The Bright Side

of
Complex Analysis

version
[Link]
February 26, 2025

Julian P. Großmann
[Link]
[Link]@[Link]
Contents
1 Introduction 6

2 Complex Differentiability 8

3 Complex Derivative and Examples 10

4 Holomorphic and Entire Functions 13

5 Totally Differentiability in R2 15

6 Cauchy-Riemann Equations 17

7 Cauchy-Riemann Equations Examples 19

8 Wirtinger Derivatives 21

9 Power Series 23

10 Uniform Convergence 28

11 Power Series Are Holomorphic - Proof 29

12 Exp, Cos and Sin as Power Series 30

13 Complex Logarithm 31

14 Powers 36

15 Laurent Series 39

16 Isolated Singularities 40

17 Complex Integration on Real Intervals 41

18 Complex Contour Integral 42

19 Properties of the Complex Contour Integral 43

20 Antiderivatives 44

21 Closed curves and antiderivatives 46

22 Goursat’s Theorem 47
Contents 3

23 Cauchy’s theorem 48

24 Winding Number 49

25 Cauchy’s Theorem (general version) 52

26 Keyhole contour 53

27 Cauchy’s Integral Formula 54

28 Holomorphic Functions are C-infinity Functions 55

29 Liouville’s Theorem 56

30 Identity Theorem 57

31 Application of the Identity Theorem 58

32 Residue 59

33 Residue for Poles 60

34 Residue theorem 61

35 Application of the Residue Theorem 62

Index 63
Some words
This book was developed along my video series which is freely available on YouTube and
can be used while reading this written form here. For this, you find QR codes at the
beginning of each chapter of the book or clickable links in the digital version.

Overview for the whole course

Here you find all videos for
the Complex Analysis Course.

Learning a whole mathematical topic like Complex Analysis can be seen as a whole hiking
trip, which requires stamina and strength.
We start in the valley of mathematics and will shortly scale the first hills. Always stay
in shape, practise and don’t hesitate to ask about the ways up. It is not an easy trip but
you can do it. Maybe the following tips can guide you:
• You will need a lot of time for this book if you really want to understand every-
thing you learn. Hence, make sure that you have enough time each week to do
mathematics and keep these time slots clear of everything else.
• Work in groups, solve problems together and discuss your solutions. Learning math-
ematics is not a competition.
• Explain the content of the lectures to your fellow students. Only the things you can
illustrate and explain to others are really understood by you.
• Learn the Greek letters that we use in mathematics:
α alpha β beta γ gamma Γ Gamma
δ delta  epsilon ε epsilon ζ zeta
η eta θ theta Θ Theta ϑ theta
ι iota κ kappa λ lambda Λ Lambda
µ mu ν nu ξ xi Ξ Xi
π pi Π Pi ρ rho σ sigma
Σ Sigma τ tau υ upsilon Υ Upsilon
φ phi Φ Phi ϕ phi χ chi
ψ psi Ψ Psi ω omega Ω Omega
Contents 5

A quick overview for Greek letters

Watch Greek letters in

YouTube.

• Choosing a book is a matter of taste. Look into different ones and choose the book
that really convinces you.
• Keep interested, fascinated and eager to learn. However, do not expect to under-
stand everything at once.
• If you are unsure if you have all the prerequisites to understand this series about
Complex Analysis, check my website to see the network for your learning path:
[Link] Also do this whenever a used
notion seems too unfamiliar to you.

DON’T PANIC
1 Introduction
Complex Analysis is nice and also fun. We will have a lot of videos here and you find
them at the beginning of each chapter.

Introduction to Complex Analysis

Watch part 1 on YouTube.

https:// [Link]/ dEu5ie25U0Y

The numbering of the chapters coincides with the numbering of the videos and all defini-
tions and theorems are numbered with a leading chapter number. This should help you
to navigate this book and the additional video material.
In short, we can say that Complex Analysis is the analysis of differentiable functions
f : C → C. This is in contrast to Real Analysis, where only functions f : R → R
are considered. It turns out that increasing the number set in such a way adds a lot of
different properties to the differentiable functions. We will also see that even real-valued
problems can be solved with this detour over the complex numbers.
Example 1.1. Improper Riemann integrals
The following integral might be hard to solve with methods from real analysis:

x sin(x)
Z ∞
dx .
∞ 1 + x2

However, at the end of the course, we can easily conclude that the value is πe .

You do not need a lot of prior knowledge to understand most of this course, but you
definitely need to know the complex numbers as we have them introduced in the Start
Learning Mathematics course, which you can also find on my webpage. On the other
hand, we will talk about derivatives, power series, and integrals in the complex realm such
that some knowledge about these topics from real analysis is helpful but not extremely
necessary since we will explain a lot anyway. Just make sure that you don’t get lost and
always check out some videos from the Real Analysis series if needed.
So let’s start with the basic definitions we need:
1 Introduction 7

Definition 1.2. Distance function

The set of the complex numbers C becomes a so-called metric space with the distance
function
d(z, w) := |z − w|
where we just have the absolute value in complex numbers.

Definition 1.3. Convergence

A sequence of complex numbers, denoted by (zn )n∈N , is called convergent to a complex
number a ∈ C if

∀ε > 0 ∃N ∈ N ∀n ≥ N : |zn − a| < ε .

In this case, the number a is called the limit of the sequence (zn )n∈N and we write:
n→∞
zn −−−→ a

We say that the sequence lies in each small ε-ball around a, eventually.

So no matter how small the ball is chosen, eventually, all infinitely many remaining
sequence members lie in it. We say ball in the general meaning, but in this picture, it is
a two-dimensional disc. However, we keep the general name:
Definition 1.4. Epsilon-Ball
For a real number ε > 0 and complex number a ∈ C, we define the ε-ball as

Bε (a) := {z ∈ C | |z − a| < ε}

Please note that convergence of a complex sequence (zn )n∈N with limit a is equivalent to
saying that the real sequence
(|zn − a|)n∈N
is a convergent in the real numbers with limit 0.
Since we can measure distances between complex numbers as we can do it for real numbers,
all notions that just need a distance immediately generalize from real functions to complex
functions. For example, we know that continuous functions send close points to close
points again. This means we can directly define such functions:
Definition 1.5. Continuity
A function f : C → C is called continuous at z0 ∈ C if, for all sequence (zn )n∈N in
the complex numbers, we have:

=⇒ f (zn ) −−−→ f (z0 ) .

n→∞ n→∞
zn −−−→ z0
2 Complex Differentiability
One could think that differentiability in the complex numbers means the same as in the
real numbers or at least the same as for functions defined on R2 . However, this is not
true simply because C offers more structure which we will use.

