Complex Numbers

1 Enlarging sets to solve equations

Suppose we start off with the set of natural numbers

N = {1, 2, 3, 4, . . .}

which are the standard counting numbers. Using these numbers we can solve some equations:

x+2=7

but not others:

x + 8 = 5.

We can’t solve this last equation, because its solution requires a negative number, and our
starting set N doesn’t have negative numbers. So what we need to do is to enlarge our set to
include negative numbers. Our new set is the set of integers:

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

and now we can certainly solve more equations. But not all! What about

3x + 4 = 9?

Well, to solve this equation we need fractions. So—enlarge the set yet again, to obtain the set
of rational numbers:
 
p
Q= , p, q ∈ Z, q 6= 0 .
q
Even with this new set, we still can’t solve all equations; for example:

x2 − 2 = 0.
√
The solution for this is x = 2, and it can be shown that this is not a rational number. So we
need the irrationals as well, and the new set which includes all the rationals and irrationals is
the set of real numbers. We can think of the real numbers as all possible numbers which can be
expressed as finite or infinite decimals.
The real numbers are a fine and dandy set of numbers, and we need them for lots of things
in our mathematical lives. But are they good enough for solving equations? For example:

x2 + 4 = 0.

This can’t be solved using the reals, because it can be written as

x2 = −4

and the real numbers have the property that all squares are positive.
You know the drill now—if we can’t solve an equation with the numbers we have, enlarge
the set to include some more numbers to help us. The trouble is that we’ve gone as far as we
can using the symbols 0, 1, 2, 3, 4, 5, 6, 7, 8, 9. So if we are to have any new numbers, we’re going
to have some new symbols for them. In fact we only need one new number, denoted i
and which satisfies

i2 = −1.

1
You see why we need a new symbol—no real number can satisfy this property, and the real
numbers already use all the digit symbols. So here is a new number.
We now define the complex numbers to be

C = {a + bi, a, b ∈ R}.

We can now solve the pesky equation above; noticing that

(2i)2 = (22 )(i2 )

= (4)(−1)
= −4

and so a solution to the equation is x = 2j.

The remarkable thing about complex numbers is that in terms of solving equations, we need
go no further: every polynomial equation can be solved using complex numbers.
So in terms of enlarging sets of numbers to include more numbers, complex numbers represent
the end of the line.

2 Definitions and arithmetic

Real and imaginary parts
Here are some examples of complex numbers:
√ 2
2 + 3i, 4− 3i, −π + i, 4, −25i
7
We notice that a complex number has two parts (one of which may be zero, like in the last two
examples above), a bit with an “i” next to it, and a bit without. In fact, if we have

z = x + yi

then x is called the real part of z, and is denoted

x = Re(z).

The value y is called the imaginary part of z, and is denoted

y = Im(z).

For example, if

z = 3 + 5i

then

Re(z) = 3, Im(z) = 5.

Equality
Two complex numbers are equal if and only if both their real parts and imaginary parts are
equals. In symbols:

w = z ⇐⇒ Re(z) = Re(w) and Im(z) = Im(w).

2
Complex conjugate
For any complex number

z = a + bi

its complex conjugate, denoted z, is defined to be

z = a − bi.

So, for example, if

z = 4 + 7i

then

z = 4 − 7i.

And if

z = 8 − 3i

then

z = 8 + 3i.

Notice that if you add a number and its complex conjugate, the imaginary parts cancel out and
you end up with a real number.

Arithmetic: addition and subtraction

These are easy: to add or subtract complex numbers, we just add or subtract their real and
imaginary parts. For example:

(2 + 3i) + (7 + 4i) = (2 + 7) + (3 + 4)i

= 9 + 7i

and

(5 − 2i) − (3 − 6i) = (5 − 3) + (−2 − (−6))i

= 2 + (−2 + 6)i
= 2 + 4i

Arithmetic: multiplication
To multiply two complex numbers, just multiply out the brackets, and at the end substitute −1
for i2 :

(2 + 5i)(3 − 2i) = 2(3 − 2i) + 5i(3 − 2i)

= 6 − 4i + 15i − 10i2
= 6 + 11i − 10(−1)
= 6 + 11i + 10
= 16 + 11i.

3
Arithmetic: division
We have noted above the that the sum of a complex number z and its complex conjugate z is
real. But what about their product? Let’s try an example:

(2 + 5i)(2 − 5i) = 2(2 − 5i) + 5i(2 − 5i)

= 4 − 10i + 10i − 15i2
= 4 − 15(−1)
= 19.

Note that the i’s again cancel out. To prove that this always happens, try the same thing with
symbols:

(a + bi)(a − bi) = a(a − bi) + bi(a − bi)

= a2 − abi + abi − b2 i2
= a2 − b2 (−1)
= a2 + b2 .

We have just shown that for any complex number z, that zz is real. We use this fact to perform
division; given
w
z
multiply both top and bottom by z. Here’s an example:
13 − 18i (13 − 18i)(2 + 5i)
=
2 − 5i (2 − 5i)(2 + 5i)
116 + 29i
=
29
= 4 + i.

And another example:

7 − 5i (7 − 5i)(3 − 2i)
=
3 + 2i (3 + 2i)(3 − 2i)
11 − 29i
=
13
11 29
= − i.
13 13

Arithmetic: square roots

Square roots are one of those things we can’t perform always within the set of real numbers, since
all squares of real numbers are positive, we “can’t take the square root of a negative number”.
(In fact we can, we just need the complex number to do it!). But there are no such restrictions
on complex numbers, and every complex number has a square root which is another complex
number. Calculating square roots requires a bit of algebraic fiddling, (see 5.1 for more formal
solution). Here we will make an educated guess based on the following considerations:
Here’s an example; to find the square root of

z = 16 + 30i.

Suppose that the square root we want is

w = x + yi.

4
Then

w2 = (x + yi)2 = (x2 − y 2 ) + 2xyi = 16 + 30i.

