MT1001 Introductory Mathematics

Integration Lecture Notes 1

Tom Coleman

November 7, 2018

1 This work is licensed using a CC BY-NC-SA 4.0 license.

Contents

1 Definite integrals 3
1.1 What is integration? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 Definite integrals as areas . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.3 Properties of definite integrals . . . . . . . . . . . . . . . . . . . . . . . . 9

2 The Fundamental Theorem Of Calculus 12

2.1 From definite to indefinite . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.2 The theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.3 Some antiderivatives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.4 Summary of first two chapters . . . . . . . . . . . . . . . . . . . . . . . . 21

3 Integration by substitution 22
3.1 Techniques of calculus . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
3.2 How integration by substitution works . . . . . . . . . . . . . . . . . . . . 23
3.2.1 Method for integration by substitution . . . . . . . . . . . . . . . 24
3.3 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

4 Trigonometry and integration 31

4.1 Previously in trigonometry... . . . . . . . . . . . . . . . . . . . . . . . . . 31
4.2 Some more antiderivatives . . . . . . . . . . . . . . . . . . . . . . . . . . 32
4.3 Using trigonometric identities . . . . . . . . . . . . . . . . . . . . . . . . 35
4.3.1 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

1
 4.4 Inverse functions: a different kind of substitution . . . . . . . . . . . . . . 39
4.4.1 Method for integration by trigonometric substitution . . . . . . . . 42
4.4.2 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42

5 Integration by parts 48
5.1 What is integration by parts? . . . . . . . . . . . . . . . . . . . . . . . . 48
5.2 How integration by parts works . . . . . . . . . . . . . . . . . . . . . . . 48
5.2.1 How to use integration by parts effectively . . . . . . . . . . . . . 50
5.3 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

6 Integration using partial fractions 62

6.1 What are partial fractions? . . . . . . . . . . . . . . . . . . . . . . . . . . 62
6.2 Rational functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
6.3 Integration using partial fractions . . . . . . . . . . . . . . . . . . . . . . 67
6.3.1 And finally... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74

2
Chapter 1

Definite integrals

1.1 What is integration?

You may have seen that differential calculus is the study of the rate of change of quantities.
In differential calculus, the derivative f 0 (x) of a function f (x) is a function that measures
the rate of change of f (x). The derivative of a function has an application in geometry:
evaluating f 0 (x) at a point a allows you to find the gradient of the tangent to the curve of
f (x) at a.
This part of the course is concerned with integration of functions. There are two types of
integration on a function f (x):

• the indefinite integral of f (x) with respect to x, written as

Z
f (x) dx

• the definite integral of f (x) between limits a and b with respect to x, written as
Z b
f (x) dx
a

(It will be explained what this notation means shortly!)

The difference between the two types of integral is given by their uses. The indefinite integral
is precisely the reverse process of differentiation; this result is known as the Fundamental
Theorem of Calculus and is covered in Chapter 2 of the course. The definite integral

3
 function derivative w.r.t x
f (x) = a f 0 (x) = 0
f (x) = axn f 0 (x) = naxn−1
f (x) = a sin(kx) f 0 (x) = ka cos(kx)
f (x) = a cos(kx) f 0 (x) = −ka sin(kx)
f (x) = aekx f 0 (x) = akekx
f (x) = a ln(kx) f 0 (x) = a
x

Table 1.1: Some derivatives, with both a, k constants

represents (amongst other things) the signed area bounded by the curve f (x) and the lines
x = a and x = b.
Before learning about how integration works, you should make sure that you know (and can
find) the derivatives of the functions given in Table 1.1. Here, a, k are both constants.

1.2 Definite integrals as areas

Let’s begin with an initial definition of a definite integral.

Definition 1.2.1. Let f (x) be a positive (so for all a ≤ x ≤ b, then f (x) ≥ 0) and
continuous (informally, f can be drawn without taking your pen off the paper) function,
defined on the interval [a, b]. Let A be the area bounded by f (x), the lines x = a and x = b
and the x-axis (see Figure 1.1). Then A equals the definite integral of f (x) between the
limits a and b with respect to x; in symbols,
Z b
A= f (x) dx
a

Remark. Here, it’s very important to note that this is only true if the function is positive.
This condition can be removed, but the definition is not the same: see Definition 1.2.4.

Terminology. There are four pieces of notation that make up a definite integral, and each
of them have different names.

• The symbol is called the integral sign. This is adapted from a letter S for sum;
R

as you will see later, an integral is actually a sum of an infinite amount of terms.

4
 f (x)

x
a b

Figure 1.1: The area of the shaded region is equal to the integral
Rb
a f (x) dx.

• The value a is called the lower limit and the value b is known as the upper limit.

• The function f (x) is called the integrand.

• The dx denotes the variable you are integrating with respect to. Here, the variable x
doesn’t have to be called x; for instance it can be called t or u. The important thing
to remember is that the integral doesn’t change its value if you change the name of
the variable throughout.

It is important to remember that in every integral, f (x) and dx are multiplied together.
In every definite integral you write, you should include every one of these pieces of notation.

Example 1.2.2. Consider the area bounded by f (x) = 2, the x-axis, and the lines x = 1
and x = 3 (see Figure 1.2).

f (x)

x
1 3

Figure 1.2: The area of the shaded region is equal to the integral
R3
1 2 dx

As the shaded region is a square with side length 2, the area of this region is A = 2 · 2 = 4.
You can use Definition 1.2.1 to say that
Z 3
2 dx = 4
1

5
Example 1.2.3. You are given the area bounded by f (x) = x, the x-axis, and the lines
x = 0 and x = 2 (see Figure 1.2).

f (x)

x
0 2

Figure 1.3: The area of the shaded region is equal to the integral
R2
0 x dx.

As the shaded region is a triangle with base and height both 2, the area is A = (2·2)/2 = 2.
You can use Definition 1.2.1 to say that
Z 2
x dx = 2
0

However, evaluating areas gets more complicated if your curve is not a straight line. For
instance, how would you work out the area bounded by f (x) = x2 , the limits x = 0 and
x = 1 and the x-axis, as seen in Figure 1.4? In other words, what is the value of the definite
integral of f (x) = x2 between 0 and 1 with respect to x?

f (x)

x
0 1

Figure 1.4: What is

R1 2
0 x dx?

The idea is to approximate the area underneath the curve f (x) = x2 by dividing it into n
rectangular strips, each of width 1/n. So for i = 1, . . . , n, the ith strip is a rectangle with
width 1/n and height given by f ((i − 1)/n) = [(i − 1)/n]2 . This means that the area of
the ith strip is (1/n) · [(i − 1)/n]2 (see Figure 1.5).

6
 f (x)

x
0 1/n 2/n 3/n 4/n · · · (n − 1)/n 1

Figure 1.5: Approximating with n strips.

R1 2
0x dx

You can add all of these strips together to approximate the area underneath the curve. Here,
the sum of all the ith strips for i = 1, . . . , n is given by

1 02 1 12 1 22 1 (n − 1)2
S= · 2 + · 2 + · 2 +... + · 2
|n {zn } |n {zn } |n {zn } |n {zn }
i=1 i=2 i=3 i=n

and you can factorise to get

1  2 2 2

S= 1 + 2 + . . . + (n − 1)
n3

The problem is that this is just an approximation. In order to find a value for the definite
integral 01 x2 dx, you need to divide the area into as many strips as possible. This
R

is achieved by letting the number of strips n tend to infinity. When you do this, the
approximation S gets closer and closer to the true value of 01 x2 dx. You can write this
R

mathematically by
Z 1
1  2 
x2 dx = lim 1 + 2 2
+ . . . + (n − 1) 2
0 n→∞ n3

You can work this limit out by using the result

k
k(k + 1)(2k + 1)
r2 = 12 + 22 + . . . + k 2 =
X

r=1 6

7
By setting k = (n − 1) in this result, you can see that
Z 1
1 (n − 1)(n)(2n − 1)
x2 dx = n→∞
lim
0 n3 6

You can then reduce this in the same way you would any limit to get
Z 1
(1 − 1/n)(2 − 1/n) 1
x2 dx = n→∞
lim =
0 6 3

However, this is just one example. It would be extremely useful to work out the definite
integral of any positive, continuous function f (x) between any limits a and b. This can
be done like the process above: by dividing the interval [a, b] into n strips, working out the
area of each strip, adding up the area of all the strips and then letting n tend to infinity
and working out the limit.

Definition 1.2.4. Let f (x) be a positive and continuous function, defined on the interval
[a, b]. Set x0 = a and xk = a + k(b − a)/n for all k = 1, 2, . . . , n. Then the definite
integral of f (x) between the limits a and b with respect to x is given by
n
Z b
b−a X
f (x) dx = lim f (xk )
a n→∞ n k=1

where

• (b − a)/n is the width of every strip;

• f (xk ) is the height of the kth strip (the strip between xk and xk+1 ), and;

• f (xk ) is the sum of the areas of all n strips.

b−a Pn
n k=1

See Figure 1.6 for a diagram.

Example 1.2.5. You are asked to evaluate the definite integral

R1 x
0e dx.
Using the fact that a = 0 and b = 1 and f (x) = ex , and setting xk = k/n, the sum in
Definition 1.2.4 becomes n
1X
S= ek/n
n k=1
This is a geometric series. Using the formula for the sum of a geometric series

rn − 1
r + r2 + ... + rn = r
r−1

8
 f (x)

x
a x1 x2 x3 x4 ··· xn−1 b

Figure 1.6: Partitioning the interval [a, b] into n pieces to find the definite integral
dx. Not a very good approximation!
Rb
a f (x)

with r = e1/n you can obtain

1 1/n e − 1
e S=
n e1/n − 1
The goal now is to work out the limit of this expression as n tends to infinity. Here, you
can get
Z 1
e1/n e − 1
ex dx = lim S = lim
0 n→∞ n→∞ n e1/n − 1

You can take (e − 1) out of the limit and rearrange and, using the expression
 
lim n e1/n − 1 = 1
n→∞

gives
Z 1
e1/n
ex dx = (e − 1) n→∞
lim =e−1
0 n (e1/n − 1)

1.3 Properties of definite integrals

You can use Definition 1.2.4 and the properties of sums and limits to show the following
properties of definite integrals.
Rb
Proposition 1.3.1. Let a f (x) dx be a definite integral. Then the following properties
hold:

(1) For c such that a < c < b, then:

Z b Z c Z b
f (x) dx = f (x) dx + f (x) dx
a a c

9
(2) If k is some constant, then:
Z b Z b
kf (x) dx = k f (x) dx
a a

(3) For a function g(x) where the definite integral exists on the interval [a, b]:
Z b Z b Z b
f (x) + g(x) dx = f (x) dx + g(x) dx
a a a

(4) If f (x) ≤ g(x) for all x between a and b, then:

Z b Z b
f (x) dx ≤ g(x) dx
a a

(5) If a is some real number, then:

Z a
f (x) dx = 0
a

(6) Finally if a < b, then:

Z b Z a
f (x) dx = − f (x) dx
a b

Using properties (1) and (2) of Proposition 1.3.1, you can define a definite integral for
functions that are negative. For instance, let f (x) be a function where f (x) ≥ 0 for all x
between a and c, and f (x) ≤ 0 for all x between c and b.
You can use property (1) of Proposition 1.3.1 to get:
Z b Z c Z b
f (x) dx = f (x) dx + f (x) dx.
a a c

As f (x) ≤ 0, it follows that −|f (x)| ≤ 0 between c and b. So using property (2) gives:
Z b Z c Z b
f (x) dx = f (x) dx + −|f (x)| dx
a a c
Z c Z b
= f (x) dx − |f (x)| dx
a c

This shows that the definite integral is the area above the axis minus the area below the
axis. So regions where f (x) < 0 correspond to negative areas (see Figure 1.7).

Example 1.3.2. Let f (x) = sin(x) be defined between −π and π. Using property (1) of

10
 f (x)

b
x
a c

Figure 1.7: The curve f (x) crosses the x-axis at x = c, so the shaded region below the
x-axis has a negative area.

Proposition 1.3.1 gives:

Z π Z 0 Z π
sin(x) dx = sin(x) dx + sin(x) dx (1.1)
−π −π 0

As sin(x) is an odd function, then sin(x) = − sin(x). You can use this, along with property
(2) of Proposition 1.3.1 to see that
Z 0 Z 0 Z 0
sin(x) dx = − sin(−x) dx = − sin(−x) dx
−π −π −π

Here, −x is contained in the interval −π < −x < 0. Multiplying through by −1 gives

π > x > 0. So Z 0 Z π
− sin(−x) dx = − sin(x) dx
−π 0

Putting this in to Equation 1.1 gives

Z π Z π Z π
sin(x) dx = − sin(x) dx + sin(x) dx = 0
−π 0 0

This result is illustrated in Figure 1.8; you can see that the areas are equal in size, but on
opposite sides of the x-axis.

f (x)

−π
x
π

Figure 1.8:
Rπ
−π sin(x) dx

11
Chapter 2

The Fundamental Theorem Of

Calculus

2.1 From definite to indefinite

You may have wondered why the term ‘definite’ has been attached to integrals throughout
Chapter 1. This is because the definite integral ab f (x) dx is an expression of a definite
R

signed area bounded by f (x), the x axis, the lower limit a and the upper limit b.
Now suppose instead of setting an upper limit b, this limit is allowed to vary. So here you
can write: Z x
F (x) = f (t) dt
a

It is important to say that the variable in the integral has changed to t; this doesn’t affect
the value of the integral as mentioned in the Terminology section after Definition 1.2.1.
The function F (x) represents the change of area expressed by the integral when x is allowed
to change. This means that the area bounded by f (t), the t-axis, the lower limit a and
the upper limit x changes as x changes. In fact, you could say this area is not definite.
Therefore, F (x) is known as the indefinite integral of f .
In this chapter, the Fundamental Theorem of Calculus (Theorem 2.2.2) says that F (x) =
a f (t) dt is the antiderivative of f (x); that is, a function whose derivative is f (x). This
Rx

will show that integration is a reverse process of differentiation, and that you can evaluate
definite integrals ab f (x) dx by considering the antiderivative of the integrand f (x) and
R

calculating the difference of the antiderivative at the two limits a and b.