Differentiability in the complex realm

Watch part 2 on YouTube.

https:// [Link]/ Jk5hxibwW80

So first let’s start with definition we will need in the following. We already know from
real analysis that differentiability can only be defined at points of the domain where there
are enough neighbours since, otherwise, the limit process would not make sense. For
this reason, one focuses on open sets, which guarantee that a whole epsilon ball forms a
neighbourhood inside the domain. So let’s formulate this:
Definition 2.1. Open sets in the complex numbers
A subset U ⊆ C is called open if for each point z ∈ U there is an ε > 0 such that

Bε (z) ⊆ U .

From now on, we will always consider complex-valued functions that are defined on an
open set. In the best case, the function has the whole complex numbers as its domain.
So let’s fix differentiability:
Definition 2.2. Differentiability in the complex realm
Let U ⊆ C be an open set, z0 ∈ U , and f : U → C be a complex-valued function.
2 Complex Differentiability 9

Then f is called (complex) differentiable at z0 if the following limit exists

f (z) − f (z0 )
lim .
z→z0 z − z0

Of course, the limit is understood as we learnt in the Real Analysis course.

So more precisely, the limit above means that for every choice of a sequence (zn )n∈N in
U \ {z0 } that is convergent to z0 , we have that the right-hand side is also a convergent
sequence, where the limit is always the same complex number.
So we see that the definition looks exactly like the differentiability for functions f : R → R
and not like the total differentiability for functions f : R2 → R2 . In fact, we heavily use
the multiplication and division of complex numbers here, something we cannot do with
vectors in Rn . And this makes the whole difference. The definition respects the whole
structure of the complex numbers and, therefore, it is more restrictive. However, this
leads to a lot of properties complex differentiable functions have.
3 Complex Derivative and Examples
Let’s go to the explanation of the complex derivatives.

Derivatives in the complex realm

Watch part 3 on YouTube.

https:// [Link]/ _Q4pGe6GyR4

Of course, we could visualize the complex derivative like an ordinary derivative in a graph
where it is the linear approximation in form of a tangent.

The picture still works for that but it hides the fact that the complex numbers are usually
seen as something two-dimensional. In that sense, the picture above should be a four-
dimensional picture and we only look at some projection of that to the two-dimensional
paper plane. It’s still helpful because it tells us how to reformulate our original definition.
Proposition 3.1. Equivalence for Differentiability
Let U ⊆ C be an open set, z0 ∈ U , and f : U → C be a complex-valued function.
Then f is (complex) differentiable at z0 if and only if there is function

∆f,z0 : U → C

with two properties:

(1) f (z) = f (z0 ) + (z − z0 )∆f,z0 (z) for all z ∈ U ,

(2) ∆f,z0 is continuous at z0 .

3 Complex Derivative and Examples 11

The proposition exactly tells us that we have this linear approximation of the function at
the point z0 . Obviously, this function ∆f,z0 is not hard to find because the only possibility
for z 6= z0 is the following:

f (z) − f (z0 )
∆f,z0 (z) =
z − z0
And there we see that the existence of the limit z → z0 connects the continuity of ∆f,z0
at z0 with the differentiability of f at z0 . So we can just remember this formula:

for z 6= z0
 f (z)−f (z )
 z−z 0
0
∆f,z0 (z) =
 lim f (w)−f (z0 )
w−z0
for z = z0
w→z0

Moreover, it should not be a surprise that we choose a short notation for this limit (in
the case that it exists):
Definition 3.2. Complex Derivative
Let U ⊆ C be an open set, z0 ∈ U , and f : U → C be (complex) differentiable at z0 .
Then we set
f (z) − f (z0 )
f 0 (z0 ) := ∆f,z0 (z0 ) = lim
z→z0 z − z0
and call the complex number f 0 (z0 ) the (complex) derivative of f at z0 .

Using this complex number, we definitely get the complex version of the linear approxi-
mation given in the tangent equation
f (z0 ) + (z − z0 )f 0 (z0 )
which is exactly f (z0 ) at the point z = z0 .
Let’s look at some examples. Obviously, every linear function f (z) = mz + c with fixed
m, c ∈ C is complex differentiable at any point and we get m = f 0 (z0 ). So what about a
very similar function?

Example 3.3. Complex Conjugation as Function

Is the function f : C → C with f (z) = z (complex) differentiable at the point z0 = 0?
If it was differentiable, the the following limit would exist:
f (z) − f (0) z
f 0 (0) = lim = lim
z→0 z−0 z→0 z
3 Complex Derivative and Examples 12

However, it cannot exist because we can approach the limit either on the imaginary
axis or on the real axis and we get different results:
i 1
n n
i = −1 6= 1= 1
n n

If you look at the calculation above, you see that for no point z0 we can get differentiability.
So the function f : C → C given by z 7→ z is nowhere (complex) differentiable. This is an
interesting result because in the real numbers it is quite hard to construct a function that
is nowhere differentiable. However, here we immediately get a very simple function with
this property. That already shows that the term complex differentiable is a very strong
requirement.
4 Holomorphic and Entire Functions
Having the complex derivative at a single point is not so interesting and it turns out that
we get a lot of nice results for functions that a (complex) differentiable on their whole
domain. This is usually known as holomorphic.

Holomorphic and Entire Maps

Watch part 4 on YouTube.

So the definition is quite simple.

Definition 4.1. Holomorphicity
Let U ⊆ C be an open set and f : U → C be a complex-valued function. The
function f is called holomorphic if f (complex) differentiable at z0 for every z0 ∈ U .
If, in addition, U = C, then the holomorphic function is called entire.

Please remember here, that the term holomorphic implies that the function is defined on
an open set U .

By knowing how the differentiability acts under operations of functions, we get the fol-
lowing:
Proposition 4.2. Properties of holomorphic functions

(a) If f is holomorphic, then f is also continuous.

(b) If f, g : U → C are holomorphic, then f + g and f · g are holomorphic.

(c) For the complex derivative, we also have the sum rule, product rule, and chain
rule. In short:
4 Holomorphic and Entire Functions 14

(1) (f + g)0 (z0 ) = f 0 (z0 ) + g 0 (z0 )

(2) (f · g)0 (z0 ) = f 0 (z0 ) · g(z0 ) + f (z0 ) · g 0 (z0 )
(3) (f ◦ g)0 (z0 ) = f 0 (g(z0 )) · g 0 (z0 )

The proofs look exactly the same as in Real Analysis. Let’s look at some simple examples
of holomorphic functions.
Example 4.3. Holomorphic Functions

(a) Every polynomial function f : C → C is holomorphic. For f (z) =

Pn
k=0 ak zk
with ak ∈ C, the derivative is given by f (z) = k=1 kak zk−1 .
0
Pn

(b) The function f : C \ {0} → C given by f (z) = 1

z
is holomorphic but not an
entire function.