Equating real and imaginary parts:

x2 − y 2 = 16
2xy = 30.

Therefore we are looking for x and y such that:

30
• their product xy = 2 ,

• and the difference of their squares x2 − y 2 = 16

Clearly a pair 3, 5 fits the bill. Now we can write:

w2 = 16 + 30i = 25 + 30i − 9 = 25 + 30i + 9i = (5 + 3i)2

Thus the square roots we require are

w = 5 + 3i, w = −(5 + 3i) = −5 − 3i.

3 The Argand diagram and polar form

The real numbers can be represented as a line stretching off in both directions. A single line isn’t
enough for the complex numbers; we need a plane, on which a complex number corresponds to
a point. On this plane, the horizontal axis is called the real axis and the vertical axis is called
the imaginary axis. A point (x, y) on this plane corresponds to the complex number x + yi.
Here’s a picture of the plane, with a few numbers on it:
yi

−1 + 3i
3i

2i 3 + 2i

x
−4 −3 −2 −1 1 2 3 4

−i

−2i
−2 − 2i

−3i
2 − 3i

5
Such a picture, of numbers plotted on the complex plane, is also called an Argand diagram1
Given a plot of complex numbers, we can introduce the polar form, which involves the
distance from the origin, and the angle to the positive x-axis:

x + yi

In order to obtain relationships between the values x, y, r and θ, first redraw the above diagram
as a right-angled triangle:

r
y

x
By Pythagoras’ theorem, we have

r 2 = x2 + y 2

so that
p
r= x2 + y 2 .

And by trigonometric definitions:

x = r cos θ
y = r sin θ
tan θ = y/x

The value r is called the absolute value or the modulus of the complex number z = x + yi and
is denoted |z|. The value θ is called the argument of z, and is denoted arg(z).
From the above equations, we can write

x + yi = r cos θ + ri sin θ
= r(cos θ + i sin θ)

This last expression is the polar form of the complex number z = x + yi. The original “x + yi”
form is called the Cartesian form of z. Because the expression cos θ + i sin θ occurs so often in
discussing complex numbers, it is often abbreviated to cis θ, so the polar form can be written as
r cis θ.

1
Named after the French mathematician Jean-Robert Argand (1768–1822).

6
 There is another way of writing the complex number in polar form. And this is because cis θ
is an exponential function. In fact (see 5.2 Euler formula):

cos θ + i sin θ = eiθ

and so we can express any complex number as an exponential:

z = reiθ

where r and θ are the modulus and argument, as before. This “new form” of a complex num-
ber is sometimes called its exponential form. In fact, it isn’t really a new form, it’s just a new
way of writing the polar form, but using an exponential function instead of the abbreviation cis θ.

If we consider the complex number z = −1, which has r = 1 and θ = π, the exponential
form produces
eiπ + 1 = 0
which is known as Euler’s formula.
From now on we shall use the exponential version of the form, alternative usage of cis θ is
shown in the appendix.

Convert Cartesian to polar

Let’s consider a complex number z = (x, y) that is customarily written as z = x + yi.

We have seen that the number can be represented in a polar form as z = (r, θ) and written
as
z = reiθ .
θ is the principal argument z - the unique value of the argument that is in the range −π < θ ≤ π
sometimes denoted as Arg z
And now:

p
r = |z| = x2 + y 2
y
Let β = arctan  
 
x
Then θ value depends on a quadrant of z.


 β for z in the first quadrant
π − β for z in the second quadrant

θ=

 β − π for z in the third quadrant
−β for z in the fourth quadrant


For example, suppose

z = −2 + 2i.

Then
p √ √ √
r= (−1)2 + (2)2 = 4 + 4 = 8 = 2 2.

 
−1 
 2 π
β = tan − 2  = 4


7
 Since x < 0 and y > 0 then z is located in the second quadrant (You can check this by
drawing a diagram.)
π 3π
θ=π− = .
4 4
Thus:
3π
√ i
z= reiθ = 2 2e 4

Example
Show that:
√ π
1. If z1 = 1 + i then z1 = 2ei 4
√ 3π
2. If z2 = −1 − i then z2 = 2e−i 4
√ 2π
3. if z3 = −1 + 3 then z3 = 2ei 3
√ √ π
4. If z4 = 2 − i 2 then z3 = 2e−i 4
Solution:
1.
z1 = 1 + i, x = 1 and y = 1
p p √
r = x2 + y 2 = 12 + 1 2 = 2
y π
arctan   =
 
x 4
π
As z1 is located in the first quadrant θ =
 π 4
i π4 ◦
z1 = e 1, or (1, 45 )
4
2.
z2 = −1 − i, x = −1 and y = −1
p q √
r = x + y = (−1)2 + (−1)2 = 2
2 2
y π
arctan   =
 
x 4
π 3π
As z2 is located in the third quadrant θ = − π = −
4 4
√ −i 3π √ √
 
3π  
z3 = 2e 4 2, − or 2, −150◦
4

3.
√ √
z3 = −1 + i 3, x = −1 and y = 3
√ 2 √
p q
r = x + y = (−1)2 + ( 3) = 4 = 2
2 2
y π
arctan   =
 
x 3
π 2π
As z3 is located in the second quadrant θ = π - =
  3 3
2π 2π
z3 = 2ei 3 2,
3

8
 4.
√ √ √ √
z4 = 2 − i 2, x = 2 and y = − 2
√ 2 √ 2 √
p q
2 2
r = x + y = ( 2) + (− 2) = 2 + 2 = 2
y π
arctan   =
 
x 4
π
As z4 is located in the fourth quadrant θ = -
4
−i π4
 π
z4 = 2e 2, −
4

Convert any angle to the principle argument

As both sin x and cos x are periodic functions with the period of 2π then

eiθ = ei(θ+2πn) where n = 0, −1, 1, −2, 2 . . .

