→∞ x − 1
2
 
−1 1
= lim +
a→∞ a − 1 2−1
= 0+1
= 1.

Therefore the integral converges and has the value 1.

 91 MAT1613/GD001

5.3 INTEGRALS FOR WHICH BOTH (5.1(S)) AND (5.2(S)) HOLD

We start with a worked example.

Worked Example
Determine whether
Z ∞
1
√ dx
0 x

converges and, if so, calculate the integral.

Solution
This integral has both an unbounded integrand over the integration interval, and an unbounded integration
interval. In this case, we express the integral as the sum of improper integrals, each of which is of one of
the types mentioned above.
Consider Z ∞
1
√ dx.
0 x
The integrand is unbounded at 0, therefore we choose any number greater than 0, for example 1, and we
write Z ∞ Z 1 Z ∞
1 1 1
√ dx = √ dx + √ dx. (*)
0 x 0 x 1 x
The first integral on the right-hand side of (*) is of type (5.1(S)) and therefore we do it as follows:
Z 1 Z 1
1 1
√ dx = √ dx
lim
0 x b→0+ b x
 √ 1
= lim 2 x b
b→0+
h √ i
= lim 2 − 2 b
b→0+
= 2.

The second integral on the right-hand side of (*) is of type (5.2(S)), and therefore we do it as follows:
Z ∞ Z b
1 1
√ dx = √ dx
lim
1 x x
b→∞ 1
 √ b
= lim 2 x 1
b→∞
h √ i
= lim 2 b − 2 .
b→∞
R∞
It is clear that this integral diverges and thus √1 dx also diverges.
0 x

You must be very careful when determining an improper integral. Before you determine any integral,
first determine whether the integrand and the integration interval are bounded. If that is not the case,
you must use the method for improper integrals.
 92

Extra Worked Example

Determine
Z 3
dx
.
0 (1 − x)2

If you do not consider the fact that the integrand is unbounded at x = 1, you will get the wrong answer of
− 32 . If you take this as an improper integral you will get the correct answer, i.e., that the integral diverges.

Solution Z 3 Z t Z 3
dx 1 1
= lim dx + lim dx (*)
0 (1 − x)2 t→1− 0 (1 − x)2 u→1+ u (1 − x)2
Let v = 1 − x then
dv
= −1
dx
so that dx = (−1) dv.

If x = 0 then v = 1
If x = t then v = 1 − t
If x = u then v = 1 − u
If x = 3 then v = −2.

Then (*) becomes

Z 3 Z 1−t
dx
= (−1) lim v −2 dv
0 (1 − x)2 t→1− 1
Z −2
+ (−1) lim v −2 dv
u→1 +
1−u
    −2 
−1 1−t
= lim v 1 + lim v −1 1−u
t→1− u→1+
h i
= lim (1 − t)−1 − 1
t→1−
 
1 −1
+ lim − − (1 − u) .
u→1+ 2

Now  
1
lim −1 = ∞.
t→1− 1−t
Also  
1 1
lim − − = ∞.
u→1+ 2 (1 − u)
Thus this integral diverges.

5.4 THE LIMIT COMPARISON TEST

You have to study the relevant section in your prescribed textbook.

Summary
 93 MAT1613/GD001

The improper integral can sometimes be evaluated directly using the methods specified in paragraph
5.1(S), 5.2(S) and 5.3(S). However, when an improper integral cannot be evaluated directly we use the
comparison test. The idea of the comparison test is to compare a given improper integral to another
improper integral whose convergence or divergence is already known.
 94

Chapter 6

Infinite Sequences and the Taylor Series

Introduction
Informally, a sequence is any list (or ordered collection) of things, but in this chapter the things will usually
be numbers.

It is also explained here what is meant by convergence and divergence of a sequence. We also develop one of
the most remarkable formulas in all of mathematics, a formula that enables us to express many
functions as “infinite polynomials”. In addition to providing effective polynomial approximations
of differentiable functions, these infinite polynomials (called power series) have many uses. For instance,
they provide an efficient way to evaluate non–elementary integrals and they solve differential equations
that give insight into vibration, chemical diffusion and signal transmission.

Outcomes
After studying this chapter the learner should be able to:

• use limit techniques to determine the limit of a sequence.

• write out the Taylor series generated by f at a.

• obtain the Taylor polynomial Pn (x) of any given order n generated by f at a point, where f is, for
example of the form f (x) = ex , f (x) = ln x.

6.1 LIMITS OF INFINITE SEQUENCES

Study this section in your textbook.

6.2 TAYLOR SERIES

Please note that you don’t need to be able to find the remainder for a Taylor series. This
will be done in the second year.

In the previous section, the concept of an infinite sequence was discussed. When the terms of such a
sequence are added, we obtain an infinite series, one of which we discuss in more detail in this section.
 95 MAT1613/GD001

The values of polynomial functions can easily be calculated in a given point. If, for example,

f (x) = a0 + a1 x + a2 x2 + · · · + an xn ,

we can easily calculate f (r) for any number r. However, at this stage we do not yet have a method of
calculating the values of functions such as `n x, sin x, ex , etc. In this section, we become acquainted with
infinite series. We convert functions such as `n x, sin x, ex , etc., to infinite series, which then reduces the
approximate determination of the value of such functions in a given point to the calculation of a polynomial.
Computers, for example, also employ this method for calculating the above-mentioned functions.