(c) Similary we can look at general rational functions f : C \ S → C defined as a

quotient of two polynomials p, q, which means f (z) = p(z)
q(z)
where S is the set of
zeros for the polynomial q.
5 Totally Differentiability in R2
Do you know the difference between R2 and C? It shouldn’t be so big and in both we
have the differentiability definition. However these definitions don’t coincide completely.

Total Derivative
Watch part 5 on YouTube.

As you might remember, in order to define the complex numbers, we just took the real
vector space R2 and endowed it with a multiplication. This procedure lead to a field, which
has more structure than the bare vector space R2 . Indeed having this multiplication on C
makes it possible to define the complex differentiability as we did. Now, let’s discuss how
this is different to the notion of total differentiability as we have in every vector space Rn .
Definition 5.1. Induced maps
For a complex function f : C → C, we can define the induced map fR : R2 → R2 by

Re(f (x + iy))
 
fR (x, y) :=
Im(f (x + iy))

Note that this also works the other way around.

Example 5.2. Quadratic function
Let f : C → C be given by f (z) = z 2 . So we have:

(x + iy)2 = x2 + 2xyi − y 2 = (x2 − y 2 ) + i(2xy) .

This means:  
x2 − y 2
fR (x, y) =
2xy
5 Totally Differentiability in R2 16

From multivariable calculus, we know what differentiability means for the function fR .
Let’s recall this here:
Definition 5.3. Total Differentiability
A map fR : R2 → R2 is called totally differentiable at (x0 , y0 ) ∈ R2 if there is a
matrix J ∈ R2×2 and a map φ : R2 → R2 such that
   !
x x0
fR (x, y) = fR (x0 , y0 ) + J − + φ(x, y)
y y0

for all (x, y) ∈ R2 and such that

   
x x
 → 0 
φ(x, y) y y0
    −−−−−−−−→ 0
x x0
−
y y0

where k · k denotes the standard euclidean norm in R2 .

The definition just explains that we have a linear approximation at (x0 , y0 ) given by the
matrix J because the remainder term φ goes fast enough to zero. With this definition of a
local linear approximation, we can only at most one. Hence there can only be one matrix
J that satisfies everything from above and this matrix is called the Jacobian matrix of fR
at (x0 , y0 ) ∈ R2 .
It turns out that one can easily calculate that one with partial derivatives which we put
as columns in the matrix
| |
 

J =  ∂f∂x
R ∂fR 
∂y evaluated at (x0 , y0 )
| |

Example 5.4.
 2 
x − y2
Consider fR given by (x, y) 7→ . Then at position (x0 , y0 ) we have
2xy
 
2x0 −2y0
J=
2y0 2x0
6 Cauchy-Riemann Equations
By comparing the differentiability definition in R2 with the one in C, we reach the so-called
Cauchy-Riemann differential equations.

Cauchy-Riemann equations

Watch part 6 on YouTube.

https:// [Link]/ BvvFdmOrob4

For the complex function f : C → C, the linear approximation at z0 is written as:

f (z0 ) + f 0 (z0 )(z − z0 ) .
On the other hand, the corresponding function fR : R2 → R2 has the linear approximation
as    !
x x0
fR (x0 , y0 ) + J − .
y y0
This means that, on the complex level, we have the multiplication of complex numbers,
and, on the real R2 level, we have a matrix-vector multiplication. These are completely
different things but for complex differentiable functions they have to be connected. So we
immediately get the following:
Question:
In which cases does a matrix-vector multiplication represent a multiplication of
complex numbers?

To answer this, let’s say that a complex number w = a + ib ∈ C should represent a

2 × 2-matrix, and another complex number z = x + iy ∈ C represents a column vector in
R2 as always. First the multiplication of complex numbers is given by
w · z = (a + ib) · (x + iy) = ax − by + i(bx + ay)
and we can see the same with the following matrix-vector multiplication:
    
a −b x ax − by
= .
b a y bx + ay

With that we have answered the question: the matrix has to be in this form. Hence, for
the differentiability notion, the Jacobian of fR cannot be any matrix but it has to be in
this form. Let’s formulate this as equivalences.
6 Cauchy-Riemann Equations 18

Theorem 6.1. Cauchy-Riemann equations

For a function f : C → C, the following claims are equivalent:

(a) f is complex differentiable at z0 = x0 + iy0 .

(b) The induced map fR : R2 →  R isdifferentiable at (x0 , y0 ) where the Jacobian

2

a −b
at this point has the form .
b a

(c) The induced  map fR : R → R is differentiable at (x0 , y0 ) and if we write

2 2

u(x, y)
fR (x, y) = with two real-valued functions, then they satisfy:
v(x, y)

∂u ∂v ∂u ∂v
(x0 , y0 ) = (x0 , y0 ) , (x0 , y0 ) = − (x0 , y0 ) .
∂x ∂y ∂y ∂x

Proof. Essentially, we have already finish the proof. One just has to note that the Jacobian
of fR can be written as !
∂u ∂u
∂x ∂y
∂v ∂v
.
∂x ∂y
 
a −b
Together with the knowledge that the matrix has to be of the form , we get the
b a
differential equations.

Definition 6.2. Cauchy-Riemann equations

The two equations for real-valued functions u : R2 → R, v : R2 → R given by
∂u ∂v ∂u ∂v
= , =− ,
∂x ∂y ∂y ∂x
are called Cauchy-Riemann equations.
7 Cauchy-Riemann Equations Examples
Let’s now look at examples for the Cauchy-Riemann equations.

Examples for Cauchy-Riemann

Watch part 7 on YouTube.

From the last chapter we already now the equivalences for complex differentiability at
a given point. We can extend this to a claim for differentiability at every point. This
directly explains the following formulation for holomorphic functions.
Theorem 7.1. Cauchy-Riemann equations for holomorphic functions
Let U ⊆ C be open and UR ⊆ R2 the corresponding domain in R2 . f : U → C. We
can take the real part of f as a function u : UR → R and the imaginary part of f
as a function v : UR → R. Then f is holomorphic if and only if
∂u ∂v ∂u ∂v
= , =− ,
∂x ∂y ∂y ∂x
at all points in UR .

Let’s consider some examples of complex-valued functions where we can immediately

check the two Cauchy-Riemann equations from above.
Example 7.2. Holomorphic functions

(a) f : C → C with f (z) = z. This means u(x, y) = x and v(x, y) = y. So we get:

∂u ∂v ∂u ∂v
=1= and =0=−
∂x ∂y ∂y ∂x

(b) f : C → C with f (z) = z. This means u(x, y) = x and v(x, y) = −y. So we

get:
∂u ∂v
= 1 6= −1 =
∂x ∂y
This means, the function cannot be holomorphic.