For example:
3π 3π 3π 27π
ei 4 = ei( 4
+2π·3)
= ei( 4
+6π)
= ei 4

3π 27π
Both ei 4 and ei 4 refer to the same point −0.707 + 0.707i on the Argand diagram.

Example
77π
Find the principle argument for ei 3

Solution

The principle argument will be in a form

kπ
θ=
3
, where k is an integer. We are looking for two integer numbers k and n such as:
kπ 77π
+ 2πn =
3 3
k + 2n · 3 = 77

Step 1 : Find both n and k.2

77/(2 · 3) = 12 ← integer division
k = 77 − 12 · 6 = 5
5π
θ=
3
Step 2 :If θ is outside −π < θ 6 π perform one extra turn:
5π π
θ= − 2π = −
3 3
Answer:
77π π
ei 3 = e−i 3
2
Please note that majority of scientific calculators can help you finding the answer for 77/6 by changing display
mode to fraction display 77/6=12/5/6=12r5. Please check the functionality of your calculator. In case of negative
numbers interpret the remainder as negative as well. Be careful of fraction simplification. For example 1234/18=
68/5/9 so the answer is 68r10.

9
Try it by yourself
Find the principle argument for:

177π
ei 5

and
317π
e−i 4

Arithmetic with polar forms

The polar form of a complex number can be used to provide very powerful and efficient methods
for performing multiplication and division, powers and n-th roots.

Multiplication with polar forms

That is, if
z = reiθ
w = seiφ
then
zw = rseiθ eiφ = rsei(θ+φ) .

This means that the product of two complex numbers can be obtained (in polar form) by mul-
tiplying their moduli, and adding their arguments.

For example, let’s try to compute

(−2 + 2j)(1 + j).
√
The first number has modulus 8 and argument 3π/4, so that its polar form is
√ i 3π
8e 4
It is easy to work out that the polar form of the second number is
√ iπ
2e 4 .
So their product is
√ i 3π √ i π √ √ 3π π
8e 4 2e 4 = 8 2ei( 4 + 4 )
√
= 16eiπ
= 4(−1 + 0j)
= −4.

Division with polar forms

You can probably guess what going to happen here: to perform a division we divide the moduli,
and subtract the arguments. That is, if
z = reiθ
w = seiφ
then
z r
= ei(θ−φ) .
w s

10
Powers with polar forms
z = reiθ
z n = (reiθ )n = rn einθ

Or written in more traditional form as de Moivre’s theorem,

(cos(θ) + i sin(θ))n = cos(nθ) + i sin(nθ)

Example
 −99
1+i
Express √
2
in a Cartesian x + iy form.
√ π
From previous example we know that:1 + i = 2ei 4
 
1+i π
z= √ = ei 4
2
 π −99 99π 3π 3π
z −99
= ei 4 = e−i 4 = ei(− 4 −2π·12) = e−i 4

Converting to Cartesian form:

   
−i 3π 3π 3π
z=e 4 = cos + i sin = −0.707 + 0.707i
4 4

Example
Express

4
√

9
3 + 43 i
3
z= √ √
19
2−i 2
in a Cartesian x + iy form.

Let’s start with the numerator:

4√
 
4
z1 = 3+ i
3 3
s
4√ 2
  2
4 8
r= 3 + =
3 3 3
 
1 π
θ = arctan √ =
3 6
9 9 227 iπ

8 iπ 8 i9π
z19 = e6 = 9 e 6 = 9 e− 2
3 3 3
And denominator:
√ √ 
z2 = 2−i 2
r 
√ 2 √ 2
r= 2 + 2 =2
π π
β = arctan (1) = as the number is in the fourth quarter, we have θ = −
4 4
iπ 19
  i19π 3π 3π
z219 = 2e− 4 = 219 e− 4 = 219 ei(− 4 −2π·2) = 219 e−i( 4 )

11
 Combining the answers, we have:
27 iπ
2 −2
z1 9 e 28 − iπ +i 3π 28 i π
z= = 3 = e 2 4 = e 4
z2 2
3π
19 e−i( 4 ) 39 39
√ √
28 2 2
z = 9( + i) ≈ 0.009(1 + i)
3 2 2

n-th roots with polar forms

The first thing to note is that every complex number z has n distinct n-th roots. So z will have
two square roots, three distinct cube roots, and so on.
Let’s consider all complex numbers in the form of:

z = rei(θ+k2π) (k = 0, 1, 2, . . . , n − 1)

Now we can find roots of z by performing basic algebraic operations on z:

1 1
z n = (rei(θ+k2π) ) n
1 1 θ+2kπ
z n = r n ei n (k = 0, 1, 2, . . . , n − 1)

Example
Find all cube roots of 8
We will be looking for 3 solutions. Convert 8 to polar form. It’s principle argument is 0.

z = 8 = 8ei(0) = 8ei(0+2kπ) , (k = 0, 1, 2)
1 1 2kπ
z 3 = 8 3 ei 3
 1
 (8) 3 ei(0) = 2, k = 0
1

1 (0+2kπ) 2π √
(8) 3 ei 3 = 2ei 3 = 2 cos 2π + sin 2π


z3 = 3 3 = −1 + i 3, k = 1
 1 (0+2·2π) 4π 2π √
(8) 3 ei 3 = 2ei 3 = 2e−i 3 = 2 cos − 2π 2π




3 + sin − 3 = −1 − i 3, k = 2

12
 Im z

k=1

k=0
1 Re z

k=2

Example
Find all roots of
 1
−1 − i 5
√
2

 
−1 − i 3π 3π
5
z = √ = e−i 4 = e−i( 4
+2kπ ) , k = 0, 1, 2, 3, 4
2

1  3π
1 3+8k
z 5 = e−i( 4 +2kπ) = e−iπ( 20 ) , k = 0, 1, 2, 3, 4
5

3

e−iπ 20 = cos − 3π 3π



20 + i sin − 20 = 0.89 − 0.45i, k = 0


−iπ ( 3+8·1
20 ) = e
11
−iπ ( 20 ) = −0.15 − 0.98i, k = 1


 e


1
3+8·2 19
z5 = e−iπ( 20 ) = e−iπ( 20 ) = −0.98 − 0.15i, k = 2
3+8·3 27 13
e−iπ( 20 ) = e−iπ( 20 ) = eiπ( 20 ) = −0.45 + 0.89i, k = 3




 3+8·4 35 5
e−iπ( 20 ) = e−iπ( 20 ) = eiπ( 20 ) = 0.71 + 0.71i, k = 4



13
 Im z

k=3 i
k=4

1 Re z
k=2

k=0

k=1

4 Applications of complex numbers

Complex numbers crop up all over the place in both pure and applied mathematics. We shall
just investigate some simple applications.

