COMPLEX ANALYSIS

Engineering Physics Applied to Industry

Mathematical Methods Applied to Engineering
Welcome to a new course on the topic of mathematical analysis. In this case we plunge into the fascinating
branch of Complex Calculus. In previous courses, either of math or physics, you have dealt with complex
numbers already so you have a taste of the importance they suppose. In this part of the course we take all the
knowledge we already have and bring it a step (or rather a few steps) forward.

You probably have a bitter memory of complex numbers, mainly because you see them as something abstract
you don’t really comprehend and probably hard to deal with both algebraically and in their application. It is time
we make amends and become allies with them since complex numbers are of vital importance regarding both
physics and engineering (not to mention mathematics themselves).

In the brief time we have to study complex analysis we will try to cover the basics of complex functions and the
main results that will help us solve as many types of problems as possible. To do so we accompany these notes
with different examples and practical applications.

But first let’s talk a bit about history.

Brief History of Complex Numbers

The first appearence of a complex number was in the first century a.d. on the work Stereometrica of the greek
mathematician
p Hero of Alexandria where in certain calculations regarding pyramid construction he arrived at
the term 81 ¡ 144 . Roots of negative numbers were not conceived at that time so Hero merely changed it to
p
144 ¡ 81

It wasn’t until the 16th century when complex numbers began to sink
into the minds of mathematicians. The might to solve cubic equations
led to the conclusion that real numbers were not sufficient and it was
by hands of italian mathematicians, especially Gerolamo Cardano (on
the picture), that complex numbers were finally on the game.

Cardano found a way to solve cubic and quadric equations. This was
the biggest breakthrough since babilonic times. Things were still
unclear though since most of the big minds of the time didn’t really
rely on how consistant this was. Bare in mind that it took many years
for negative numbers to be accepted and now they were talking about
computing their square root (!).

A century later the french philosopher and mathematician René

Descartes first used the concept of imaginary numbers. Descartes
applied algebra to work on geometry and considered that imaginary
numbers led to geometric imposibilities. He was also the founder of
what we nowadays know as cartesian geometry. Girolamo Cardano (1501 - 1576)

About this time was also postulated the fact that polynomials of degree n should have n roots. Descartes was
one of the minds that correctly pointed this out, also assuming that not all roots must be real. This result was
later established as the Fundamental Theorem of Algebra.

During the following years many mathematicians joined the party and slowly started to accept complex
numbers and work with them. In the XVIII century they were broadly used, among others, by Huygens and
Leibniz, remarkable scientists of the time, the latter being the father of infinitesimal calculus (even though
Newton accused him of cheating).

One of the main utility of complex numbers at the time was at resolution of different integrals such as
Z Z Z µ ¶
1 1 1 1 1 1
dx = dx = ¡ ¡ dx = ¡ (log (x + ai) ¡ log (x ¡ ai))
x + a2
2 (x + ai) (x ¡ ai) 2ai x + ai x ¡ ai 2ai
Controversy came along with this type of calculations since the logarythm was not defined for negative values let
alone for complex numbers. Controversy was solved by Leonhard Euler who came up with the identity

ei¼ + 1 = 0

Probably the most important identity in Calculus since it involves the use of the fundamental numbers 0, 1, e, π
and the imaginary unit i. Euler was the first one naming i as the square root of negative 1 and got even further by
making the connection between the exponential and trigonometric functions:

eix = cos x + i sin x

The years to come were a bit messy since complex numbers fell into philosophycal arguments and the algebra
required to operate with them was nowhere close to being defined correctly. Big minds of Cambridge were
esceptical of teaching such topic cause the fundamentals were unstable.

Geometric representation of the complex numbers came to throw some light upon this obscure picture. It was the
norwegian Caspar Wessel who treated the geometry of complex numbers in his work then the swiss Jean-Robert
Argand also contributed. Nontheless it was Carl Friedrich Gauss’ word which stood out. It had been said that
Gauss was in possesion of the geometric representation of complex numbers far before any other person and it
was kept unpublished until 1831 or so.

If this subjet has hitherto been considered from the wrong viewpoint and thus
enveloped in mystery and surrounded by darkness, it is largely an unsuitable
terminology which should be blamed. Had +1, -1 and √−1, instead of being
called positive, negative and imaginary (or worse still, impossible) unity, been
given the names say,of direct, inverse and lateral unity, there would hardly have
been any scope for such obscurity

Carl Friedric Gauss

(1777 - 1855)

Gauss was also the one who proved the Fundamental Theorem of Algebra, which, as mentioned above, assures
that every polynomial of degree n has exactly n roots.

By the second half of the XIX century complex numbers were already pretty much accepted. At this time the
contribution to the topic by Agoustin-Louis Cauchy, french mathematician, was largely applauded and
established certain results that nowadays are widely used. Cauchy initiated complex function theory, the one we
are going to study in this course.
Algebra and geometry of Complex numbers

Before going into the main topics of the course we may as well recall the basic properties and notions of complex
numbers such as, how to compute basic operations, graphic representation, etcetera. The very beginning of this
story shall of course be the definition of a generic complex number, written in its binomial form:

z = a + bi; a; b 2 R
where i is the unit imaginary number, i.e. the square root of minus one √-1. The real number a is the real part of z
and b would be the imaginary part, also denoted as

Re (z) = a
Im (z) = b

The set of complex numbers is denoted by ℂ and contains all the rest of sets of numbers we know forming a
chain of inclusions

N½Z½Q½R½C
The set of complex numbers is a field which grants it with nice algebraic properties such as:

• The set of complex numbers is closed with respect to addition and multiplication, i.e. if we add, substract,
multiply or divide any pair of complex numbers we obtain yet another complex number.
• The set is commutative both under addition and multiplication:

z1 + z2 = z2 + z1
z1 ¢ z2 = z2 ¢ z1
• The set is associative both under addition and multiplication:

z1 + (z2 + z3) = (z1 + z2) + z3

z1 ¢ (z2 ¢ z3) = (z1 ¢ z2) ¢ z3
• The set is distributive:

z1 ¢ (z2 + z3) = z1 ¢ z2 + z1 ¢ z3

• The neutral element with respect to addition is 0 and 1 for the multiplication.
• Every complex number has an inverse both with respect to addition and multiplication.

The main difference with the field of real numbers ℝ is that the field of complex numbers is not ordered. This
means that there is no such thing as z_1 < z_2. If we think about it, it makes sense since complex numbers can be
represented on the plane.
For a complex number z = a + bi we can assign a point on the plane, say (a,b) so that there is a bijective
correspondance with the Euclidean 2-dimensional space. This correspondance is actually an isomorphism. In
mathematical terms, this means that both spaces share properties and it is denoted as

2
C»
= R
Thus the complex field can be equipped with all the geometric tools we know. Further so, we can represent a
complex number using polar coordinates:

f a = jzj cos µ
b = jzj sin µ
a b
b ooo

jzj : modulus
iz
µ : argument
10
a

The modulus is calculated using Pythagoras’ Theorem and the calculus of the argument depends on the quadrant
where the complex number lives:
p b
jzj = a2 + b2 ; µ = arctan [+¼ if z is in 2nd or -¼ if 3rd quadrant]
a
This leads to the trigonometric representation of a complex number:

z = a + bi = jzj (cos µ + i sin µ)

Or the more compact exponential form:

z = jzj eiµ
Recall also that for a complex number z we define its conjugate z as z = a - bi, which geometrically is the
symmetric with respect to the real axis.

b z at bi
Observe that:
a
z = jzj eiµ ) z = jzj e¡iµ

b I a bi
The conjugate behaves very nicely with algebraic operations, say:

z1 + z2 = z1 + z2
z1 ¢ z2 = z1 ¢ z2
Let’s recall now how we operate with complex numbers. Let z = a + bi and w = c + di. Then:

• Addition:

z + w = (a + bi) + (c + di) = (a + c) + (b + d) i

So basically we add together the real and imaginary parts respectively. Addition (likewise substraction) cannot be
operated in exponential form so first we must write the expression in its binomial form then compute the
operation.

• Multiplication:

z ¢ w = (a + bi) ¢ (c + di) = ac + adi + bci + bdi2 = (ac ¡ bd) + (ad + bc) i

Computation of a multiplication in binomial form can be a bit tedious since we have to multiply every term.
However, the calculus becomes much easier if we do it in its exponential form. Suppose first that

z = jzj eiµ; w = jwj ei'

Then:
¡ ¢ ¡ ¢
z ¢ w = jzj eiµ ¢ jwj ei' = jzj ¢ jwj ei(µ+')

So in exponential form, we multiply the modulus of the complex numbers and add up the arguments.

• Division:

z a + bi (a + bi) ¢ (c ¡ di) ac + bd + cbi ¡ adi ac + bd (cb ¡ ad)

= = = = + 2 i
w c + di (c + di) ¢ (c ¡ di) c 2 + d2 c 2 + d2 c + d2

As in the multiplication, division is also a long computation in binomial form. However in exponential form we
have

z jzj i(µ¡')
= e ; w6
=0
! jwj
• Powers

Calculating powers in binomial form can become an unbearable calculation if the power is greater than 3 so we
will avoid doing so. Instead, the calculation of powers of a complex number in its exponential form is a rather
natural computation due to de Moivre’s formula (which can be easily deduced using properties of the
exponential function). Thus, for a number n∈ℤ we have:

n n
z n = (jzjeiµ)n = jzj einµ = jzj (cos nµ + i sin nµ)

• Roots

For a complex number z and a natural number n we have that

8 p
> n
<½ = jzj
p
n
z = ! = ½ei' ) µ + 2k¼
>
:' = ; k = 0; 1; : : : ; n ¡ 1
n

Observe that calculating the n-th root of a complex number z is equivalent to solving the equation:

wn = z
And by the Fundamental Theorem of Algebra we know that this type of equation has n complex roots, which is
consistant with the solution given above. In fact, the solutions are complex numbers of same modulus but
different arguments. Geometrically speaking, the n-th roots of a complex number are complex numbers set on a
p
circunference of radius jzj scattered in a homogeneous pattern.
n

Calculate F where 2 1 i

First we write 2 in exponential form

Itt Ft F and O
act I I 4thquadrant

so 2 Re and hence if w a peil

p
Fi 2
y 752kt R o e e

th t
wa 2 e w 2 e w 2 e
 We Circumference of
radius 2

a
W

W3

Recall some indentities that might be very helpful and also very easy to prove.