(c) f : C → C with f (z) = z 2 + iz. This means u(x, y) = x2 − y 2 − y and

7 Cauchy-Riemann Equations Examples 20

v(x, y) = 2xy + x. So we get:

∂u ∂v ∂u ∂v
= 2x = and = −2y − 1 = −
∂x ∂y ∂y ∂x
So the quadratic function is holomorphic.
8 Wirtinger Derivatives
The following Wirtinger derivatives can be very helpful in quick calculations and give a
nice short formula for the Cauchy-Riemann equations.

Wirtinger Calculus

Watch part 8 on YouTube.

For a complex-valued function f : U → C, we can define two new symbols

∂f ∂f
(z0 ) , (z0 ) ,
∂z ∂z
and these are called the Wirtinger derivatives. For a holomorphic function, the first one
should just be the complex derivative and the second one should be zero because, roughly
speaking, a holomorphic function should not contain z. Recall that we already know that
z 7→ z is not holomorphic. Let’s see how we can actually define these strange looking
partial derivatives.
Reminder: Connection to R2
Any holomorphic function f : U → C can be written as f (x + iy)= u(x, y) + iv(x, y)
u(x, y)
which also induces a two-dimensional real function fR (x, y) := , where u
v(x, y)
and v are real-valued functions. We also know that for each z0 = x0 +iy0 , the
 complex 
a −b
number f (z0 ) = a + ib corresponds to a Jacobian for fR in the form
0
.
b a
Moreover, there we know, by looking at the first column, that a = ∂u ∂x
(x0 , y0 ) and
b = ∂x (x0 , y0 ). Hence we get
∂v

∂u ∂v
f 0 (z0 ) = f 0 (x0 + iy0 ) = (x0 , y0 ) + i (x0 , y0 )
∂x ∂x

Now, in order to bring in derivatives with respect to y as well, we can use the Cauchy-
Riemann equations, given by
∂u ∂v ∂u ∂v
= , =− ,
∂x ∂y ∂y ∂x
8 Wirtinger Derivatives 22

to get the following:

1  ∂u ∂v ∂v ∂u 
f 0 (z0 ) = (x0 , y0 ) + i (x0 , y0 ) + (x0 , y0 ) − i (x0 , y0 )
2 ∂x ∂x ∂y ∂y
1 ∂
 ∂ 
= (u + iv) − i (u + iv)
2 ∂x ∂y (x0 ,y0 )

This explains the following notation:

Definition 8.1. Wirtinger derivatives
For the complex variable z = x + iy, we use the following abbreviations:
∂ 1 ∂ ∂  ∂ 1 ∂ ∂ 
:= −i , := +i .
∂z 2 ∂x ∂y ∂z 2 ∂x ∂y

These new partial derivatives use z and z as the two variables for the differentiation.
What first might seem confusing is actually a helpful mnemonic device. For example, we
get the following derivatives for the complex functions.
Example 8.2. Wirtinger derivatives applied
∂ 3 ∂ 1 1 ∂
z = 3z 2 , =− 2, z = 1.
∂z ∂z z z ∂z

Moreover, since the partial derivative with respect to z gives a non-vanishing result if
and only if the Cauchy-Riemann equations are not satisfied, we have a new criterion for
complex differentiability.
Proposition 8.3. Criterion for Holomorphic Functions
Let U ⊆ C be open and f : U → C. Then:
∂f
f is holomorphic ⇐⇒ (z) = 0 for all z ∈ U
∂z
In this case, the complex derivative of f is given by f 0 (z) = ∂f
∂z
(z)
9 Power Series
Let’s discuss the general concept of a power series in Complex Analysis. It turns out that
we can do exactly the same thing as in the real numbers but the complex numbers will
explain the behaviour some power series have in a really clear way.

Power Series in the complex realm

Watch part 9 on YouTube.

Note that we the most important function in this course given as a power series, namely
the exponential function:

∞
zk
exp(z) :=
X
.
k=0
k!

This looks exactly like the real counterpart but we can also put complex numbers into the
formula, which means we need the calculation rules of comlex numbers. Then the outcome
is a complex number as well. Here please don’t forget, there is a symbol infinity which
means there is a limit involved. This means that we have a limit of complex numbers,
which is defined by the metric given by the absolute value.
Now with this philosophy we can also talk about the general definition of a power series
in the complex field.
Definition 9.1. Power Series in the Complex Realm
For a sequence of complex numbers (ak )k∈N , the function
∞
X
f : D → C, z 7→ ak (z − z0 )k
k=0

is called a power series. The complex number z0 is called expansion point and the
set D ⊆ C is the domain of convergence and given by:
 N 
lim ak (z − z0 ) exists in C .
X
k
D := z ∈ C
N →∞
k=0

You see that the domain is chosen in a maximal way. We define the complex function f for
every complex number for which it makes sense. This means that we just require that the
9 Power Series 24

series involved is actually convergent. Soon we will seen that the domain of convergence
is always in the shape of a disc. In the worst case, D only contains one point, which
is the expansion point z0 . Obviously, at this point the series collapses to one term and
is, therefore, always convergent. Often, we will just consider the expansion point z0 = 0
because it’s just a translation in the domain of definition.
Let’s verify these ideas with examples. The first one is an important one and you already
know it from real analysis. It’s the famous geometric series.
Example 9.2. Geometric Series
Let’s consider the power series where all coefficients ak are given by the number 1:
∞
1
for |z| < 1
X
zk =
k=0
1−z

This power series is divergent for |z| ≥ 1, so D = B1 (0).

Let’s prove this facts about the geometric series like we have done it in the real analysis
course.
Now, consider the partial sum sN := N k=0 z for z ∈ C. We can calculate can do some
k
P
telescoping:
XN N
X +1 XN
N +1 k k
1−z = z − z = (1 − z) zk
k=0 k=1 k=0

So we can distiguish two cases and we get the follwing geometric sum formula:

if z 6= 1,
N
( N +1
X 1−z
sN = zk = 1−z

k=0
N +1 if z = 1.

Hence, the sequence of partial sums (sN )n∈N is convergent if and only if |z| < 1. In this
case we have ∞
1 − z N +1 1
z k = lim sN = lim
X
= .
k=0
N →∞ N →∞ 1 − z 1 − z

So for this example we get a disc as the domain of convergence, and since the radius
of this disc is 1, we say that the radius of convergence for the power series is 1. In the
following we will see that we can find such a radius for every power series. From now on
we will use the following useful abbreviation [0, ∞] := [0, ∞) ∪ {∞}.
9 Power Series 25

Theorem 9.3. Domain of convergence is always disc-shaped

For a power series f : D → C given by z 7→ ∞ k=0 ak (z − z0 ) , there is a maximal
k
P
r ∈ [0, ∞] such that the open disc around z0 with radius r, written as Br (z0 ), where
B∞ (z0 ) := C, satisfies:

(1) Br (z0 ) ⊆ D (there we also have absolute convergence of the series)

(2) for z ∈/ Br (z0 ), the series ∞

k=0 ak (z − z0 ) is divergent.
k
P

This maximal radius is called the radius of convergence and can be calculated by the
Cauchy-Hadamard formula:
1
= lim sup k |ak | ,
p
r k→∞

where we formally define 1/∞ := 0 and 1/0 := ∞.

