NKay Study

ams 10/10A supplementary notes ucsc

A Primer on Complex Numbers

c 2009, Yonatan Katznelson

1. Imaginary and complex numbers.

One of the fundamental properties of the real numbers is that the square of a real
number is always nonnegative. I.e., if x is a real number, then x2 ≥ 0. This implies,
among other things, that certain quadratic equations don’t have real solutions. In
particular, the equation x2 = −1, has no solution in real numbers. This doesn’t
mean however, that the equation cannot have a solution. √
By the 16th century, it became apparent to various mathematicians that −1
would be very useful and this new number, and its multiples, entered slowly into
common use. As with √ all important constants, a special symbol was eventually
designated to represent −1.

Definition 1.
An imaginary number is a number of the form bi, where b is real and
√
(1.1) i = −1.

With the set of imaginary numbers in hand, we can find a square root for every real
number, positive or negative. Indeed, following the usual algebraic rules, we have

(bi)2 = b2 · i 2 = b2 · (−1) = −b2 ,

p
for any real number b. Now, if α < 0 and b = |α| (remember, |α| > 0 so it has a
real square root), then −b2 = α, so
√
bi = α.
√ √ √ √
Thus, for example, −9 = 3i, −100 = 10i and −2 = 2 · i.

More is true. By combining real and imaginary numbers, we can solve any qua-
dratic equation.
Example 1.1. Solve the equation x2 + 2x + 2 = 0. Using the quadratic formula,
we find the two solutions
√ √
−2 + 4 − 8 −2 − 4 − 8
z1 = = −1 + 2i and z2 = = −1 − 2i.
2 2
Note that the two solutions are neither real numbers, nor are they purely imaginary.
1

Nana Kwabena
 NKay Study

Definition 2.
A complex number is √ a number of the form z = a + bi, , where a and b are
real numbers, and i = −1. The (real) numbers a and b are called the real and
imaginary parts of z, respectively, and we often use the notation
a = Re(z) and b = Im(z).

If Re(z) = 0, then z is an imaginary number and if Im(z) = 0, then z is a real

number. In other words, the real numbers and the imaginary numbers are subsets
of the complex numbers. The boldfaced letter R is used to denote the set of real
numbers and the boldfaced letter C is used to denote the set of all complex numbers.
The two solutions of the quadratic equation in Example 1.1 have the same real
part and their imaginary parts are opposite. I.e., using the notation above, we have
Re(z1 ) = Re(z2 ) and Im(z1 ) = −Im(z2 ). This is not a coincidence, and there is
also a name for this.

Definition 3.
The complex conjugate of a + bi is the number a − bi. We use a bar over the
number to denote the conjugate, i.e., a + bi = a − bi.

If a, b and c are real numbers, and b2 − 4ac < 0, then the quadratic equation
(1.2) ax2 + bx + c = 0
has no real solutions. But, as in Example 1.1, there are always complex solutions
and these solutions come in conjugate pairs. Namely, the solutions to equation (1.2)
are given by α + βi and α − βi, where (from the quadratic formula)
p
b |b2 − 4ac|
α=− and β = .
2a 2a
Observe that if z = a = a + 0 · i is a real number, then z̄ = a − 0 · i = a = z.
In other words, real numbers are their own complex conjugates. Another thing to
note is that z̄¯ = z for any complex number z (check!), that is, the conjugate of the
conjugate is the original number. We’ll see a geometric interpretation of these two
observations a little later.
Comments:
a. Complex numbers first appeared explicitly in the work of the 16th century Italian
mathematician Cardano. The term ‘imaginary number’ was introduced later by
Descartes who did not think highly of the concept. His opinion notwithstanding,
complex numbers have important real applications, as we will see.
b. Introducing a ‘new’ number as the solution of an equation that didn’t already
have a solution among the accepted set of numbers was not new, even in the 16th

Nana Kwabena
 NKay Study

3
√
century. The irrational number 2 was introduced around 2500 years ago, when
it became apparent that the equation x2 = 2 had no rational solutions.

Exercises
1.1. Find the solutions of the equation 2x2 + 4x + 5 = 0.
1.2. Find the real and imaginary parts of the solutions of the equation x2 +3x+5 = 0.
1.3. Show that z = z for any complex number z.

2. Complex arithmetic.
Complex numbers may be added and multiplied, just like real numbers, and the
usual properties (commutativity, associativity and distributivity) continue to hold.
2.1 Addition and subtraction.
To add two complex numbers, we add their real and imaginary components sep-
arately and use the distributive rule bi + di = (b + d)i for the imaginary parts. In
other words,
(a + bi) + (c + di) = (a + c) + (b + d)i.
Subtraction is defined similarly:
(a + bi) − (c + di) = (a − c) + (b − d)i.

Example 2.1. If z1 = 3 + 2i, z2 = 4 − 3i and z3 = −3 + 5i, then

z1 + z2 = (3 + 2i) + (4 − 3i) = (3 + 4) + (2 − 3)i = 7 − i

and
z2 − z3 = (4 − 3i) − (−3 + 5i) = (4 + 3) + (−3 − 5)i = 7 − 8i.
Addition of complex numbers satisfies the usual rules of commutativity and asso-
ciativity. I.e., if u, v and w are complex numbers then
u+v =v+u and (u + v) + w = u + (v + w).
To see this, we write u = a + bi, v = c + di and w = e + f i, and check, using the
fact that these properties hold for addition of real numbers:
(a + bi) + (c + di) = (a + c) + (b + d)i = (c + a) + (d + b)i = (c + di) + (a + bi)
and
[(a + bi) + (c + di)] + (e + f i) = [(a + c) + e] + [(b + d) + f ]i
= [a + (c + e)] + [b + (d + f )]i
= (a + bi) + [(c + di) + (e + f i)].
The number 0 continues to be the additive identity, i.e., z + 0 = z for every
complex number z, as you can check for yourself. Also, every complex number has

Nana Kwabena
 NKay Study

an opposite with respect to addition. Specifically, −(a + bi) = −a − bi, since

a + bi + (−a − bi) = (a − a) + (b − b)i = 0.
Proposition 2.1. For any complex number z,
z + z = 2 · Re(z) and z − z = 2 · Im(z)i.
Proof: Suppose that z = a + bi, then
z + z = (a + bi) + (a − bi) = (a + a) + (b − b)i = 2a = 2Re(z).
Likewise,
z − z = (a + bi) − (a − bi) = 2bi = 2Im(z)i.