2
jzj = z ¢ z
z + z = 2Re (z)
z ¡ z = 2Im (z)

To finish with this brief introduction we make a small note about the argument of a complex number. Observe
that two a priori different complex numbers say z and w can actually be representing the same due to the fact
that the arguments differ on a multiple of 2π.

In complex analysis, the argument of a complex number is a set and we denote it by Arg(z):

Arg (z) = f' 2 R j z = jzj ei' g

So, z and w define the same complex number if and only if

jzj = jwj ^ ® = ¯ + 2k¼; k 2 Z

where α and β are the angles that z and w form with the horizontal axis respectively.

However there is one and only one argument for each complex number that lives in the interval (-π, π]. This
argument is called the principal argument and we denote it by arg(z).
Complex Functions

The main tool we are going to be working with in this course are complex functions. In previous courses of
calculus we have studied real functions defined in real euclidean spaces. The last step in this direction is to extend
the notions seen so far to the complex context.

Recall that the complex field is isomorphic (same shape) to the euclidean 2-dimensional space and so we can view
it as such. So a complex function can be seen as a function defined on the plane, with the slight addition of the
complex structure. This might sound as something complicated but actually complex functions have very nice
properties and behave nicely, even better than real ones!

But not only we are going to extend the notion of real function but see how are these very useful not only
mathematically speaking but in application. In particular, complex analysis is a fundamental tool to study quantum
mechanics, which is nowadays one of the most important branches in physics.

Let’s thus define what a complex function is. Formally in mathematical notation a complex function is defined
from a complex set A⊆ℂ to the complex plane, say

f :AµC!C
We will see later that depending on the type of function A must fullfil certain conditions, but for now we just
consider it as a subset on the complex plane. Under this definition then it makes sense that f takes complex
values and gives also a complex number, so f(z) = w = α + βi. If we consider that z = x + iy then we have that

® = u (x; y) = Re (w) ; ¯ = v (x; y) = Im (w)

So a complex function can be seen as its real part plus its imaginary part both of which are real functions. We
can also denote them as Re(f) and Im(f) respectively. Let’s see our first example of a complex function:

Let f be defined as

flat z i

Write f ar u too

Let 2 xtiy be a complex number with Xy

E IR Then

flat z i x
toy i x 2xyity i i

x y't ay 1 i

so uh yl Reff x
y and vex l Im f 2xy 1
Continuity of a Complex Function

As mentioned above complex functions are an extension of real functions and thus most of the concepts we
already know have similar if not the same definitions. This is the case for continuity and limits. The problem with
complex functions is that we cannot visualize them graphically so we might not be able to see at first what
discontinuities it has. We can however plot the real and imaginary parts of a complex function since they are seen
as two variable functions defining a surface.

In any case, analitically speaking, we can use the ε-δ definition of continuity as we did for real functions. So we
say f is continuous at some complex number a if

8" > 0; 9± = ± (") : jz ¡ aj < ± ) jf (z) ¡ f (a)j < "

The expression |z - a| < δ means that z lives in a ball (circunference) of center a and radius δ. In mathematics
this is usually expressed as

z 2 B (a; ±)
So graphically continuity can be expressed as:

n n

i
s y
a

tzeBlais title Blfial E

In other words, f is continuous on a if for any ε there exists a δ = δ(ε) such that:

f (B (a; ±)) ½ B (f (a) ; ")

We can also characterize continuity of a complex function by means of its real and imaginary parts. Thus it can be
proved that f is continuous on a if and only if Re(f) and Im(f) are continuous on a. Going thus back to the
comments above about the graphical interpretation of continuity, we can always plot the surfaces defined for both
the real and imaginary parts and deduce continuity from there.

The concept of continuity relies on the concept of the limit at certain point. By definition we say f tends to a limit
point L∈ℂ when z approaches a if

8" > 0; 9± = ± (") : jz ¡ aj < ± ) jf (z) ¡ Lj < "

So basically a function is continuous on a if the limit at that point exists and L = f(a), which is the definition we
already know. The limit can also be characterized by means of the real and imaginary parts of the function.
Assume thus L = α + βi, then

lim f (z) = L () lim Re (f) = ® ^ lim Im (f) = ¯

z!a z!a z!a
 Let flt Z i Decide when f is continuous at a e ti

We saw on a previous example that

Relf x
y In f 2xy 1
Then
Ewa rect II 7

1
Exact II dig 1

ties flat
Hence Ott i i

Further flat ati i f if 2i i i

so f is continuous on a Ati

We could have argued also that Re t and Jen f are continuous

on 1,11

Differentiation of a complex function

Let’s start this section by giving the definition of the derivative of a complex function and then discuss it. We say f
is differentiable at w = a if the following limit exists:

f (z) ¡ f (a)
lim 2C
z!a z¡a

In that case we denote that limit as f’(a). The first observation we make is that the definition is the same as it was
for real functions. The only difference is that we now work on the complex plane and that makes the concept of
differentiability somehow stronger than usual. Observe that the limit is a quotient limit where the denominator is a
complex number so it needs more work to be understood and then computed properly.

We see this fact clearly next. Suppose that we have a real valued complex function, i.e. f(t) = u(t) + iv(t) where t is
real. Then for a∈ℝ we have:

f (t) ¡ f (a) u (t) ¡ u (a) v (t) ¡ v (a)

= +i
t¡a t¡a t¡a
The equality above tells us that a real valued complex function is differentiable at a if and only if Re(f) and Im(f)
are differentiable at a and in this case we have

f 0 (a) = u0 (a) + iv 0 (a)

We see much analogy between real and complex differentiability. In fact, differentiation rules are inherited from
real to complex analysis as we know them:

Proposition (Differentiation rules). Let f and g be complex functions and suppose both are differentiable at
some point w = a. Then

• f + g is differentiable at a and (f + g)’(a) = f’(a) + g’(a).

• f·g is differentiable at a and (f·g)’(a) = f’(a)·g(a) + f(a)·g’(a).

• If g(z) ≠ 0 for every z in the domain of f then f/g is differentiable at a and

µ ¶0
f f 0 (a) g (a) ¡ f (a) g 0 (a)
(a) = 2
g g (a)

• (The chain rule). Let f : A → ℂ and g : B → ℂ such that f(A) ⊆ B and consider the composition h = g∘f.
Suppose f is differentiable at a and g is differentiable at f(a). Then f is differentiable at a and

h0 (a) = g 0 (f (a)) ¢ f 0 (a)

n Let flt t't it 1 Compute f t

Observe that
m t th 1 v El t

Clearly both are differentiable functions with

It
If 1

Then

If flt n t t in't l ZE t

We reach the same conclusion

by direct differentiation

If fit t't it all LE t

The Cauchy-Riemann Equations

We have talked about the definition and properties of complex functions. However, we would like to have an
easier way to decide whether a complex function is differentiable or not and in the affirmative case, be able to
compute the derivative. This process remarks the fact that complex differentiation is more restrictive (hence
stronger) than real differentiation.

Theorem (Relation between complex and real differentiation). Let Ω ⊆ ℂ be an open set, a a point in Ω and f
: Ω → ℂ a complex function. Let a = α + iβ, u(x,y) = Re(f(x+iy)), v(x,y) = Im(f(x+iy)). The following
statements are equivalent:

• f is differentiable in the complex sense at a.

• The functions u(x,y) and v(x,y) are differentiable at (α, β) and they satisfy the Cauchy-Riemann conditions:

@u @v
(®; ¯) = (®; ¯)
@x @y
@u @v
(®; ¯) = ¡ (®; ¯)
@y @x

In case the above statements are true we have

@u @v
f 0 (a) = (®; ¯) + i (®; ¯)
@x @x

Now we are more aware about how complex differentiation is much more restrictive than the real one. Observe
that even if the functions u and v are very nice it might happen that f = u + iv is not differentiable because the
Cauchy-Riemann equations are not satisfied.

We see next some examples of different complex functions and the analysis of their differentiability on the
complex plane.

Study differentiability of the following functions

al fletiy x

b
fixtiyl xtiy x
ty
c
flxtiyl 2xy t ily x

d fixtiyl e
cosytisiny

a Observe that flt Real in terms of complex variables

We see if the C R equations are satisfied

 1 I o
IF f is notdifferentiable

b Observe that
x't y Z E feel z E
The function depends on E as well so we
might expect f to be
non differentiable unless 2 0 We compute

xtiy atty I x'try t i

x'y t y l u t in

The

E 3x't y
E Y es e y

of etsy
Hence f is differentiable at 2 0 as expected
only
c u e
ay re
y

Ey and
Y
2x
Ey
Ly
Y
So f is differentiable everywhere and

fletiy 2y Ki 2 g ee

d Observe that fled et so

clearly f is differentiable everywhere

Easy Yy f is differentiable everywhere and

e'sing IT fixtiy excory t ie sing e osytising

f x
tiy f
In the examples above we got the sense that functions depending only on z are candidates to be good
differentiable complex functions. This is indeed so. Observe that if we consider z = x + iy then

z+z z¡z
x = Re (z) = ; y = Im (z) =
2 2i
So µ ¶ µ ¶
z+z z¡z z+z z¡z
f (z) = u ; + iv ;
2 2i 2 2i

And thus
@f 1 @u 1 @u i @v 1 @v
= ¡ + ¡ =0
@z 2 @x 2i @y 2 @x 2 @y

From where we easily deduce the Cauchy-Riemann equations.

