Complex Numbers: Vicky Neale Michaelmas Term 2019
Complex Numbers: Vicky Neale Michaelmas Term 2019
Complex Numbers: Vicky Neale Michaelmas Term 2019
Vicky Neale
Michaelmas Term 2019
Contents
1 Introduction 2
5 Complex conjugation 6
6 Modulus 7
7 Argument 10
8 Historical interlude 13
9 De Moivre’s Theorem 13
10 Roots of unity 15
12 Polynomials 17
13 Historical interlude 19
1
1 Introduction
This course on complex numbers will introduce new ideas to some of you,
refresh ideas for others, and make sure that we all agree on notation and
terminology as you begin your course in Oxford.
There are several resources that will help you as you study the course:
- the lectures
- these notes
- each other.
You will not generally receive solutions to problems sheets. But, just for this
course, the department understands that tutors and students might not have
time to schedule a tutorial, and so I’ll provide solutions online. I strongly
encourage you to talk with your fellow students about your solutions, their
solutions and my solutions, and more generally to help each other out if there
are parts of the course that are unfamiliar to you.
Acknowledgements
These notes, and the lectures they accompany, are extremely closely based
on those produced by Dr Peter Neumann, which in turn built on notes by
previous lecturers. The same applies to the problems sheet.
I would like these notes to be as useful as possible. If you (whether
student or tutor) think that you’ve noticed a typo, or mistake, or part that
is unclear, please check the current, up-to-date, notes on the website, to see
whether I’ve already fixed it. If not, please email me (vicky.neale@maths)
and I’ll do something about it, and (with your permission) thank you here.
Thanks to Ken Cao, Yukai Chou and Afaq Mikaiil Tahir for helping to
fix glitches in these notes and in the solutions for the problems sheet.
2
This is not the place for a formal definition of R, the set of real numbers.
You will explore fundamental properties of R in detail in the Analysis I course
this term. Here, we can think of them informally as the familiar numbers on
the number line, including both the rationals (of the form pq where p, q are
integers and q 6= 0) and the irrationals (the
√ many, many numbers that fill in
the gaps between the rationals, such as 2, π and e).
Our current vagueness about our notion of real numbers
√ may be worrying,
but perhaps more concerning is the appearance of −1 in the definition of
a complex number, since famously the square of any real number is√non-
2
√ is, x ≥ 0 for all real x. So what does it mean to write −1?
negative: that
Certainly −1 is not a ‘real number’, in the sense that it is not in the set
R, but
√ as mathematicians we may still study and work with it. We define
i = −1 to be an object with the property that i2 = −1. As we shall see,
allowing ourselves to work with i gives rise to lots of useful and fascinating
mathematics, and does√not introduce any difficulties beyond our instinctive
anxiety about writing −1.
One happy consequence of extending our horizons to include complex
numbers is that we can suddenly solve many more equations than we could
previously. Negative numbers allow us to solve equations such as x + 1 = 0.
Complex numbers allow us to solve equations such as x2 + 1 = 0. In fact,
more is true: by introducing i, a solution of x2 + 1 = 0, we can solve not
just this one quadratic equation, but also every quadratic equation with real
coefficients.
For real numbers a, b and c, with a 6= 0, we know that the equation
ax2 + bx + c = 0
can be rearranged as
2
2 b
4a x+ = b2 − 4ac
2a
3
solutions, since we can have
√
b
2a x + = ± 4ac − b2 i
2a
(note that 4ac − b2 > 0 inside the square root here). We’ll consider more
general polynomial equations later in the course.
and
(a + bi)(c + di) = (ac − bd) + (ad + bc)i.
We can use this to see how to subtract too.
Addition and multiplication in C are binary operations, and have many
nice properties that you will explore in other courses.
We can also divide a complex number by a non-zero complex number. If
a + bi 6= 0, then we see that
1 a − bi a − bi a b
= = 2 2
= 2 2
− 2 i.
a + bi (a + bi)(a − bi) a +b a +b a + b2
This is perhaps a little mysterious at the moment, but will be more natural
when we have studied complex numbers further.
4
Figure 1: Argand diagram marked with the complex numbers 3i, 3 + 4i, 2,
3 − i and −1 − 2i.
When we visualise C using the Argand diagram, we find that the real
numbers (those complex numbers with imaginary part 0) form the horizontal
axis, called the real axis, and the purely imaginary numbers (those complex
numbers with real part 0) form the vertical axis, called the imaginary axis.