Proof. We can just define the number r ∈ [0, ∞] by r−1 = lim supk→∞ |ak |. Then we
p
k

have to show the two statements:

(i) For all z ∈ C with |z − z0 |r−1 < 1, the power series is convergent.
(ii) For all z ∈ C with |z − z0 |r−1 > 1, the power series is divergent.
We can prove these statements byP using what we know from the geometric series. So let’s
fix z ∈ C and consider the series ∞k=0 ak (z − z0 ) . For this, we define
k

q := |z − z0 | lim sup k |ak | ∈ [0, ∞] .

p
k→∞

For (i), we have q < 1, which means we can also make it slightly larger q < qe < 1 and
still keep it under 1. Now by the definition of lim sup, we get that there is k0 such that
for all k ≥ k0 the sequence members satisfy:
p
|z − z0 | k |ak | ≤ qe .

Now we can take k-th power and obtain that for all k ≥ k0

|ak ||z − z0 |k ≤ qe k
9 Power Series 26

Therefore, the convergent geometric series ∞ e k is a majorant of ∞

k=0 ak (z −z0 ) and,
k
P P
k=0 q
therefore, we have absolute convergence of this series.
For showing (ii), we now assume q > 1. Using the definition of the limit superior, we find
infinitely many sequence members such that
p
|z − z0 | k |ak | ≥ 1 .

Hence by taking the k-th power, we see that the coefficients satisfy

|ak ||z − z0 |k ≥ 1 for infinitely many k .

So thePleft-hand side cannot be a sequence that converges to 0, which is required for the
series ∞ k=0 ak (z − z0 ) being convergent.
k

The theorem above with the Cauchy-Hadamard formula characterizes convergence and
divergence of the power series in dependence of z whether it is inside or outside the disc
around z0 with radius r. In the case that z lies on the boundary, meaning |z − z0 | = r,
this result does not tell us anything. Indeed, we may have points on the circle with z ∈ D
and also points on the circle with z ∈ / D.
Example 9.4. Exponential Series
The power series given by
∞
zk
exp(z) :=
X

k=0
k!
has a radius of convergence of ∞.

It’s definitely possible to use the Cauchy-Hadamard formula to calculate this radius of
convergence, but especially for this example, we can use the following result.
Proposition 9.5. Radius of Convergence with Ratio Test
Let z 7→ ∞ k=0 ak (z − z0 ) be a power series with radius of convergence r ∈ [0, ∞].
k
P
If the following limit exists in [0, ∞], we have

|ak |
lim =r
k→∞ |ak+1 |

Proof. Let’s call ρ := limk→∞ |a|ak+1

k|
|
and we know ρ ∈ [0, ∞]. To show that ρ is actually
the radius of convergence r. For that, consider:

|ak+1 ||z − z0 |k+1 |z − z0 |

lim k
= =: q
k→∞ |ak ||z − z0 | ρ

The ratio test or quotient criterion for series tells out that ∞ k=0 ak (z − z0 ) is absolutely
k
P
convergent if q < 1 and divergent for q > 1. Therefore, ρ has to be the radius of
convergence r.

With that formula we can easily calculate the radius of convergence for the exponential
function above:
9 Power Series 27

1
|ak | k! k→∞
= 1 = k + 1 −−−→ ∞
|ak+1 | (k+1)k!
10 Uniform Convergence
When we talk about power series, the notion of uniform convergence is very important.

Power series convergence uniformly

Watch part 10 on YouTube.

https:// [Link]/ qSBdxI3rI04

The notion of uniform convergence for a sequence of functions is the same as in real
analysis. First we need a distance function for two functions f, g : U → C, which we can
choose as the largest distance between the two graphs, that means the distance you can
measure at a given point:

sup |f (z) − g(z)| =: kf − gk∞

z∈U

For a sequence of functions (fn )n∈N , we have the uniform convergence to a limit function
f : U → C if this measured distance between fn and f is convergent to zero.
Definition 10.1. Uniform convergence
Let U ⊆ C and fn : U → C for every n ∈ N. We say the the sequence of functions
(fn )n∈N converges uniformly to a function f : U → C if
n→∞
kfn − f k∞ −−−→ 0 .
11 Power Series Are Holomorphic -
Proof

Mhh
Watch part 11 on YouTube.

https:// [Link]/ 2_4W_XyADfQ

12 Exp, Cos and Sin as Power Series

Mhh
Watch part 12 on YouTube.

https:// [Link]/ YP_htwpIB7Q

13 Complex Logarithm
In this part, we will learn a lot about the exponential function in the complex realm and
also about its inverse, the logarithm function.

Logarithms in the complex realm

Watch part 13 on YouTube.

First let’s see what we usually do in the real numbers. There the exponential function
exp : R → (0, ∞) can be restricted in the codomain to its range, which is (0, ∞). Since
it’s a strictly monotonically increasing function, this makes the whole map into a bijective
map. So it’s possible to define the inverse function and call it the logarithm function in
the real numbers:
log : (0, ∞) → R
We write log(x) to denote the natural logarithm but other authors might use ln(x) to
make it really clear that the basis is chosen as Euler’s number e. Since all logarithms in
this course will be natural ones, there is no need for this notation.

Now, let’s see how the exponential function works in the complex numbers. The definition
is the same as in the real numbers, given by a power series:

∞
zk
exp(z) :=
X
.
k=0
k!

From this definition, and by using the Cauchy product of series, we get the important
formula:
13 Complex Logarithm 32

exp(z + w) = exp(z) exp(w) for all w, z ∈ C .

We call it the fundamental multiplicative identity for the exponential function. Let’s
collect some important facts:
Proposition 13.1. On the unit circle
For every y ∈ R, we have
| exp(iy)| = 1

Proof. Note that the conjugate of is obviously given by k=0 k! . Hence we have:
P∞ zk
P∞ zk
k=0 k!

| exp(iy)|2 = exp(iy) · exp(−iy)

Now we can use the fundamental multiplicative identity from above and get:

| exp(iy)|2 = exp(iy − iy) = exp(0) = 1

This proposition tells us that exp(iy) always lies on the unit circle if y is a real number.
We can use this together with the fundamental multiplicative identity to describe each
value of the exponential function:

exp(z) = exp(x + iy) = exp(x) exp(iy) for z = x + iy

So we learn that the absolute value of exp(z) is given by the real exponential function
exp(x). And on the other hand, the position on the unit circle is given by exp(iy). This
means it’s possible to split that into real and imaginary part:

exp(iy) = cos(y) + i sin(y)

We can see this as the definition of the two real functions cos and sin but we can also use
the power series description which also implies the formula above:
Definition 13.2. Sine and Cosine functions
∞ ∞
z 2k z 2k+1
cos(z) := sin(z) :=
X X
(−1)k , (−1)k .
k=0
(2k)! k=0
(2k + 1)!
13 Complex Logarithm 33

Proposition 13.3. Euler’s formula

For any complex number z ∈ C, we have

exp(iz) = cos(z) + i sin(z) .