Definition (Holomorphic functions). Let Ω be a non-empty open sunbset of ℂ. A function f : Ω → ℂ is called

holomorphic in Ω if it is differentiable on the whole set. In that case the function defined as z ↦ f’(z) is called
the derivative of f. The functions that are holomorphic on the whole complex plane are called entire functions.

Let’s see a few examples of entire functions:

• Complex polynomials, i.e. functions of the form

p (z) = c0 + c1z + c2z 2 + : : : + cnz n

with all coefficients c∈ℂ, are entire functions. The derivative of p is

p0 (z) = c1 + 2c2z + 3c3z 2 + : : : + ncnz n¡1

• Rational functions, i.e. functions of the form R(z) = p(z)/q(z) where p and q are polynomial, are holomorphic
on their natural domain Ω = { z∈ℂ : q(z) ≠ 0}. The derivative of R(z) is

0 p0 (z) q (z) ¡ p (z) q 0 (z)

R (z) = 2
q (z)

Proposition. Let f be a holomorphic function on some domain Ω. If f’(z) = 0 on the whole domain then f is
constant.

As a direct consequence of the previous result we see that:

Corollary. Let f and g be holomorphic functions on some domain Ω. If f’(z) = g’(z) for ∀z∈Ω and there exists
some point ω such that f(ω) = g(ω) then f = g everywhere on Ω.
For elementary complex functions the derivatives have the same outcome as in real analysis:

In general, for any holomorphic function we can differentiate directly with respect to z using the rules we already
know. The outcome will be the same as using the Cauchy-Riemann equations.

Let fled ez Show that f is holomorphic and compute fled

Let a then
xtiy
2 x
tiny x
y t
2xyi
And
It In
f is
E Ey
y g
2x
Ly holomorphic
By the differentiation theorem we have

fled Yy t
if 2x
ily 2
atty 22

As we had expected since t Zz

Show that the components of a holomorphic function satisfy

Laplace's equation

t
35 o n

Let fled mix tier x y be holomorphic so

É Y
and
27 35
By differentiating both equations in x and
y
we
get

o
Elementary Complex Functions
In this section we discuss the extensions of the elementary functions we know to the complex plane. It may seem
like something trivial but there are some nuances underneath that need some explanation. It is not enough to just
substitute the variables x and y by z and w.

The exponential function

We know that the real valued exponential can be expressed as

1
X xn
x
e =
n=0
n!

So the most sensible thing to do is to express the same series but using complex variables. This is indeed the
case. Further, as in the real case, the series converges everywhere so we can simply write the Taylor series for the
complex exponential function as:

X zn
z
e =
n¸0
n!

Let’s see some of the properties of the complex exponential. Most of them does not change with respect to the
real case as we will see:

1. exp’(z) = exp(z) for every z ∈ ℂ

2. exp(0) = 1
3. exp(x) with x ∈ ℝ matches the real exponential. So clearly the complex exponential is an extension of the
real case.
4. exp(z + w) = exp(z)exp(w)
5. Euler’s formula. For every t ∈ ℝ we have that exp(it) = cost + i sint. In particular we have the identity

ei¼ + 1 = 0

6. For z = Re(z) + i Im(z) ∈ ℂ we have

ez = eRez (cos (Imz) + i sin (Imz))

Thus we deduce that

jez j = eRez ; Imz 2 Arg (ez )

7. The complex exponential never vanishes

ez 6
= 0; 8z 2 C
8. The complex exponential is a periodic function of period 2πi. In fact we have that

ez = ew () z = w + 2k¼i; k 2 Z

9. The complex exponential is holomorphic.

The complex logarithm

The logarithm, as we know it, is the inverse of the exponential function. Recall that, in real analysis we have

loga b = c () ac = b
However, in real analysis the exponential is a one-to-one function which makes the logarithm a well-defined
function. Observe now that since the complex exponential is a periodic function, the logarithm needs to be
redefined somehow. Let’s see how.

Observe that the equation exp(w) = z where z is a non-zero complex number has infinite solutions for w.
Indeed, we see that

ew = eRew (cos(Imw) + i sin(Imw))

So in order to satisfy the equation it suffices:

1. |exp(w)| = |z|, this is, exp(Rew) = |z|, which means Re(w) = log|z| (natural logarithm of the positive real
number |z|.
2. Arg(exp(w)) = Arg(exp(z)), this is, Imw ∈ Arg z and this is true if and only if Im w = arg(z) + 2kπ, with
k ∈ ℤ.

We have thus proved the following:

fw 2 C : ew = zg = fln jzj + i (arg (z) + 2k¼) ; k 2 Zg

So we have infinite solutions to the above equation and any of them is a logarithm of z. The set of all of them
is represented by Log z.

Logz = fln jzj + i(arg (z) + 2k¼); k 2 Zg = ln jzj + iArg (z)

Within this set, we take one of them and name it principal logarithm (or principal branch), defined as:

log z = ln jzj + i arg (z)

When working with the complex logarithm we need to be careful because the properties we know from real
analysis may fail in this case. For instance the following identity

log (zw) = log z + log w

is no longer true. Let’s see an example:

Let 2 1 t it and w B ti
i
Observe that 2 Ze w ze so

log t In 2 t
if log w In 2 t i

Now

loglaw log entire l Eti logl Yi en 4 it

But
en 4 if I eat if t en z ti In 4 i

So the usual identity is not satisfied

However, the following IS true:

log z + log w 2 Log (zw)

that is, log z + log w is a logarithm of zw but not necessarilly the principal.

Definition. Let f : A → ℂ*, with A ⊆ ℂ a non-empty set in ℂ.

• Any function g : A → ℂ such that g(z) ∈ Log f(z) for every z ∈ A is called a logarithm of f in A or branch of
the logarithm of f in A. When f is the identity we simply say g is a logarithm in A or branch of the
logarithm in A.
• Any function θ : A → ℂ such that θ(z) ∈ Arg f(z) for every z ∈ A is called an argument of f in A or a branch
of the argument of f in A. When f is the identity we simply call it argument in A or branch of the
argument in A.

The notation ℂ* means ℂ-{0}. Note that the logarithm is not defined for 0.
 Here is another example on how to compute the logarithm of a complex number. Unless otherwise stated we will
be working with the principal branch of the multi-valued functions so that everything works out smoothly.

Calculate dog l i
Observe that i é I it 1 argtid E

Ther

dog C il en l il tiarg til YI if E if

For mathematicians the next question to be answered would be: given a holomorphic function f that does not vanish
on some set A ⊆ ℂ*, are there holomorphic logarithms of f in A? Using the previous notation, this means that we
look for holomorphic branches g of the logarithm of f in A.

The answer to this question is not trivial and requires many mathematical tools. For the purposes of this course
however, we will not go deep into this matter. The following proposition would be a first step to comprehending the
previous idea:

Proposition. Let f : A → ℂ* with A ⊆ ℂ and a ∈ A. Let g : A → ℂ be a logarithm of f in A. Assume further that f

is differentiable in a and g is continuous in a. Then g is differentiable in a and

0 f 0 (a)
g (a) =
f (a)
Consequently if f is a holomorphic function in A, 0 ∉ f(A) and g is continuous un A such that exp(g(z)) = f(z) for
every z ∈ A then g is holomorphic in A with g’(z) = f’(z)/f(z) for every z ∈ A.

It should be mentioned that the notation log we are using makes reference to the natural logarithm so we might
as well use the usual ln i.e. the logarithm with base e. In this scenario we know that

1
f (z) = w = ln z =) f 0 (z) = ; z6
=0
z
The feeling we might be getting at this point is that whenever we work with holomorphic functions everything
seems to work smoothly. This is pretty much the point since holomorphic functions are basically infinitely
differentiable analitic functions, probably the most versatile kind.

Definition (n-th roots of a function). Given a complex function f defined on a complex set A and a natural
number n ≥ 2, any function h : A →ℂ such that h(z)^n = f(z) for every z ∈ A is called a root (function) of f of
order n in A or a branch of the n-th root of f in A. If f is the identity we simply call it a root of order n in A or a
branch of the n-th root in A.

Suppose that we are given the function ω = z^(1/2). Suppose further that we allow z to make a complete circuit
around the origin starting from some point A. We have

iµ p i 2µ
z = re ) != re
After a complete circuit back to A, θ + 2π, we have

p i
¡ µ+2¼ ¢ p µ p µ
!= re 2 = rei 2 ei¼ = ¡ rei 2

Thus we have not achieved the same value of ω with which we started. However, by making a second complete
circuit back to A, i.e. θ + 4π we get

p i
¡ µ+4¼ ¢ p µ p µ
!= re 2 = rei 2 ei2¼ = rei 2

and we do then obtain the same value ω with which we started. Each of the circuits describes a branch of the
function and it clearly is a single-valued function on it.