2.2 Multiplication.
To multiply the numbers a + bi and c + di, we use the familiar ‘foil’ rule from
elementary algebra while keeping in mind that i 2 = −1, and collecting real and
imaginary terms:
(2.1) (a + bi) · (c + di) = ac + adi + cbi + bdi 2 = (ac − bd) + (ad + bc)i.
Example 2.2. Using the numbers from Example 2.1, we have
z1 · z2 = (3 + 2i)(4 − 3i)
= 12 − 9i + 8i − 6i 2
= (12 − (−6)) + (−9 + 8)i = 18 − i
and
z3 · z1 = (−3 + 5i)(3 + 2i)
= −9 − 6i + 15i + 10i 2
= (−9 + (−10)) + (15 + 10)i
= −19 + 25i.
When one of the two factors is real, then multiplication is particularly simple,
since real numbers have no imaginary parts. E.g., 7 · z2 = 7(4 − 3i) = 28 − 21i.
As with addition, the usual properties of commutativity and associativity hold
for multiplication of complex numbers, as you can verify for yourself (see exercises).
And, multiplication distributes over addition, as usual:



(a + bi) · (c + di) + (e + f i) = (a + bi) · (c + e) + (d + f )i



= a(c + e) − b(d + f ) + a(d + f )) + b(c + e) i
= (ac + ae − bd − bf ) + (ad + af + bc + be)i



= (ac − bd) + (ad + bc)i + (ae − bf ) + (af + be)i
= (a + bi)(c + di) + (a + bi)(e + f i).

Nana Kwabena
 NKay Study

2.3 Division.
Division is properly thought of as multiplication by the inverse, where the inverse
of a complex number z is the complex number z −1 satisfying z · z −1 = 1. As with
the real numbers, a complex number has a multiplicative inverse if and only if it is
not zero. Finding the multiplicative inverse is simply a matter of solving a pair of
linear equations in two unknowns.
Specifically, given a (nonzero) complex number a + bi, we want to find a complex
number x + yi such that (a + bi)(x + yi) = 1. Using the rule for multiplication of
complex numbers this gives
(ax − by) + (ay + bx)i = 1 = 1 + 0i,
which reduces to a pair of linear equations in the variables x and y:
(2.2) ax − by = 1
(2.3) bx + ay = 0.
Multiplying equation (2.2) by a and equation (2.3) by b and adding the results
together gives
a
(a2 x − aby) + (b2 x + aby) = a + 0 =⇒ (a2 + b2 )x = a =⇒ x= .
a2 + b2
Note that division by (a2 + b2 ) is allowed since a + bi 6= 0, which means that either
a 6= 0, b 6= 0 or both, so a2 + b2 > 0.
In similar fashion, multiplying equation (2.2) by −b and equation (2.3) by a and
adding the results together gives
b
(−abx + b2 y) + (abx + a2 y) = −b + 0 =⇒ (a2 + b2 )y = −b =⇒ y=− .
a2 + b2
We can summarize these calculations as follows:
Proposition 2.2. If a + bi 6= 0, then
a b
(a + bi)−1 = − 2 i.
a2 +b2 a + b2

Now, to divide c + di by a + bi, we multiply by (a + bi)−1 :

 
c + di −1 a b ac + bd ad − bc
= (c + di) · (a + bi) = (c + di) · 2 2
− 2 2
i = 2 + i.
a + bi a +b a +b a + b2 a2 + b2
z1 z1
Example 2.3. Compute and , for the numbers z1 , z2 and z3 from Example
z2 z3
2.1.
z1 3 + 2i 12 − 6 −9 − 8 6 17
= = + i= − i,
z2 4 − 3i 25 25 25 25
and
z1 3 + 2i −9 + 10 −15 − 6 1 21
= = + i= − i.
z3 −3 + 5i 34 34 34 34

Nana Kwabena
 NKay Study

2.4 Arithmetic and conjugation.

The complex conjugate plays nicely with addition, multiplication and division.
Proposition 2.3. For complex numbers z and w, z + w = z + w, z · w = z · w;
z z
and if w 6= 0, then = .
w w
Proof: Let z = a + bi and w = c + di (where a, b, c and d are real, as usual). Then
for addition we have
(a + bi) + (c + di) = (a + c) + (b + d)i
= (a + c) − (b + d)i
= (a − bi) + (c − di) = (a + bi) + (c + di);
and for multiplication we have
(a + bi) · (c + di) = (ac − bd) + (ad + bc)i
= (ac − bd) − (ad + bc)i
= (a − bi) · (c − di) = (a + bi) · (c + di)
The proof for the case of division is left as an exercise. 
Exercises
2.1. Let u = 3 + 2i and v = 5 − 2i. Compute u + v, u · v, u/v and v/u.
2.2. Compute the following products, quotients and powers:
(a) (1 + 4i) · (2 − 3i) =
(b) 2/(3 + i) =
(c) (1 + i)3 =
(d) (2 − i)−2 =

2.3. Compute (1 + i)−1 and (3 − 4i)−1 .

2.4. Let ρ = √1 + √1 i. Compute ρ2 , ρ3 and ρ4 . Can you guess what ρ25 is?
2 2

2.5. Solve the pair of equations

iu + v = 1
4u − 2iv = 2.
The solution will be a pair of complex numbers, u and v.
z z
2.6. Show that if w 6= 0, then = .
w w

n
2.7. Show that if n is a positive integer and ζ is a complex number, then ζ n = ζ .†
2.8. Show that u · v = v · u for any two complex numbers u and v.
2.9. Show that u · (v · w) = (u · v) · w for any three complex numbers u, v and w.
†
ζ is the greek letter zeta.