Proof. By using the power series representation and using the fact i4 = 1, we get
∞ ∞ ∞ ∞ ∞
zk z 4m z 4m+1 z 4m+2 z 4m+3
exp(iz) =
X X X X X
k
i = +i − −i .
k=0
k! m=0
(4m)! m=0
(4m + 1)! m=0
(4m + 2)! m=0
(4m + 3)!

Then we can put the two real parts together and, in the same way, we put the two
imaginary parts together:
∞ ∞ ∞
zk ` z
2`
z 2`+1
exp(iz) = = cos(z) + i sin(z) .
X X X
k
i = (−1) +i (−1)`
k=0
k! `=0
(2`)! `=0
(2` + 1)!

At this point, we can play dumb and act as we have never seen the functions Cosine and
Sine before and that we only know the power series definition and Euler’s formula above.
This means that we also don’t know that both functions are periodic when we see them
as real functions R → R. However, some things are really easy to prove by just using the
power series. For example, the derivatives are directly given by differentiation under the
sum symbol:
sin0 = cos , cos0 = − sin .
Moreover, we also already know cos(0) = 1 and sin(0) = 0. Also by Proposition 13.1, we
have cos(y)2 + sin(y)2 = 1 for every real number y ∈ R. We can also put a larger real
number into the Cosine function:
1 1 1 1
cos(3) = 1 − 32 + 34 − 36 + 38 − · · ·
2 4! 6! 8!
Looking at the absolute value of each term, we see that they are decreasing. However,
the alternating sign tells us the everything summed up after the first three terms will be
negative. Hence, we can estimate the cos(3) by using just the first three parts:
1 1 36 27 9
cos(3) ≤ 1 − 32 + 34 = 1 − + =1− <0
2 4! 8 8 8
This means that the continuous function cos : R → R must have at least one zero inside
the interval [0, 3]. There must a smallest one and this one we call π2 .
Definition 13.4. Pi
For the real function cos : R → R, there exists a smallest positive number x0 ∈ [0, 3]
such that cos(x0 ) = 0. We define

π := 2x0 .

Hence we can say, by definition: cos( π2 ) = 0. This helps immediately to get the value of
the exponential function at i π2 :
13 Complex Logarithm 34

 π π  π  π 
exp i = cos + i sin = i sin
2 2 2 2
We already know that this number has to lie on the unit circle and now we also know
that it lies on the imaginary axis. Which means:
π 
sin ∈ {−1, 1} .
2
But which of the cases actually is it? We already know that sin(0) = 0 and that sin0 (x) =
cos(x) So the sine function is definitely strictly monotonically increasing for x ∈ [0, π/2)
because the cosine function is positive there. So the value of sine has to be positive for
all x ∈ [0, π/2] and we have expliclity showed:
π   π
sin = 1 and exp i = i.
2 2
Now let’s use the fundamental multiplicative identity again to get:

exp(iπ) = −1 and exp(i2π) = 1 .

This proves the following:

Proposition 13.5. Periodicity of the complex exponentional
For z ∈ C and k ∈ Z, we have

exp(z + i2πk) = exp(z) .

The crucial thing here is that the exponential function in the complex plane cannot be
injective. We can only make it injective if we restrict it to a strip.

If we want that the strip is an open set, then it’s clear that also a whole line in in codomain
is missing. We have to take this into account in order to achieve surjectivity. We define
the slotted complex plane
Dπ := C \ (−∞, 0] ,
and the symmetric strip

S(−iπ,iπ) := {z ∈ C | −iπ < Im(z) < iπ} .

With these sets, we have a bijective map exp : S(−iπ,iπ) → Dπ and we can define the
inverse.
13 Complex Logarithm 35

Definition 13.6. Logarithm - principal value

The inverse of the exponential function exp : S(−iπ,iπ) → Dπ is called the principal
value of the logarithm function and denoted by

log : Dπ → S(−iπ,iπ) .

We call it the principal value since we have a lot of different possibilities to define an inverse
function if we choose different restrictions for the exponential function. For example, we
can consider a different slotted plane for an angle ϕ ∈ R by setting

Dϕ := C \ {r exp(iϕ) | r ≥ 0}

and we can consider the corresponding strip

S(iϕ−2πi,iϕ) := {z ∈ C | iϕ − 2πi < Im(z) < iϕ} .

Also there the exponential function is bijective and lead an inverse function, which we
could denote by:
logϕ : Dϕ → S(iϕ−2πi,iϕ) .
Note that the principal value is, using this special notation, just given by logπ .
Proposition 13.7. Properties of the complex logarithm
For real numbers r > 0 and y ∈ R with −π < y < π, the principal value of the
logarithm functions satisfies

log(r exp(iy)) = log(r) + iy

where log(r) is just the common real logarithm.

We get a similar result for the other logarithm parts:

logϕ (r exp(iy)) = log(r) + iy , for ϕ − 2π < y < ϕ .

What we can see in this formula is that we cannot extend the domain Dπ more to also
include the missing line without losing the continuity of the function. We see that there
is a jump of 2πi at the line:
log(exp(iy)) −−→ iπ
y→π

log(exp(iy)) −−−→ −iπ

y→−π

For the rest of this book, please never forget this property. It will be essential for all the
integrations we will do. Keep the following remark in your heart and internalise it so that
you never forget it.
Remark:
A slotted plane Dϕ is largest possible domain such that a complex logarithm function
can be a continuous functions. There is no continuos logarithm function defined on
C \ {0}.
14 Powers
We can use the logarithm in the complex plane to define general powers of complex
numbers.

Powers with complex numbers

Watch part 14 on YouTube.

Of course, we don’t have a problem to define natural powers for complex numbers because
z 3 := z · z · z makes immediately sense. However, we really don’t know what ii should
represent since complex powers is a completely new subject for us.
Let’s start be recalling the general power definition with real numbers: take a real number
a > 0 and then, for any integer n ∈ N, an is naturally defined by
1 1 1
an := a · a · · · · · a and a−n := · · ··· ·
a a a
where each product has exactly n factors.
Moreover we can also define a to the power 1/n for every n ∈ N:
1 √
n
a n := a

And finally we can combine these definitions to get a power for every rational exponent:
 1 m
, for m ∈ Z , n ∈ N .
m
a n := a n

At this point we can use the real exponential function together with its inverse function,
the real logarithm, to get a nice relation:
  1 m     1   
m
= exp log a = exp m · log a = exp · log(a) (14.1)
m
a n n n
n

Here we just applied two times the property of the logarithm which follows from the
fundamental multiplicative identity of the exponential function:

log(xy) = log(x) + log(y) .