12
2.2 The theorem

This section aims to show that the function F (x) as stated above acts as the antiderivative
of f (x). You have seen from the part of the course on differentiation that the derivative of
a function is defined to be

f (x + h) − f (x)
f 0 (x) = lim
h→0 h

The aim is to demonstrate that F 0 (x) = f (x) using this definition of derivative. So take
Z x
F (x) = f (t) dt
a

as above, representing the area under the curve f (t) between the lines t = a and t = x. It
follows that Z x+h
F (x + h) = f (t) dt
a

(see Figure 2.1 for a picture).

f (x)

x
a x x+h

Figure 2.1: An illustration of the areas f (t) dt and

Rx R x+h
a a f (t) dt.

By Proposition 1.3.1 (1), you can write:

Z x+h Z x Z x+h Z x+h
F (x + h) = f (t) dt = f (t) dt + f (t) dt = F (x) + f (t) dt
a a x x

and so Z x+h
F (x + h) − F (x) = f (t) dt
x

The idea now is to ‘bound’ the area underneath f (t) between x and x + h between two
areas. To do this, set the minimum value of f (t) between x and x + h as m, and the similar
maximum value by M . The minimum area of xx+h f (t) dt is given by the area of rectangle
R

with side lengths h and m; and the maximum area is given by the area of the rectangle with

13
side lengths h and M . You can see a diagram of this in Figure 2.2.

f (x)

M
m

x
x x+h

Figure 2.2: Describing the rectangles hm (red rectangles,R minimum area) and hM (blue and
red rectangles together, maximum area). The value of xx+h f (t) dt is somewhere between
the two.

So you can write Z x+h

hm ≤ f (t) dt ≤ hM.
x

Dividing through each term by h gives

F (x + h) − F (x)
m≤ ≤ M.
h

As f (t) is continuous, when h tends towards 0 both the upper and the lower bounds tend
to the value f (x). So, letting h → 0 gives

F (x + h) − F (x)
f (x) ≤ lim ≤ f (x)
h→0 h

and so
F (x + h) − F (x)
f (x) = lim = F 0 (x)
h→0 h
This statement is the first of the two parts of the Fundamental Theorem of Calculus. Here
is an example of how it can be used, illustrating the way forward for a proof of the second
part.

Example 2.2.1. You are given the function

Z x
F (x) = et dt (2.1)
0

Using the above result, you can write

F 0 (x) = ex

14
The idea now is to find F (x) such that F 0 (x) = ex . You can look at Table 1.1 to see that
the derivative of ex is ex , which is on the right hand side. However, you can also see in
Table 1.1 that the derivative of a constant C is 0. So when you are finding F (x) such that
F 0 (x) = ex , you need to be careful to include a constant term! So you can take

F (x) = ex + C

as a function whose derivative is ex .

You can substitute this into Equation 2.1 to get
Z x
ex + C = et dt
0

You now have enough information to find the constant C. You can set x = 0 to see that
Z 0
0
e +C =1+C = et dt = 0
0

and so C = −1, giving F (x) = ex − 1. Finally, evaluating F (x) at x = 1 gives:

Z 1
F (1) = et dt = e − 1
0

which is exactly the same result as you found in Example 1.2.5.

As has just been shown, the integral

Z x
F (x) = f (t) dt
a

represents an antiderivative of f (x). Following the discussion in Example 2.2.1, this an-
tiderivative is not unique. In fact, F (x) + C, where C is some constant, is also an
antiderivative of f (x), as

d
(F (x) + C) = F 0 (x) + 0 = f (x)
dx

So you can write that Z x

F (x) + C = f (t) dt
a

To work out C, you can evaluate this function at x = a to get that:

Z a
F (a) + C = f (t) dt = 0
a

15
and so C = −F (a). Therefore F (x) − F (a) = f (t) dt; you can set x = b to get
Rx
a

Z b
f (t) dt = F (b) − F (a)
a

This is the second part of the Fundamental Theorem of Calculus.

Theorem 2.2.2. [Fundamental Theorem of Calculus]

Let f (x) be a continuous function on the interval [a, b]. For x between a and b define F (x)
by Z x
F (x) = f (t) dt
a

Then F (x) is differentiable on [a, b] with derivative F 0 (x) = f (x). Furthermore,

Z b
f (t) dt = F (b) − F (a)
a

Both parts of the theorem are important:

• The first part says that finding the antiderivative of a function is a reverse process of
differentiation.

• The second part says that you can find a definite integral by considering antiderivatives,
rather than working through limits.

The Fundamental Theorem of Calculus also outlines the difference between definite and
indefinite integrals. To evaluate a definite integral ab f (x) dx, the first (and most important)
R

step is to find an antiderivative F (x) of f (x), and the second involves finding the difference
F (b) − F (a). The first of these steps does not involve the limits a and b in any way. So
here, you can use the notation Z
F (x) = f (x) dx

to denote antiderivatives of f (x). As stated in Section 2.1, this is the indefinite integral
of f (x).
There are two important differences between finding definite and indefinite integrals.

• When you work out a indefinite integral f (x) dx, your final answer should be
R

a function plus a constant of integration C. This expression represents all the

possible antiderivatives of f (x). You should always remember to add a +C at the
very end when finding an indefinite integral.

16
 • When you work out a definite integral f (x) dx, you could follow this method:
Rb
a

Step 1: Find an antiderivative F (x) of f (x). You do not need a +C.

Step 2: The result F (x) can be put into square brackets, with the upper limit at the top
right and the lower limit at the bottom right. So you can write:
Z b
f (x) dx = [F (x)]ba
a

Step 3: Work out F (x) at x = b and then work out F (x) at x = a. You can then
subtract F (a) from F (b) to get
Z b
f (x) dx = [F (x)]ba = F (b) − F (a)
a

The number at the end is the answer.

Your final answer should be a number. This quantity represents the signed area
underneath the curve of f (x) between the lines x = a and x = b. There should be
no variables in your answer, and you do not need a +C at the end.

2.3 Some antiderivatives

This section details some examples of finding antiderivatives. If you are unsure about
derivatives of functions here, you could take the opportunity to re-familiarise yourself with
Table 1.1 before continuing.
The idea of Table 1.1 is to provide a list of derivatives for common functions. As indef-
inite integration is the reverse process of differentiation, it makes sense to write a list of
antiderivatives.
For instance, suppose you wanted to find the antiderivative of the function f (x) = axn ,
where a is a constant and n 6= −1. You can use the fact that integration is the opposite of
differentiation to say that the solution will be a function F (x) such that

d
(F (x)) = axn
dx

17
 axn+1
Taking F (x) = n+1
, and using the power rule for differentiation gives:

axn+1
!
d a
= · (n + 1)xn = axn
dx n+1 n+1

which is f (x). Whenever you differentiate a constant, it goes to 0; so you could take
n+1
F (x) = axn+1 + C, differentiate it with respect to x, and still get f (x). So this means that
you can write:
Z
axn+1
axn dx = + C for n 6= −1 (2.2)
n+1

A special case happens when n = 0; if this happens, then f (x) = ax0 = a is a constant.
1
Integrating this using the formula above gives a dx = ax0 dx = ax1 + C = ax + C and
R R

so Z
a dx = ax + C (2.3)

You may have asked yourself already why doesn’t this process work for n = −1? These are
functions of the form f (x) = xa . It is because if you do the process with n = −1, you end
up dividing by 0; which is not good. However, you can look at Table 1.1 and see that

d a
(a ln(x)) =
dx x

However, you need to make sure that ln(x) is defined on only positive x. You can get
around this by writing ln |x| instead of ln(x); this ensures the value of the input is positive.
This means you can write
Z
ax−1 dx = a ln |x| + C (2.4)

You have already seen in Example 2.2.1 that an antiderivative for f (x) = ex is F (x) =
ex + C. But what about f (x) = aekx ? You are looking for a function F (x) such that
d
dx
(F (x)) = aekx . Here, taking F (x) = k1 aekx + C, and differentiating with respect to x
using the rules in Table 1.1 gives

d 1 kx
 
ae + C = aekx
dx k

and so you can write

Z
1 kx
aekx dx = ae + C (2.5)
k

Finally, you can consider antiderivatives of a sin(kx) and a cos(kx). Using rules of differen-

18
 function antiderivative w.r.t x notes
Z
f (x) = a f (x) dx = ax + C a constant (2.3)

Z
ax(n+1)
f (x) = ax n
f (x) dx = +C n 6= −1 (2.2)
n+1
Z
f (x) = ax −1
f (x) dx = a ln |x| + C (2.4)

Z
1 kx
f (x) = ae kx
f (x) dx = ae + C (2.5)
k
Z
1
f (x) = a cos(kx) f (x) dx = a sin(kx) + C (2.6)
k
Z
1
f (x) = a sin(kx) f (x) dx = − a cos(kx) + C (2.7)
k

Table 2.1: Some antiderivatives

tiation, you can say that

d 1 d 1
   
− a cos kx = a sin(kx) and a sin kx = a cos(kx)
dx k dx k

This means that you can write:

Z
1
a cos(kx) dx = a sin kx + C (2.6)
k

and Z
1
a sin(kx) dx = − a cos kx + C (2.7)
k

Summarising the boxed results, Table 2.1 gives a list of common antiderivatives that will be
useful throughout your mathematical career. Throughout, a, k, C are constants.
You can use the examples given in Table 2.1 to find some integrals that previously were
harder to find.

Example 2.3.1. You are asked to find the definite integral 02 x dx. To do this, you need
R

to find an antiderivative F (x) of f (x) = x, and then find out F (b) − F (a) where b = 2

19
and a = 0. You can use Equation 2.2 with a = 1 and n = 1 to get
" #2 " #2
Z 2
1 · x1+1 x2
x dx = =
0 1+1 0
2 0

Using F (x) = x2 /2, you can write that F (2) = 4/2 = 2 and F (0) = 0, so now you can
evaluate the integral: " #2
Z 2
x2 4
x dx = = −0=2
0 2 0 2
This is the same answer you got in Example 1.2.3.

It can be shown that properties (2) and (3) of Proposition 1.3.1 hold for indefinite integrals.
So for any constant k (so nothing that involves a variable!) you can write that
Z Z
kf (x) dx = k f (x) dx

and for a continuous function g(x):

Z Z Z
f (x) + g(x) dx = f (x) dx + g(x) dx

Example 2.3.2. You are asked to find the indefinite integral 16 − x2 dx. Here, you can
R

use the properties of indefinite integrals stated above (and a remark in Example 1.2.2) to
write Z Z Z Z Z
16 − x2 dx = 16 dx + −x2 dx = 16 dx − x2 dx

So the problem is reduced to working out antiderivatives of 1 and x2 . Using Equation 2.3
with f (x) = 1 gives an antiderivative of 1x = x. Using Equation 2.2 with n = 2 gives x3 /3
as an antiderivative of x2 . So you can write:
Z
2
Z Z
x3
16 − x dx = 16 1 dx − x2 dx = 16x − +C
3

You can notice here that you only need to add one constant of integration +C at the very
end of the working.

Example 2.3.3. You are asked to find the definite integral −π

π
sin(2x) + 2 cos(x) dx. You
R

can use Equation 2.7 with a = 1 and k = 2 to say that an antiderivative for sin(2x) is
− 21 cos(2x). You can use Equation 2.6 with a = 2 and k = 1 to get that an antiderivative

20
for 2 cos(x) is 2 sin(x). Therefore, you can write
Z π π
1

sin(2x) + 2 cos(x) dx = − cos(2x) + 2 sin(x)
−π 2 −π

You can now evaluate the value of this antiderivative at the limits π and −π. Here

1 1 1
− cos(2π) + 2 sin(π) = − · (1) + 2 · 0 = −
2 2 2

and
1 1 1
− cos(−2π) + 2 sin(−π) = − · (1) + 2 · 0 = −
2 2 2
You can put this all together to get that
π
Z π
1 1 1
    
sin(2x) + 2 cos(x) dx = − cos(2x) + 2 sin(x) = − − − =0
−π 2 −π 2 2

This example shows that you do not have to write F (x) at any point when evaluating an
integral, nor do you have to split up the integral of two functions added together into two
parts before proceeding.

2.4 Summary of first two chapters

• The Fundamental Theorem of Calculus states that indefinite integration is the reverse
process to differentiation, and that you can work out definite integrals (the limit of
sums of areas) by using the antiderivative of the integrand.

• Working out both indefinite and definite integrals involve using antiderivatives; you
can find these for common functions in Table 2.1.

• When you work out a indefinite integral f (x) dx, your answer should be a function
R

F (x) plus a constant of integration C. This expression represents all the possible
antiderivatives of f (x). You should always remember to add a +C at the end when
working out an indefinite integral.

• When you work out a definite integral, your answer should be a number. This
number represents the signed area underneath the curve of f (x) between the lines
x = a and x = b. There should be no variables in your answer, and you do not need
a +C.

21
Chapter 3

Integration by substitution

3.1 Techniques of calculus

You may have studied techniques for finding derivatives of composite functions using the
chain rule, the product rule and the quotient rule. These rules are detailed in Table 3.1 for
various operations on functions f (x) and g(x).

function derivative w.r.t x name

d d
y(x) = f (g(x)) y 0 (x) = (f (g(x))) · (g(x)) chain rule
dg dx

y(x) = f (x) · g(x) y 0 (x) = g(x) · f 0 (x) + f (x) · g 0 (x) product rule

g(x) · f 0 (x) − f (x) · g 0 (x)

y(x) = f (x)
g(x)
y 0 (x) = quotient rule (g 6= 0)
(g(x))2

Table 3.1: Helpful rules for differentiation

Similar techniques exist for when you want to find antiderivatives of composite functions.
This chapter is on integration by substitution, which is similar to the technique of the
chain rule. This technique is discussed further in Chapter 4, when considering trigonometric
functions in more detail. Chapter 5 is on integration by parts, which is similar to the
technique of the product rule. Both techniques are very important for finding antiderivatives
of a variety of functions.