Nana Kwabena
 NKay Study

3. The geometry of complex numbers.

3.1 Complex numbers as points in the plane.

Every complex number has two real coordinates, namely its real and imaginary
parts. It is natural therefore to represent complex numbers as points in what is called
the complex plane. In this representation, the convention is to plot the real part of
the complex number on the horizontal axis and the imaginary part on the vertical
axis. For this reason, the horizontal axis of the complex plane is called the real
axis and the vertical axis is called the imaginary axis. In this representation, the
real numbers correspond to the points on the real axis and the (purely) imaginary
numbers correspond to the points on the imaginary axis. The real and imaginary
parts of a complex number are also called the rectangular coordinates of that number.
To each complex number we can also associate a vector : the directed line segment
(arrow) whose tail is at the origin and whose head is at the point in the plane cor-
responding to the complex number. Addition of complex numbers can be described
geometrically, using vectors. Given complex numbers u and v, thought of as vectors
in the complex plane, translate the vector v so that it’s tail is at the head of u (or
vice versa, translate the tail of u to the head of v), then the head of the translated
vector is the point corresponding to the sum u+v, see Figure 1.† We’ll learn a useful
geometric description of multiplication once we have studied the representation of
complex numbers using polar-coordinates.

u+v
(copy of) v
(copy of) u
u

Figure 1. Addition, viewed geometrically.

†
There is a ‘coordinate-free’ description of addition as well. The sum u+v is the complex number
such that the triangle with vertices 0, u and v in the complex plane is congruent to the triangle
with vertices u, v and u + v.

Nana Kwabena
 NKay Study

3.2 The modulus of a complex number.

The magnitude of a real number a is given by its absolute value, |a|, which may
be described as the distance of the number from 0. This idea is easily extended to
all complex numbers.

Definition 4.
The modulus, or absolute value, of a complex number z = a + bi, is denoted
by |z| and defined to be the distance in the complex plane between the point
z and the point 0, (see Figure 2). Equivalently, |z| is the length of the vector
corresponding to z.

bi z=a+bi

|z|=(a2+b2)1/2

Figure 2. The modulus (absolute value) of z.

Comments:
a. The modulus of a complex number may be computed from its rectangular coor-
dinates (the real and imaginary parts) using the Pythagorean theorem. I.e.,
p
|a + bi| = a2 + b2 .
b. It follows directly from the definition that |z| = |z|, since
p p
|a − bi| = a2 + (−b)2 = a2 + b2 = |a + bi|.
Proposition 3.1. For any complex number z
|z|2 = z · z.
Proof: Let z = a + bi, then |z|2 = a2 + b2 and
z · z = (a + bi)(a − bi) = a2 − abi + abi − b2 i 2 = a2 + b2 ,
since i 2 = −1. 

Nana Kwabena
 NKay Study

Using the modulus and the conjugate, we can also shorten the expression for z −1 .
Namely, if z 6= 0, then
z
(3.1) z −1 = ,
|z|2
as you can check by comparing to Proposition 2.2. It is also important to note that
the modulus is a multiplicative function. I.e., the modulus of a product is equal to
the product of the moduli.
Proposition 3.2. For any complex numbers z and w, |z · w| = |z| · |w|.
Proof: It follows from Propositions 3.1 and 2.3 and the commutative property of
multiplication that
|z · w|2 = (z · w) · (z · w) = z · w · z · w = (z · z) · (w · w) = |z|2 · |w|2 .
Taking square roots of both sides then gives
p p p p
|z · w| = |z · w|2 = |z|2 · |w|2 = |z|2 · |w|2 = |z| · |w|,
as claimed. 
6 0 then z −1  = (|z|)−1 .
 
Proposition 3.3. If z =
Proof: See exercises.
3.3 The argument of a complex number.
The modulus of a complex number defines its magnitude. We can also associate
an angle to each nonzero complex number.

Definition 5.
The argument (or phase) of z = a + bi (z 6= 0) is ‘the’ angle, φ, between
the positive real axis and the line segment connecting z to 0, measured in the
counterclockwise direction, (see Figure 3). The argument of z is denoted by
arg(z).

Comments:
a. Unless specified otherwise, we’ll measure angles in radians. Recall that there
are 2π radians in a circle.
b. In the definition of the argument, the second definite article (‘the’) appears in
quotation marks because for any complex number z, arg(z) is only determined
up to a multiple of 2π. This is also illustrated in Figure 3.
c. Among the (infinitely many) possible choices for arg(z), we’ll usually choose the
one that lies between 0 and 2π. This value of arg(z) is called the principal
value of the argument, e.g., the angle φ in Figure 3.
The principal value of the argument of z = a+bi can be found from the rectangular
coordinates of z, a = Re(z) and b = Im(z), by using the arctangent function (or
inverse tangent function) and some basic trigonometry. This is done as follows.

Nana Kwabena
 NKay Study

10

z=a+bi
bi

Φ+4π

Φ+2π

-a Φ
a
Φ+π

-bi
-z=-a-bi

Figure 3. z, −z and their arguments.