The Argand diagram gives a bridge between algebraic and geometric
thinking. For example, we can interpret addition of complex numbers ge-
ometrically, as shown in Figure 2.
Note how this resembles addition of vectors. We can interpret multipli-
cation geometrically too—we’ll return to this later.
One really crucial difference between R and C relates to ordering. We
have a notion of < for real numbers: we can say that −17.2 < 3 and that
0 < 0.001, for example. This ordering behaves rather nicely: for real numbers
a and b, we see that
5
Figure 2: Argand diagram with the complex numbers 0, z, w and z + w
forming the vertices of a parallelogram.
You will explore these properties of the binary relation < on R in detail in
the Analysis I course. What we note here is that there is no such ordering
on C. I don’t mean “the ordering you might first think of doesn’t work”.
I mean that it is not possible to define a binary relation on C that has all
three of the properties above.
5 Complex conjugation
Definition. The complex conjugate of the complex number z = a + bi is
z = a − bi. (Some people write the complex conjugate of z as z ∗ instead of
z.)
6
Figure 3: Argand diagram showing the complex numbers −1 + 4i, i and
3 + 2i, and the complex conjugates of these, namely −1 − 4i, −i and 3 − 2i
respectively.
6 Modulus
Definition.
√ The modulus of the complex number z = a + bi is defined to be
|z| = a + b2 . (This is also sometimes called the absolute value of z.)
2
2 2
Remark. Notice that a and b are real, so √ a + b ≥ 0, so |z| is a real number.
By convention, for real x when we write x we mean the non-negative square
root. So |z| ≥ 0 for all z ∈ C, and |z| = 0 if and only if z = 0.
Geometrically, |z| is the length of z, that is, the distance of z from the
origin 0.
7
Lemma 2. For z = a + bi ∈ C, we have
(ii) (
a if a ≥ 0
|a| =
−a if a < 0
—so the modulus of a real number is just the usual absolute value;
(iii) |z| ≥ |a|, that is, |z| ≥ | Re(z)|, and similarly |z| ≥ | Im(z)|.
Proof. These are immediate from the definition. (Can you write out the
details?).
(i) zz = |z|2 ;
by definition of |z|.
Now we can take square roots, since we know that |zw| and |z||w| are
both non-negative.
Tip We could have proved Proposition 3(ii) by writing out real and imaginary
parts. Try it yourself! I predict that it will take longer, and more importantly
give less insight, than the proof here. It is often a good idea to work with
complex numbers as single entities rather than real and imaginary parts,
using properties such as those we have proved so far. This can take some
practice to feel natural, but it is worth investing the effort until it becomes
habit.
8
Figure 4: Argand diagram showing a parallelogram with vertices 0, z, w and
z +w, and with line segments of lengths |w|, |z| and |z +w| forming a triangle.
9
Exercise. Prove this!
Now is a good time to revisit division of complex numbers.
If z ∈ C, then zz = |z|2 , and if z 6= 0 then |z|2 6= 0, so
1 z
= 2.
z |z|
Writing z = a + bi, this gives
1 1 a − bi
= = 2 ,
z a + bi a + b2
exactly as we saw previously.
Now that we can compute z1 for any z ∈ C \ {0}, we can more generally
compute wz for any w, z ∈ C with z 6= 0 (by multiplying w by z1 ).
So we can add, subtract and multiply complex numbers, and we can divide
by nonzero complex numbers, and these operations behave in the usual ‘nice’
ways. Formally, C is a field (as are R and Q, and other examples too). You’ll
study fields in other courses.
7 Argument
So far, we have written complex numbers as z = a + bi. We have given
the real and imaginary parts. Geometrically, we interpret this number us-
ing Cartesian coordinates (a, b). Often it is more convenient to work with
complex numbers when they are written in modulus-argument form, which
corresponds geometrically to polar coordinates. Rather than recording a point
using distances relative to the horizontal and vertical axes, we measure the
distance from the origin 0 and the angle anticlockwise from the positive hor-
izontal axis. We illustrate this in Figure 5.
We have already defined the modulus |z| of a complex number z to be
the distance of z from the origin.
Definition. For a nonzero complex number z, the argument of z is the angle
anticlockwise from the positive real axis to the line segment joining 0 and z.