22
3.2 How integration by substitution works

Suppose you are asked to integrate a composite function f (u(x)). You can write the
function u(x) as a variable u here (see the terminology on page 4), and use the terms
interchangeably; but you should always remember that u is a function of x. Finding this
integral of f (u(x)) = f (u) can be done in two different ways; one with respect to u and
one with respect to x. Firstly, using the Fundamental Theorem of Calculus, you are looking
for a function F (u) such that
d
(F (u)) = f (u)
du
and so you can write Z
F (u) = f (u) du

As u is a function of x, you can differentiate F (u(x)) with respect to x using the chain
rule, to get

d d d
(F (u(x))) = (F (u(x))) · (u(x)) = F 0 (u(x)) · u0 (x) = f (u(x)) · u0 (x)
dx du dx

You can see from this that F (u(x)) is an antiderivative (with respect to x) of f (u(x))·u0 (x).
This means that you can write
Z
F (u) = f (u) · u0 (x) dx

This gives two expressions for F (u), which you can compare to get:
Z Z
f (u(x)) · u0 (x) dx = f (u) du (3.1)

Equation 3.1 demonstrates the principle of integration by substitution for indefinite

integrals. Essentially, you can choose a function u = u(x) of x, and ‘substitute’ this into
the left hand side of Equation 3.1. This is then equal to the right hand side which, on
the correct choice of u(x), should be an easier integral to solve. You should consider using
integration by substitution (with substitution u = u(x)) where the integrand is a composite
function f (u(x)) of x multiplied by a term that ‘looks like’ the derivative u0 (x) of u(x).
The difficulty of integration by substitution lies in the correct choice of u = u(x). There is
no general rule for choosing the ‘right’ substitution in a given integrand, so you will have
to decide on a correct substitution for every integration by substitution that you do. You
will know that you have made the correct substitution when there is no mention of x in

23
 your expression. A useful thing to remember is that the idea of this method is to make
integration easier; so the integral on the right is something you know you can work out.
This should motivate your choice for u.
One thing that you can notice is that integration by substitution at this stage is only used
in finding antiderivatives of composite functions. Suppose that you are using integration by
substitution to evaluate a definite integral ab f (u(x)) dx. You need to be careful here as
R

the limits of the integral are defined in terms of x, and not in terms of u. To correct this,
you need to evaluate the limits a and b in your chosen substitution u = u(x), and then
substitute these values in the correct place. In other words:

Z b Z u(b)
f (u(x)) · u0 (x) dx = f (u) du (3.2)
a u(a)

This is integration by substitution for definite integrals.

Another thing you can notice is that f (u) is present in both integrands in Equation 3.1 (and
Equation 3.2), and that this implies that u0 (x)dx = du. This expression can be rearranged;
so you can write dx = du/u0 (x). This technique is very useful; particularly if there is no
u0 (x) term immediately visible in the integrand.
You can follow these steps whenever you would need to use integration by substitution. This
method will be used throughout the examples given in Section 3.3.

3.2.1 Method for integration by substitution

Step 1: Choose a suitable u = u(x). Your choice of substitution should not be a constant
function u(x) = a.

Step 2: Work out u0 (x), and write down an expression for dx = du/u0 (x). If you are
considering a definite integral, work out u(a) and u(b) where a and b are the limits
of the integral.

Step 3: Now, you should

• replace every instance of u(x) with the letter u

• replace dx with du/u0 (x), and cancel;
• (for definite integrals only) replace a with the value u(a) and b with u(b).

Warning: At this stage, the integral should be solely in terms of u. If there are still terms
containing x at this stage, stop and consider another choice of u.

24
Step 4: If you can, work out the integral. If you are considering a definite integral, then the
method stops here with the answer. if you are considering an indefinite integral,
don’t forget the +C!

Step 5: (For indefinite integrals only) Your antiderivative should be in terms of u. Replace
every instance of u with the original function u(x). The method stops here for
indefinite integrals.

3.3 Examples

Throughout these examples, the properties of integrals as outlined in Proposition 1.3.1 are
used without reference.

Example 3.3.1. You are asked to evaluate sin(3x + 9) dx. Since this is an indefinite
R

integral, you have no limits to change but you must remember to do Step 5 of the method
in Subsection 3.2.1.

Step 1: You can integrate sin, so a good choice of substitution here would be u = 3x + 9.

Step 2: Here, u0 (x) = 3, and so dx = du/3. As this is an indefinite integral, you have no
limits to change.

Step 3: Replacing each instance of u(x) = 3x + 9 with u, and replacing dx by du/3 gives
Z Z
1 1Z
sin(3x + 9) dx = sin(u) du = sin(u) du
3 3

Since there is no x left in the integral, you are OK to continue.

Step 4: Not forgetting the constant of integration, evaluating the integral using the appro-
priate rule from Table 2.1 gives

1Z 1
sin(u) du = − cos(u) + C
3 3

Step 5: You need to substitute in 3x + 9 = u into your solution to get 1

3
cos(3x + 9). So
you can write Z
1
sin(3x + 9) dx = − cos(3x + 9) + C
3
which is the final answer.

25
Example 3.3.2. Say you are asked to evaluate 04 (x/2−4) 1
Since this is a definite
R
3 dx.

integral, you must remember to change the limits of integration, but you don’t have to do
Step 5 of the method in Subsection 3.2.1.

Step 1: You can integrate any function of the form xn , so a good choice of substitution
here would be u = x/2 − 4.

Step 2: Here, u0 (x) = 1/2, and so dx = du/(1/2) = 2du. As this is a definite integral,
you must remember to change the limits. You can do this by evaluating u at x = 0
and x = 4. So here u(0) = −4 and u(4) = −2.

Step 3: Replacing each instance of u(x) = x/2 − 4 with u, replacing dx by 2du, and
replacing the lower limit 0 with −4 and the upper limit 4 with −2 gives
Z −2 Z −2
Z 4
1 1
dx = · 2 du = 2 u−3 du
0 (x/2 − 4)3 −4 u3 −4

As there are no terms involving x left in the integral, you are OK to continue and
you can now evaluate the integral.

Step 4: Evaluating the integral using the corresponding rule from Table 2.1 gives
#−2
u−2
Z −2 "
h i−2
2 u−3 du = 2 = −u−2
−4 −2 −4
−4

So this means that

" # " #
h i−2 1 1 1 1 3
−u−2 = − 2
− − 2
=− + =−
−4 (−2) (−4) 4 16 16

which is the final answer.

Example 3.3.3. You are asked to evaluate sin3 (x) cos(x) dx. Since this is an indefinite
R

integral, you have no limits to change but you must remember to do Step 5 of the method
in Subsection 3.2.1.

Step 1: You can integrate un , so a good choice of substitution here would be u = sin(x).

Step 2: Here, u0 (x) = cos(x), so you can write dx = du/ cos(x). Again, as this is an
indefinite integral, you have no limits to change.

26
Step 3: Replacing each instance of u(x) = sin(x) with u, and replacing dx by du/ cos(x)
gives
3 cos(x)
Z Z Z
3
sin (x) cos(x) dx = u du = u3 du
cos(x)
Since there is no x left in the integral, you are OK to continue.

Step 4: Not forgetting the constant of integration, evaluating the integral using the corre-
sponding rule from Table 2.1 gives
Z
u4
u3 du = +C
4

Step 5: You need to substitute in sin(x) = u into your solution to get 1

4
sin4 (x). You can
write Z
1
sin(3x + 9) dx = − cos(3x + 9) + C
3
which is the final answer.
√
x
Example 3.3.4. You are asked to evaluate x1 + e√x dx. Since this is an indefinite integral,
R

you have no limits to change but you must remember to do Step 5 of the method in
Subsection 3.2.1. Here, you can integrate 1/x using a rule in Table 2.1 (giving ln |x|, but
√ √
you cannot integrate e x / x so easily. The idea here is to split the integrand into two, using
integration by substitution on one of the parts. Using the techniques from Proposition 1.3.1
can make your life easier in these circumstances. So by saying
√ √ √
Z
1 e x Z
1 Z
e x Z
e x
+ √ dx = dx + √ dx = ln |x| + √ dx
x x x x x

you can focus on evaluating the remaining integral by substitution.

√
Step 1: You can integrate ex , so a good choice of substitution here would be u = x=
x1/2 .

Step 2: Here, u0 (x) = 21 x−1/2 = 1

2x1/2
. This means that

dx = du/(1/2x1/2 ) = 2x1/2 du

As x1/2 = u, you can write dx = 2udu. Since this is an indefinite integral, you
have no limits to change.

27
 √
Step 3: Replacing each instance of u(x) = x with u, and replacing dx by 2udu gives
√
Z
e x Z u
e Z
√ dx = · 2u du = 2 eu du
x u

Since there is no x left in the integral, you are OK to continue.

Step 4: Not forgetting the constant of integration, evaluating the integral using the corre-
sponding rule from Table 2.1 gives
Z
2 eu du = 2eu + C

√ √
Step 5: You need to substitute in x = u into your solution to get 2e x . So you can write
√
Z
e x √
√ dx = 2e x + C
x

which is the final answer.

So therefore, you can now integrate the entire expression. The final answer is
√ √
Z
1 e x Z
1 Z
e x √
+ √ dx = dx + √ dx = ln |x| + 2e x + C.
x x x x
R π/4
Example 3.3.5. Say you are asked to evaluate 0 x + cos(2x)2 sin(2x) dx. This is a
definite integral with two parts. Similar to the previous example, you can write
Z π/4 Z π/4 Z π/4
2
x + cos (2x) sin(2x) dx = x dx + cos2 (2x) sin(2x) dx
0
| 0 {z } |0 {z }
(1) (2)

and work out both integrals one at a time.

So for (1), you can evaluate this using the rules in Table 2.1. Here
" #π/4
Z π/4
x2 (π/4)2 0 π2
x dx = = − =
0 2 0
2 2 32

For part (2) however, you need to use integration by substitution. As this is a definite
integral, you must remember to change the limits of integration.

Step 1: You can integrate any function of the form xn , so a good choice of substitution
here would be u = cos(2x).

28
Step 2: Here, u0 (x) = −2 sin(2x), and so dx = du/(−2 sin(x)). As this is a definite
integral, you must remember to change the limits. You can do this by evaluating
u at x = 0 and x = π/4. So here u(0) = cos(0) = 1 and u(π/4) = cos(π/2) = 0.

Step 3: By replacing each instance of cos(2x) with u, the term dx with du/(−2 sin(x)),
and the lower limit 0 with 1 and the upper limit π/4 with 0 gives
Z π/4
2
Z 0
u2 sin(2x) 1Z 0 2
cos (2x) sin(2x) dx = du = − u du
0 1 −2 sin(2x) 2 1

As there are no terms involving x left in the integral, you are OK to continue and
you can now evaluate the integral.

Step 4: Using Proposition 1.3.1 (6), you can see that

1Z 0 2 1Z 1 2
− u du = u du
2 1 2 0

Evaluating this integral using the corresponding rule from Table 2.1 gives
" #1 " #1
1Z 1 2 1 u3 u3
u du = =
2 0 2 3 0
6 0

So this means that " #1

u3 1 1
 
= − − [0] =
6 0
6 6
which is the final answer for (2).

Putting the results of (1) and (2) together give

Z π/4
π2 1
x + cos2 (2x) sin(2x) dx = +
0 32 6

Remark. In these examples, the tradition has been to substitute dx with du/u0 (x) and
then cancel through. However, sometimes it is convenient to replace u0 (x)dx with du. For
instance, you could have used this technique in Example 3.3.3 and Example 3.3.5.

Finally in this chapter, you can use integration by substitution in a more general setting,
and demonstrate that it works as the ‘inverse’ of the chain rule of differentiation. Consider
the function y = ln(h(x)), where h is some non-zero function of x. Using the chain rule in

29
Table 3.1 with f = ln(x) and g = h(x) gives

dy 1 h0 (x)
= ·h0 (x) =
dx h(x) h(x)
| {z }
f 0 (g(x))

Now, for some non-zero function h, you can think about solving
Z
h0 (x)
dx
h(x)

You can use integration by substitution to do this, with a choice of substitution u = h(x).
This means that u0 (x) = h0 (x), and so dx = du/h0 (x). Replacing h(x) with u and dx with
du/h0 (x) gives
Z 0
h (x) Z
h0 (x) Z
1
dx = 0
du = du
h(x) u · h (x) u
You can integrate this using the corresponding rule from Table 2.1 to get
Z
1
du = ln |u| + C
u

Substituting h(x) back in for u gives the following useful identity

Z
h0 (x)
dx = ln |h(x)| + C (3.3)
h(x)

You can notice that this reverses the differentiation performed above.
This equation is a valuable tool, as you can ‘spot’ the patterns of a function in the denom-
inator and its derivative on the top. This will save you time in performing an integration by
substitution. For instance, you can consider the indefinite integral
Z
2x
dx
x2 +1

Here, you can see that the numerator h0 (x) = 2x is precisely the derivative of the denomi-
nator h(x) = x2 + 1. Therefore, you can use Equation 3.3 to say that
Z
2x
dx = ln |x2 + 1| + C.
x2 +1

30
Chapter 4

Trigonometry and integration

4.1 Previously in trigonometry...

So far in this portion on integration, you have seen the cos and the sin functions... but not
much else. You may have seen at different points that there are four other trigonometric
functions to consider; tan, sec, csc and cot. The definitions of these functions are given in
Table 4.1 along with their derivatives. You can find the derivatives on your own by using
the chain rule on each of the definitions of the functions.

function definition

sin(x)
f (x) = tan(x) f (x) =
cos(x)
1
f (x) = sec(x) f (x) =
cos(x)
1
f (x) = csc(x) f (x) =
sin(x)
1 cos(x)
f (x) = cot(x) f (x) = =
tan(x) sin(x)

Table 4.1: Definitions of tan, sec, csc and cot

You may have studied the derivatives of all of the functions in Table 4.1 in previous parts
of the course. Table 4.2 provides the derivatives of the six trigonometric functions you have
seen so far, as well as all the known antiderivatives.