Recall that if −∞ < x < ∞, then arctan(x) is defined to be the angle φ, satisfying
− π2 < φ < π2 and tan(φ) = x.‡ Now, as you can see in Figure 3, if φ = arg(a + bi),
then tan(φ) = b/a (tangent = opposite/adjacent). This might lead you to the
conclusion that arg(a + bi) = arctan(b/a), however this is not necessarily so.
For example, (−b)/(−a) = b/a, so that the arctangent function doesn’t distin-
guish the arguments of z and −z, even though they are different, since arg(−z) =
arg(z) ± π, as illustrated in Figure 3. Moreover, arctan(x) produces values between
−π/2 and π/2, and (for the principal value of the argument) we want values between
0 and π.
The correct approach is to adjust the value of arctan(b/a) according to the quad-
rant of the complex plane in which a + bi lies.
a. If a+bi is in the first quadrant, then the arctangent function produces the correct
value, so no adjustment is necessary. This is the case of z1 in Figure 4. I.e.,
A1. if a > 0 and b > 0, then arg(a + bi) = arctan(b/a).
b. If a + bi is in the second quadrant, then −π/2 < arctan(b/a) < 0, and the correct
value of the argument is π + arctan(b/a). In Figure 4, this corresponds to the
case of z2 , in which case arctan(b/ − a) = −φ (in red), but arg(z2 ) = π − φ. I.e.,
A2. if a < 0 and b > 0, then arg(a + bi) = π + arctan(b/a) = π − arctan(|b/a|).
c. If a + bi is in the third quadrant, then 0 < arctan(b/a) < π/2, and the correct
value of the argument is once again given by π + arctan(b/a). In Figure 4,
this corresponds to the case of z3 , in which case arctan(−b/ − a) = φ, but
arg(z3 ) = π + φ. I.e.,
‡
By specifying that the angle fall in this range, we are technically defining what is called the
principal branch of the arctangent function. These are the values produced by (most) calculators.

Nana Kwabena
 NKay Study

11

z2 bi z1

π-Φ
Φ
-a a
-Φ

π+Φ
2π-Φ

z3 -bi z4

Figure 4. The argument, by quadrant.

A3. if a < 0 and b < 0, then arg(a + bi) = π + arctan(b/a).

d. If a + bi is in the fourth quadrant, then −π/2 < arctan(b/a) < 0, and the correct
value of the argument is given by 2π + arctan(b/a). In Figure 4, this corresponds
to the case of z4 , in which case arctan(−b/a) = −φ, but arg(z4 ) = 2π − φ. I.e.,

A4. if a > 0 and b < 0, then arg(a + bi) = 2π + arctan(b/a) = 2π − arctan(|b/a|).

e. Finally, if a = 0, then arctan(b/a) is not defined (why?), but the argument is

easily determined without the arctangent function. Specifically, If Re(z) = 0
and Im(z) > 0 then arg(z) = π/2 (= 90◦ ), and if Re(z) = 0 and Im(z) < 0 then
arg(z) = 3π/2 (= 270◦ ).
Example 3.1. Find the moduli and the arguments of u = 3 + 4i and v = 12 − 5i.§
The moduli are easy to compute, even without a calculator:
√ √
|u| = 9 + 16 = 5 and |v| = 144 + 25 = 13.
Since u lies in the first quadrant (both Re(u) > 0 and Im(u) > 0), it follows from
rule A1 that
arg(u) = arctan(4/3) ≈ 0.9273.
Likewise, since v lies in the fourth quadrant (Re(v) > 0 and Im(v) < 0), it follows
from rule A4 that
arg(v) = 2π + arctan(−5/12) ≈ 5.8884.

§
The plural of modulus is moduli.

Nana Kwabena
 NKay Study

12

(Remember — angles are measured here in radians. To convert to degrees, you

multiply by 180/π. For example, measuring the arguments above in degrees, we
have arg(u) ≈ 53.13◦ and arg(v) ≈ 337.38◦ .)
5

4 u
3

arg(u)≈0.9273 rad.s
1

-3 -2 -1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

-1

arg(v)≈5.8884 rad.s
-2

-3

-4

-5 v
-6

Figure 5. Arguments of the numbers in Example 3.1.

3.4 The polar-coordinate representation of complex numbers.

The modulus and argument of a complex number are called the polar coordi-
nates of the number, and they determine that number completely. I.e., given the
polar coordinates of z, we can easily find Re(z) and Im(z) using basic trigonometry.
Proposition 3.4. If |z| = r and arg(z) = θ, then Re(z) = r cos θ and Im(z) =
r sin θ. In other words, z = r cos θ + r sin θi, or more succinctly
(3.2) z = r(cos θ + sin θi).

Proof: See Figure 6, in which r = |z| and θ = arg(z). 

Comment: When expressing a complex number in polar coordinates, as in equation
(3.2), it is common to write ... + sin θi instead of ... sin θi.
The addition of complex numbers is easiest to express in terms of their rectan-
gular coordinates. Multiplication however, is simplest to express in terms of polar
coordinates. This follows from Proposition 3.2 and the following fact.
Proposition 3.5. For complex numbers z1 and z2 , we have
arg(z1 z2 ) = arg(z1 ) + arg(z2 ).

Nana Kwabena
 NKay Study

13

r
r⋅sinθ

θ
r⋅cosθ

Figure 6. ‘Proof’ of Proposition 3.4.

Proof: Let z1 = r(cos θ + sin θi) and z2 = ρ(cos φ + sin φi), then it follows from
equation (2.1) and the rules for sine and cosine of a sum of angles that



z1 z2 = (r cos θ)(ρ cos φ) − (r sin θ)(ρ sin φ) + (r cos θ)(ρ sin φ) + (r sin θ)(ρ cos φ) i


= r · ρ (cos θ cos φ − sin θ sin φ) + (cos θ sin φ + sin θ cos φ)i


= r · ρ cos(θ + φ) + sin(θ + φ)i ,
Which implies that θ + φ is equal to (one of the possible values of) arg(z1 z2 ). 
Comments:
a. If arg(z1 ) and arg(z2 ) are the principal values of z1 and z2 and if their sum is
greater than 2π, then we need to subtract 2π from arg(z1 ) + arg(z2 ) to obtain
the principal value of arg(z1 z2 ).
b. Some authors use the shorthand cis(θ) = cos θ + sin θi, but we shall not. Instead,
we’ll use exponential notation.