14 Powers 37

√
One just need to do an induction to get log(an ) = n · log(a) and for b := n
a:
√
log(a) = log(bn ) = n · log(b) = n log( n a)

which is exactly what we needed. Also note that log(a−1 ) = − log(a) since we always
have exp(−b) exp(b) = 1 for b ∈ R.
Now the idea is to use the formula in (14.1) to lift the power definition from the rational
numbers to all real numbers:
Definition 14.1. Power in the real numbers
For every real number a > 0 and x ∈ R, we set:

ax := exp(x · log(a))

This one makes the power function x 7→ ax into a continuous function defined on the
whole real number line which fits with the common understanding of natural powers, see
(14.1).
Now the question is: can we use the same definition for complex numbers? At least if we
keep a real base, this is not a problem at all:
Definition 14.2. Power with complex exponent
For every real number a > 0 and complex number z ∈ C, we set

az := exp(z · log(a))

where log denotes the natural real logarithm function.

For example, for Euler’s number e, we immediately get ez = exp(z) with this definition,
which explains why we can also use the abbreviation for the exponential function for
complex numbers. Indeed, know we are allowed to write eiπ = −1, for example.
However, this definition can only work for a complex basis a ∈ C if we go to the complex
logarithm function on the right-hand side. But as we already know, there is no unique
choice for that. We have a lot of continuous logarithm functions and none is defined on
the whole complex plane.
So, it might happen that we have a lot of power functions on the complex plane. Hence it
makes sense to fix the one that comes from the principal value of the logarithm function.
14 Powers 38

Definition 14.3. Principal value of power

For every complex number a ∈ D−π and complex number z ∈ C, we set

az := exp(z · log(a))

where log denotes the principal value of logarithm function.

Note that something like (−1)i is not included in this definition! For that reason some
people interpret the power definition as a whole set:
Remark: Alternative interpretation
One can see az , for a, z ∈ C \ {0}, as the set of all possible branches of the power
function:
set of az := {exp(z · logϕ (a)) | ϕ ∈ R possible}
Note that this means:

set of az = {exp(z · w) | w ∈ C such that exp(w) = a}

= {exp(z · (logϕ (a) + i2πk)) | k ∈ Z}

where ϕ ∈ R is fixed such that logϕ (a) is defined.

This can be helpful because we get:

set of (−1) 2 = {−i, i} .

1

Moreover, this also extends to roots of positive numbers:

n  2π   4π o
set of 1 = 1, exp i , exp i
1
3 .
3 3

However, usually we will not work with this interpretation of the power as a set but rather
taking the principal value as given in Definition 14.3. The disadvantage of this is that
not all the power rules are satisfied.
Attention! Power rules
For the principal value of the power we always have

az · aw = az+w , but az 6= az·w in general .

w
15 Laurent Series
The properties of holomorphic functions can be described with power series and more
generally with so-called Laurent series.

Laurent Series in the Complex Realm

Watch part 15 on YouTube.

16 Isolated Singularities

Mhh
Watch part 16 on YouTube.

https:// [Link]/ 4LzhMll8O8U

17 Complex Integration on Real
Intervals
Now we are finally ready to extend the integral into the complex realm. We start with
defining the integral for functions on real intervals.

Complex-valued integrals

Watch part 17 on YouTube.

So what we want is to define an integral for a function γ where the codomain is given by
the complex-numbers.

For us is sufficient to only consider continuous functions such that we don’t have a problem
with integrability.
Definition 17.1. Extend integral to the complex realm
For a continuous function γ : [a, b] → C, we have two real-valued functions given
by real and imaginary part,

Re(γ) : [a, b] → R , Im(γ) : [a, b] → R ,

which are continuous and Riemann-integrable. Hence, we define the complex inte-
gral of γ by linearity:
Z Z Z
γ(t) dt := Re(γ)(t) dt + i Im(γ)(t) dt .
[a,b] [a,b] [a,b]

Note that this definiton works because the function γ is defined on a real interval such
that we can use the ordinary Riemann-integral on R.
18 Complex Contour Integral

Mhh
Watch part 18 on YouTube.
19 Properties of the Complex Contour
Integral

Mhh
Watch part 19 on YouTube.
20 Antiderivatives
In the next part, we introduce primitives for complex-valued functions. A more suitable
name is just antiderivatives. They can be used to calculate contour integrals.

Antiderivatives for complex functions

Watch part 20 on YouTube.

The definition of such an antiderivative is not very surprising and, actually, exactly the
same as in our Real Analysis course.
Definition 20.1. Antiderivative/Primitive
Let U ⊆ C open and f : U → C be a complex-valued function. Then a function
F : U → C is called an antiderivative of f or primitive of f if F is complex
differentiable at each point and satisfies

F 0 (x) = f (x) for all x ∈ U .

Obviously, if there is at least one antiderivative, we get infinitely many by adding a

constant to F . In the case that U is a connected domain, these are all the antiderivatives
we can find.
Now, from real analysis, you might know the famous fundamental theorem of calculus.
This one we can use the get a similar result for complex functions as well. The key
ingredient is just the chain rule for the complex derivative:
d
F (γ(t)) = F 0 (γ(t)) γ 0 (t) (20.1)
dt

Proposition 20.2. Antiderivatives in curve integrals

Let U ⊆ C open and f : U → C be a complex-valued function and γ : [a, b] → U
a parametrized curve, which continuously differentiable. If f has an antiderivative
F : U → C, then Z
f (z) dz = F (γ(b)) − F (γ(a))
γ

So you can remember that an antiderivative immediately gives the result of a curve inte-
gral, no matter which way the curve takes. Only the start point and the end point of the
curve determine the value of the whole integral.
20 Antiderivatives 45

Proof. We can immediately calculate it:

21 Closed curves and antiderivatives
The last chapter showed that the existence of an antiderivative implies that any closed
curve integral is zero. Now we will discuss the converse.

Closed curves of holomorphic functions

Watch part 21 on YouTube.

To formulate the claim of the video we will need a special manner of speaking which is
common in Complex Analysis.
Definition 21.1. Open region/ open domain
A subset D ⊆ C is called an open domain or open region if it is open and path-
connected, which means that any two points z, w ∈ D can be connected by a con-
tinuos curve γ : [a, b] → D with γ(a) = z and γ(b) = w.

This is easy to illustrate because such an open region can only have one connected com-
ponent. It’s not allowed to have separate islands as the set D.