31
 function derivative w.r.t x
R
dx

f (x) = a cos(kx) f 0 (x) = −ak sin(kx) (2.6)

f (x) = a sin(kx) f 0 (x) = ak cos(kx) (2.7)

ak
f (x) = a tan(kx) f 0 (x) = 2
= ak sec2 (kx) (4.1)
cos (kx)
sin(kx)
f (x) = a sec(kx) f 0 (x) = ak = ak sec(kx) tan(kx) (6.3)
cos2 (kx)
cos(kx)
f (x) = a csc(kx) f 0 (x) = −ak = −ak csc(kx) cot(kx) (4.3)
sin2 (kx)
ak
f (x) = a cot(kx) f 0 (x) = − 2 = −ak csc2 (kx) (4.2)
sin (kx)

Table 4.2: Derivatives and antiderivatives of trigonometric functions

Here, the references given in the antiderivative column refers to the equation in which they
are first presented.
The aim of this chapter is for you to expand your knowledge about integration of trigonomet-
ric functions, using trigonometric identities to calculate examples, and the use of trigono-
metric functions in integration by substitution.
To begin with, you should make sure that you know the derivative and antiderivative of
both sin and cos functions (listed in Table 4.2).

4.2 Some more antiderivatives

This section is designed to find some more antiderivatives that you can use throughout your
integration course. A good place to start would be to look at the functions outlined in
Table 4.1, and work out antiderivatives for those.
Let’s begin with a tan(kx). Here, you can write
Z Z
sin(kx)
a tan(kx) dx = a dx.
cos(kx)

32
You can work this integral out using integration by substitution.
Here, you can integrate this expression using integration by substitution, with u = cos(kx).
It follows that u0 (x) = −k sin(kx) and so dx = du/ − k sin(kx). Making the substitution
gives
Z
sin(kx) Z
sin(kx) Z
a aZ 1
a dx = −a du = − du = − du
cos(kx) u · k sin(kx) ku k u
You can integrate this expression using Equation 2.4 to get

aZ 1 a
− du = − ln |u| + C
k u k

Substituting u = cos(kx) back into this antiderivative, and using the laws of logarithms
together with the fact that sec(kx) = cos(kx)
1
gives

Z
a a 1 a
a tan(kx) dx = − ln | cos(kx)| + C = ln + C = ln | sec(kx)| + C
k k cos(kx) k

and so you can write

Z
a
a tan(kx) dx = ln | sec(kx)| + C (4.1)
k

There is another way in which you can obtain this result. By Table 4.2, the derivative of
cos(kx) is − sin(kx), which is nearly equal to the numerator. So here, you can use the fact
that −a/k · −k = a, together with Proposition 1.3.1 (2) to write that
Z
sin(kx) Z
sin(kx) a Z −k sin(kx)
a dx = (−ak) dx = − dx
cos(kx) −k cos(kx) k cos(kx)

You can recognise that this integrand is a fraction with a function h(x) on the bottom and
its derivative h0 (x) on the top. Therefore, you can use Equation 3.3 to write

a Z −k sin(kx) a a
− dx = − ln | cos(kx)| + C = ln | sec(kx)| + C
k cos(kx) k k

and recover the result from Equation 4.1.

You can use a similar method to work out the antiderivative of a cot(kx). Here, you can
use Table 4.1 to write that
cos(kx)
a cot(kx) = a
sin(kx)
You can integrate this by substitution, using u = sin(kx) as your choice of substitution.

33
This means that u0 (x) = k cos(kx) and so dx = du/k cos(kx). Substituting these in and
cancelling gives
Z
cos(kx) Z
cos(kx) aZ 1
a dx = a du = du
sin(kx) u · k cos(kx) k u

You can integrate this using Equation 2.4 to write

aZ 1 a
du = ln |u| + C
k u k

Finally, you can substitute u = sin(kx) into this to write

Z
a
a cot(kx) dx = ln | sin(kx)| + C (4.2)
k

Remark. While the antiderivatives of tan and cot have been worked out with techniques
that you have already studied up to this point, finding the indefinite integrals of sec(x) and
csc(x) with respect to x need some more work.

You can use the Fundamental Theorem of Calculus (Theorem 2.2.2) to state the antideriva-
tives for those functions given in the second column of Table 4.2. These results are collected
in Table 4.3. You can check these antiderivatives by differentiating them and getting f (x)
as an answer.

function antiderivative w.r.t x

Z
a
f (x) = a sec2 (kx) f (x) dx = tan(kx) + C
k
Z
a
f (x) = a sec(kx) tan(kx) f (x) dx = sec(kx) + C
k
Z
a
f (x) = a csc(kx) cot(kx) f (x) dx = − csc(kx) + C
k

2
Z
a
f (x) = a csc (kx) f (x) dx = − cot(kx) + C
k

Table 4.3: Antiderivatives of some derivatives found in Table 4.2

34
Example 4.2.1. You are asked to integrate
Z
−4 tan(2x) + 4 sec2 (x) dx

Here, you can use Proposition 1.3.1 (3) to say that

Z Z Z
2
−4 tan(2x) + 4 sec (x) dx = −4 tan(2x) dx + 4 sec2 (x) dx

Using Equation 4.1 with a = −4 and k = 2 for the first integral, and using the result in
Table 4.3 with a = 4 and k = 1 for the second integral, you can write
Z
−4 4
−4 tan(2x) + 4 sec2 (x) dx = − ln | sec(2x)| + tan(x) + C
2 1
=2 ln | sec(2x)| + 4 tan(x) + C
= ln | sec2 (2x)| + 4 tan(x) + C

4.3 Using trigonometric identities

In your mathematical career, you may have come across a range of identities that relate two
trigonometric functions together. These trigonometric identities prove to be very useful in
integrating some functions that cannot be integrated directly or by substitution. Techniques
to do this are discussed in the next two sections.
The first set of these pairs together trigonometric functions in the square identities (or
Pythagorean identities; see Table 4.4):

functions identity

sin, cos cos2 (x) + sin2 (x) = 1

tan, sec 1 + tan2 (x) = sec2 (x)

cot, csc cot2 (x) + 1 = csc2 (x)

Table 4.4: Square identities of trigonometric functions

These are so called because you can derive the first of these from Pythagoras’ Theorem,

35
and the second and third of these identities from the first by dividing through by cos2 (x)
and sin2 (x) respectively.
The second set of these are known as angle sum rules, and are particularly useful for
dealing with powers of trigonometric functions. These are very useful when you take the
angles A and B to be the same angle, giving the related double angle formulas (see
Table 4.5):

Sum rule double angle formula

sin(A ± B) = sin(A) cos(B) ± cos(A) sin(B) sin(2A) = 2 sin(A) cos(A)

cos(A ± B) = cos(A) cos(B) ∓ sin(A) sin(B) cos(2A) = cos2 (A) − sin2 (A)

tan(A) ± tan(B) 2 tan(A)

tan(A ± B) = tan(2A) =
1 ∓ tan(A) tan(B) 1 − tan2 (A)

Table 4.5: Sum rules of trigonometric functions

You can choose the angles A and B carefully in these rules in order to find expressions to
help you solve integrals. For instance, taking A = 2x and B = x in the sum rules yield
the triple angle formulas. Here, for instance, is the derivation of the formula for cos3 (x),
using the double angle formulas and the fact that sin2 (x) = 1 − cos2 (x):

cos(3x) = cos(2x + x) = cos(2x) cos(x) − sin(2x) sin(x)

= (cos2 (x) − sin2 (x)) cos(x) − (2 sin(x) cos(x)) sin(x)
= cos3 (x) − sin2 (x) cos(x) − 2 sin2 (x) cos(x)
= cos3 (x) − 3 sin2 (x) cos(x)
= cos3 (x) − 3(1 − cos2 (x)) cos(x)
= 4 cos3 (x) − 3 cos(x). (*)

You should know, and be confident with manipulating the identities in Tables 4.4 and 4.5
before continuing.

36
4.3.1 Examples
R π/4
Example 4.3.1. You are asked to find 0 tan2 (x) dx. You can notice that you cannot
integrate this using substitution, as it is not of the form f (u(x)) · u0 (x) for some function
u(x). However, you can see that you can write down the antiderivative of sec2 (x) from
Table 4.3, and that tan2 (x) = sec2 (x) − 1 from Table 4.4. This means you can write
Z π/4 Z π/4
2
tan (x) dx = sec2 (x) − 1 dx
0 0

You can integrate this to get

Z π/4
sec2 (x) − 1 dx = [tan(x) − x]π/4
0
0
π π
   
= tan − − (0 + 0)
4 4
π
=1 −
4

Example 4.3.2. You are asked to find sin2 (x) dx. Again, you cannot integrate this using
R

substitution, and this is not something you can write down the antiderivative for. Here, you
need to use trigonometric identities to find a way to write sin2 (x) in a form you know you
can integrate.
Here, you can notice that sin2 (x) appears in the double angle formula for cos(2x), which is
cos2 (x)−sin2 (x). Furthermore, sin2 (x) appears in the square identity cos2 (x)+sin2 (x) = 1,
which you can rewrite as cos2 (x) = 1 − sin2 (x). Combining these gives

cos(2x) = cos2 (x) − sin2 (x)

= (1 − sin2 (x)) − sin2 (x)
= 1 − 2 sin2 (x)

Rearranging to make sin2 (x) the subject gives

1 1
sin2 (x) = − cos(2x)
2 2

and the right hand side of this equation is something that you can integrate. So now, you
can write Z Z
1 1
sin2 (x) dx = − cos(2x) dx
2 2

37
and integrate using Equation 2.3 and Equation 2.6 to get
Z
2
Z
1 1 1 1
sin (x) dx = − cos(2x) dx = x − sin(2x) + C
2 2 2 4

Example 4.3.3. Now, you are asked to find 2 cos3 (x) dx. Once again, you cannot use
R

substitution or integrate directly to find the answer to this problem. This integrand is a
little harder to represent as something you can integrate directly as there is no trigonometric
identity in Tables 4.4 and 4.5 to use with cos3 (x) in it. However, at the start of the section,
it was demonstrated in Table * that

cos(3x) = 4 cos3 (x) − 3 cos(x)

You can rearrange this to say that

1 3
cos3 (x) = cos(3x) + cos(x)
4 4

which is something that you can integrate. Replacing the integrand gives
Z
3
Z
1 3
2 cos (x) dx = cos(3x) + cos(x) dx
2 2

and you can integrate this using Equation 2.6 to get

Z Z
1 3 1 3
2 cos3 (x) dx = cos(3x) + cos(x) dx = sin(3x) + sin(x) + C
2 2 6 2

Finally in this section, nou can use trigonometric identities, together with some previous
results, to prove one of the final two antiderivatives of the functions from Table 4.1 not yet
considered.

Example 4.3.4. You are asked to find a csc(kx) dx. You can use Table 4.1 and Propo-
R

sition 1.3.1 (2) to write the integral as

Z Z
a
a csc(kx) dx = dx
sin(kx)

Now, you can use the sum rule for sin in Table 4.5 with A = B = kx/2 to write that

1 1
=
sin(kx) 2 sin(kx/2) cos(kx/2)

38
Next, you can use the fact that 1 = cos2 (kx/2) + sin2 (kx/2), and cancelling gives to write

1 cos2 (kx/2) + sin2 (kx/2) cos(kx/2) sin(kx/2)

= = +
2 sin(kx/2) cos(kx/2) 2 sin(kx/2) cos(kx/2) 2 sin(kx/2) 2 cos(kx/2)

Finally, you can use Table 4.1 to write that

1 1 1
= cot(kx/2) + tan(kx/2)
sin(kx) 2 2

You know how to integrate cot and tan by Equations (4.2) and (4.1) respectively. So
replacing the integrand gives
Z
a Z
a a
dx = cot(kx/2) + tan(kx/2) dx
sin(kx) 2 2

and integrating using Equation 4.2 and Equation 4.1 gives

Z
a a 2a 2a
cot(kx/2) + tan(kx/2) dx = ln | sin(kx/2)| + ln | sec(kx/2)| + C
2 2 2k 2k

Cancelling the 2’s in each term, and using the laws of logarithms together with the fact
that sin(kx) sec(kx) = tan(kx) gives:
Z
a
a csc(kx) dx = ln | tan(kx/2)| + C (4.3)
k

Remark. This technique does not work in the case of integrating a sec(kx). This is because
using the double angle formula on cos(kx) in the same fashion as in Example 4.3.4 causes
the denominator to become cos2 (kx/2) − sin2 (kx/2); an expression that you cannot cancel.
The technique to integrate a sec(kx) will be explored later.