Combining Propositions 3.2 and 3.5, we obtain a simple geometric characteriza-

tion of multiplication by a (nonzero) complex number.
Proposition 3.6. If z 6= 0 and w is any complex number, then multiplying w by z
rotates w by the angle arg(z) and scales (stretches or shrinks) the modulus of w by
the factor |z|.
Example 3.2. Returning to the numbers u and v from Example 3.1, and using
formula (2.1), we have
uv = (36 + 20) + (−15 + 48)i = 56 + 33i.

Nana Kwabena
 NKay Study

14

It follows that p √
|uv| = 562 + 332 = 4225 = 65
and
arg(uv) = arctan(33/56) ≈ 0.5325.
Now, |u||v| = 5 · 13 = 65, as it should, but
arg(u) + arg(v) ≈ 0.9273 + 5.8884 = 6.8157 6= 0.5325.
This discrepancy is immediately cleared up, however, when we subtract 2π (≈
6.2832) from the sum of the arguments, as you should verify.
3.5 Exponential notation.
The polar coordinate representation of a complex number, equation (3.2), can be
simplified by using Euler’s formula:¶
(3.3) cos θ + sin θi = eθi ,
where θ is real and ez is the familiar exponential function defined by the Taylor
series,k
∞
X zn
(3.4) ez = 1 + .
n!
n=1
In most calculus classes, the variable in the series is assumed to be a real number.
But the operations of addition and multiplication make perfectly good sense for
complex numbers too and it is not hard to show that the series above converges for
every complex number z. Furthermore, the resulting function has all of its familiar
properties, most notably
(3.5) ez1 +z2 = ez1 · ez2
and
ez1 z2 = (ez1 )z2
for all complex numbers z1 and z2 .
Using Euler’s formula (3.3), we can restate Proposition 3.4 as
Proposition 3.7. If |z| = r > 0 and θ = arg(z), then
(3.6) z = r · eθi .

Proposition 3.5 (the fact that the argument of a product is equal to the sum of the
arguments) can now be justified by using property (3.5) of the exponential function:
if z1 = r1 eθ1 i and z2 = r2 eθ2 i , then
   
z1 · z2 = r1 eθi 1 · r2 eθi 2 = r1 r2 eθ1 i+θ2 i = r1 r2 e(θ1 +θ2 )i .

¶
Leonhard Euler (pronounced ‘oiler’), 1707 – 1783, is one of the most prolific and important
mathematicians of all time. His work has influenced (almost) every field in mathematics and
mathematical physics.
k
If Taylor series are unfamiliar or they fill you with a vague sense of dread, then you may skip
the explanation that follows and accept formula (3.3) and Proposition(3.6) on faith.

Nana Kwabena
 NKay Study

15

Note that cos(2π) = cos(0) = 1 and sin(2π) = sin(0) = 0, and this together with
Euler’s formula (3.3) implies that
e2πi = 1.
Furthermore, since ekα = (eα )k the previous identity generalizes to
 k
(3.7) e2kπi = e2πi = 1k = 1

for any integer k.

Combining the identities (3.7) and (3.4) proves the following useful fact.
Proposition 3.8. For any number θ and any integer k,
e(θ+2kπ)i = eθi .

The rest of this subsection explains Euler’s formula. You may skip it without
loss of continuity, but I recommend that you at least skim it.

To understand Euler’s formula, we set z = θi in the series on the right-hand side

of (3.4) (with a real number θ), and manipulate the series a little as follows.
To begin, it helps to recognize that the powers of i follow a simple pattern, namely
i 0 = 1, i 1 = i, i 2 = −1, i 3 = −i, i 4 = 1, i 5 = i, i 6 = −1, i 7 = −i, etc. In other
words, for any integer n,


 1 : if n leaves remainder 0 when divided by 4;
i : if n leaves remainder 1 when divided by 4;

in =

 −1 : if n leaves remainder 2 when divided by 4;
−i : if n leaves remainder 3 when divided by 4.


We can summarize this information even more usefully by distinguishing even and
odd values of n. Specifically, note that if n = 2k for some integer k (n is even), then
i n = i 2k = (i 2 )k = (−1)k ,
and if n = 2k + 1 for some integer k (n is odd), then
i n = i 2k+1 = i 2k · i = (−1)k · i.
Now, if θ is a real number, then when n = 2k we have
(3.8) (θi)n = (θi)2k = (−1)k θ2k ;
and if n = 2k + 1, we have
(3.9) (θi)n = (θi)2k+1 = (−1)k θi 2k+1 .
Next, we substitute z = θi in the Taylor series for ez in (3.4), split the series into
two sub-series based on the parity (evenness or oddness) of the power n, and use
the formulas (3.8) and (3.9) to simplify.

Nana Kwabena
 NKay Study

16

∞
X (θi)n
eθi = 1 +
n!
n=1
∞ ∞
X (θi)2k X (θi)2k+1
= +
(2k)! (2k + 1)!
k=0 k=0
∞ ∞
!
X (−1)k θ2k X (−1)k θ2k+1
(3.10) = + i.
(2k)! (2k + 1)!
k=0 k=0
At this point, we are done because the series on the left-hand side of line (3.10) is
the Taylor series for cos θ, and the series on the right-hand side is the Taylor series
for sin θ, showing that
eθi = cos θ + sin θi,
as claimed.
Exercises
3.1. Compute the moduli and arguments of the numbers
√
z1 = −1 − i, z2 = 3 − 4i, z3 = −8 + 6i, z4 = 3/2 + 0.5i.

3.2. Compute the arguments and moduli of (z1 · z2 ), (z1 · z4 ) and (z2 · z3 ), with
z1 , z2 , z3 and z4 as above.
3.3. Express the four numbers in exercise 3.1 using exponential notation.
3.4. Show that multiplication by i corresponds to counterclockwise rotation by an
angle of π/2.
3.5. Show that if z 6= 0, then z −1  = |z|−1 .
 