Solution of quadratic equations

This is where we started! Quadratics can be solved using the quadratic formula, noting that
any negative square roots can now be managed easily with complex numbers.
For example:
x2 + 6x + 13 = 0.
The quadratic formula produces
p
−6 ± (6)2 − 4(1)(13)
x=
√ 2
−6 ± 36 − 52
=
√2
−6 ± −16
=
2
This is the point where complex numbers come in. Since (4i)2 = 16i2 = −16, we can continue
our working as:
−6 ± 4i
x=
2
= −3 ± 2i.

14
This is the solution we require.
The same can be achieved by the completion of the square method.
z 2 + 6z + 13 = 0
(z + 3)2 − 9 + 13 = 0
(z + 3)2 = −4
√
z + 3 = ± −4 = ±2i
z = −3 ± 2i
There’s nothing in the formula which requires the coefficients a, b and c to be real. We can
use the formula to solve equations such as:
x2 − (1 + 7i)x + (11i − 16) = 0.
We can apply the quadratic formula to this, using
a=1
b = −1 − 7i
c = 11i − 16
Here’s the formula:
p
1 + 7i ± (−1 − 7i)2 − 4(1)(11i − 16)
x=
√ 2
1 + 7i ± −48 + 14i − 44i + 64
=
√ 2
1 + 7i ± 16 − 30i
=
2
Using the method presented earlier, we can find that
√
16 − 30i = 5 − 3i.
Then the solution can be continued as
1 + 7i ± (5 − 3i)
x=
2
1 + 7i + (5 − 3i) 1 + 7i − (5 − 3i)
= ,
2 2
6 + 4i −4 + 10i
= ,
2 2
= 3 + 2i, −2 + 5i.

Graphs on the Argand plane

We can draw circles, ellipses and hyperbolas using Cartesian coordinates. For example, a circle
with centre (a, b) and radius r can be described by
(x − a)2 + (y − b)2 = r2 .
Similarly, an ellipse can be described by
(x − a)2 (y − b)2
+ =1
A2 B2
and a hyperbola by
(x − a)2 (y − b)2
− = ±1.
A2 B2

15
Circles
Complex numbers provide a very neat way of describing such curves. Since for any complex
number z, its modulus |z| is its distance from the origin, then the distance between any two
points w and z on the Argand plane is given by |w − z|.
Suppose w = a + bi. Then the circle with centre w and radius r consists of all complex
numbers z for which
|z − w| = r.
As for any z = x + iy we have:

|z − w| = r
q
|x + iy − a − bi| = |(x − a) + i (y − b)| = (x − a)2 + (y − b)2
q
(x − a)2 + (y − b)2 = r
(x − a)2 + (y − b)2 = r2

Lines
Example
What is represented by all z satisfying Im(z(2 − i))?

As previously z = x + iy. So we have:

(x + iy)(2 − i) = 2x − ix + 2iy + y = 2x + y + i(2y − x)
Im(z(2 − i)) = 2y − x = 5
x 5
Therefore, we get an equation of a line: y = +
2 2

Example
 
z +i
z − 3 = 1
 

Using the same technique as above:

|z + i| = |z − 3|
|x + i(1 + y)| = |(x − 3) + iy|
q q
x2 + (1 + y)2 = (x − 3)2 + y 2
x2 + 1 + 2y + y 2 = x2 − 6x + 9 + y 2
2y = −6x + 8
y = −3x + 4

yi

16
Try by yourself
Determine the shape of:
 
z + i
z − i = 2
 

Hint: it is not a line.  

5 3
Answer: a circle in a centre of 0, and radius of .
3 4
Note: This is a classical problem of Circles of Apollonius. Wikipedia explains: Apollonius
showed that a circle can be defined as the set of points in a plane that have a specified ratio of
distances to two fixed points known as foci. In the previous example the ratio is 1, therefore we
get a line.

Ellipses (Optional)
An ellipse can be described by two points, called foci, such that the sum of the distance to the
foci is constant. For example, suppose we have the ellipse
x2 y 2
+ = 1.
9 4
√ √
It can be shown that the foci are at points (− 5, 0) and ( 5, 0), and the common distance sum
to these foci is 6:

a
b

a+b=6

This ellipse can be described in the Argand plane as being the locus of all points z for which
√ √
|z − 5| + |z + 5| = 6.

That can be proven algebraically, although it is very tedious. The following outlines the
major steps for those who would like to follow.

√ √
|x − 5 + iy| + |x + 5 + iy| = 6.