4.4 Inverse functions: a different kind of substitution

Every trigonometric function has an inverse. For instance, the inverse of sin(x) is sin−1 (x).
You may have seen in the differentiation part of the course that you can use implicit dif-
ferentiation to find derivatives of these functions. A list of inverse trigonometric functions
and their derivatives are given in Table 4.6; this means that you can find antiderivatives for
each of these derivatives by the Fundamental Theorem of Calculus.
However, sometimes the integral you are considering does not come in the same form as

39
 function derivative w.r.t x
R
dx

−ak
f (x) = a cos−1 (kx) f 0 (x) = √
1 − k 2 x2
ak
f (x) = a sin−1 (kx) f 0 (x) = √
1 − k 2 x2
ak
f (x) = a tan−1 (kx) f 0 (x) =
1 + k 2 x2
a
f (x) = a sec−1 (kx) f 0 (x) = q
kx2 1 − (kx)−2
−a
f (x) = a csc−1 (kx) f 0 (x) = q
kx2 1 − (kx)−2
−ak
f (x) = a cot−1 (kx) f 0 (x) =
1 + k 2 x2

Table 4.6: Derivatives and antiderivatives of inverse trigonometric functions

one of the derivatives in Table 4.6. For instance, how would you go about integrating
√
4
dx? Or integrating something like x2 x2 − 16 dx, which looks nothing like one
R R
4+9x2
of the derivatives above and can’t be integrated by substitution? You can notice that the
parts of the integrals that look like
√ √
x 2 − a2 , a2 − x 2 , a2 + x2

can be rearranged to look like constant multiples of the square identities given in Table 4.4.
To make use of these identities, you need to make a trigonometric substitution in terms
of some angle θ.
This type of substitution is different to the integration by substitution that you already know.
This is because instead of substituting u as a function of x, you are having to substitute x
with a function of some variable θ. So here, your substitution will look something like
x = g(θ), where g is some trigonometric function of θ.
√
For instance, what substitution could you consider for a function that contains a2 − x2 ?
By substituting x = a sin(θ), and knowing that 1 − sin2 (θ) = cos2 (θ) from Table 4.4, you

40
can see that
√ q q
a2 − x2 = a2 − a2 sin2 (θ) = a2 (1 − sin2 (θ)) = a cos(θ)

Similarly, for functions that contain x2 + a2 , you could consider the substitution a tan(θ).
Then, using the fact that 1 + tan2 (θ) = sec2 (θ), you can write

x2 + a2 = a2 tan2 (θ) + a2 = a2 (tan2 (θ) + 1) = a2 sec2 (θ)

√
Finally, for functions that contain x2 − a2 you could consider the substitution a sec(θ).
Then, using the identity tan2 (θ) = sec2 (θ) − 1, you can say that
√ q q
x2 − a2 = a2 sec2 (θ) − a2 = a2 (sec2 (θ) − 1) = a tan(θ)

Warning. Of course, there are functions containing these types of expressions that do not
require a trigonometric substitution. For instance,
Z
x3
x2 + a2 dx = + a2 x + C
3

Some integrands containing these types of expression can be solved using a conventional
substitution rather than a trigonometric one. For instance, using the substitution u = x2 −16
gives that
Z √
(x2 − 16)3/2
x x2 − 16 dx = +C
3
Therefore, you should only use a trigonometric substitution unless you really have to! This
is usually when one of the three expressions above is in the denominator of the integrand.

The choice of substitution is not the only thing you need to consider before making a
trigonometric substitution. Because you are considering a function of θ, this has the effect
of changing how you handle the dx term in the integral. Here, the correct substitution for
dx is dx = g 0 (θ)dθ.
Furthermore, if you are considering a definite integral, you need to change the limits; again,
this is slightly different to the usual process. As your limits a and b will be in terms of x,
you need to change this into terms of θ. As x = g(θ), it follows that g −1 (x) = θ, where
g −1 is the inverse trigonometric function of g. This means that you should replace a with
g −1 (a) and replace b with g −1 (b).
This leads to a modified method for integration via a trigonometric substitution. The

41
 principle is the same, with a few key differences.

4.4.1 Method for integration by trigonometric substitution

Step 1: Choose a suitable x = g(θ).

Step 2: Work out g 0 (θ) (the derivative of g(θ) with respect to θ, and write down the
expression dx = g 0 (θ)dθ. If you are considering a definite integral, work out an
expression for θ in terms of x (usually, this is the inverse function g −1 (x) = θ),
and then evaluate this expression at a and b where a and b are the limits of the
integral.

Step 3: Now, you should

• replace every instance of x with the function g(θ)

• replace dx with g 0 (θ)dθ, and cancel if you need to;
• (for definite integrals only) replace a with the value g −1 (a) and b with g −1 (b).

Warning: At this stage, the integral should be solely in terms of θ. If there are still terms
containing x at this stage, stop and consider another choice of g(θ).

Step 4: If you can, work out the integral. If you are considering a definite integral, then the
method stops here with the answer. if you are considering an indefinite integral,
don’t forget the +C!

Step 5: (For indefinite integrals only) Your antiderivative should be in terms of θ. You can
replace every instance of θ with the inverse function g −1 (x), or find an expression
for the antiderivative h(θ) in terms of x if the alternative is too difficult (see
Example 4.4.2). The method stops here for indefinite integrals.

If you do require a trigonometric substitution, then the information about the choice you
could make is summarised in Table 4.7. The reason why these substitutions are suggested
is the fact that what results after the substitution x = g(θ) looks like a constant multiple
of the derivative g 0 (θ) of g(θ).

4.4.2 Examples
R √2
Example 4.4.1. You are asked to find the integral 1
√1
x2 4−x2
dx.

42
 function contains... suggested substitution expression for dx
√
a2 − x 2 g(θ) = a sin(θ) dx = a cos(θ)dθ
√
x 2 − a2 g(θ) = a sec(θ) dx = a sec(θ) tan(θ)dθ

a2 + x 2 g(θ) = a tan(θ) dx = a sec2 (θ)dθ

Table 4.7: Suggested trigonometric substitutions

Since this is a definite integral, you must remember to change the limits of integration, but
you don’t have to do Step 5 of the method in Subsection 4.4.1.

Step 1: Following the advice of Table 4.7, and the fact that 22 = 4, a good choice of
substitution here would be x = g(θ) = 2 sin(θ).

Step 2: Here, g 0 (θ) = 2 cos(θ), and so dx = 2 cos(θ)dθ. As this is a definite integral, you
must remember to change the limits. You can do this by evaluating g −1 (x) at x = 1
√
and x = 2. Now, as x = 2 sin(θ), this means that θ = sin−1 (x/2) = g −1 (x).
This means that
g −1 (1) = sin−1 (1/2) = π/6

and
√ √
g −1 ( 2) = sin−1 ( 2/2) = π/4

Step 3: Replacing each instance of x with 2 sin(θ), replacing dx by 2 cos(θ)dθ, and replacing
√
the lower limit 1 with π/6 and the upper limit 2 with π/4 gives
Z √2
1 Z π/4
2 cos(θ)
√ dx = q dθ
1 x 2 4 − x2 π/6 (2 sin(θ))2 4 − (2 sin(θ))2

q
Using the fact that 4 − (2 sin(θ))2 = 2 cos(θ) you can simplify to get

Z π/4
2 cos(θ) Z π/4
2 cos(θ)
q du = 2 dθ
π/6 (2 sin(θ))2 4 − (2 sin(θ))2 π/6 4 sin (θ) · 2 cos(θ)
Z π/4
1
= dθ
π/6 4 sin2 (θ)
1 Z π/4 2
= csc (θ) dθ
4 π/6

43
 As there are no terms involving x left in the integral, you are OK to continue and
you can now evaluate the integral.

Step 4: You can evaluate the integral using the corresponding rule from Table 4.3, which
gives
1 Z π/4 2 1 π/4
csc (θ) dθ = [− cot(θ)]π/6
4 π/6 4
√
Using the fact that cot(π/6) = 3 and cot(π/4) = 1, you can evaluate this to get
√
1 π/4 1h √ i 3−1
[− cot(θ)]π/6 = −1 − (− 3) =
4 4 4

which is the final answer.

Example 4.4.2. You are asked to find the integral 4
R
(x2 +9)3/2
dx.
Since this is an indefinite integral, you have no limits to change but you must remember to
do Step 5 of the method in Subsection 4.4.1.

Step 1: Following the advice of Table 4.7, and the fact that 32 = 9, a good choice of
substitution here would be x = g(θ) = 3 tan(θ).

Step 2: Here, g 0 (θ) = 3 sec2 (θ), and so dx = 3 sec2 (θ)dθ. As this is an indefinite integral,
you have no limits to change.

Step 3: Replacing each instance of x with 3 tan(θ), and replacing dx by 3 sec2 (θ)dθ gives
Z
4 Z
3 sec2 (θ)
dx = 4 dθ
(x2 + 9)3/2 (9 tan2 (θ) + 9)3/2

You can use the fact that tan2 (x) + 1 = sec2 (x) to simplify the integral to:
Z
3 sec2 (θ) Z
sec2 (θ)
4 dθ =12 dθ
(9 tan2 (θ) + 9)3/2 93/2 · (sec2 (θ))3/2
12 Z 1
= dθ
27 (sec (θ))1/2
2

4Z 1 4Z
= dθ = cos(θ) dθ
9 sec(θ) 9

Since there is no x left in the integral, you are OK to continue.

Step 4: Not forgetting the constant of integration, evaluating the integral using Equa-
tion 2.6 gives
4Z 4
cos(θ) dθ = sin(θ) + C (*)
9 9

44
Step 5: You now need to express your answer in terms of the original variable x. It is not
wise to use the inverse function g −1 (x) = θ here, as although it is correct, putting
this expression into a sin function will look less than pleasing. Here, you can use
the fact that x = 3 tan(θ) to write that

3 sin(θ) = x · cos(θ)

Now, you can square this expression to get that

9 sin2 (θ) = x2 cos2 (θ)

Using the fact that cos2 (θ) = 1 − sin2 (θ), you can write that

9 sin2 (θ) = x2 (1 − sin2 (θ))

Therefore, as 9 sin2 (θ) = x2 − x2 sin2 (θ), you can write that (9 + x2 ) sin2 (θ) = x2
and so
x2
sin2 (θ) =
9 + x2
√
You can square root this equation to get sin(θ) = x/ 9 + x2 and substitute this
in to the antiderivative in Equation (*) to say that
Z
4 4x
dx = √ +C
(x2 + 9) 3/2
9 9 + x2

and this is your final answer.

Finally, you can use trigonometric substitution to find expressions for antiderivatives for the
functions
a a a
√ , √ ,
b2 − x 2 x x 2 − b2 x 2 + b2

For an example, you can work out the antiderivative of √ a

x x2 −b2
by using trigonometric
substitution.

Step 1: Following the advice of Table 4.7, a good choice of substitution here would be
x = g(θ) = b sec(θ).

Step 2: Here, g 0 (θ) = b sec(θ) tan(θ) by Table 4.2, and so dx = b sec(θ) tan(θ)dθ. As this
is an indefinite integral, you have no limits to change.

45
Step 3: Replacing each instance of x with b sec(θ), and replacing dx by b sec(θ) tan(θ)dθ
gives
Z
a Z
b sec(θ) tan(θ)
√ dx = a q dθ
x x −b
2 2
b sec(θ) b2 sec2 (θ) − b2

You can use the fact that sec2 (x) − 1 = tan2 (x) to simplify the integral to:
Z
b sec(θ) tan(θ) Z
tan(θ)
a q dθ =a q dθ
b sec(θ) b2 sec2 (θ) − b2 b2 tan2 (θ)
a Z tan(θ)
= dθ
b tan(θ)
aZ
= dθ
b

Since there is no x left in the integral, you are OK to continue.

Step 4: Not forgetting the constant of integration, evaluating the integral using Equa-
tion 2.3 gives
aZ a
dθ = θ + C
b b
Step 5: You now need to express your answer in terms of the original variable x. Here, as
you only have to find an expression for θ, finding θ = g −1 (x) will do. Here, you
can use the fact that x = b sec(θ) to write that

θ = sec−1 (x/b)

And so
a a x
Z  
√ dx = sec−1 +C (4.4)
x x −b
2 2 b b

You can use similar techniques to show that

a −1 x
Z  
√ dx = a sin +C (4.5)
b2 − x 2 b

and
a a −1 x
Z  
dx = tan +C (4.6)
b2 + x 2 b b

These expressions could also be found by using the Fundamental Theorem of Calculus on the
derivatives in Table 4.6. Finally in this chapter, the information is summed up in Table 4.8.

46
function antiderivative w.r.t x

a x
Z  
−1
f (x) = √ 2 f (x) dx = a sin +C
b − x2 b
a a x
Z  
f (x) = 2 f (x) dx = tan−1 +C
b + x2 b b
a a x
Z  
f (x) = √ 2 f (x) dx = − sec−1 +C
x x − b2 b b

Table 4.8: Summary of Equations (4.4), (4.5), and (4.6)

47
Chapter 5

Integration by parts

5.1 What is integration by parts?

Say you are given the integration Z

xex dx

How would you go about evaluating it? You can’t use Equation 2.5 as x is not a constant.
Since the derivative of x is 1, you can’t use integration by substitution as x is not a constant
multiple of 1. This means a new approach has to be developed.
You can notice here that the integrand is the product of two functions, u(x) = x and v(x) =
ex . The deep connection between differentiation and integration implies that evaluating this
integral may be related to the product rule.
Chapter 3 asserted that integration by substitution is the integral version of the chain rule. In
that chapter, it was mentioned that the integral version of the product rule of differentiation
(see Table 3.1) is known as integration by parts. This is the second main technique of
finding antiderivatives of functions, and is the technique to use on xex dx.
R

This chapter is dedicated to the technique of integration by parts: where it comes from,
how to use it, and several examples detailing different ways to use it.

5.2 How integration by parts works

Let u(x) and v(x) be functions of x, and say that f (x) = u(x)v(x), the product of u(x)
and v(x). Table 3.1 says that the derivative of f with respect to x is given by the product

48
rule, which is
f 0 (x) = u(x)v 0 (x) + v(x)u0 (x)

To save time, you can write u = u(x), u0 = u0 (x), v = v(x) and v 0 = v 0 (x), and so the
product rule becomes
f 0 (x) = uv 0 + vu0

You can integrate both sides of this equation to say that

Z Z Z
0 0
f (x) dx = uv dx + vu0 dx

By the Fundamental Theorem of Calculus (Theorem 2.2.2), it follows that

Z
f 0 (x) dx = f (x) = u(x)v(x) = uv

and so Z Z
0
uv = uv dx + vu0 dx

You can rearrange this to get

Z Z
uv 0 dx = uv − vu0 dx (5.1)

This equation is the principle of integration by parts.

When you do integration by parts, you can see that you are left with an integral vu0 dx
R

of a product of two functions; the general idea is that this integral is somehow “easier” to
solve than uv 0 dx.
R

If you cannot integrate this immediately, you can use either integration by substitution or
integration by parts again until you find an integral you know you can solve. You can repeat
this as many times as you need in order to find an integral you can solve. Alternatively, the
remaining integral vu0 dx may be a constant multiple of your original integral uv 0 dx, in
R R

which case you can rearrange to find an expression for uv 0 dx in terms of anything you
R

have found.
If you are evaluating a definite integral using integration by parts, you should evaluate the
limits at the very end of your calculations of the antiderivative. This saves any confusion in
the middle of your working.