3.6. Show that eπi + 1 = 0.

3.7. Show that if z 6= 0, and 0 ≤ arg(z) < π, then arg(−z) = arg(z) + π. How
does this claim need to be adjusted if π ≤ arg(z) < 2π, assuming that we want the
principal value of arg(−z)?
 
3.8. Show that eθi  = 1 for any real number θ.

4. Roots of polynomials.

4.1 The fundamental theorem of algebra.

Let P (z) = cn z n + cn−1 z n−1 + · · · + c1 z + c0 be a polynomial with degree n > 0
and complex coefficients, c0 , c1 , . . . , cn . The Fundamental Theorem of Algebra
states that P (z) has a complex root, i.e., the equation
(4.1) P (z) = 0,

Nana Kwabena
 NKay Study

17

has a solution in the complex numbers.†

If ζ1 is a solution of equation (4.1),‡ then the polynomial P (z) may be factored
as
P (z) = (z − ζ1 ) · P1 (z),
where P1 (z) is a polynomial of degree n − 1.§ If n − 1 ≥ 1, then we can apply the
fundamental theorem of algebra again to conclude that the equation P1 (z) = 0 has
a solution ζ2 , from which it follows that P1 (z) = (z − ζ2 ) · P2 (z) and so
P (z) = (z − ζ1 ) · (z − ζ2 ) · P2 (z),
where P2 (z) is a polynomial of degree n − 2. Continuing in this way, we can factor
the polynomial completely, which gives the following proposition.
Proposition 4.1. If P (z) is a polynomial of degree n ≥ 1 with complex coefficients,
then P may be expressed as a product of linear factors,
(4.2) P (z) = cn · (z − ζ1 ) · (z − ζ2 ) · · · (z − ζn ),
where ζ1 , ζ2 , . . . , ζn are solutions of the equation P (z) = 0.
Note: It is not necessarily the case that the roots ζ1 , . . . , ζn are all different from
each other, so that some of the factors in (4.2) may be repeated. For example, the
polynomial Q(z) = z 4 − 2z 3 + 2z 2 − 2z + 1 factors as
Q(z) = z 4 − 2z 3 + 2z 2 − 2z + 1 = (z − 1)(z − 1)(z − i)(z + i),
as you should verify.
It is also important to note that while the roots ζ1 , . . . , ζn that appear in (4.2)
may not all be different from each other, they do include all the roots of P (z). This
implies that a polynomial of degree n has no more than n roots. We can summarize
all of this as follows.
Proposition 4.2. If P (z) is a polynomial of degree n ≥ 1 with complex coefficients,
then P has at least one root, but no more than n roots.

While the fundamental theorem of algebra guarantees that every polynomial has
roots, it says nothing about the important, practical problem of actually finding
these roots.
For quadratic equations we have the quadratic formula which works just as well
when the coefficients are complex numbers (as we will see below). There are anal-
ogous formulas for finding the solutions of cubic and quartic equations (degrees 3
and 4), but a deep theorem in algebra states that there are no general formulas for
finding the roots of polynomials of degree 5 or more.
While there are ad hoc methods that can be used to find roots in special cases,
it is generally very difficult to find the roots of a polynomial of degree greater than
four. In most cases, we have to settle for approximate solutions using methods from
calculus, e.g., Newton’s method and the like.
†
The proof of this theorem is beyond the scope of this course.
‡
The symbol ζ is the greek letter zeta.
§
This claim can be proved using long division of polynomials. See the exercises.

Nana Kwabena
 NKay Study

18

In the last few sections below, we’ll make a couple of general observations about
polynomials with real coefficients and then look at some simple types of polynomials
whose roots are relatively easy to find.
4.2 Polynomials with real coefficients.
Since the real numbers are a subset of the complex numbers, it follows that any
polynomial with real coefficients also has a root, and such polynomials may be
factored as in equation (4.2). On the other hand, even though the coefficients of
the polynomial may be real, there is no guarantee that its roots will be real, as in
Example 1.1.
The complex roots of a real polynomial do have a special property however: they
occur in complex conjugate pairs, as proved below.
Proposition 4.3. If a0 , a1 , . . . , an are real numbers and ζ is a root of the polynomial
P (z) = an z n + · · · + a1 z + a0 , then ζ is also a root of P (z).
Proof: ζ is a root of P (z) if and only if an ζ n + · · · + a1 ζ + a0 = 0. Taking
conjugates of both sides of this equation gives
(4.3) an ζ n + · · · + a1 ζ + a0 = 0 = 0,
since 0 is real. Now, applying Proposition 2.3 (and exercise 2.7) to the left-hand
side of equation (4.3) gives
an ζ n + · · · + a1 ζ + a0 = an ζ n + · · · + a1 ζ + a0
= an (ζ)n + · · · + a1 (ζ) + a0
(4.4) = an (ζ)n + · · · + a1 (ζ) + a0 ,
since the coefficients a0 , . . . , an are real. Finally, replacing the left-hand side of (4.3)
by the right-hand side of (4.4) shows that
P (ζ) = an (ζ)n + · · · + a1 (ζ) + a0 = 0,
so that ζ is also a root of P (z), as claimed. 
Example 4.1. Consider the polynomial P (z) = z 4 − 2z 3 + z 2 + 2z − 2. Suppose
that a mystical experience reveals that ζ = 1 + i is a root of P , then because the
coefficients of P are all real, it follows that ζ = 1 − i is also a root of P .
This means that both (z − (1 + i)) and (z − (1 − i)) are factors of P (z), so that
P (z) = (z − (1 + i)) · (z − (1 − i)) · Q(z) = (z 2 − 2z + 2) · Q(z),
where Q(z) is a polynomial of degree 2. Dividing P (z) by z 2 − 2z + 2 gives
P (z)
Q(z) = = z 2 − 1,
z 2 − 2z + 2
so that
P (z) = (z − (1 + i)) · (z − (1 − i)) · (z 2 − 1) = (z − (1 + i)) · (z − (1 − i)) · (z − 1)(z + 1),
which means that the roots of P (z) are ζ1 = 1 + i, ζ2 = 1 − i, ζ3 = 1 and ζ4 = −1.