We will therefore try to reduce the evaluation of f (x) for certain functions f to the computation of the
function values of polynomial functions. The method entails the determination of polynomial functions
that are good approximations of f on an interval that is contained in the domain of f.

To begin with, assume that f is differentiable at the point A = (a, f (a)). We expect that the tangent to
the graph of f at A provides us with a good approximation of f in the immediate neighbourhood of A.
This tangent is the graph of a polynomial function of the first-degree that passes through A and possesses
the same derivative as f at A.
y T1
T2
f

f (a )

x
0 a

The equation of the tangent is

P1 (x) = f (a) + f 0 (a)(x − a).
We cannot expect that the tangent at A will continue to yield good approximations of the function values of
f for values of x further away from a. On the other hand, we can however try to get a better approximation
of f by means of a second-degree polynomial. Suppose f possesses first and second derivatives at A. Then
we choose this second-degree polynomial in such a way that its graph passes through A = (a, f (a)) and,
furthermore, is such that its first and second derivatives at A coincide with those of f . Suppose this
polynomial is
P2 (x) = a0 + a1 (x − a) + a2 (x − a)2 .
Then
P2 (x) = a0 + a1 (x − a) + a2 (x − a)2 , and hence P2 (a) = a0 ;
P20 (x) = a1 + 2a2 (x − a), and hence P20 (a) = a1 ;
P200 (x) = 2a2 , and hence P200 (a) = 2a2 .
 96

Now we choose a0 , a1 and a2 , such that

P2 (a) = f (a), in other words, we choose a0 = f (a);

P20 (a) = f 0 (a), in other words, we choose a1 = f 0 (a);
P200 (a) = f 00 (a), in other words, we choose a2 = 12 f 00 (a).

Therefore the polynomial is

1
P2 (x) = f (a) + f 0 (a)(x − a) + f 00 (a)(x − a)2 .
2

We can now proceed in this manner and find higher degree polynomials of which we hope that they will
be still better approximations of f . Then we find that the polynomial Pn (x) is as follows:

Polynomial Pn (x)

f 00 (a) f (n) (a)

Pn (x) = f (a) + f 0 (a)(x − a) + 2! (x − a)2 + . . . + n! (x − a)n .

The polynomial Pn (x) is called the nth degree Taylor polynomial of f at the point a. When a = 0, the
polynomial is said to be the nth degree Maclaurin polynomial of f .

When working with infinite series, the convergence of such series must be considered. At this stage we
will discuss the convergence of series – this will only be done in your second year Mathematics course.
We now show the approximations for `n(1 + x) and sin x, respectively, at a = 0 where P1 (x), P2 (x) and
P3 (x) were obtained from the method above.

y P1 ( x ) = x

f ( x )= ln (1+ x)
P3 ( x )

x
0 _ x_2 + _x3
P3 ( x ) = x 2 3
P2 ( x )
2
P2 ( x ) = x _ _x2
P1 ( x ) f ( x )= ln (1+ x)
 97 MAT1613/GD001

y P1 ( x ) = x

_ _x3
P3 ( x ) = x 3!
f ( x ) = sin x

x
0
f ( x ) = sin x
_ _x3
P3 ( x ) = x 3!

Worked Example

Determine the Taylor polynomial of order 3 generated

√
by f (x) = x about the point a = 4.

Solution
First obtain the derivatives f 0 (x) , f 00 (x) and f 000 (x) and then the values in the point 4.
1
f 22 = 2


f (x) = x 2

1
f 0 (x) = 21 x− 2 f 0 22 = 1


4

3
f 00 (x) = − 41 x− 2 f 00 22 = − 32
1

5 000
f 000 (x) = 38 x− 2 22 = 3


f 256

Write out the Taylor polynomial of order 3 and substitute the values above:

(x − 4) f 0 (4) (x − 4)2 f 00 (4) (x − 4)3 f 000 (4)

P3 (x) = f (4) + + +
1! 2! 3!
1 1 3
= 2 + (x − 4) − (x − 4)2 + (x − 4)3
4 32 × 2 256 × 3 × 2
1 1 1
= 2 + (x − 4) − (x − 4)2 + (x − 4)3 .
4 64 512
 98

Exercises:
Find the Taylor polynomial of any order (say 3) or of order n for as many functions as possible. See the
exercises in your prescribed textbook. Remember, you do not need to be able to find the remainder for
the Taylor series. This will be done in the second year.
Also, try the following:

1. Expand the given polynomial f (x) = 10 − 20x + 15x2 − 4x3 as a polynomial about the point 1 (or
as a polynomial in (x − 1)).
1
2. If f (x) = 1−x , determine P5 (x) about the point 0.

Answers

1. 1 − 2(x − 1) + 3(x − 1)2 − 4(x − 1)3 .

2. 1 + x + x2 + x3 + x4 + x5 .

Summary
The learner should now have an intuitive understanding of sequences and be able to determine the limit
of a sequence when it exists. The learner should also be able to demonstrate a basic understanding of the
Taylor series and how to construct Taylor polynomials for certain functions.