Let fled 2 and consider 2 Teton Of O c 21T

Then

re
zit Fei
gym
F eiQi
g
function f are
So graphically how both branches of the
we see

defined The first corresponds to 0,2T and the second one to

21T YT
Complex powers

Given real numbers a > 0 and b ∈ ℝ we can write

ab = eb ln a

Now suppose a and b are complex numbers with a different from 0. We know there are infinitely many
logarithms of a, all of them under the form ln |a| + i(θ+2kπ) with k ∈ ℤ. Thus any complex number of the form

eb(lnjaj+i(µ+2k¼)) ; k2Z

is a power of base a and exponent b. We represent by [a^b] the set of all of them:

£ b¤ bLoga © b! ª
a =e = e : ! 2 Loga

Among all of them we define the principal value (or principal branch) as:

ab = eb log a

Calculate i and i

til
fi eid Considering principal branch

Observe that
th
i e es en fit ith
And thus
l
e
i eientil ei.fi etz
eni
it e

i e as eni ith
Thus
hi i
it e e e wet tisine
Complex trigonometric functions

Based on Euler’s formula we can easily deduce the following definitions of the real trigonometric functions:

eit + e¡it eit ¡ e¡it

cos t = ; sin t =
2 2i

To do so it suffices to compute exp(it) + exp(-it) and exp(it) - exp(-it). These identities are valid for every
value t ∈ ℝ. So for every z ∈ℂ we can likewise define complex trigonometric functions as:

eiz + e¡iz eiz ¡ e¡iz

cos z = ; sin z =
2 2i

It is clear that complex trigonometric functions extend the real ones. Since these definitions rely on the
exponential, properties are easily deduced:

• The fundamental identity holds in the complex plane:

cos2 z + sin2 z = 1

• The cosine is even cos(-z) = cos(z) and the sine is odd sin(-z) = -sin(z). Both functions are periodic of period
2π.

• The addition formulas hold:

sin (z + w) = sin z cos w + sin w cos z

cos(z + w) = cos z cos w ¡ sin z sin w

• Complex trigonometric functions are entire and analitic on the whole complex plane and:

1
X (¡1)
n
cos0 (z) = ¡ sin z; cos z = z 2n
n=0
(2n) !
1
X (¡1)
n
0
sin (z) = cos z; sin z = z 2n+1
n=0
(2n + 1) !

• Complex trigonometric functions, alike real ones, are not bounded. However if we restrict ourselves to a
bounded horizontal stripe then trigonometric functions are also bounded. Indeed:

¯ ¯
eiz = ei(a+bi) = e¡beai =) ¯eit¯ = e¡b
• Complex trigonometric functions have the same amount of zeros as real ones, that is:

sin z = 0 () z = k¼; k 2 Z
¼
cos z = 0 () z = + k¼; k 2 Z
2

• The complex tangent is defined as usual:

sin z
tan z =
cos z

and it is a holomorphic function on its domain z 2 Cnfcos z = 0g.

Calculate the real and imaginary parts of sin Ati cos i i

By definition
eie
É iz.ee
e
sinati

e tlwsntisi.nlisi
glosn

cosieIe tsin1ez Y

sine
etÉ t i cost
ez y
And likewise
eim.it itI ie
cos g is ze eietze

cosnetzItsinie I i
e
 The previous computation can be simplified by using the real hyperbolic functions:

ex + e¡x
cosh x =
2
e ¡ e¡x
x
sinh x =
2
For which we have the following identities:

cos (ix) = cosh x; sin (ix) = ¡i sinh x; x2R

And for z = x + iy:

sin z = sin x cosh y + i cos x sinh y

cos z = cos x cosh y ¡ i sin x sinh y

Hyperbolic functions can also be defined in terms of complex variables:

ez + e¡z
cosh z =
2
e ¡ e¡z
z
sinh z =
2

All these relations are very useful when we have to solve complex trigonometric equations. The way to proceed
here is a bit different from the techniques we used in real analysis. However, a solid knowledge of trigonometry is
again required if we want to solve this type of equations. Let’s see an example:

Find every solution to the following equation

COS2 2

Let's solve this equation using two different

ways
1 We use the identity cost cost
coshy i sinxsinhy

the
Cosa cosxcoshy isimxsinhy 2 t i0

Matching real and imaginary parts we have

cost coshy 2
0
sinesinny
 e e
We solve first the second equation

sinesinhy o a x ka Ke z v
y
É
And analyze both cases on the first equation

4 0 cost
ashy Osx 2

which is not possible so o is not a

possible solution
y s o
cushy
X kit KEE coshy 2 w coshy

From this we conclude that x 2kt KEI and solve

coshy 2 a Etd 2 e te 7 4

multiply eh yet 1
by a t 0

ex
Utf YETI 2153

yeen Irs

Note that

enle ol en
CIII en
Ig en ar

so
y ten 2 03

All in all we
get that the solution to the equation is

2 2kt ien 2 83
KEIL
I
 2 Now we solve the equation
directly using the definition
of the complex cosine

e
cosz tze 2 eitté y

So

feet Ye't t y o as we did before

Then

MIMI
it
412ft 2153

So the equation now is

eit 2153 E é osxtisinx 218 to

And we can derive the following system

é cost 2 IF
e sin x O X KIT KEI

If K is odd then e 2 IF I e ZEB b

which is not possible

If k is ever then e ZEB y en IB

2 53
y ten
Hence we reach the same answer

KE TE
2 2kt I ik 12th
For the complex hyperbolic functions the following identities might be usefull:

sinh (iz) = i sin z; sin (iz) = i sinh z

cosh (iz) = cos z; cos (iz) = cosh z
sinh z = sinh x cos y + i cosh x sin y
cosh z = cosh x cos y + i sinh x sin y
And:

sinh (¡z) = ¡ sinh z

cosh (¡z) = cosh z
cosh2 z ¡ sinh2 z = 1

The inverse functions of the complex trigonometric functions can also be defined although it takes some work to
deduce them. For the purposes of this course we don’t need to dive that deep so we only focus on the main
definition of each of them:
p ³ ´
Arc cos z = ¡iLog z + 1 ¡ z 2

³ p ´
Arc sin z = ¡iLog iz + 1 ¡ z 2

For which the principal branches are defined by taking the principal branch of the logarithm log. For the inverse
of the tangent we have the following expression:

µ ¶
1 1 + iz
Arc tan z = Log ; z6
= ¡i
2i 1 ¡ iz

Compute Arcsini
By definition we have
Arcsini i log i it fit ilog ate
For 1 E we have
i log l e r i enate i it 2kt

T t 2kt ien Atr KE ZE

For 1 tf we have
i log f Atf i en fetal i 2kt 2kt ien Itf KEE
Complex Integration

Up to this point we have dealt with integration in different situations, the definition, real functions, double and
triple integrals, vector integrals, and so on… but there is yet one last step we should cope with and that is complex
integration. There is no reason to think that integration was not possible on the complex plane. In fact, as we will
see real soon, complex integration has very powerful implications and it is very elegantly built.

Recall that for a real valued function we use the notation

Z b
f (x) dx
a

meaning that we integrate the function f over an interval [a,b] and dx is the differential. Recall also that the
integral sign can be understood as an infinite (continuous) sum. The integral above is also called a Riemann
Integral from the way it is built (Riemann sums based on partitions).

However this type of integral gives us no real lead on how to integrate complex functions since these can be seen
as functions defined on a plane and delivering values also on a plane. We might think that a double integral might
be the answer but this is neither the case. The main reason behind this is that even though we identify each
complex number with a point on the plane, a complex number is a single object and should be treated as such.
Further, the outcome of a complex number is still another complex number and thus a double integral would make
little geometric sense.

So how can we integrate a single valued complex function that can be understood as a piece of plane?
The answer is: (complex) line integrals. Indeed, points on a plane can be joined by (regular) curves and we know
how to compute integrals over such mathematical objects. Recall that a curve on a plane is a function of the form

° : [a; b] ¡! R2
t 7¡! °(t) = (x (t) ; y (t))
which we also referred to as parametrization. Thus for a function f(x,y) we define a line integral as:

Z Z b
f (x; y) dC = f (° (t)) ¢ ° 0 (t) dt
C a

So we need to define what a curve on the complex plane is. There is nothing fancy about it; in fact, curves on the
complex plane are defined exactly as we did on the two dimensional euclidean space. The most used ones are
lines and circles.

• A line on the complex plane joining z_1 and z_2 is defined as:

° (t) = z1 + t ¢ (z2 ¡ z1) ; 0 · t · 1

Which is the exact same definition as in the euclidean context. Observe that when t = 0 we get the initial point
z_1 and for t = 1 we get the endpoint of the segment z_2.
• A circle (circunference) of center a ∈ ℂ and radius r > 0 has the following parametrization:

° (t) = a + reit; 0 · t · 2¼

Describe the line

joining 2 h and Zz Iti
definition
By
Nlt att er and t iti

that'll
t t Catti i i O f t 31
z

Fits
pg

Describe the circle centered at 2 i and radius r 1

By definition
Jlt it te't it cost is int

cost Atsint i O f t e 21T

part In
part
ol Ati is the imma
rio nti
i point of the circuit
So now we can define the integral of a complex function f(z) over some curve C as:
Z Z b
f (z) dz = f (° (t)) ¢ ° 0 (t) dt
C a

where γ is the parametrization of the curve C. In particular, when a curve is closed we use the notation:
I
f (z) dz
C

We will see how this kind of integrals are of major importance. But let’s see a few examples of integrals that can
be solved using the definition above:

Solve the following integral where c Jlt 3 t t it l et en

Edt

Observe that flat E so

f r tell DIET 3Etit It it

and
y t 3 Zit

Thus

Edt 3t it 3 2 it Idt

I at it ti 6th 3 t l do

att 2t dt if 3 Edt
 Et't It J tilt

195 65

Calculate
Idk where C is the unit circle in counterclockwise

direction
single spin
First we parametrize C as
y
it
8 t O 1 e eit O f t e 21T

center
I radius

Observe also that 8 El ie't so

iet
Ids It zit

We didn’t mention the integration properties of complex functions but you can see in the examples that are quite
sensible-built from the rules we already know. Let’s mention some of them:

• Let k ∈ ℂ and f continuous on some domain D that includes C. Then

Z Z
kf (z) dz = k f (z) dz
C C

• Let f and g be continuous complex functions over some domain D that includes C. Then:

Z Z Z
(f (z) + g (z)) dz = f (z) dz + g (z) dz
C C C

• Let C be a curve that can be divided in sub-curves, say C_1 and C_2 then:

Z Z Z
f (z) dz = f (z) dz + f (z) dz
C C1 C2

where the end point of C_1 coincides with the start point of C_2.
• For a curve C consider -C to be the same path in the opposite direction, then:

Z Z
f (z) dz = ¡ f (z) dz
C ¡c

• If f is a real valued complex function, say f(t) = u(t) + i v(t) ∈ ℂ with t ∈ ℝ then:

Z Z Z
f (t) dt = u (t) dt + i v (t) dt
C C C

Calculate G't iy dz where C is the following curve

At Zi

G C C V Cz
Iti
C

We have flat x't iy and by the complex integration properties

I
ft dt feel da t feelda

so we calculate the integral over each of the line segments that

compose C

We parametrize C by 8 t t ati tt ti Tilt Ati

so

fields t Ceti
GI
at Zi
I fi
e Xlt t t't it t ai
yet t x'tip
 As for G we have
Nlt it i t t lathi Ceti
At it it It att i o e t en

And tilt i so

fields it att if i dt

att dt ti
fat

I i
I I i
ft i

All together we thus get

fads i
f ti f tf i

Thus far we see that the calculus of complex integrals over curves follows the same mechanic as in usual line
integration. You may also recall that line integrals were path-dependant in some cases and independant in others.
The same situation happens here. Let’s look again at one of the examples used above where we integrated the
function 1/z over the circunference of radius 1 and see what happens when we use a different parametrization:

Consider the parametrization of the unit circle

it it
8 t e r Ctl e Oe t a zit

In the first case we ride the circle in clockwise direction

and in the second we make three complete circuits

Let's see what happens when we compute the integrals
 it
Tilt ie the
21T

So fat f ftp.f ieftldt ai

c

it
And 8 ti Bie thus

3ieitdt
t da
fit biti

so all the obtained results are different and we clearly

deduce that the integral in this case is
path dependant
The obvious question now arises: does this always happen? Are there functions for which integration is path
independant?

Observe that the function we integrated in the previous example is not defined for z = 0, further, it’s not even
holomorphic on that point. You may argue that the integral didn’t even go through that point. That is one of the most
powerful features of complex integration over contours or closed curves. The value of the integral can vary
depending on the singularities (which we will cover later on the course) the function might have inside a path.

In particular the function 1/z is not holomorphic z = 0 and that makes the integral path dependant. But as mentioned,
we will come around this topic soon enough and discuss the details in depth.

As a remark, like we defined in previous courses of calculus, the arclength of a curve on the complex plane can be
calculated by the formula:
Z b
L= j° 0 (t)j dt
a

which is a particular case of the contour or line integral defined above taking f = 1. Further, if the function f is
continuous over some curve C and for every z ∈ C we have |f(z)| ≤ M then:

¯Z ¯
¯ ¯
¯ f (z) dz ¯ · M L
¯ C ¯
Observe that the result of integrating a complex function gives as a result another complex number so there is not
much we can discuss about the geometry of the definition. However the above bound tells us that the result of the
integral is a complex number whose modulus is upper bounded by the length of the curve and the maximum value
of the function itself.

This bound may not be useful for us but actually is a nice way to carry out estimations when the integral is rather
complicated.

Find an upper bound for of II da where C 121 4

Let flat II which is continuous over C circumference centred

at origin and radius 4 Observe that ZIV 21T 4 81T

Also
let let let e ell
d
12 11 2 11 1 a 1 3 121 1 4 1 3

So all together we have

l e M

Hence
e 8
I fo ta e Me 8T T
r
Cauchy-Goursat’s Theorem

This section is centered around contour integrals over closed curves C of the form
I
f (z) dz
C

If nothing is said we assume the curve takes the counter clockwise direction. The main objective is to see that when
f is holomorphic on some connected domain D the integral does not depend on the curve C contained in the
domain. A connected domain is nothing but an open set that consists in one whole piece, i.e. an open set that does
not have holes in it. We further say a domain D is simply connected if for any pair of curves with same intial and
end points one can be continuously deformed into the other.

On the other hand we can find multiply connected domains. To avoid all the topological details about this
definition we simply say that multiply connected domains are simply connected domains that might contain holes
in it.

Simply connected domain Multiply connected domain

Let’s get back to integration over closed curves and let also f(z) = u + iv be a holomorphic function over some
simply connected domain D. Observe that:
I I
f (z) dz = [u (x; y) + iv (x; y)] (dx + idy)
C C

where x, y are Re(z) and Im(z) respectively and the complex differential dz is broke into dx + idy. Now by
multiplying everything out we get the following expression:
I I
f (z) dz = [udx ¡ vdy] + i [udy + vdx]
C C
So the real and imaginary parts are separated and we can treat them individually:
I I
udx ¡ vdy; udy + vdx
C C

Assuming u and v are continuous we could apply Green’s Theorem so that the line integrals defined above
become double integrals over the domain enclosed by C, say D:

Z Z µ ¶ Z Z µ ¶
@v @u @u @v
¡ ¡ dxdy; ¡ dxdy
D @x @y D @x @y

Observe now that if f is holomorphic then by the Cauchy-Riemann equations the above integrals are both zero.
As a consequence we have the following theorem:

Theorem (Cauchy - Goursat). Let D be a simply connected domain and let f be holomorphic in D. Then for
every closed path C we have that
I
f (z) dz = 0
C

Calculate where C is the curve shown below

fetch

We know that the exponential function is an entire function die

it is holomorphic on the entire complex plane so

by the

C G theorem we harm

feeds O
 Calculate
fo
da where C is the ellipse x 2
hit p

The ellipse has the form

O
The function flat Yz
holomorphic everywhere except for
is

2 0 which lies outside the ellipse hence

da O

Now what if the domain D is multiply connected? Or the function f is not holomorphic on some point within the
domain?

In these cases we cannot assure that the contour integral over a closed path will be zero. Let’s thus see what
happens. To enlighten the procedure as much as possible let’s make it geometrically visual. Suppose our domain
has the following form:

And let C be a curve over which we want to calculate the integral for some holomorphic f. However we don’t
know whether f is holomorphic outside D so we cannot conclude anything since the curve is surrounding a hole of
the domain. Let’s thus consider another path, say γ such that we generate a new path as shown below:
So we connect both C and γ by the segment AB and so we generate a closed circuit

C [ AB [ ¡° [ ¡AB
that starts and ends at point A. Observe further that the domain enclosed by this path skips the hole and so f is
holomorphic in it so by the C-G theorem we conclude that:

I Z I Z
f (z) dz + f (z) dz + f (z) dz + f (z) dz = 0
C AB ¡° ¡AB

or
I Z I Z
f (z) dz + f (z) dz ¡ f (z) dz ¡ f (z) dz = 0
C AB ° AB

from where we conclude that:

I I
f (z) dz = f (z) dz
C °

This result basically tells us that whenever we have to calculate a contour integral over a curve on a multiply
connected domain and the curve surrounds a hole, we can choose any other (more suitable) path to do the
calculation. The identity above is called the contour deformation principle.

Calculate the following contour integral

ft
where C is the path shown below
 It i
2 35

Cv i
2 n

2 21
L

Observe that the function fled

f is not holomorphic

on 2 i which is enclosed within the area limited C

by
To calculate the contour integral which cannot deduce to be 0
we should integrate over each of the segments that compose C

Instead we choose another curve that surrounds 2 i and use

the deformation principle Let than Nlt ite't i.e the

unit circle centered at t i Under this we have
saramethfation

IF I dz
at

r'Aldt
2

ieitdt
And the calculation is done z i ft e
it
I e't

This example can be generalized as follows:

Proposition. If ω ∈ ℂ is interior to some closed path C then for each n ∈ ℤ we have:

I ½
dz 2¼i; n = 1
n =
C (z ¡ !) 0; n 6
=1
This result also tells us that f being holomorphic is not a necessary condition to conclude that a closed contour
integral is zero. In fact, according to the previous proposition we have that
I
1
2
dz = 0
jzj=1 z

Even though f is not holomorphic at the origin. Let’s see another example on how to use C-G’s theorem:

Calculate
III an

when C I 2 21 2

Observe that z't Lt 3 has zeros at t 1 and t 3

T
C z 21 2

z 3
É
2 2

Observe that f is holomorphic everywhere except for 2 1 within

C so if we decompose f and use C G's theorem we have

fat
fit I In
so

fields
21
t
31 yo t 3 Ziti Gti

P holomorphic I use the

in it 21 2 proposition above
so
ff o
The previous examples tell us that the contour integrals reduce to calculating contour integrals around the points
where the function is not holomorphic since we know that outside these domains the value of the integral is zero.
In fact we can generalize Cauchy-Goursat’s theorem as follows:

Theorem. Let C be a closed curve and consider C_1, C_2, … , C_n be a set of closed curves interior to C and
such that they don’t share interior points (i.e. they do not intersect). If f is holomorphic on each contour and
every interior point to C but exterior to each C_k then:

I n I
X
f (z) dz = f (z) dz
C k=1 Ck

Example:

I I Z
f dz = f dz + f dz:
C C1 C2

calculate
af
Observe that z't n t o et il so the function f is not holomorphic

on those points Observe further that

C It 4

z i

i
And consider the circles around 2 Ii
say
C 1
E Iz i 1 2

G I let it 12

The function f can be decomposed as

far
It I I fi
So we can write

finds
If f ftp.jds
And by the previous theorem

feel Ii at
f on

p
iÉ I
a

holingphic holomorphic
in G

f
If
wi
If ai n t
The Fundamental Theorem of Calculus for Complex Functions

We are familiar with the Fundamental Theorem of Calculus for real functions so it is sensible to think that there
must be an analogue version for the complex setting. Linked to this idea we might think that integrals should also
be path independant under some circumstances whenever the path or contour is not closed as we have seen so far.
Being path independant means that the value of the integral does not depend on the itinerary itself but just on the
initial and end points.