17
 q √ q √
(x − 5) + y + (x + 5)2 + y 2 = 6.
2 2

q √ q √
(x − 5) + y = 6 − (x + 5)2 + y 2 .
2 2

Squaring both sides and simplifying the sides separately, we get:

p 2 2
√ 2 2
q √
5 − 2x (5) + x + y = 41 + 2 5x + x + y − 12 (x + 5)2 + y 2 )

After further simplifications and dividing by both sides by 4, we have:

√ q √
9 + x 5 = 3 (x + 5)2 + y 2 )
√
We need to square both sides and simplify again to get rid of all 5:

4x2 + 9y 2 = 36

The nice thing about this approach is that we can draw ellipses at all sorts of angles. For
example:

|z − (1 + j)| + |z + i| = 3

looks like this:

1+i

−i

Here’s how to sketch an ellipse given by the equation

|z − w| + |z − v| = s

18
where w and v are complex numbers, and s is a real number. First, draw w and v on an Argand
plane and draw a line through them. Find the midpoint (w + v)/2 and draw a perpendicular
line through it. These lines will be the axes of the ellipse. Determine f = |w − v| which is the
distance between w and v. Then determine the values

a = s/2
p
b = ( s2 − f 2 )/2

These are the lengths of the ellipse in the directions of the focal line and perependicular line.
One more example:

|z − (2 + j)| + |z − (−1 + 2i)| = 5.

p2 + i and v = −1 + 2i and√the line between them, and determine their distance

First, draw w =
f = |w − v| = (2 − (−1))2 + (1 − 2)2 = 10:

√
10
−1 + 2i

2+i

Next find the middle point m = (w + v)/2 = 0.5 + 1.5i and draw a perpendicular line through
it:

−1 + 2i
m

2+i

Now compute

a = s/2 = 2.5,
p √ √
b = ( s2 − f 2 )/2 = ( 25 − 10)/2 = 15/2 ≈ 1.936.

19
So we measure 2.5 from the centre along the focal line, and 1.936 in each direction along the
perpendicular line:

1.936
2.5

2.5

w
1.936

Finally, draw the ellipse to fit the four new outer points:

20
Trigonometric identities
Recall from earlier that we used the identities:

sin(a + b) = sin a cos b + sin b cos a

cos(a + b) = cos a cos b − sin a sin b

to establish that multiplying in polar form corresponded to adding the arguments. We can go
the other way, and use the polar form of complex numbers to establish trigonometric identites.
For example, we have

(cos 3θ + i sin 3θ) = (cos θ + i sin θ)3

by de Moivre’s formula. We can expand the right hand side using the binomial theorem:

(cos θ + i sin θ)3 = cos3 θ + 3i cos2 θ sin θ + 3i2 cos θ sin2 θ + i3 sin3 θ

and simplify:

(cos θ + i sin θ)3 = (cos3 θ − 3 cos θ sin2 θ) + i(3 cos2 θ sin θ − 3 sin3 θ).

If we equate real and imaginary parts, we have the two identities:

cos 3θ = cos3 θ − 3 cos θ sin2 θ

sin 3θ = 3 cos2 θ sin θ − 3 sin3 θ

Clearly this approach can be used to generate identities for cos nθ and sin nθ for any n.
We can also perform what’s called power reduction, in which a power of a trigonometric
function can be represented as a sum of cos kθ and sin kθ for various values of k. For example,
let’s see if we can express sin4 θ as a sum.
Start with

eiθ = cos θ + i sin θ

e−iθ = cos θ − i sin θ

so that if we subtract these equations we obtain:

eiθ − e−iθ
sin θ =
2i
and if we add them we obtain:
eiθ + e−iθ
cos θ = .
2
Raising the sin θ expression to the fourth power, and expanding the right hand side using the
binomial theorem produces

ei4θ − 4e3iθ e−iθ + 6e2iθ e−2iθ − 4eiθ e−3iθ + e−4iθ

sin4 θ = .
16
Since

eiθ e−iθ = 1

the huge numerator above can be first simplified by removing all powers of eiθ e−iθ , to produce:

ei4θ − 4ei2θ + 6 − 4e−2iθ + e−4iθ .

21
Collecting like terms produces:

ei4θ + e−4iθ − 4(ei2θ + e−2iθ ) + 6

which can be written as

2 cos 4θ − 8 cos 2θ + 6.

Finally dividing by 16:

cos 4θ − 4 cos 2θ + 3
sin4 θ = .
8
This sort of identity can be used for integrating expressions involving powers of sin θ and cos θ.

Integration
Prove that:

eax
Z
I= eax cos(bx)dx = (a cos(bx) + b sin(bx))
a2 + b2
This could be done by integration by parts, twice, but is messy and fiddly. Here’s a way with
complex numbers. Main idea:

cos(bx) = Re(eibx )
Z Z Z  Z 
ax ax ibx ax ibx x(a+ib)
I = e cos(bx)dx = e Re(e )dx = Re e e dx = Re e dx

Assuming that normal integration rules apply to complex numbers we have:

Z
1
ex(a+ib) dx = ex(a+ib) + C
a + ib
1 (a − ib) (a − ib)eax
ex(a+ib) = ex(a+ib) = (cos(bx) + i sin(bx)) =
a + ib (a + ib)(a − ib) a2 + b2
eax
a cos(bx) + ia sin(bx) − ib cos(bx) − i2 b sin(bx) =


2
a +b 2
eax
((a cos(bx) + b sin(bx)) + i(a sin(bx) − b cos(bx)))
a2 + b2
Taking Re() from the above expression we get:

eax
I= (a cos(bx) + b sin(bx)) + C
a2 + b2

Example
Evaluate the integral
Z
e3x sin 4x dx.

This time we will use Im().

e3x sin 4x = Im(e3x ei4x ) = Im(e(3+4i)x ).

22
Then the integral is
Z Z
(3+4i)x
Im(e ) = Im e(3+4i)x dx
 
1
= Im e(3+4i)x + C .
3 + 4i

The above expression can now be rewritten:

(3 − 4i)(e(3+4i)x )
32 + 42
for which the imaginary part (which is all we want) is
1
(3e3x sin 4x − 4e3x cos 4x).
25
So we have found that
Z
1
e3x sin 4x dx = (3e3x sin 4x − 4e3x cos 4x) + C.
25

Exercises
1. Compute the complex number corresponding to each of the following:
(i) (3 + 4j) + (6 − 3j) (ii) 3(5 − j) − 2(−1 + 2j) (iii) (2 + j)(3 − 2j)
(iv) (4 − 3j)(−2 + 5j) (v) (17 + 6j)/(4 − 3j) (vi) (1 + j)/(2 − 5j)
2. Compute each of the following:
(i) (1 − 2j)2 , (1 + 2j)2 (ii) (4 + 3j)2 , (4 − 3j)2 (iii) (5 − 3j)2 , (5 + 3j)2
Can you guess at a general rule about squares and squares of conjugates? Can you prove
your rule?