49
5.2.1 How to use integration by parts effectively

The idea is to take the product function you are asked to evaluate, name one of the functions
u and the other v 0 , and then work through the formula to find uv − vu0 dx, where hopefully
R

the integral vu0 dx is easier to integrate than uv 0 dx.

R R

The difficulty in integration by parts is in the naming of the functions at the start; that
is, which function u to differentiate and which one v 0 to integrate. As is the case with
integration by substitution, there is no general rule for choosing the functions u(x) and
v 0 (x). However, here are a few hints you can use when doing integration by parts.

• If the function you are trying to integrate is of the form xn g(x) (for n ≥ 1), the
idea is to take u = xn and v 0 = g(x) (as in Examples 5.2.1, 5.3.1 and 5.3.2). This
way, the integral vu0 dx = xn−1 v dx, which is theoretically easier to evaluate than
R R

uv 0 dx. There are exceptions to this case; particularly where you cannot integrate
R

v(x) easily (see Example 5.3.3). Sometimes, it is not easy to decide which of the
functions you can integrate; however, integrating by parts can often yield a solution
(see Example 5.3.5).

• Sometimes, you may be confronted with a function f (x) that you cannot integrate at
all, but you are able to differentiate it without much difficulty. In this case, you can
try integrating by parts with u = f (x) and v 0 = 1; this may make the integral easier
to evaluate (see Example 5.3.4).

• If you find that you do not know what to do with the integral vu0 dx, or that it looks
R

‘more complicated’ than your original integral uv 0 dx, it is best to stop immediately,
R

and change your choices of u and v 0 . If this does not make the integration simpler,
try another technique.

As with integration by substitution, there is a method that you can use to evaluate an
integral using integration by parts.

Step 1: Identify your choice of u(x) = u and v 0 (x) = v 0 in the integrand.

Step 2: Work out u0 by differentiating u with respect to x, and work out v by integrating
v 0 with respect to x (you do not need a +C for this integral).

Step 3: Write down Z Z

uv 0 dx = uv − vu0 dx

with your values for u, u0 , v and v 0 . Simplify if you can.

50
Step 4: There are a number of cases for vu0 dx, which are listed below:
R

Case 1: You can integrate vu0 dx directly or by substitution: In this case,

R

do what you need to do to find vu0 dx. You can then write your answer
R

and proceed to Step 5 (see Example 5.2.1).

Case 2: You can integrate vu0 dx by parts: In this case, you need to start the
R

process over with vu0 dx, from Step 1. Here, using different letters for
R

these functions (such as f, g or a, b) at this stage is highly recommended.

You can repeat this case as many times as you need (see Example 5.3.2).
(You will find that if you have to repeat integration by parts more than
once, you will have a lot of minus signs in your working. Take care to
make sure that your signs are correct!)
vu0 dx is a constant multiple of uv 0 dx:
R R
Case 3: Some part of the integral
In this case, you can write
Z
K uv 0 d = [uv + · · · ]

where the number of terms on the left hand side corresponds to how
many times you needed to integrate by parts. You can now proceed to
Step 5 (see Example 5.3.5 and Example 5.3.6).
Case 4: After trying everything, you do not know what to do with vu0 dx:
R

In this case, you can try reversing your choices of u and v 0 , or use a
different integration technique.
Remark. Sometimes, you are asked to do integration by parts on a function with
an unknown power n, in order to find a reduction formula. In this case (and this
case alone), you can leave an integral sign on the right hand side of the equation,
almost always after an occurrence of Case 3.

Before continuing to Step 5 you should ensure that in any of these cases (apart
from when you are finding a reduction formula), your answer on the right hand
side does not have an integral sign in it.

Step 5: If you are finding an indefinite integral, you should add the constant C on the
right hand side. If you are evaluating a definite integral between limits a and b,
you should evaluate this integral as normal.

To illustrate this method, you can use integration by parts on the following example.

51
Example 5.2.1. Suppose you are asked to find the integral 01 xex dx. This is a common
R

example of an integral you can solve by integration by parts. This is a definite integral, so
you should aim to find the complete antiderivative in Step 4 before evaluating the limits in
Step 5.

Step 1: Here, you can follow the advice at the beginning of Subsection 5.2.1 and take u = x
and v 0 = ex .

Step 2: Now, differentiating u with respect to x gives u0 = 1. Integrating v 0 = ex with

respect to x gives v = ex ; remember that you do not need the +C here.

Step 3: You can now write that

Z 1  Z 1
x x x
xe dx = x · e − 1 · e dx
0 0
 Z 1
= xex − ex dx
0

Step 4: You can integrate directly (Case 1). Integrating this gives
R x
e dx
Z
ex dx = ex

Substituting this into the above equation gives

Z 1
xex dx = [xex − ex ]10
0

which you can now evaluate as it does not have an integral sign in it.

Step 5: Here,

[x · ex − ex ]10 =(1 · e1 − e1 ) − (0 · e0 − e0 )
=1

Therefore you can write Z 1

xex dx = 1
0

and this is your final answer.

Remark. Suppose that in Step 1 in Example 5.2.1 you picked u = ex and v 0 = x; which
means that u0 = ex and v = x2 /2. Putting these terms into the integration by parts formula

52
(Equation 5.1) gives:
#1
x2 ex Z x2 ex
Z 1 "
x
xe dx = − dx
0 2 2 0

But this integral is harder to evaluate; and so this is not a good choice for u and v 0 .

Table 5.1 gives some recommended choices for integration by parts questions.

integral choice of u choice of v

Z
axn cos(kx) dx u = axn v 0 = cos(kx)

Z
axn sin(kx) dx u = axn v 0 = sin(kx)

Z
axn ekx dx u = axn v 0 = ekx

Z
axn ln(kx) dx u = ln(kx) v 0 = axn

Table 5.1: Some recommended choices for u and v 0 for common integration by parts ques-
tions

5.3 Examples

The rest of this chapter is devoted to a range of different examples involving integration by
parts.

Example 5.3.1. Suppose you are asked to evaluate −x csc2 (x) dx. This is a definite
R

integral, so you should aim to find the complete antiderivative in Step 4 before evaluating
the limits in Step 5.

Step 1: Here, you can follow the advice at the beginning of Subsection 5.2.1 and take u = x.
You can integrate − csc2 (x) (see Table 4.3) and so you can take v 0 = − csc2 (x).
It does not matter whether either u or v 0 contains the minus sign; as long as it is
not forgotten!

Step 2: Now, differentiating u with respect to x gives u0 = 1. Integrating v 0 = − csc2 (x)

53
 with respect to x gives v = cot(x) by Table 4.3; remember that you do not need
the +C here.

Step 3: You can now write that

Z  Z 
2
−x csc (x) dx = x cot(x) − cot(x) dx

Step 4: You can integrate cot(x) directly by Equation 4.2 with a = k = 1 (Case 1). Doing
this gives Z
cot(x) dx = ln | sin(x)|

Substituting this into the above equation gives

Z
−x csc2 (x) dx = [x cot(x) − ln | sin(x)|]

which does not have an integral sign in it; so you can proceed to Step 5.

Step 5: All that is left to do add the +C here; so you can write
Z
−x csc2 (x) dx = x cot(x) − ln | sin(x)| + C

and this is your final answer.

The next example details what happens when you need to integrate by parts more than
once.
R π/4
Example 5.3.2. Suppose you are asked to find the integral 0 4x2 cos(2x) dx. This is
a definite integral, so you should aim to find the complete antiderivative in Step 4 before
evaluating the limits in Step 5.

Step 1: Again you can follow the advice at the beginning of Subsection 5.2.1 and take
u = x2 and v 0 = 4 cos(2x). Here, it does not matter whether the 4 goes in u or
v 0 ; as long as it goes in one of them.

Step 2: Now, differentiating u with respect to x gives u0 = 2x. Using Equation 2.6 with
a = 4 and k = 2, you can integrate v 0 = 4 cos(2x) with respect to x to get
v = 2 sin(2x); remember that you do not need the +C here.

54
Step 3: You can now write that
Z π/4  Z π/4
4x2 cos(2x) dx = x2 · 2 sin(2x) − 2x · 2 sin(2x) dx
0 0
 Z π/4
= 2x2 sin(2x) − 4x sin(2x) dx (*)
0

Step 4: Here, you can notice that 4x sin(2x) dx is only integrable by parts (Case 2); so
R

this is what you need to do.

Step 1.1: Here, you are trying to integrate the indefinite integral 4x sin(2x) dx;
R

you can take a = x and b0 = 4 sin(2x). Notice here that you can’t use u
and v again in this working.
Step 1.2: Differentiating a = x with respect to x gives a0 = 1; integrating b0 =
4 sin(2x) with respect to x gives b = −2 cos(2x). You should be careful
with the minus signs here.
Step 1.3: Using the fact that − −2 cos(2x) dx = 2 cos(2x) dx, you can now
R R

write that
Z  Z 
4x sin(2x) dx = x · −2 cos(2x) − 1 · −2 cos(2x) dx
 Z 
= −2x cos(2x) + 2 cos(2x) dx

Step 1.4: You can integrate 2 cos(2x) directly (Case 1). Doing this gives
Z
2 cos(2x) dx = sin(2x)

Substituting this into the above equation gives

Z
4x sin(2x) dx = [−2x cos(2x) + sin(2x)]

which does not have an integral sign in it. As you are finding this integral
as part of another integral, you do not need a +C here.

You can substitute your expression for 4x sin(2x) dx into item * to get
R

Z π/4 h iπ/4
4x2 cos(2x) dx = 2x2 sin(2x) − [−2x cos(2x) + sin(2x)]
0 0
h iπ/4
= 2x2 sin(2x) + 2x cos(2x) − sin(2x)
0

55
 Finally, the right hand side of this integral does not have an integral sign in it; so
you can proceed to Step 5.

Step 5: Using the fact that 2 · π/4 = π/2, you can evaluate the limits to get
"   #
Z π/4
π 2 π
 
4x2 cos(2x) dx = 2 sin(π/2) + 2 cos(π/2) − sin(π/2) − [sin(0)]
0 4 4
" #
2
π
= + 0 − 1 − [0]
8
π2
= −1
8

and this is your final answer.

The next two examples are slightly different, and it shows that integration by parts can be
used in creative ways. They also provide antiderivatives for some common functions that
have not been found yet.

Example 5.3.3. Suppose you are asked to find the integral a ln(kx) dx. You do not know
R

to find the antiderivative of ln(x); but you know that the derivative of ln(kx) is x1 from
Table 1.1. You can use integration by parts to find the antiderivative of a ln(kx).

Step 1: As discussed above, you do not know how to integrate ln(kx); but you know how
to differentiate it. This means that you are forced to take u = ln(kx); here, you
can take v 0 = a. Notice here that if a = 1, then you can take v 0 = 1; this is
perfectly acceptable!

Step 2: Now, differentiating u with respect to x gives u0 = 1

x
by Table 1.1. Integrating
v 0 = a with respect to x gives v = ax.

Step 3: You can now write that

1
Z  Z 
a ln(kx) dx = ax ln(kx) − ax · dx
 Z
x

= ax ln(kx) − a dx

Step 4: You can integrate a dx directly to get ax (Case 1). Substituting this into the
R

above equation gives

Z
a ln(kx) dx = [ax ln(kx) − ax]

56
 As this expression does not have an integral sign in it, you can proceed to Step 5.

Step 5: Finally, you can write that

Z
a ln(kx) dx = a(x ln(kx) − x) + C (5.2)

Example 5.3.4. Suppose you are asked to find the integral sin−1 (x) dx. You do not know
R

to find the antiderivative of sin−1 (x); but you know that the derivative of sin−1 (x) is √1−x
1
2

from Table 4.6. You can use integration by parts to find the antiderivative of sin (x).
−1

Step 1: As discussed above, you do not know how to integrate sin−1 (x); but you know how
to differentiate it. This means that you are forced to take u = sin−1 (x); here, you
can take v 0 = 1.

Step 2: Now, differentiating u with respect to x gives u0 = √ 1

1−x2
by Table 1.1. Integrating
v 0 = 1 with respect to x gives v = x.

Step 3: You can now write that

" #
Z Z
1
sin−1 (x) dx = x sin−1 (x) − x · √ dx
1 − x2
" #
Z
x
−1
= x sin (x) − √ dx (**)
1 − x2

Step 4: You can integrate √ x dx using integration by substitution (Case 1).

R
1−x2

Here, you can take the substitution u = 1 − x2 . This means that u0 (x) = −2x, and
so dx = du/ − 2x. Replacing each instance of u(x) = 1 − x2 with u and replacing
dx by du/ − 2x in the integral gives
Z
x Z
x
√ dx = √ du
1−x 2 u · −2x
Z
1
= √ du
−2 u

Integrating this expression using Equation 2.2 with a = −1/2 and n = −1/2 gives
Z
1
√ du = −u1/2
−2u

57
 and substituting u = 1 − x2 back into this gives the integral as
Z
x √
√ dx = − 1 − x2
1 − x2

As you are finding this integral as part of another integral, you do not need a +C
here.
Finally, you can substitute this into item ** to get that

Z h √ i
sin−1 (x) dx = x sin−1 (x) + 1 − x2

As this expression does not have an integral sign in it, you can proceed to Step 5.

Step 5: Finally, you can write that

Z √
sin−1 (x) dx = x sin−1 (x) + 1 − x2 + C (5.3)

Remark. You can use a similar method to show that

Z
1
tan−1 (x) dx = x tan−1 (x) − ln |x2 + 1| + C (5.4)
2

The next example of integration by parts demonstrate the need for Case 3 in Step 4 of the
method in Subsection 5.2.1

Example 5.3.5. Suppose you are asked to find the integral ex cos(x) dx. Here, the choice
R

of u and v 0 is less clear, as neither choice is going to make the integral a lot easier to find.
However, you know that if you integrate cos(x) twice, you get to − cos(x); so perhaps you
can find an expression for ex cos(x) dx after integrating by parts twice.
R

Step 1: It doesn’t matter too much which function you take to be u and which function
you take to be v 0 in this case. Here, you can take u = ex and v 0 = cos(x).