The following theorem answers the last question:

Theorem (Path independant integrals). Let D be a simply connected domain and f be holomorphic in D. Let
further C be any path contained in D. Then the integral
Z
f (z) dz
C
does not depend on C.

Calculate where C is the path shown below

fads
C begins at t 1

and ends at E it i
f
f r

Since 22 is entire we can choose a different path that begins

at 2 1 and ends at a Ati so that the value of the

integral does not change Let thus feel be the straight

line that joins both points title e it O ft e1
ray
Then

22 do
fads i 2 fat it i dt 2iffetitldt

Rift It Zi n
A contour integral that is path independant can be written as
Z z1
f (z) dz
z0

where z_0 and z_1 are the initial and end points respectively.

Another concept that is herited from real analysis is the antiderivative function. Thus if there exists a function F
such that F’ = f then F is said to be the antiderivative of f. From this we can also deduce Barrow’s rule for
complex integration:

Theorem (Barrow’s Rule). Let f be a continuous function over some domain D and let F be an antiderivative of
f in D. Then for any contour or path C in D with start and end points z_0 and z_1 we have:
Z
f (z) dz = F (z1) ¡ F (z0)
C

The antiderivatives of the elemental functions and some others that are well known do not change with respect to
their analogues in real analysis:
 The previous example can be solved easily using Barrow's rule
t

since
i 22dm

241 22
z

n
n ti C nd 1 Zi

With all the previous tools we can write down the complex version of the Fundamental Theorem of Calculus:

Theorem (Complex FTC). If f is continuous and the integral over any C within some domain D is path
independant then f has an antiderivative in D.

This theorem does not give practical tools to ensure that a function has an antiderivative since we cannot check
whether every contour integral is path independant. Instead, we characterize the existance of the antiderivative
by the following result:

Theorem. Let D be a simply connected domain and f holomorphic in it. Then f has an antiderivative in D, i.e.
there exists F such that F’(z) = f(z) for every z in D.

Let’s see with a few examples how to use this results:

We have seen that

171 1
f f dz Ziti

In real analysis the antiderivative of Vx is tnx so we might

think that Log2 is the suitable antiderivative of Yt If this
were true though the above integral would have been zero

Observe that Ye is not holomorphic at 2 0 and further

Loge is not even continuous on IR

Zi

f for
However if we want to calculate then
 Zi

for Koga
i log Zi log 3

Observe that

logLi em2 t
it and log3 en3

Hence

fat enz ti E en en i

f I
Calculate dz where C is the segment joining Z i

and Z 9

The function flat É is holomorphic for 121 o and

it e Arg al e it and its antiderivative is given by FG 221
Consequently we have
9

fi
at 22
i 2 3 i 2 3 E Ei
G F
Ey
Integration by parts

Integration by parts is also valid for complex integration. Let f and g be holomorphic functions on some domain
D. Then
Z Z
0
f (z) g (z) dz = f (z) g (z) ¡ f 0 (z) g (z) dz

So it’s the usual formula we are used to.

Evaluate
f zet dz

Using the formula for integration by parts we have

factor zet
i
Je'd Utile't ie e't r e

mile
ally
se

The Cauchy Integral Formula

The following results are very powerful tools in the theory of complex analysis. Let f be a holomorphic function
in some connected domain D and let ω be a complex number in D as well. The quotient f(z)/(z - ω) is no longer
holomorphic in D so we cannot use Cauchy-Goursat’s Theorem for closed curves. We have the following result
instead:

Theorem (Cauchy’s Integral Formula). Let D be a simply connected domain, f holomorphic in D and C any
closed curve in D. Then for any ω interior to C we have:
I
1 f (z)
f (!) = dz
2¼i C z¡!

This result tells us that the value of a holomorphic function f at any point ω interior to a closed curve is totally
defined by the values of the function on the curve.

Calculate
2224ft out where C 121 2

Let flat z ya 4 and w i interior to C f is clearly

in fact entire C Now

holomorphic in and on
by the Cauchy
Integral formula we have

EYI.tt da ziti f f il ziti 3 4 il 21T 4 Bi

s
Evaluate where C 12 2 1 4
fogy an

Observe that the denominator can be factorized as

Z't 9 2 3 il z 3i

And if we look at the domain we observe that

L 3i

3i

2 3 is outside the curve so we write the integrand as

7
so that f is holomorphic on and in C Now we can
apply
Cauchy's integral formula

If de ziti f si ziti 3ft ti

t
ft f
The previous Theorem can be used to derive a similar result for the derivatives of a holomorphic function. This
new result is known as the Cauchy Integral formula for derivatives:

Theorem (Cauchy Integral Formula for derivatives). Let D be a simly connected domain, f holomorphic in it
and C any closed contour in D. For any interior point ω to C we have:
I
n! f (z)
f (n) (!) = n+1 dz
2¼i (z ¡ !)

We mentioned it earlier but we see by this result that holomorphic functions are infinitely differentiable with each
of the derivatives being also holomorphic.

In fact holomorphic functions are analitic and admit a Taylor series expansion.

Theorem (Taylor). Let D be a domain, f holomorphic in it and ω ∈ D. Then f admits a power series
representation
1
X f (n) (!) n
f (z) = (z ¡ !)
n=0
n!

valid within the biggest disc centered at ω and entirely contained in D.

A few observations regarding Taylor expansion of a function:

• The series corresponding to a holomorphic function is unique, no matter how we obtain it.

• The radius of convergence of the series is the smallest number R > 0 that represents the distance from the
center of the series to a singularity (a point where the function stops being holomorphic), i.e. R is the distance
to the closest point where the function is not defined.

Suppose that the function

feel
fit
has a Taylor expansion centered at w 4 Zi What is the radius
of convergence

The function f has a singularity at 2 1 ti so

R Iw fetid 15 3 it
By
z I 4 Li
Find the Taylor expansion of fat
Iz centered at w o and

w Zi
For w o we use the fact that f is the sum of a geometric

series of ratio 2 and if 121 e I we have

I
For w Zi we use a little trick Observe that

EE
E ate
Use
previous

I
result
Ez EI I I Iga E as

And this series converges for

Itt el this is

12 Lil e 11 Lil E
Observe that the first series was convergent for 121 I and the

second for Z Zi I e F

On the intersection both

series are valid

1111

Outside the intersection at

least one of the series must
diverge
After this small parenthesis we go back to Cauchy Integral Formulas. We see next a couple of examples on how to
use the theorem for derivatives:

Evaluate
go.gg where C 121 1
ztdt
Observe that the denominator can be written as 23 Zt2i so

the integrand is not holomorphic on 2 0 and 2 Zi However

only the former lies within the contour We thus write

that
and we choose flat
It w o and n 2 so we

can make use of the theorem Observe further that

2
f z I f o

Hence

2
de f lol I Ii
z

Evaluate
Iffy da where C is the curve shown below

o
First observe that C is not a simple curve it intersects itself so we

can see it as the union of two closed paths C and G the former
surrounding the origin with negative orientation and the latter
z i
surrou ding
We can thus write

III
on
f Etat

f III III t.tt

an a

To evaluate I we use Cauchy's Integral Formula with flat Efi

so

FI da ziti flo Ziti 3 GTi

For I we use the derivative formula with flt

2 1 n t and

w i so

2 1
fila f i 3 2i

So
da flip ziti 3 21 Y't t Gti

And hence

IIF do Giti Y'T 6 Hi 41T 1 3

if
At this point we should be aware of how powerful the complex integration tools are and how versatile holomorphic
functions are in terms of calculus. We have basically learnt how to compute integrals without using specific
integration techniques, but rather using the properties of holomorphic functions.

There are still some results we haven’t mentioned and that correspond to this block of the course. These are
Liouville’s Theorem and the Fundamental Theorem of Algebra. The first one is a very important result in complex
analysis and the latter is a fundamental theorem not only in a complex setting but for mathematics in general.

Theorem (Liouville). The only bounded entire functions are the constant ones.

Theorem (The Fundamental Theorem of Algebra). If p(z) is a non-constant polynomial then the equation

p (z) = 0
has, at least, one root.
Power Series and Taylor expansion of a complex function

It is by now somehow clear that complex analysis is an extension of real analysis. Other concepts that might be
extended to the complex context are sequences and series. In this notes we are not going to discuss much about
them but only scratch the tip of the iceberg. For more information about sequences and series we can always turn
to specialized bibliography.

In any case, for the purpose of the course we don’t need a deep understanding of the topic. In fact, with the
knowledge we already have is more than sufficient.