3. Compute the square roots of

(i) −5 + 12j (ii) 15 − 8j (iii) 33 − 56j

4. Sketch each of the following on an Argand diagram:

√ 1 1
(i) −1 + j (ii) 1 − 3j (iii) √ −√ j
√ 2 2
(iv) 3+j (v) −3 − 3j (vi) −4
and determine the polar form of each one.

5. Using the polar forms from the previous questions, compute the following powers:
14
√

1 1
(i) (−1 + j)7 (ii) (1 − 3j)6 (iii) √ −√ j
√ 2 2
(iv) ( 3 + j)9 (v) (−3 − 3j)5 (vi) (−4)2
6. Find
(i) all the cube roots of −2 + 2j, (ii) all the fourth roots of −4.
In each case sketch all roots on an Argand diagram.

7. Find all√the roots, provide answer in Cartesian

√ form: √
3 4 3
(i) 27 (ii) 16 (iii) −8
√ √3
p
4
√
(iv) i (v) 2i − 2 (vi) 8i 3 − 8

23
 iπ
8. Let w = e 4 . Find all solutions to z 3 = w2019 . Give the answer in polar form.

9. Solve the following quadratic equations using complex numbers:

(i) x2 − 4x + 5 = 0 (ii) x2 + 2x + 5 = 0 (iii) x2 + 6x + 10 = 0
(iv) x2 − 2x + 17 = 0 (v) x2 + 10x + 26 = 0 (vi) x2 − 10x + 41 = 0

10. (Optional (iv)-(vi) Sketch the curves defined by:

(a) |z| = 2 (ii) |z − (2 − 3j)| = 1 (iii) |z + j| = 2
(iv) |z| + |z + j| = 3 (v) |z +1|+|z +j| = 2 (vi) |z − (1 + j)| + |z + (1 + j)| = 3

11. (Optional) Express the following trigonometric expressions as powers of sin θ and cos θ:
(i) sin 4θ (ii) cos 4θ (iii) sin 2θ cos 2θ
(iv) sin 3θ cos 3θ (v) sin 4θ cos 2θ (vi) sin 2θ cos 4θ

12. (Optional) Perform the following power reductions:

(a) cos4 θ (ii) sin2 θ cos2 θ (iii) sin θ cos4 θ
(iv) sin2 θ cos3 θ (v) sin3 θ cos2 θ (vi) cos5 θ

Answers
1. (i) 9 + j, (ii) 17 − 7j, (iii) 8 − j, (iv) 7 + 26j, (v) 2 + 3j, (vi) −(3/29) + (7/29)j

2. (i) −3 − 4j, , −3 + 4j, (ii) 7 + 24j, 7 − 24j, (iii) 16 − 30j, 16 + 30j

In general the square of a conjugate is equal to the conjugate of the square.

3. (i) 2 + 3j, (ii) 4 − j, (iii) 7 − 4j

√ 3π π π √
4. (i) 2ei 4 , (ii) 2e−i 3 , (iii) e−i 4 , (iv) 2e π6 , (v) 3 2 cis(−3π/4), (vi) 4eiπ

5. (i) −8 − 8j, (ii) 64, (iii) j, (iv) −512j, (v) 972 + 972j, (vi) 16
√ √ √ √
1+ 3 1+ 3 1 + 3 −1 + 3
6. (i) − − j, − + j, 1 + j, (ii) ±1 ± j
2 2 2 2
√ √ 1+i −1−i
7. (i) 3, 23 (−1 + i 3), 23 (−1 − i 3) (ii) 2, 2i, −2, −2i, (iii) −2, 1 ± i (3), (iv) √
p
, √ , (v)
(2) (2)
p
1 + i, −1.366 + 0.366i, 0.366 − 1.366i, (vi) ±(sqrt(3) + i), ±(1 − i )
5π 17π 9π
8. e−i 33 , ei 33 , e−i 11

9. (i) 2 ± j, (ii) −1 ± 2j, (iii) −3 ± j, (iv) 1 ± 4j, (v) −5 ± j, (vi) 5 ± 4j

1
0.5 1

0.5

0 0.5
0

10. (iv) -0.5 (v) -0.5 (vi) 0

y

-1
-1 -0.5

-1.5

-1.5 -1
-2

-2.5 -2 -1.5
-2 -1.5 -1 -0.5 0 0.5 1 1.5 2 -2 -1.5 -1 -0.5 0 0.5 1 -1.5 -1 -0.5 0 0.5 1 1.5
x x x

11. (i) 4 cos3 θ sin θ − 4 cos θ sin3 θ, (ii) sin4 θ − 6 cos2 θ sin2 θ + cos4 θ, (iii) 2 cos3 θ sin θ −
2 cos θ sin3 θ, (iv) 3 cos θ sin5 θ−10 cos3 θ sin3 θ+3 cos5 θ sin θ, (v) 4 cos θ sin5 θ−8 cos3 θ sin3 θ+
4 cos5 θ sin θ,
(vi) 2 cos θ sin5 θ − 12 cos3 θ sin3 θ + 2 cos5 θ sin θ

24
12. (i) (cos(4 θ) + 4 cos(2 θ) + 3)/8, (ii) (1 − cos(4 θ))/8,
(iii) (sin(5 θ) + 3 sin(3 θ) + 2 sin θ)/16, (iv) (− cos(5 θ) − cos(3 θ) + 2 cos θ)/16,
(v) (− sin(5 θ) + sin(3 θ) + 2 sin θ)/16, (vi) (cos(5 θ) + 5 cos(3 θ) + 10 cos θ)/16

25
5 Appendix
5.1 Arithmetic: square roots
Square roots are one of those things we can’t perform always within the set of real numbers, since
all squares of real numbers are positive, we “can’t take the square root of a negative number”.
(In fact we can, we just need the complex number to do it!). But there are no such restrictions
on complex numbers, and every complex number has a square root which is another complex
number. Calculating square roots requires a bit of algebraic fiddling, and solving some quadratic
equations. Here’s an example; to find the square root of

z = 16 + 30i.