Step 2: Now, differentiating u with respect to x gives u0 = ex , and integrating v 0 = cos(x)

with respect to x to get v = sin(x); remember that you do not need the +C here.

Step 3: You can now write that

Z  Z 
ex cos(x) dx = ex sin(x) − ex sin(x) dx (†)

58
Step 4: Here, you can notice that is only integrable by parts (Case 2); so this
R x
e sin(x) dx
is what you need to do.

Step 1.1: You are trying to integrate the indefinite integral ex sin(x) dx; so here
R

you can take a = ex and b0 = sin(x). Notice here that you can’t use u
and v again in this working. It is also best to be consistent with your
choices; so this means differentiating ex throughout and integrating the
other term.
Step 1.2: Differentiating a = ex with respect to x gives a0 = ex ; integrating b0 =
sin(x) with respect to x gives b = − cos(x). You should be careful with
the minus signs here.
Step 1.3: Using the fact that − − cos(x) dx = cos(x) dx, you can now write
R R

that
Z  Z 
x x x
e sin(x) dx = −e cos(x) − −e cos(x) dx
 Z 
x x
= −e cos(x) + e cos(x) dx

Step 1.4: You can notice here that the integral ex cos(x) dx is a constant multiple
R

of your original integral (Case 3). Therefore, you can return to the original
integral.

You can substitute your expression for into item † to get

R x
e sin(x) dx
Z   Z 
x x x x
e cos(x) dx = e sin(x) − −e cos(x) + e cos(x) dx
Z
=ex sin(x) + ex cos(x) − ex cos(x) dx

and so collecting the terms together gives

R x
e cos(x) dx
Z
2 ex cos(x) dx = ex (sin(x) + cos(x))

There is no integral sign on the right hand side; so you can continue to Step 5.

Step 5: Dividing by 2 on both sides and adding the constant of integration gives

Z
ex
ex cos(x) dx = (sin(x) + cos(x)) + C
2
and this is your final answer.

59
There are some instances of integration by parts where you do not know how many times
you need to do this. This is usually when the power of the function is some number n; for
instance, in the functions sinn (x) and xn ex . You can do integration by parts to find what is
called a reduction formula; this formula provides an iterative process to find the integral.

Example 5.3.6. You are asked to find a reduction formula for the integral cosn (x) dx.
R

This involves doing integration by parts. Even though this is an indefinite integral, you do
not need a +C at the end unless you are working out the integral for a given n.

Step 1: You can follow the advice at the beginning of Subsection 5.2.1; as you don’t know
how to integrate cosn (x), but you do know how to integrate cos(x), you should
take u = cosn−1 (x) and v 0 = cos(x). You can notice here that uv 0 = cosn (x).

Step 2: Now, you can use the chain rule to differentiate u with respect to x; this gives
u0 = −(n − 1) cosn−2 (x) sin(x). You can integrate v 0 = cos(x) with respect to x
to get v = sin(x); remember that you do not need the +C here.

Step 3: You can now write that

Z  Z 
n n−1 n−2
cos (x) dx = cos (x) · sin(x) − −(n − 1) cos (x) sin(x) · sin(x) dx
 Z 
n−1 n−2 2
= cos (x) sin(x) + (n − 1) cos (x) sin (x) dx

Here, you can use the fact that 1 − cos2 (x) = sin2 (x) to write that
Z  Z 
n n−1 n−2 2
cos (x) dx = cos (x) sin(x) + (n − 1) cos (x)(1 − cos (x)) dx
 Z Z 
n−1 n−2 n
= cos (x) sin(x) + (n − 1) cos (x) dx − cos (x) dx

Step 4: After expanding the brackets, you can see that the integral −(n − 1) cosn (x) dx
R

is a constant multiple of your original integral (Case 3). You can take this term
over to the left hand side and write that
Z Z
n n−1
n cos (x) dx = cos (x) sin(x) + (n − 1) cosn−2 (x) dx

Since you are finding a reduction formula, it is OK to leave the integral sign on the
right hand side and proceed to Step 5.

60
Step 5: Dividing through by n gives your final answer as
Z
n 1 n−1 n−1Z
cos (x) dx = cos (x) sin(x) + cosn−2 (x) dx
n n

You can use this reduction formula with n = 5 (and again, with n = 3 for cos3 (x) dx) to
R

say that
Z
1 4Z
cos5 (x) dx = cos4 (x) sin(x) + cos3 (x) dx
5 5
1 4 1 2Z

= cos4 (x) sin(x) + cos2 (x) sin(x) + cos(x) dx
5 5 3 3
1 4 8
= cos4 (x) sin(x) + cos2 (x) sin(x) + sin(x) + C
5 15 15

61
Chapter 6

Integration using partial fractions

6.1 What are partial fractions?

In your mathematical education, you may have learned how to add and subtract algebraic
fractions from each other. This is acheived by the technique of ‘cross-multiplication’. For
instance
2 3 2(x + 2) + 3(x − 1) 5x + 1
+ = = 2
x−1 x+2 (x − 1)(x + 2) x +x−2
Now, suppose you were asked to integrate this function. Which would you rather integrate:
Z
2 3 Z
5x + 1
+ dx or dx ?
x−1 x+2 x2 + x − 2

They are the same function, and so the integrals are the same. However, the integral on
the left seems easier to integrate than the one on the right; you can use integration by
substitution on each of the terms, and your answer should involve ln. In fact:
Z
2 3
+ dx = 2 ln |x − 1| + 3 ln |x + 2| + C
x−1 x+2

However, this answer would not seem at all plausible if you started with 5x+1
R
x2 +x−2
dx.
The idea is to develop a method to reverse the technique of cross-multiplying algebraic
fractions. This way, if you are asked to integrate a function like x25x+1
+x−2
, you can split it up
into smaller fractions that are easier to integrate. This technique is called partial fraction
decomposition. The final chapter in this course deals with the technique of integration
using partial fractions.

62
6.2 Rational functions

You may remember from a previous course that a polynomial in x is a function f (x) that
is a finite sum of powers of x. For example,

f (x) = x3 + 3x2 − x + 6

is a polynomial. If f (x) and g(x) are two polynomials with g(x) 6= 0, then the function
h(x) = fg(x)
(x)
is called a rational function. You should be confident with how to factorise
and manipulate polynomial functions before continuing with this chapter of the course.
There are two types of polynomial that will appear as factors in the denominators of partial
fractions.

• A linear polynomial is a function of the form ax + b, for some real numbers a, b

where a is non zero.

• A quadratic polynomial is a function of the form ax2 +bx+c, for some real numbers
a, b, c with a not equal to 0.

You may have been asked to find roots of quadratic polynomials in the past. If a quadratic
polynomial f (x) has a real root, then it can be written as f (x) = (x − a)(x − b), where a
and b are the real roots of the polynomial. If a quadratic polynomial g(x) has no real roots
(so it has complex roots), it cannot be written in this way. If this happens, you can call the
quadratic equation g(x) irreducible.
In this course, you will only be asked to consider partial fraction decomposition for rational
functions h(x) = fg(x)
(x)
where the polynomial g(x) in the denominator has a larger power in
it than the polynomial f (x) in the numerator. In addition to this, you will only be asked
to perform partial fraction decomposition on partial fractions whose denominator can be
factorised into one of the following three cases:

(1) Fractions containing distinct, linear factors in the denominator: These include
functions like
3x − 3 −x2 − 1
and
(x − 4)(x + 2) (x − 5)(x + 1)(x − 2)

(2) Fractions containing distinct, possibly repeated linear factors in the denomi-
nator: These include functions like

2x + 1 −5x + 2
and
(x − 1)2 (x − 2)(x + 1)2

63
(3) Fractions containing distinct linear factors and a distinct or repeated irre-
ducible quadratic factor: These include fractions like

2x − 1 −4x + 1
and
(x2 + 1)(x − 1) (x2 + 9)2

To split rational functions into partial fractions, you need to:

• write the denominators into separate fractions with unknown numerators, then;

• then solve a series of equations to find the numerators.

You need to be careful when writing the denominators into separate fractions, as you do
not want to lose any information. To do this, you could follow the following guidelines.

• For every linear factor, repeated n times in your denominator, you need one
partial fraction for each power less than or equal to n in your expansion; so

A1 A2 An
+ + ... +
(ax + b) (ax + b) 2 (ax + b)n

This means that for every distinct linear factor in your denominator, you need a
term that looks like
A
(ax + b)

• For every irreducible factor, repeated n times in your denominator, you need one
partial fraction for each power less than or equal to n in your expansion; so

A1 x + B1 A2 x + B2 An x + Bn
+ + ... +
2 2
(ax + bx + c) (ax + bx + c)2 (ax2 + bx + c)n

This means that for every distinct irreducible factor in your denominator, you need
a term that looks like
Ax + B
(ax2 + bx + c)

In this course, the factors will be repeated no more than twice.

Table 6.1 gives a brief list of common types of partial fractions you may encounter, and a
suggested decomposition for them.
To find the unknown numerators of a partial fraction decomposition of a rational function
f (x)
g(x)
, you could follow these steps:

64
 rational function looks like partial fractions look like

f (x) A B
+
(ax + b)(cx + d) (ax + b) (cx + d)
f (x) A1 A2 B
+ +
(ax + b)2 (cx + d) (ax + b) (ax + b)2 (cx + d)
f (x) A B C
+ +
(ax + b)(cx + d)(mx + n) (ax + b) (cx + d) (mx + n)
f (x) Ax + B D
+
(ax2 + bx + c)(mx + n) 2
(ax + bx + c) (mx + n)
f (x) Ax + B D1 D2
+ +
(ax2 + bx + c)(mx + n)2 (ax + bx + c) (mx + n) (mx + n)2
2

f (x) A1 x + B1 A2 x + B2 D
+ +
(ax2 + bx + c)2 (mx + n) 2 2
(ax + bx + c) (ax + bx + c)2 (mx + n)

Table 6.1: Common layouts for partial fraction decomposition

Step 1: Add together the fractions in your partial fraction decomposition. If necessary,
cancel terms on top and bottom so the denominator of the added fractions is the
same as the original rational function fg(x)
(x)
.

Step 2: As the denominators are the same and the fractions are equal, the numerators
f (x) and your expression for unknowns are the same. Write out these numerators
again.

Step 3: For every (x − a) factor on the expression involving unknowns, set x = a and solve
the resulting equation. This should give you a solution to the unknown. If you
find all the unknowns, you can stop here.

Step 4: For each remaining unknown, you need to generate an equation for that unknown.
For instance, two remaining unknowns require two equations.
You can generate these by letting x = n1 , n2 , ... where each of these n’s is not
equal to some a used in Step 3. You should then get a system of equations, which
you can solve to find the remaining unknowns.

Let’s see how this works in practice.

65
Example 6.2.1. Suppose you are asked to decompose

−3x2 + 8x − 1
(x − 1)2 (x2 + 1)

into partial fractions. You can use the guidelines above to write your decomposition as

−3x2 + 8x − 1 A1 A2 Bx + D
= + + 2
(x − 1) (x + 1)
2 2 (x − 1) (x − 1)2 (x + 1)

in an attempt to split the left hand side into partial fractions. You now need to find the
numerators A1 , A2 , B and B; you can do this using the method above.

Step 1: You need to add together the partial decomposition back into a fraction. This can
be done by the usual method of adding fractions. So this becomes

−3x2 + 8x − 1 A1 A2 Bx + D
= + +
(x − 1)2 (x2 + 1) (x − 1) (x − 1)2 (x2 + 1)
A1 (x − 1)2 (x2 + 1) + A2 (x − 1)(x2 + 1) + (Bx + D)(x − 1)3
=
(x − 1)3 (x2 + 1)

At this stage, you need to make sure the denominators are the same. You can
notice there is an (x − 1) common to each of the terms, and an (x − 1)3 term on
the bottom. This means that you can write

−3x2 + 8x − 1 A1 (x − 1)2 (x2 + 1) + A2 (x − 1)(x2 + 1) + (Bx + D)(x − 1)3

=
(x − 1)2 (x2 + 1) (x − 1)3 (x2 + 1)
A1 (x − 1)(x2 + 1) + A2 (x2 + 1) + (Bx + D)(x − 1)2
=
(x − 1)2 (x2 + 1)

Step 2: Now the denominators are the same, the numerators must also be the same. So
you can write

−3x2 + 8x − 1 = A1 (x − 1)(x2 + 1) + A2 (x2 + 1) + (Bx + D)(x − 1)2

Step 3: To find some of the numerators, you can set each factor to equal 0 in turn. For
instance, setting x = 1 throughout this equation gives

−3 + 8 − 1 = 0 + A2 (2) + 0 + 0

and so 4 = 2A2 , giving A2 = 2. There are no more ways to set a linear factor

66
 equal to 0 here, so you should proceed to Step 4.

Step 4: You have three numerators left to find in this case, and so you need to generate
three equations to find them. So setting x = 0, x = −1 and x = 2 (notice that
you can’t set x = 1 here) gives

−1 = − A1 + 2 + D (x = 0)
−12 = − 4A1 + 4 − 4B + 4D (x = −1)
3 =5A1 + 10 + 2B + D (x = 2)

Solving this gives A1 = −1, B = 1, D = 4. So this means that

−3x2 + 8x − 1 −1 2 x−4
= + + 2
(x − 1) (x + 1)
2 2 (x − 1) (x − 1) 2 (x + 1)

Don’t worry; it is unlikely that you will be asked to find anything so complicated on a
tutorial sheet or exam! This example is just to illustrate the ways in which you can find the
numerators. You can find some easier examples in the integrations that follow in the next
section.

6.3 Integration using partial fractions

The two most common types of partial fraction you are likely to integrate are of the form

a a
and
bx + c (bx + c)n

where n 6= 1 in the second term.