We write arg z for the argument of z. We do not define the argument of 0.
Remark. This definition is not entirely satisfactory, because the angle is
not uniquely defined. If the angle θ satisfies the definition, then so does the
angle θ + 2kπ for any integer k. We say that arg z is only determined modulo
2π. To address this, we can define the principal value of the argument, by
requiring it to lie in a particular interval. Traditionally we require either
10
Figure 5: Two sets of axes, with the same point marked on both. In one, it
is described using Cartesian coordinates, in the other with polar coordinates.
11
Figure 6: Axes are shown, with a triangle with vertices at the origin, at the
point (a, b), and at the point on the horizontal axis vertically below (a, b).
The hypotenuse has length r, and the angle at the origin is θ. The other two
sides have lengths r cos θ (horizontal) and r sin θ (vertical).
Now
zw = |z||w|(cos θ + i sin θ)(cos φ + i sin φ).
But
and so arg(zw) = θ + φ.
Remark. Notice that it was important here that we worked with arguments
modulo 2π. If we had insisted for example that −π < arg z, arg w ≤ π, then
we might have had arg z + arg w outside this interval, and then we’d have
had to worry about ‘wrapping round’.
12
Remark. We can now interpret multiplication of complex numbers geomet-
rically. Multiplying by the nonzero complex number z rotates the complex
plane anticlockwise by arg z, and enlarges by a factor of |z| centred on the
origin.
8 Historical interlude
It’s good to know at least a bit about the history of our subject. Mathemat-
ical ideas don’t just turn up: mathematicians think of them! But who are
the mathematicians, and what else do we know about their lives and work
beyond their most famous theorems?
One useful resource is the MacTutor History of Mathematics Archive:
http://www-history.mcs.st-and.ac.uk/. There’s a list of biographies of
mathematicians there. Why not look up the mathematicians we’ve met so
far in this course?
Jean Argand
http://www-history.mcs.st-and.ac.uk/Biographies/Argand.html
René Descartes
http://www-history.mcs.st-and.ac.uk/Biographies/Descartes.html
Of course, we don’t meet every significant mathematician in an under-
graduate course. Plenty of mathematicians did really important work, but
we don’t hear their names very often. So here are a couple of additional
suggestions of early mathematicians you might like to read about.
Brahmagupta
http://www-history.mcs.st-and.ac.uk/Biographies/Brahmagupta.html
Sun Zi
http://www-history.mcs.st-and.ac.uk/Biographies/Sun_Zi.html
9 De Moivre’s Theorem
If z ∈ C has |z| = 1 and arg z = θ, then we have
√ seen that2 z = cos θ + i sin θ.
Conversely, if z = cos θ + i sin θ, then |z| = cos2 θ + sin θ = 1. So we can
nicely understand the unit circle in C.
Definition. The unit circle in C is S 1 := {z ∈ C : |z| = 1}.
We have just noted that also S 1 = {cos θ + i sin θ : θ ∈ R}.
Fun fact If z ∈ S 1 then z −1 = z, and z −1 ∈ S 1 .
Remark. If z, w ∈ S 1 then |zw| = |z||w| = 1 so zw ∈ S 1 . We say that S 1 is
closed under multiplication. It turns out that the set S 1 forms a group under
13
multiplication (the fun fact above is relevant here too). You’ll study groups
in the course Groups and Group Actions. This group, the unit circle in C
under multiplication, is a very important group. Watch out for it appearing
in other courses!
Theorem 6 (De Moivre’s Theorem). If z ∈ S 1 and n ∈ Z, then
arg(z n ) = n arg z.
arg(z n ) = arg(z m )
= m arg z
= m(− arg z)
= (−m) arg z
= n arg z.
14
Taking real parts gives
10 Roots of unity
Definition. If z ∈ C, n ∈ Z>0 and z n = 1, then we say that z is a root of
unity (an nth root of unity).
Example. The square roots of 1 are 1, −1. The fourth roots of 1 are 1, i,
−1, −i.
arg 1 2kπ
arg z = =
n n
for some k ∈ Z.
(⇐) Suppose that |z| = 1 and that arg z = 2kπn
for some k ∈ Z.
Then |z n | = |z|n = 1 and, by de Moivre, arg(z n ) = n arg z = 2kπ, so z n
has the same modulus and argument as 1, so z n = 1.