What we do cover on these notes are the power series of complex numbers. Let a∈ℂ and {c_n} be a sequence of
complex numbers. A power series centered at z = a is defined as

1
X n
cn (z ¡ a) = c0 + c1(z ¡ a) + c2(z ¡ a)2 + ¢ ¢ ¢ + cn(z ¡ a)n + ¢ ¢ ¢
n=0

As you may recall a series can either converge or diverge. The nature of the convergence depends in this case
pretty much on the neighborhood of the complex plane where we look. In other words, what is the behaviour
of the series around z = a or how far can we move from this point so that convergence is still attained.

The goal of this course is not to analyze convergence. However it is important to keep in mind how this works
so that we make sense out of it. We have the following result to throw some light upon this:

Abel’s lemma. Let ρ > 0 a positive number such that the sequence fcn½n g is bounded. Then the series
1
X n
cn (z ¡ a)
n=0
converges absolutely on the disk D(a,ρ) and converges uniformly on compact sets K⊆D(a,ρ).

Uniform convergence means that the convergence does not depend on the points, i.e. that the convergence rate
is the same for every point within the set K.

The biggest number R for which a power series converges

i is called radius of convergence as we saw in
previous courses.

Once convergence is settled it is natural to think that the limiting result should be a function f(z). This is indeed
so and whenever this happens we refer to f as an analitic function. A function of this type is differentiable
infinitely many times within the radius of convergence. In fact we have

1 1
X n
X n¡1
f (z) = cn (z ¡ a) =) f 0 (z) = ncn (z ¡ a)
n=0 n=1

This process can be repeated as many times as we want so that we obtain higuer order derivatives of the
function f.
Theorem (differentiation of a power series). Let a∈ℂ be the center of a power series and let Ω be the domain of
convergence of the power series. Let f : Ω → ℂ be the limit function of the series, i.e.
1
X n
f (z) = cn (z ¡ a) ; z 2 «
n=0

Then f is infinitely differentiable in Ω and for every k∈ℕ the k’th derivative can be obtained by differentiating the
series term by term, this is:

1
X n¡k
f (k) (z) = n (n ¡ 1) ¢ ¢ ¢ (n ¡ k + 1) cn (z ¡ a)
n=k
In particular

(k) f (k) (a)

f (a) = k!ck =) ck =
k!

Under the conditions of the previous theorem we can give the following definition:

Definition (Taylor series). Let f be an infinitely differentiable function over some domain Ω ⊆ ℂ and let a ∈ Ω.
The power series
X f (n) (a) n
(z ¡ a)
n¸0
n!
is called the Taylor series of f at z = a.

So the way to define the Taylor series of an analitic function is pretty much the same as we did for real functions.
In this course we are interested in holomorphic functions. The good news is that every holomorphic function is
also analitic and hence we can represent them as Taylor series.

In the literature, the concepts of analitic and holomorphic are not distinguished some times and use the former
one to refer to the latter. Analitic functions were defined much before and were used in several applications in
mathematical analysis that involved real functions. Then the same definition was extended in a pretty natural
way to complex functions and that is why some writers have not changed the terminology.

In essence both concepts are the same but the term holomorphic is closer to the concept of entire which makes
complex functions somehow particular.

We do not need to worry about the terminology in this course since we are going to focus on holomorphic
functions and so we can use that term freely.
Here you can find the Taylor Series of the fundamental functions or the functions that we will be using more often
in this course:
Meromorphic Functions

We have seen that a holomorphic function is also analytic and that it can be represented by a power series, better
known as the Taylor expansion of the function. This expansion is valid within a domain for which the function is
holomorphic (always around some point ω). The next question that needs to be adressed is what happens if a
function f is not holomorphic at some point? Can we no longer represent it by a series? This seems unreasonable if
we consider a domain D where the function is holomorphic except for (at least) one point.

In this chapter we deal with singularities of complex functions. Singularities are basically points where a function
f is not holomorphic and we will see how we can define a different type of series to represent it, the so called
Laurent Series. This concept carries along the concept of residue, which we will work out as well and end up with
the Residue Theorem and its very nice applications.

Laurent Series

Definition (Singularity). We say a point ω is a singularity or singular point of a function f if this is not
differentiable on that point. Further, the point ω is said to be an isolated singularity if there exists a reduced
neighborhood 0 < | z - ω | < R where f is holomorphic.

As an easy example, we see that the points z = ±2i are isolated singularities of the function
z
f (z) =
z2 + 4

f is holomorphic inside the

annulli 0 < |z ± 2i| < 1

I
On the other hand (as another example) the point ω = 0 is not an isolated singularity for the function f(z) = Log(z)
since for every neighborhood of the origin there are points of the negative real line, points for which f is not
holomorphic.

Theorem (Laurent). Let f be holomorphic on a ring-shaped domain D say r < |z - ω| < R. Then f admits a series
expansion of the form: 1
X k
f (z) = ak (z ¡ !)
k=¡1

valid for r < |z - ω| < R. The coefficients are given by

I
1 f (s)
ak = k+1 ds; k2Z
2¼i C (s ¡ !)
where C is any closed contour in D surrounding ω.
Definition (Laurent Series). The series in the theorem above is called the Laurent Series or Laurent Expansion of
f centered at ω in the annullus r < |z - ω| < R. The series
1
X a¡k
k
k=1 (z ¡ !)
is called the principal part of the expansion whereas the series
1
X k
ak (z ¡ w)
k=0
is the analitic part.

As an observation, the ring for which the Laurent Series is defined does not necessarilly need to have a ring shape.
We consider the following possible cases:

• r = 0, R < ∞. In this case we have a punctured disc.

• r > 0, R = ∞. In this case we have the exterior part of a closed disc centered at ω.

• r = 0, R = ∞. In this case we have the punctured complex plane ℂ\{ω}

I
Calculate the Laurent series of the function

flat sing
centered at w 0

We know that the sine function is entire and its McLaurin expansion

is given by

sine
É 22kt 2 e e

If we divide this series by 2 we obtain

El
É
226 3
flat si It

Is IT t
É É t

So the principal part of the expansion is

É I
which is defined for 121 0 the analytic
whereas
part

É É t

converges everywhere Hence the Laurent Series is valid

for o e Iz le t o
As an observation, note that in the Laurent Theorem, if we consider k = -1 for the calculation of the coefficients we
get:
I
1
a¡1 = f (z) dz
2¼i C

or rather
I
f (z) dz = 2¼ia¡1
C

The coefficient a_{-1} is important as we will see later on the course.

Zeros and poles

We have mentioned that points ω where a function is not holomorphic is called a singular point. Singular points
can be classified into different categories depending its nature. This classification is based on how the Laurent
Series of the function centered on that singularity is.

• If the principal part of the series is zero then ω is a removable singularity.

• If the principal part of the series has a finite number of non-zero terms, then ω is called a pole. In this case, if
the last non-zero term of the series is a_{-n} we say that ω is a pole of order n. For instance, if ω is a pole or
order 1 then the principal part has exactly one term a_{-1}. A pole of order 1 is commonly known as simple
pole, a pole of order 2 is a double pole, and so on.

• If the principal part of the series has infinitely many non-zero terms then ω is an essential singularity.

Definition (Meromorhic function). A function f that is holomorphic except for poles is called meromorphic.

Show that 2 0 is a removable singularity of flat 557

We know that
sinz
E.EE zLht

so if divide the series

we
by t we
get
sing I z n

Since every a 0 the principal part is zero we conclude

that 2 0 is a removable singularity
If a function f has a removable singularity we can make it holomorphic by defining the function on that point. For
instance, in the previous example, it makes sense to define f(0) = 1 if we look at the series. By doing so we make f
entire since it becomes holomorphic on the whole complex plane.

Using the same argumentation as in the previous example we can prove that z = 0 is a simple pole of the function

sin z
f (z) =
z2

Definition (Zeros). We say a function f has a zero of order n in ω if it is a zero of f and the first n - 1 derivatives,
i.e.:
f (!) = f 0 (!) = : : : = f (n¡1) (!) = 0

but
f (n) (!) 6
=0

Theorem. An analitic function f has a zero of order n in ω on a disc |z - ω| < R if and only if f can be written in
the following way:
n
f (z) = (z ¡ !) Á (z)

where the function Φ is analitic on ω with Φ(ω) ≠ 0.

Show that 2 0 is a zero of order 3 of the function

flz Zsin z

We know that
sine
E Etf Int

Thus
sine
É Egg aunt z
I gig e

And so
flat zsin t 23 z

where

I
a a
act
Theorem. A function f has a pole of order n in ω on the punctured disc 0 < |z - ω| < R if and only if f can be
written as
Á (z)
f (z) = n
(z ¡ !)
where Φ is an analitic function on ω with Φ(ω) ≠ 0.

In complex analysis zeros of functions are isolated, i.e. the function does not vanish on a reduced neighborhood
of a zero. As a consequence if ω is a zero of some analitic function f then 1/f has an isolated singularity on ω.
The following result characterizes poles of quotient functions.

Theorem. If g and h are analitic functions in ω and h has a zero of order n in ω but g(ω) ≠ 0, then the quotient
f=g/h has a pole of order n in ω.

Determinethe zeros and poles of the following functions

22 1
il fiz
z 1 ats z 2T
ii glad
the
if The numerator and denominator have no common zeros so

f has a simple zero at

f has simple poles at 2 1 and 2 5
i f has a pole of order 4 at 2 2

iil We saw on a
previous example that

2 sin22 23 0 t o 0

Hence 2 0 is a zero of order three

By
the previous theorem thus
g has
a
pole of order 3
at 2 0
The Residue Theorem

This section introduces the concept of the residue of a function, which we already mentioned. We see further how
the calculation of residues is all we need to calculate integrals on the complex plane over closed contours. Let’s
thus first give the formal definition of a residue:

Definition (Residue). Suppose the function f has an isolated singularity on ω. The coefficient a_{-1} on the
Laurent series of the function valid on a punctured disc 0 < |z - ω| < R is called residue of f in ω and it is denoted
by:
a¡1 = Res (f ; !)