You can integrate both of these expressions this using the substitution u = bx + c. Differ-
entiating this gives u0 (x) = b, and so dx = du/b.
Integrating the first term here gives
Z
a Z
a aZ 1 a
dx = du = du = ln |u| + C
bx + c b·u b u b

and so Z
a a
dx = ln |bx + c| + C (6.1)
bx + c b

67
Now, integrating the second term here gives
Z
a Z
a a Z −n au−n+1
dx = du = u du = +C
(bx + c)n b · un b b(−n + 1)

and so
Z
a a(bx + c)1−n
dx = +C (6.2)
(bx + c)n b(1 − n)
for n 6= 1. You can use these results throughout the chapter. In addition to these, you also
need to be able to integrate by substitution effectively; see Chapter 3 for more details.
Note: In this examples section, all integrals considered are indefinite. If you are asked to
find a definite integral using partial fractions, you should evaluate the limits after calculating
the antiderivative; in a similar fashion to integration by parts.

Example 6.3.1. You are asked to find the integral x25x+1 dx. Here, you can factorise
R
+x−2
the denominator to get
x2 + x − 2 = (x + 2)(x − 1)

This means that the integral becomes

Z
5x + 1 Z
5x + 1
dx = dx
x +x−2
2 (x + 2)(x − 1)

You can use the technique of partial fraction decomposition to find this integral. Using the
recommended layout for the partial fraction decomposition as seen in Table 6.1, you can
write
5x + 1 A B
= +
(x + 2)(x − 1) x+2 x−1
Adding these fractions gives

5x + 1 A(x − 1) + B(x + 2)
=
(x + 2)(x − 1) (x + 2)(x − 1)

As the denominators are the same, the numerators are the same and so you can write

5x + 1 = A(x − 1) + B(x + 2) (*)

Here, you can set each factor to be equal to 0 in order to find A and B. Setting x = 1 and
substituting it into Equation * gives 6 = 3B, and so B = 2. You can now set x = −2 and

68
substitute into Equation * to give −9 = −3A, and so A = 3. This means that

5x + 1 3 2
= +
(x + 2)(x − 1) x+2 x−1

and this is the same answer as in the introduction. Following this, the integral becomes
Z
5x + 1 Z
3 2
dx = + dx
x +x−2
2 x+2 x−1

You can see from the introduction that

Z
5x + 1 Z
3 2
dx = + dx = 3 ln |x + 2| + 2 ln |x − 1| + C
x +x−2
2 x+2 x−1

and this is your final answer.

Example 6.3.2. You are asked to find

Z
−3x2 + 8x − 1
dx
(x − 1)2 (x2 + 1)

You can split the integrand into partial fractions... which you have already done in Exam-
ple 6.2.1. So this means the integral is
Z
−3x2 + 8x − 1 Z
−1 2 x−4
dx = + + 2 dx
(x − 1) (x + 1)
2 2 (x − 1) (x − 1)2 (x + 1)

You can integrate the first term using Equation 6.1 with a = −1 = c and b = 1 to get
Z
−1
dx = − ln |x − 1|
x−1

Integrating the second term using Equation 6.2 with a = 2, b = 1, c = −1 and n = 2 gives
Z
2 2
dx = −
(x − 1) 2 (x − 1)

You cannot integrate the third term directly. In this case, you have to split the integral into
two separate parts. You can do this by saying that
Z
x−4 Z
x Z
−4
2
dx = 2
dx + 2
dx
x +1 x +1 x +1

You can integrate the first of these terms using integration by substitution. Taking u =

69
x2 + 1, you can work through the method of integration by substitution to get
Z
x 1
dx = ln |x2 + 1|
x2 +1 2

In the second of these integrals, you can use Equation 4.6 with a = −4 and b = 1 to get
Z
−4
dx = −4 tan−1 (x)
x2+1

Collecting together these four antiderivatives, you can write that

Z
−3x2 + 8x − 1 2 1
dx = − ln |x − 1| − + ln |x2 + 1| − 4 tan−1 (x) + C
(x − 1)2 (x2 + 1) (x − 1) 2

not forgetting the +C!

Example 6.3.3. You are asked to find

Z
2x2 + x − 7
dx
(x − 2)(x + 1)(x − 1)

You can do this using partial fractions. Following the recommended layout for the partial
fraction decomposition as seen in Table 6.1, you can write

2x2 + x − 7 A B D
= + +
(x − 2)(x + 1)(x − 1) x−2 x+1 x−1

Adding these fractions gives

2x2 + x − 7 A(x + 1)(x − 1) + B(x − 2)(x − 1) + D(x − 2)(x + 1)

=
(x − 2)(x + 1)(x − 1) (x − 2)(x + 1)(x − 1)

As the denominators are the same, the numerators are the same and so you can write

2x2 + x − 7 = A(x + 1)(x − 1) + B(x − 2)(x − 1) + D(x − 2)(x + 1) (†)

Here, you can set each factor to be equal to 0 in order to find A, B and D. So, setting
x = 2 and substituting this into Equation † gives 3 = A(3)(1), and so A = 1. You can
now set x = −1 and substitute this into Equation † to get −6 = B(−3)(−2) = 6B, and
so B = −1. Finally, you can set x = 1 and substitute this into Equation † to get that
−4 = D(−1)(2) = −2D, and so D = 2. This means that

2x2 + x − 7 1 1 2
= − +
(x − 2)(x + 1)(x − 1) x−2 x+1 x−1

70
Following this, the integral becomes
Z
2x2 + x − 7 Z
1 1 2
dx = − + dx
(x − 2)(x + 1)(x − 1) x−2 x+1 x−1

You can use Equation 6.1 to integrate each of these terms to say that
Z
2x2 + x − 7 Z
1 1 2
dx = − + dx
(x − 2)(x + 1)(x − 1) x−2 x+1 x−1
= ln |x − 2| − ln |x + 1| + 2 ln |x − 1| + C

Now, let’s look at an example with a repeated linear factor.

Example 6.3.4. You are asked to evaluate

Z
7x2 − 25x − 2
dx
(x − 3)2 (2x + 1)

You can do this using partial fractions. Following the recommended layout for the partial
fraction decomposition as seen in Table 6.1, you can write

7x2 − 25x − 2 A1 A2 B
= + +
(x − 3) (2x + 1)
2 (x − 3) (x − 3)2 2x + 1

Adding these fractions gives

7x2 − 25x − 2 A1 (x − 3)2 (2x + 1) + A2 (x − 3)(2x + 1) + B(x − 3)3

=
(x − 3)2 (2x + 1) (x − 3)3 (2x + 1)

Here, the denominators are not the same. However, there is an (x − 3) term on the top
and bottom of the right hand side, so you can cancel to get

7x2 − 25x − 2 A1 (x − 3)(2x + 1) + A2 (2x + 1) + B(x − 3)2

=
(x − 3)2 (2x + 1) (x − 3)2 (2x + 1)

So as the denominators are the same, the numerators are the same and therefore you can
write
7x2 − 25x − 2 = A1 (x − 3)(2x + 1) + A2 (2x + 1) + B(x − 3)2 (‡)

Here, you can set each factor to be equal to 0 in order to find A1 , A2 and B. So, setting
x = 3 and substituting this into Equation ‡ gives

63 − 75 − 2 = 0 + A2 (7) + 0

71
giving −14 = 7A2 ; so A2 = −2. You can now set x = −1/2 and substitute this into
Equation ‡ to get
7 25 49
+ −2= B
4 2 4
Multiplying through by 4 on both sides gives

7 + 50 − 8 = 49B

giving 49 = 49B; this means B = 1. There are no other factors you can set to be 0 in
order to find A1 . However, you only need to generate one more equation. So here, you can
set x = 0 and substitute this into Equation ‡ to get that −2 = A1 (−3)(1) − 2 + 9, and so
−9 = −3A1 , giving A1 = 3. This means that

7x2 − 25x − 2 3 2 1
= − +
(x − 3) (2x + 1)
2 (x − 3) (x − 3)2 2x + 1

Following this, the integral becomes

Z
7x2 − 25x − 2 Z
3 2 1
dx = − + dx
(x − 3)2 (2x + 1) (x − 3) (x − 3)2 2x + 1

You can use Equation 6.1 to integrate the first and third of these terms, and Equation 6.2
to integrate the second to get
Z
7x2 − 25x − 2 Z
3 2 1
dx = − + dx
(x − 3) (2x + 1)
2 (x − 3) (x − 3) 2 2x + 1
2 1
=3 ln |x − 3| + + ln |2x + 1| + C
x−3 2

Example 6.3.5. Suppose you are asked to find the integral

Z
x2 + 3x + 2
dx
x(x2 + 1)

You can use the technique of partial fraction decomposition to find this integral. Using the
recommended layout for the partial fraction decomposition as seen in Table 6.1, you can
write
x2 + 3x + 2 A Bx + D
2
= + 2
x(x + 1) x x +1

72
Adding these fractions gives

x2 + 3x + 2 A(x2 + 1) + (Bx + D)x

=
x(x2 + 1) x(x2 + 1)

As the denominators are the same, the numerators are the same and so you can write

x2 + 3x + 2 = A(x2 + 1) + (Bx + D)x (F)

Here, you can set one factor to be equal to 0 in order to find A. Setting x = 0 and
substituting it into Equation F gives 2 = A. There are no other factors you can set to be
0 in order to find B and D. As you need to find two unknowns, you need to generate two
equations. So taking x = 1 and x = −1 and substituting these into Equation F gives the
simultaneous equations

6 = 4 + (B + D) (x = 1)
0 = 4 − (−B + D) (x = −1)

Adding these two equations together cancels the D’s to give 6 = 8 + 2B; so B = −1. This
means that D = 3. You can therefore write:

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

So the integral becomes

Z
x2 + 3x + 2 Z
2 −x + 3
2
dx = + 2 dx
x(x + 1) x x +1

The first of these you can integrate directly by Equation 6.1 to get
Z
2
dx = 2 ln |x|
x

However, the second term is not integrable directly. Like in Example 6.3.2, you can split
the integral up into two parts to get
Z
−x + 3 Z
−x Z
3
2
dx = 2
dx + 2
dx
x +1 x +1 x +1

You can integrate the first of these using integration by substitution (with u = x2 + 1). The

73
integrand of the second term is the derivative of 3 tan−1 (x) (see Equation 4.6); and so you
can write (with some working omitted):
Z
−x + 3 1
2
dx = − ln |x2 + 1| + 3 tan−1 (x)
x +1 2

At last, you can write your answer as

Z
x2 + 3x + 2 1
dx = 2 ln |x| − ln |x2 + 1| + 3 tan−1 (x) + C
x(x2 + 1) 2

6.3.1 And finally...

You can use the technique of partial fractions to find the as yet unconsidered antiderivative
of a sec(kx). You know that sec(x) = cos(x)
1
and so you can write
Z Z
a Z
1
a sec(kx) dx = dx = a dx
cos(kx) cos(kx)

Multiplying top and bottom of this fraction by cos(kx) gives

Z
cos(kx)
a dx
cos2 (kx)

As cos2 (kx) = 1 − sin2 (kx), you can write

Z Z
cos(kx)
a sec(kx) dx = a dx
1 − sin2 (kx)

Now, you can integrate this by substitution by taking u = sin(kx). Using the chain rule, it
follows that u0 (x) = k cos(kx) and so dx = du/k cos(kx). So the integral becomes
Z
cos(kx) Z
cos(kx)
a 2 dx =a du
1 − sin (kx) (1 − u2 ) · k cos(kx)
aZ 1
= du
k 1 − u2

As 1 − u2 = (1 − u)(1 + u), you can integrate 1

1−u2
using partial fractions. Following the
advice in Table 6.1, you can write

1 A B
= +
1−u 2 1−u 1+u

74
Adding these fractions together gives

1 A(1 + u) + B(1 − u)
=
1−u2 1 − u2

As the denominators are the same, the numerators are the same and so you can write

1 = A(1 + u) + B(1 − u) ()

Here, you can set each factor to be equal to 0 in order to find A and B. Setting x = 1 and
substituting it into Equation  gives 1 = 2A, and so A = 1/2. You can now set x = −1
and substitute into Equation  to give 1 = −2B, and so B = −1/2. This means that

1 1 1
= −
1−u 2 2(1 − u) 2(1 + u)

Following this, the integral becomes

aZ 1 aZ 1 1 aZ 1 1
du = − du = − du
k 1 − u2 k 2(1 − u) 2(1 + u) k 2(u + 1) 2(u − 1)

Integrating gives

aZ 1 a a
du = ln |u + 1| − ln |u − 1| + C
k 1−u2 2k 2k

Combining the two logarithms using the laws of logs gives

aZ 1 a u+1
du = ln +C
k 1 − u2 2k u−1

Substituting u = sin(x) back into this equation gives the integral as

Z
a sin(x) + 1
a sec(kx) dx = ln +C
2k sin(x) − 1

While this is a perfectly valid answer, you could find a way to simplify this. Here, multiplying
top and bottom of the term inside the natural log by sin(x) + 1 and simplifying gives

(sin(x) + 1) · (sin(x) + 1) (sin(x) + 1)2

=
(sin(x) − 1) · (sin(x) + 1) (sin2 (x) − 1)

75
As sin2 (x) − 1 = − cos2 (x), you can write

sin(x) + 1 (sin(x) + 1)2

=
sin(x) − 1 − cos2 (x)

Putting this expression into the natural log gives

Z
a (sin(x) + 1)2
a sec(kx) dx = ln +C
2k − cos2 (x)

You can use the properties of the absolute value to get rid of the minus sign in the denom-
inator, giving that
Z
a (sin(x) + 1)2
a sec(kx) dx = ln +C
2k cos2 (x)
Using the laws of logarithms, you can cancel the 1
2
by taking the square root of everything
inside the logarithm, writing that
Z
a (sin(x) + 1)
a sec(kx) dx = ln +C
k cos(x)

The final step in the course is to simplify the fraction to get

Z
a
a sec(kx) dx = ln |tan(x) + sec(x)| + C (6.3)
k

76