Theorem 21.2. Closed curves imply antiderivative

Let
H D ⊆ C be an open region and f : D → C be a holomorphic function. If
γ
f (z) dz = 0 for all possible closed curves, then f has an antiderivative.
22 Goursat’s Theorem

Mhh
Watch part 22 on YouTube.
23 Cauchy’s theorem

Mhh
Watch part 23 on YouTube.
24 Winding Number
The winding number, sometimes also called index, measures how many turns a curve does
around a given point.

Winding number for complex curves

Watch part 24 on YouTube.

If we have a circle around the origin, like t → exp(it), then we say the the curve takes
one turn around the circle in the case that t covers the whole interval [0, 2π]. If t goes
beyond that interval, we immediately have more than just turn. So the natural question
is: how can we measure the number of turns for any curve around the origin? Is there a
simple formula for that?

This picture shows a closed curve which goes around z0 two times but only once around
z1 . Of course, the picture does not show all the details but the idea should be clear. So
let’s start putting a concrete definition together. We begin by consindering circles:
Remark: Circles around the origin
For a circle given by γ : t 7→ eit we get the following:

2πi , for γ : [0, 2π] → C


I 
1
dz = 4πi , for γ : [0, 4π] → C
γ z
6πi , for γ : [0, 6π] → C


24 Winding Number 50

So, at least for such closed curve integrals given by circles, we have that
I
1 1
dz
2πi γ z
gives us an integer that corresponds to number one times we go around the origin in the
positive sense. Obviously, we can also translate this to any other point z0 in the complex
plane. All these things lead to the following.
Definition 24.1. Winding number
The winding number of a complex curve γ : [a, b] → C \ {z0 } around a point z0 is
defined as Z
1 1
wind(γ, z0 ) := dz .
2πi γ z − z0

The definition shows that we can talk about the winding number also for non-closed
curves but these will not be the important cases. Indeed, we have to show that we get an
integer out for closed curves, like we have it for the circles above.
Proposition 24.2. Winding numbers counts
For a closed curve γ : [a, b] → C \ {z0 }, we always have wind(γ, z0 ) ∈ Z.

Proof. Without loss of generality, we can choose z0 = 0 because it’s just a translation in
the plane. Now, the idea is to split the curve around the origin into two continuous and
piecewise continuously differentiable functions such that we have γ(t) = ρ(t)eiϕ(t) for all
t ∈ [a, b].

The first function is ρ : [a, b] → (0, ∞) defined by ρ(t) := |γ(t)|. Since γ is piecewise
continuously differentiable, our real function ρ is it as well. On the other hand, the
argument function ϕ : [a, b] → R is not so easy to define. For the first step let’s decompose
the interval [a, b] into finite many segments
a =: a0 < a1 < a2 < . . . < an < an+1 := b
such that each restriction γ|[aj ,aj+1 ] has image completely inside a slotted plane, denoted
by
Dθ := C \ {r exp(iθ) | r ≥ 0} .
24 Winding Number 51

Hence, from part 13, we already know that we have a continuos and holomorphic logarithm
function on such a slotted plane:

logθ : Dθ → C , logθ (r exp(iy)) = log(r) + iy , θ − 2π < y < θ .

So, we can just define, while using t ∈ [aj , aj+1 ],

 
1
ϕ(t) := logθ (γ(t)) − log(ρ(t))
i

Obviously, it’s possible to choose the next theta such that we have an overlap for the next
logarithm function. After that, the only thing we have to do is to glue these logarithm
functions together such that we get a continuous argument function ϕ : [a, b] → R,
defined on the whole interval [a, b]. By looking at the subintervals, we also see that we
have piecewise continuously differentiable function.
Finally, we can solve the contour integral while using the product rule:
I Z b 0 Z b
1 γ (t) 1 
0 iϕ(t) 0 iϕ(t)

dz = dt = iϕ(t)
ρ (t)e + ρ(t)iϕ (t)e dt
γ z a γ(t) a ρ(t)e
Z b 0 Z b
ρ (t) b
= dt + i ϕ0 (t) dt = log(ρ(t)) + i(ϕ(b) − ϕ(a))
a ρ(t) a a

Since γ is a closed curve, we ρ(b) = ρ(a) and eiϕ(a) = eiϕ(b) , which immediately implies

ϕ(b) = ϕ(a) + 2πk , for k ∈ Z .

This integer k is exactly our winding number.

25 Cauchy’s Theorem (general version)

Mhh
Watch part 25 on YouTube.
26 Keyhole contour

Mhh
Watch part 26 on YouTube.
27 Cauchy’s Integral Formula

Mhh
Watch part 27 on YouTube.
28 Holomorphic Functions are
C-infinity Functions
In the next part, we will generalise Cauchy’s integral formula also for derivatives. While
doing this, we also show that each holomorphic can locally be represented by power series.
In particular, this shows that each holomorphic function is a C ∞ -function.

Mhh
Watch part 28 on YouTube.
29 Liouville’s Theorem

Mhh
Watch part 29 on YouTube.
30 Identity Theorem
As a consequence from the last results, we can prove the famous identity theorem.

Mhh
Watch part 30 on YouTube.
31 Application of the Identity Theorem
Next, let’s see some nice applications of the identity theorem from above.

Mhh
Watch part 31 on YouTube.
32 Residue

Mhh
Watch part 32 on YouTube.
33 Residue for Poles

Mhh
Watch part 33 on YouTube.
34 Residue theorem

Mhh
Watch part 34 on YouTube.
35 Application of the Residue Theorem
To end this book, we show how we can use the residue theorem to calculate a real integral
with it. This is exactly what we used to motivate at the beginning and it’s interesting
application of the complex theory to solve real problems.

Solving Real integrals

Watch part 35 on YouTube.

Index

ε-ball, 7 Jacobian matrix, 16

(complex) derivative of f at z0 , 11
(complex) differentiable at z0 , 9 Laurent series, 39
limit, 7
antiderivative, 44 logarithm, 31
antiderivatives, 44
metric space, 7
ball, 7
open, 8
Cauchy-Hadamard formula, 25 open domain, 46
Cauchy-Riemann differential equations, open region, 46
17 open sets, 8
Cauchy-Riemann equations, 18
chain rule, 13 path-connected, 46
Complex Analysis, 6 power series, 23
primitive, 44
complex differentiable, 12
primitives, 44
complex integral, 41
principal value of the logarithm
complex numbers, 6
function, 35
continuous at, 7
product rule, 13
convergent, 7
converges uniformly, 28 radius of convergence, 24, 25
rational functions, 14
domain of convergence, 23
Real Analysis, 6
entire, 13
set of all possible branches of the power
expansion point, 23
function, 38
fundamental multiplicative identity, 32 Start Learning Mathematics, 6
fundamental theorem of calculus, 44 sum rule, 13

geometric sum formula, 24 total differentiability, 15

totally differentiable, 16
holomorphic, 13
winding number, 50
index, 49 Wirtinger derivatives, 21