Suppose that the square root we want is

w = x + yi.

Then

w2 = (x + yi)2 = (x2 − y 2 ) + 2xyi = 16 + 30i.

Equating real and imaginary parts:

x2 − y 2 = 16
2xy = 30.

From the last equation we have y = 15/x which can be substituted into the first equation to
obtain
 2
2 15
x − = 16.
x

Multiplying out by x2 produces

x4 − 225 = 16x2

or

x4 − 16x2 − 225 = 0

We can think of this equation as a quadratic in x2 :

(x2 )2 − 16(x2 ) − 225 = 0.

Solving this using the quadratic formula produces two possible solutions

x2 = −9, x2 = 25.

Since both x and y have to be real, we can discard the first value, and obtain

x = ±5

which produces

y = ±3.

Thus the square roots we require are

w = 5 + 3i, w = −5 − 3i.

26
5.2 Euler Formula
Euler formula (via series expansion)

eiθ = cos(θ) + i sin(θ)

We know that:
∞
x2 x3 x4 x5 X xn
ex = 1 + x + + + + + ··· =
2! 3! 4! 5! n!
n=0

Therefore:
(iθ)2 (iθ)3 (iθ)4 (iθ)5 (iθ)6 (iθ)7
eiθ = 1 + (iθ) + + + + + + + ...
2! 3! 4! 5! 6! 7!
i1 = i; i2 = −1; i3 = −i; i4 = 1; i5 = i; . . .
θ2 iθ3 θ4 iθ5 θ6 iθ7
eiθ = 1 + iθ − − + + − − + ...
2! 3! 4! 5! 6! 7!
cos x
z }| {
θ 2 θ 4 θ 6 x2 x4 x6
eiθ = 1 − + − + ... cos(x) = 1 − + − + ...
2! 4! 6! 2! 4! 6!
θ3 θ5 θ7 x3 x5 x7
 
+i θ − + − + ... sin(x) = x − + −
3! 5! 7! 3! 5! 7!
| {z }
sin x

5.3 Arithmetic with polar forms (cis version)

The polar form of a complex number can be used to provide very powerful and efficient methods
for performing multiplication and division, and powers and n-th roots.

Multiplication with polar forms

Suppose we start with two complex numbers and their polar forms:

z = x + yi = r(cos θ + i sin θ)
w = u + vi = s(cos φ + i sin φ)

Let’s now attempt to multiply their polar forms:

zw = r(cos θ + i sin θ)s(cos φ + i sin φ)

= rs(cos θ cos φ + i cos θ sin φ + i sin θ cos φ + i2 sin θ sin φ)
= rs(cos θ cos φ − sin θ sin φ + i(cos θ sin φ + sin θ cos φ))

You may recognize the expressions in the last bracket as the addition formulas for cos and sin.
Thus:

zw = rs(cos(θ + φ) + i sin(θ + φ)).

This means that the product of two complex numbers can be obtained (in polar form) by
multiplying their moduli, and adding their arguments.
For example, let’s try to compute

(−2 + 2i)(1 + i).

27
 √
The first number has modulus 8 and argument 3π/4, so that its polar form is
√
    
3π 3π
8 cos + i sin
4 4
It is easy to work out that the polar form of the second number is
√ π 
2 cis .
4
So their product is
√ 3π √ π  √ √
   
3π π
8 cis 2 cis = 8 2 cis +
4 4 4 4
√
= 16 cis(π)
= 4(−1 + 0i)
= −4.

Division with polar forms

You can probably guess what going to happen here: to perform a division we divide the moduli,
and subtract the arguments. That is, if
z = r cis θ
w = s cis φ
then
z r
= cis(θ − φ).
w s

Powers with polar forms

From the formula for multiplication, we can deduce that
z 2 = r2 cis(2θ)
which can of course be generalized to any power:
z n = rn cis(nθ).
This is based on an important result for complex numbers called de Moivre’s theorem3 , which
says that
(cos θ + i sin θ)n = cos(nθ) + i sin(nθ).
For example, suppose we wish to calculate
(1 + i)11 .
The polar form of 1 + i is
√ π 
2 cis
4
and so its 11-th power is
√ 11 √
      
11π 11π 11π
2 cis = 32 2 cos + i sin
4 4 4
√
 
1 1
= 32 2 − √ + i √
2 2
= 32(−1 + i)
= −32 + 32i
3
Named after the French mathematician Abraham de Moivre (1667–1754).

28
n-th roots with polar forms
The first thing to note is that every complex number z has n distinct n-th roots. So z will have
two square roots, three distinct cube roots, and so on.
Here is a formula for generating all the n-th roots, starting with

z = r cis θ.

Then:
 
1/n 1/n θ + 2kπ
z =r cis for k = 0, 1, 2, . . . , n − 1.
n

For example, let’s try to calculate all the cube roots of 8i. We start with its polar form, which
can be easily found to be
π 
8 cis .
2
Applying the formula above produces
π 
1/3 1/3 2 + 2kπ
(8i) = 8 cis for k = 0, 1, 2.
3

We can now see what happens for each different value of k on the right hand side of this formula.
Putting k = 0 produces
π 
2 + 0kπ
π 
1/3
8 cis = 2 cis
3 6
 π   π 
= 2 cos + i sin
6 ! 6
√
3 1
=2 +i
2 2
√
= 3 + i.