Calculate Res f o where f z eh

By definition the residue is the first coefficient of the principal
part of the Laurent Series In this case we base on
et
I tzed

So that

e
If at
It E Est
Observe that this expansion is valid for 12130 Observe
further that a 3 so that

ResCf O
Z

However sometimes the calculation of the residue using the definition can be a little too tedious since that implies
calculating the Laurent Series of the function in hands. Next we give different tools to calculate the residue
withouth having to go through the definition.

Proposition. If ω is a simple pole of f then

Res (f ; !) = lim [(z ¡ !) f (z)]

z!!
The previous result can be generalized as follows:

Proposition. If ω is a pole of order n of f then

1 dn¡1 n
Res (f ; !) = lim n¡1 [(z ¡ !) f (z)]
(n ¡ 1) ! z!! dz

In some cases the previous propositions might be a bit too exhausting to use since it involves the calculation of
derivatives and limits. Next we see a few other results that might come in handy to calculate residues.

Proposition. If g is a holomorphic function on a neighborhood of ω, m ∈ ℕ and if

g (z)
f (z) = n
(z ¡ !)
with g(ω) ≠ 0, then

g (m¡1) (w)
Res (f ; w) =
(m ¡ 1) !
Proposition. Suppose f = g/h where g, h are holomorphic function on ω. If g(ω) ≠ 0 and h has a simple pole at ω
then f has a simpole pole at ω and
g (!)
Res (f ; !) =
h0 (!)

Proposition. Let g, h be holomorphic function on ω. Suppose further that ω is respectively a zero of order m and
a zero of order m + 1 of g and h. Then f = g/h has a simple pole at ω and

g (m) (!)
Res (f ; !) = (m + 1) (m+1)
h (!)

Seems like a lot of information but each of this propositions eases the calculation of the residues greatly. We only
need to carefully learn when to use each of them by identifying the type of function we are working with. And if
non of them works we can always use the definition of a residue if we see an easy way to calculate the
corresponding Laurent Series.

Let’s see an example of calculation of residues.

Determine the poles of the function

flat II
and calculate the corresponding residues

The poles of f are the solutions to 2 t 1 0 so thefourth

roots of a that we can identify as
Z e z e z e's z ed

previous to the last theorem that

By the we see

Reset w
YET
Taking g n and h It 1 we conclude that

Res f ti typ g ft il Resff a G i

f
ResffEs I Treat it Reset z ati
All the previous results were led to the following theorem, which allows us to calculate contour integrals by just
calculating the residues of a function enclosed by the curve.

Theorem (Cauchy’s Residue Theorem). Let D be a simply connected domain and let C be a closed contour
entirely contained in D. If a function f is analitic in and on C, except for a finite number of isolated singularities
surrounded by C, then
I n
X
f (z) dz = 2¼i Res (f ; zk)
C k=1

Evaluate

12 9 3

where C is

a the rectangle defined X 0 X 4 1 1

by y y
6 the circumference 171 2

We see on the following sketch the curves al bl and the poles

of f in this case simple at 2 3 and double at t 1

C
ca
Z I
73

a This curve encloses both poles for which we

already
the residues calc lated
on
example
a previous so

flt at Ziti Res fi Reiff 3 2mi t t

t
Ca
 b In this case the curve surrounds one of the poles
only
fads ziti Res fill ziti t i I
Ch

Application of the Residue Theorem to the calculus of real integrals

We now have a real feeling that how powerful complex integration results are. Let’s see next how we can use the
previous results to apply in problems that might be familiar to us but attacked from another perspective. The goal
now is to calculate integrals of the form

Z 2¼
F (cos µ; sin µ) dµ
0

Z 1
f (x) dx
¡1

where F and f are rational functions, in the case of f irreducible. Let’s thus see different cases and how to apply
the Residue Theorem in each of them.

Trigonometric integrals

The strategy for this type of integrals is to turn them into contour integrals around the unit circle by introducing
the parametrization

z (µ) = eiµ ; 0 · µ · 2¼
Thus,

0 iµ eiµ + e¡iµ eiµ ¡ e¡iµ

z (µ) = ie ; cos µ = ; sin µ =
2 2

Or equivalently

0 z + z ¡1 z ¡ z ¡1
z = iz; cos µ = ; sin µ =
2 2

By susbstituing these changes into the original integral we thus obtain:

Z 2¼ I µ ¶
z + z ¡1 z ¡ z ¡1 dz
F (cos µ; sin µ) dµ = F ;
0 jzj=1 2 2 iz

which now can be solved using the Residue Theorem.

 ftp
Evaluate
gi
We begin
byintroducing the charges

I fy
121 1
FEET
171 1

If
zt2

We find the zeros of the denominator

z
41ft YEI 2 IF

And we sketch everything out on the complex plane

121 1

2
of 2
8

So
only one of the poles is contained in the wit circle Observe
further that this pole has order 2 since

CETI 1 FtzF
And

Res f AB let2 B flat

alley
 It afar it IIIT Er
All in all we
get

I o o

Ziti Res f Lt B

I I
if
htt

Improper integrals

Recall that improper integrals are the ones evaluated at infinity or, else, the ones evaluated for functions that blow
up at certain points. In this case we focus on the former, i.e. integrals of the form
Z 1
f (x) dx
¡1

Recall also that improper integrals can be either convergent or divergent. In this course we will not discuss
convergence and work only with improper integrals we know converge. Thus
Z 1 Z R
f (x) dx = lim f (x) dx
¡1 R!1 ¡R

We further consider f to be rational and continuous on the real line, so

p (x)
f (x) = ; q(x) 6
= 0; 8x 2 R
q (x)
 So how do we introduce complex integration in this case?
We know that even if q has no real roots, by the Fundamental Theorem of Algebra it does have complex roots so
f(z) has singularities on the complex plane. We thus consider a closed contour on the form of a semi circle big
enough so that surrounds singularities on the upper half of the plane:

Singularitiesoffled
c
t n f k en
Iti
i
m R
By the Cauchy Residue Theorem we know that
I Z R Z n
X
f (z) dt = f (x) dx + f (z) dz = 2¼i Res (f ; zk)
C ¡R °R k=1

where γ_R is the arc defining the semi circle which obviously depends on R. The trick now is to take the identity in
the limit as R tends to infinity and see what happens. The ideal scenario would be that the contour integral over γ_R
tends to zero. For that we have the following proposition:

Proposition. Suppose f(z) = p(z)/q(z) is a rational function such that deg(q) ≥ deg(p) + 2. If C_R is the arc of the
semi circle of radius R then
Z
lim f (z) dz = 0
R!1 CR

So in this circumstance we have

Z 1 n
X
f (x) dx = 2¼i Res (f ; zk)
¡1 n=1

If the conditions on the proposition are not satisfies we would have to find another way to compute the contour
integral.

is
Evaluate

Latifa
the function fix
Clearly the denominator of does not have real
roots but does have complex roots 2 Ii and 2 130
say
We thus consider

flat
gift zidt iÉ
and let R be a real number such that R 3 so that all
the singularities on the upper half of the plane are contained in
the semi circle Observe that all singularities are simple poles

Ji r
e

Now since f satisfies the condition on the previous proposition

we conclude that

king fields 0

it fit etat Ii
And thus felt
fluffy dt Ziti Res fi t Res f 3

T
ReeIfpi
EY HIII Ii
I
Ziti
fi fi
 Consider now that the functions f(z) does have a singularity on the real line, say z = c. The way to go now is to
consider a contour of the following form:

c
e
r

S
R c r ctr R

So we circle around the real singularity by considering a smaller semi circle of radius r. The following result tells us
how to compute the contour integral when we let r tend to 0 as we wish.

Proposition. Suppose f has a simple pole on z = c. If we consider the parametrization z(θ) = c + exp(iθ) for the
angle moving bewteen 0 and π then
Z
lim f (z) dz = ¼iRes (f ; c)
r!0 Cr

As a conclusion we see that if a complex rational function f has poles both real and complex, the way to calculate
the improper integral is

Z 1 n
X m
X
f (x) dx = 2¼i Res (f ; zk) ¡ ¼i Res (f ; cj)
¡1 k=1 j=1

where z_k are the complex singularities and c_j are the real (simple) poles.

Evaluate the
integral

a a

Clearly x o is a real single pole of the function

feel
ZEIT
Observe also that

2 2f n I i are singularities of f
So in this case we have a real single pole on E o and
another
singularity on the upper half plane t Iti Let thus
RS Il til E so that

Ati
r

s
R r o r R

We thus have
R

fields
i If
fat t da t
fCr fields t fled da
Cr

Res f Ea Res f Iti

Since f is a rational function when dg g 3 dg p t 2 then

lien
Red f flaids 0

Cm

And

lying fields tires f o

So the exercise now is to calculate the residues at E Ati

and 2 0
Both are single poles so

Res f ol
IF IT
Res f Iti
EI I IF It
All together we conclude

I i i
RefftiRestfo
i i

Cauchy’s Residue Theorem can be used in many other situations. We only covered the basic ones. However, the
knowledge you acquired is sufficient to understand any other use of the Theorem.

The tools in Mathematical Analysis you own at this point are more than enough to face a wide variety of problems
and also to expand the knowledge on your own if ever required.