Nana Kwabena
 NKay Study

19

In the preceding example we saw that for ζ = 1 + i, the product (z − ζ)(z − ζ)

results in a (quadratic) polynomial with real coefficients. This phenomenon is true
in general and in fact we can be more precise.
Proposition 4.4. For any complex number ζ, we have
(z − ζ)(z − ζ) = z 2 − bz + c,
where b = 2 · Re(ζ) and c = |ζ|2 are both real numbers.
Proof: Exercise.
The representation of a polynomial with real coefficients as a product of linear
factors, as in 4.2, may have complex coefficients, since the polynomial may have
complex roots. Using Propositions 4.3 and 4.4 however, we can show that every real
polynomial can be represented as a product of linear and quadratic factors, all of
which have real coefficients.
Proposition 4.5. If a0 , a1 , . . . , an are real numbers (with an 6= 0), then
(4.5) P (z) = an z n + · · · + a1 z + a0 = an (z − ξ1 ) · · · (z − ξm )Q1 (z)Q2 (z) · · · Qk (z),
where
a. ξ1 , . . . , ξm are all the real roots of P (z), repeated as necessary;
b. and for 1 ≤ j ≤ k,
Qj = z 2 − 2Re(ζj ) + |ζj |2 ,
where ζ1 , ζ 1 , ζ2 , ζ 2 , . . . , ζk , ζ k are the complex roots of P (z).
Proof: Exercise.
4.3 Square roots and quadratic equations.
Before we can solve a general quadratic equation, we need to be able to compute
the square roots of a complex number. This is easiest to do using the exponential
(or polar) representation.
Proposition 4.6. If u 6= 0 then the solutions of the equation z 2 = u are
p p
(4.6) ζ1 = |u|e(θ/2)i and ζ2 = |u|e(θ/2+π)i ,
p
where θ = arg(u) (the principal value of the argument) and |u| is the (positive)
real square root of |u|. Moreover,
0 ≤ arg(ζ1 ) < π and arg(ζ2 ) = arg(ζ1 ) + π.
Proof: First, we have
p 2
ζ22 = (−ζ1 )2 = ζ12 = |u| ei(θ/2+θ/2) = |u|eθi = u,
so ζ1 and ζ2 are certainly the two solutions of the equation z 2 = u.
Next, we have 0 ≤ θ < 2π, so 0 ≤ arg(ζ1 ) < π, because arg(ζ1 ) = θ/2. Finally, it
follows from exercise 3.7 that
arg(ζ2 ) = arg(−ζ1 ) = arg(ζ1 ) + π,
and therefore π ≤ arg(ζ2 ) < 2π. 

Nana Kwabena
 NKay Study

20

Comment: By convention ‘the’ square root of a nonzero complex number u is the

solution ζ1 above, i.e., it is the solution of z 2 = u whose principal argument is half
the principal argument of u. This generalizes with the convention that the square
root of a positive real number α is the positive solution of the equation z 2 = α.
Example 4.2. Find the square roots of i, −2i and −3 + 4i. Express your answers
in both exponential and rectangular coordinates.
First, we have |i| = 1 and arg(i) = π/2, so
√ 1 1
i = eπi/4 = cos(π/4) + sin(π/4)i = √ + √ i.
2 2
Next, we have | − 2i| = 2 and arg −2i = 3π/2, so
√ √ √
−2i = 2e3πi/4 = 2 cos(3π/4) + sin(3π/4)i = −1 + i.


√
Finally, | − 3 + 4i| = 9 + 16 = 5 and arg(−3 + 4i) = π − arctan(4/3) ≈ 2.2143,¶
√ √ √
−3 + 4i = 5e1.10715i = 5 cos(1.10715) + sin(1.10715)i = 1 + 2i.

To solve a quadratic equation with complex coefficients, we use the familiar qua-
dratic formula.
Proposition 4.7. The solutions of the equation az 2 + bz + c = 0 are given by
√ √
−b + b2 − 4ac −b − b2 − 4ac
(4.7) z1 = and z2 = ,
2a 2a
where a, b and c may be any complex numbers, as long as a 6= 0.
The proof is identical to the case in which the coefficients are real (i.e., completing
the square), because none of the steps make any use of the nature of the coefficients
(real or complex), and I leave it to you as an exercise.
Example 4.3. Find the solutions of the equation z 2 + (1 − 2i)z − 2i = 0.
Applying the quadratic formula (with a = 1, b = 1 − 2i and c = −2i) we find
that the two solutions are
p p
−(1 − 2i) + (1 − 2i)2 + 8i −(1 − 2i)2 − (1 − 2i)2 + 8i
ζ1 = and ζ2 = .
2 2
Next, note that (1 − 2i)2 + 8i = −3 − 4i + 8i = −3 + 4i, so the two solutions simplify
to
−1 + 2i + 1 + 2i −1 + 2i − 1 − 2i
ζ1 = = 2i and ζ2 = = −1,
2 2
as you can check.
4.4 The nth roots of a complex number.

If n is a positive integer and α > 0 is a real number then the nth root of α is
defined
√ to be the positive real number β satisfying β n = α, which we denote by
β = n α or β = α1/n . In other words, β is a solution of the equation z n − α = 0. In
this section, we’ll generalize this to find the nth root(s) of any complex number.
¶
See rule A2 in subsection 3.3.