Remark. Note that Proposition 5 shows that roots of unity lie on the unit
circle S 1 .
Corollary 8. Take n ∈ Z>0 . Then there are exactly n nth roots of unity.
15
Proof. From Proposition 7, we see that the nth roots of unity are
2kπ 2kπ
cos + i sin for k ∈ Z.
n n
There are n distinct such values (for example we can choose k from the set
{0, 1, . . . , n − 1}).
Happily, a quick check using the values of these trig expressions shows that
we get the same answer either way.
z2 z3 z4
ez := 1 + z + + + + ··· .
2! 3! 4!
16
(We need to be very careful because adding together infinitely many things
is a delicate business. . . .)
With that definition, we would see that
θ2 iθ3 θ4
eiθ = 1 + iθ − − + + ··· .
2! 3! 4!
Separating the real and imaginary parts gives power series that you might
recognise as those for cosine and sine—that in fact we can, with care, use to
define cosine and sine. Watch out for more on this in Analysis!
Remark. Using Euler’s formula gives a convenient way to represent a com-
plex number in polar form. If z ∈ C has modulus r and argument θ, then we
have z = reiθ .
We can use this to generalise Corollary 8, which counts roots of unity.
Proposition 10. Take z ∈ C with z 6= 0, and let n be a positive integer.
Then z has exactly n nth roots.
Proof. (Existence of at least one nth root) Write z = reiθ , where r = |z| and
θ = arg z. Then r > 0 so (by an argument from Prelims Analysis I) there is
a positive real number s such that sn = r (in fact, s is unique). Let φ = nθ .
Let w = seiφ . Then wn = sn einφ = reiθ = z.
(Existence of at least n nth roots) Let w be an nth root of z as above. Let α
be an nth root of 1. Then (αw)n = αn wn = 1 · z = z. But from Corollary 8
we know that there are n possibilities for α, which give n distinct nth roots
of z.
(There are at most n nth roots) Take w as above, and let u be any nth root
n
of z. Then wn = z = un . Since z 6= 0 we have w 6= 0, so wu = 1. So wu is
an nth root of unity, so u = βw where β is one of the n nth roots of unity (n
by Corollary 8), so there are at most n possibilities for u.
12 Polynomials
We saw earlier that every quadratic with real coefficients has exactly two
(possibly repeated) complex roots. The same argument, of completing the
square, generalises to quadratics with complex coefficients. We obtain the
same quadratic formula. By Proposition 10, we know that the (complex)
discriminant has exactly two square roots if it is nonzero (and exactly one
square root if it is zero). So a quadratic with complex coefficients also has
two roots in C—either two distinct roots, or one root with multiplicity 2.
What happens when we generalise to polynomials of higher degree?
17
Proposition 11. A complex polynomial of degree n has at most n roots in
C.
for some polynomial f with complex coefficients, where f is monic and has
degree n − 1.
By the induction hypothesis, f has at most n − 1 roots in C. Now a root
β of p must satisfy p(β) = 0, which is equivalent to (β − α)f (β) = 0. So a
root of p must either be α, or must be a root of f . So p has at most n roots
in C.
Questions Does every polynomial with complex coefficients have a root?
Does every such polynomial of degree n have n roots (counted with multi-
plicity)?
Proof. Not in this course! There are proofs using ideas of complex analysis
and topology. For now, you may assume it—but only if you really have
to.
18
13 Historical interlude
I’m not sure whether it’s still called an interlude when it comes at the end,
but I like the name ‘Historical interlude’, so let’s stick with it. Here’s another
couple of MacTutor biographies for mathematicians we’ve met in the second
part of the course.
Abraham De Moivre
http://www-history.mcs.st-and.ac.uk/Biographies/De_Moivre.html
Leonhard Euler
http://www-history.mcs.st-and.ac.uk/Biographies/Euler.html
And here are a couple of mathematicians we haven’t met in this course,
this time with Oxford connections.
Mary Cartwright
http://www-history.mcs.st-and.ac.uk/Biographies/Cartwright.html
G.H. Hardy
http://www-history.mcs.st-and.ac.uk/Biographies/Hardy.html
Finally, here’s some history of the Fundamental Theorem of Algebra:
http://www-groups.dcs.st-and.ac.uk/history/HistTopics/Fund_theorem_
of_algebra.html
19