Now put k = 1:
π   
1/3 2 + 2π 5π
8 cis = 2 cis
3 6
    
5π 5π
= 2 cos + i sin
6 6
√ !
3 1
=2 − +i
2 2
√
= − 3 + i.

And finally k = 2:
π   
1/3 2 + 4kπ 3π
8 cis = 2 cis
3 2
    
3π 3π
= 2 cos + i sin
2 2
= 2(0 − i)
= −2i.

29
If we draw all these numbers on an Argand diagram:

√ √
− 3+i 3+i

−2i

we see that they are all evenly distributed around circle with radius 2. And this is the case for
any group of n-th roots—they are always evenly distributed around a circle.
One more example,the fifth roots of 1 + i, can be found to have approximate values

1.0586 + 0.168i, 0.168 + 1.059i, −0.955 + 0.487i, −0.758 − 0.758i, 0.487 − 0.955i

and here they are on an Argand diagram:

0.168 + 1.059i

−0.955 + 0.487i

1.0586 + 0.168i

−0.758 − 0.758i

0.487 − 0.955i

30
5.4 Geometry
Here are two simple example of geometry with complex numbers. First, if z, w and t are three
complex numbers, then the area of that triangle in the Argand plane is given by
1 

Im (z − w)(z − t) 
2
We saw an example when we were looking at determinants: the triangle with vertices (1, 3),
(2, −1) and (−3, −2). Now we have three complex numbers:

z = 1 + 3i, w = 2 − i, t = −3 − 2i.

Then:

(z − w)(z − t) = (−1 + 4i)(4 + 5i)

= (−1 + 4i)(4 − 5i)
= 19 + 21i.

The absolute value of the imaginary part is 21, and so the triangle has area
1
21 = 10.5.
2
Second, we can rotate any obiect in the Argand plane by an angle θ by multiplying all points
(considered as complex numbers), by cos θ + i sin θ. Again, use the example from matrices,
and the house with vertices at (1, 0), (3, 0), (3, 1), (2, 2) and (1, 1) rotated through 60◦ . Using
complex numbers:
   
1 0.5 + 0.866i
 3   1.5 + 2.598i 
   
(0.5 + 0.866i)  3 + i  =  0.634 + 3.098i 
  

 2 + 2i   −0.732 + 2.732i 
1+i −0.366 + 1.366i

These final numbers are the vertices of the rotated house in the complex plane.

5.5 Solution of cubic equations

Historically, this is the context in which the theory of complex numbers was first developed,
about 450 years ago—to find a method which can be used to solve cubic equations. There is a
“cubic formula” which in its general form is very complicated. Here is one version of it, to solve
a class of cubics of the form

x3 + 3px + 2q = 0.

(It is possible to transform any cubic equation to remove the x2 term, so this class of equation
isn’t as restrictive as it looks.) Anyway, compute
 p 1/3
A = −q + q 2 + p3
 p 1/3
B = −q − q 2 + p3

Then the three solutions are

√
1 3
x = A + B, − (A + B) ± i (A − B).
2 2

31
For example:

x3 − 6x − 6 = 0.

Here p = −2 and q = −3 and so

 p 1/3
A = 3 + (−3)2 + (−2)3
 √ 1/3
= 3+ 1
= (4)1/3
 p 1/3
B = −3 − (−3)2 + (−2)3
 √ 1/3
= 3− 1
= (2)1/3

This means that one of the three solutions is

21/3 + 41/3 .

The other two solutions are

√
1 1/3 1/3 3 1/3
− (2 + 4 ) ± i (4 − 21/3 ).
2 2
These messy looking expressions can be simplified further.
Quadratic, cubic and quartic equations can all be solved analytically, using powers and n-th
roots, but fifth degree equations can’t be—there is no general formula for solving polynomial
equations with degrees greater than 4. Solutions can be found by approximate methods though,
and if the coefficients are real the solutions will always exist as complex conjugate pairs.
For example, the solutions of

x5 + x2 − 1 = 0

can be found to be the five complex numbers

0.80873, −0.86928 ± 0.38827j, 0.46491 ± 1.07147j

to five decimal places.

5.6 Pythagorean triples

A Pythagorean triple consists of three integers A, B and C for which A2 +B 2 = C 2 . For example,
(A, B, C) = (3, 4, 5) is a Pythagorean triple because (3)2 + (4)2 = (5)2 . Complex numbers can
be used to generate Pythagorean triples.
Recall the “difference of two squares” factorization:

X 2 − Y 2 = (X − Y )(X + Y ).

We are going to factorize A2 + B 2 by turning it into a difference:

A2 + B 2 = A2 − (−B 2 )
= A2 − (iB)2
= (A − iB)(A + iB).

32
Since we want A2 + B 2 = C 2 , we have:

C 2 = (A − iB)(A + iB).

Since the expressions on the right hand side are different, but their product is a square, this
must mean that each one is a square itself. In other words, we can write

A + iB = (x + iy)2

for some integers x and y. Expanding produces:

A + iB = (x2 − y 2 ) + i(2xy)

or

A = x2 − y 2
B = 2xy.

Then

A2 + B 2 = (x2 − y 2 )2 + (2xy)2
= (x4 − 2x2 y 2 + y 4 ) + 4x2 y 2
= x4 + 2x2 y 2 + y 4
= (x2 + y 2 )2 .

What we have shown is that for any integers x and y, the values

A = x2 − y 2
B = 2xy
C = x2 + y 2

form a Pythagorean triple: A2 + B 2 = C 2 . Here’s some values of x and y and the values of A,
B and C they generate:

x y A B C
2 1 3 4 5
3 1 8 6 10
3 2 5 12 13
4 1 15 8 17
4 3 7 24 25

33