Nana Kwabena
 NKay Study

21

If u 6= 0 is a complex number and n is a positive integer, then Proposition 4.1

tells us that there are complex numbers ζ1 , ζ2 , . . . , ζn such that
(4.8) z n − u = (z − ζ1 )(z − ζ2 ) · · · (z − ζn ).
Now, for a general polynomial of degree n, there is no guarantee that the ζ’s that
appear in the factorization (4.2) are all different from one another. But in this case,
i.e., in (4.8), the roots are distinct.
Proposition 4.8. If u 6= 0, θ = arg(u) and n is a positive integer, then the solutions
of the equation z n = u are given by
p
(4.9) ζk = n |u|e(θ+2kπ)i/n ,
for k = 0, 1, 2, . . . , n − 1. Furthermore, these n numbers are all distinct from one
another.
Proof: To see that ζkn = u for each integer k, between 0 and n − 1, we simply
evaluate:
p n
ζkn = n |u|e(θ+2kπ)i/n
 p n
= n |u| · en·(θ+2kπ)i/n
= |u| · e(θ+2kπ)i
= |u|eθi
= u,
where the transition from line 3 to line 4 is justified by Proposition 3.8.
Now suppose that 0 ≤ j < k ≤ n − 1, then
p
ζk n
|u|e(θ+2kπ)i/n e(θ+2kπ)i/n k−j
= p = (θ+2jπ)i/n
= e( n )·2πi ,
ζj n
|u|e(θ+2jπ)i/n e

as you should verify (using the fact that eα /eβ = eα−β ). It follows that
k−j
arg(ζk /ζj ) = · 2π.
n
But 0 < k −j < n, so 0 < k−j
n < 1, which implies that ζk /ζj 6= 1, since the argument
of 1 must be an integer multiple of 2π. Thus ζk 6= ζj , as claimed. 
Example 4.4. The previous proposition applies to all complex numbers, including
the real numbers. The nth √
roots of the number 1 are called the nth roots of unity.
Now, since arg(1) = 0 and n 1 = 1 (for any n), the nth roots of unity are particularly
simple to write down:
(4.10) 1, e2πi/n , e4πi/n , . . . , e2kπi/n , . . . , e2(n−1)πi/n .
For example, the 4th roots of unity are
ρ0 = e0·πi/4 = 1, ρ1 = e2πi/4 = i, ρ2 = e4πi/4 = −1, ρ3 = e6πi/4 = −i;

Nana Kwabena
 NKay Study

22

and the 6th roots of unity are

√ √
0·πi/6 1 3
2πi/6 4πi/6 1 3
ρ0 = e = 1, ρ1 = e = + i, ρ2 = e =− + i, ρ3 = e6πi/6 = −1
2 2 2 2
√ √
8πi/6 1 3 10πi/6 1 3
ρ4 = e =− − i and ρ5 = e = − i.
2 2 2 2
The arguments of the nth roots of unity increase from 0 to 2(n − 1)π/n in incre-
ments of 2π/n, and the roots of unity themselves are equally spaced points on the
unit circle in the complex plane. This is illustrated below, where the 10th roots of
unity are marked with blue dots on the unit circle.

e3πi/5 i e2πi/5

eπi/5
e4πi/5

π/5

e5πi/5=-1 1=e0πi/5

e6πi/5 e9πi/5

e7πi/5 -i e8πi/5

Figure 7. The tenth roots of unity.

If we set ρn = e2πi/n , then it follows from the basic properties of exponents that
e2kπi/n = ρnk .
In other words, all of the nth roots of unity may be expressed as powers of the root
ρn . For this reason, ρn is sometimes called a primitive nth root of unity.k
Example 4.5. If α is a positive √ real number, then arg(α) = 0 and the nth roots
of α may be expressed in terms of α and the nth roots of unity. Indeed, according
n

to formula (4.9), the nth roots of α have the form

p √
ζk = n |α| · e2kπi/n = n α · ρnk ,
for k = 0, . . . , n − 1, since |α| = α in this case.
k
ρn is called a primitive nth root of unity and not the primitive nth root of unity because for
any n, there are other nth roots of unity with the same property. E.g., it is possible to represent
the 5th roots of unity as powers of e4πi /5 .

Nana Kwabena
 NKay Study

23

Example 4.6. If β is a negativep real number, then arg(β) = π and the nth roots
of β may be expressed in terms of |β|, the nth roots of unity and an nth root of i.
Specifically, the nth roots of β are given by
p p
ζk = n |β| · e(π+2kπ)i/n = n |β| · eπi/n · ρnk ,
for k = 0, . . . , n − 1, where you should note that (eπi/n )n = i.
Exercises
4.1. Find the solutions of the quadratic equations below. Express your answers in
rectangular coordinates (i.e., in the form a + bi).
(a) z 2 + iz + 1 = 0.
(b) 2z 2 − 3iz + 2 = 0.
(c) iz 2 + 2z − 1 = 0.

4.2. Write down the 8th roots

√ of unity using rectangular coordinates. Do not round
your answers, instead use 2 as needed. Hint: first express your answer in polar
coordinates.
4.3. Write down the 12th roots
√ of unity using rectangular coordinates. Do not round
your answers, instead use 3 as needed. Hint: first express your answer in polar
coordinates.
4.4. Prove Proposition 4.4.
4.5. Prove Proposition 4.5.
4.6. Use Proposition 4.5 to show that a polynomial with real coefficients and odd
degree always has at least one real root.
4.7. Show that if n is even then −1 is an nth root of unity and if n is divisible by 4,
then i is an nth roots of unity.
4.8. Show that if i is an nth root of unity, then n is divisible by 4.
4.9. Show that if η is an nth root of unity, then so is η k for any integer k.
4.10. Show that if ζ is a root of the polynomial P (z) = cn z n + · · · + c1 z + c0 , then
P (z) = (z − ζ) · P1 (z),
where P1 (z) is a polynomial of degree n − 1.
Hints: If you use ‘long division’ to divide (z − ζ) into P (z) you obtain an identity
of the form
P (z) = (z − ζ) · P1 (z) + R(z),
where R(z) is a polynomial whose degree is less than the degree of (z − ζ). What
is the degree of (z − ζ) and what is the nature of a ‘polynomial’ whose degree is less
than this? Finally, use the fact that P (ζ) = 0.

Nana Kwabena