0% found this document useful (0 votes)
26 views99 pages

Applied Mathematics III-part III

This document is a course outline for Applied Mathematics III, focusing on Complex Analysis, which includes chapters on complex differentiation, integration, power series, and advanced topics. It discusses the historical development of complex numbers, their definitions, and basic operations, emphasizing their significance in solving equations. The document also covers the properties of complex numbers and their representation in the complex plane.

Uploaded by

tadesse.bekeshie
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
26 views99 pages

Applied Mathematics III-part III

This document is a course outline for Applied Mathematics III, focusing on Complex Analysis, which includes chapters on complex differentiation, integration, power series, and advanced topics. It discusses the historical development of complex numbers, their definitions, and basic operations, emphasizing their significance in solving equations. The document also covers the properties of complex numbers and their representation in the complex plane.

Uploaded by

tadesse.bekeshie
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 99

AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B.

(PhD)

PART III: CALCULUS OF FUNCTIONS OF COMPLEX


VARIABLES [COMPLEX ANALYSIS]
This part of the course Applied Mathematics III deals with calculus of functions of the form
𝑓: 𝐷 ⊆ ℂ → ℂ and it is comprised of four main chapters and one optional chapter. They are
titled respectively complex Differentiation, Complex Integration, Power Series and Laurent
Series, Calculus of Residues and Miscellaneous Advanced Topics (eg, conformal mapping).

CHAPTER 5: COMPLEX DIFFERENTIATION

5.1 Complex Numbers

A. The Genesis and Development of Complex Numbers

The real number system is not an adequate universal set to solve certain equations such as: x2 + 1 = 0, for
instance. The need to solve such equations forced mathematicians to introduce a number system more
general than the real number system. The second (and probably the main) reason to study complex
numbers is the fact that real problems with real solutions themselves can be determined by computations
in the complex domain. Let’s delve more into a short history of the genesis and development of complex
numbers.

Historical Motivation: Solving Cubic and Quartic Equations

In the 16th century, Italian mathematicians Del Ferro, Ferrarri, Tartagila, Cardano and
Bombelli developed formulas to solve cubic and quartic equations. Del Ferro and Tartaglia
found a formula for the solution of cubic equations and Ferrari found a formula for the quartic
equation. Cardano is credited with publishing both formulas. Cardano’a cubic formula can find
only a single solution (at the time). In 1824 Abel, Norwegian mathematician proved that for 𝑛 ≥
5, in general, it is impossible to express the roots of 𝑎𝑛 𝑥 𝑛 + ⋯ + 𝑎1 𝑥 + 𝑎0 in terms of only the
coefficients 𝑎𝑛 , 𝑎𝑛−1, … , 𝑎1 , 𝑎0 and the operations of addition, subtraction, multiplication,
division and radicals (nth root extraction).

The roots of 𝑎𝑥 3 + 𝑏𝑥 2 + 𝑐𝑥 + 𝑑 = 0 can be found using the formula:

1
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

Cardano and Bombelli noticed that the cubic formula sometimes yield solutions involving the square
root of negative numbers (weird at the time), even when the equation has only real solutions. If a cubic
equation has three real roots (this case is called the "irreducible" case), the cubic formula requires the
evaluation of cube roots of "complex numbers", which can be tricky. For example, while solving 𝑥 3 −
15𝑥 = 4, Bombelli encountered a very unusual value. Though all solutions of the equation are
real numbers (one of them is 𝑥 = 4 and the other roots can easily be found by factorizing), but
Cardano’s formula (with 𝑎 = 1, 𝑏 = 0, 𝑐 = −15 and 𝑑 = − 4) gives

3 3
𝑥= √ 2 + √ −121 + √ 2 − √ −121

The value of x yielded from cardano’s formula was unusual at the time. Bombelli overcome this
difficulty by simply treating √−1 as a number and by operating with it following some specific
arithmetic rules (same rule we use today). Thus he discovered that

3 3
𝑥= √ 2 + 11√ −1 + √ −2 − 11√ −1 =2+√−1 +2 -√−1 = 4

This way Bombelli revolutionized mathematics by accepting square roots of negative numbers as
legitimate numbers and by developing the first arithmetic rules of them.

Later Developments

Many mathematicians initially viewed complex numbers with skepticism or out rightly rejected
them, considering them "imaginary" and "useless" before their utility and importance in solving
equations and other mathematical problems became apparent. Wallis showed how to represent
complex roots of quadratic equations geometrically. Wessel and Argand provided geometric
representations of complex numbers as vectors, which helped to visualize and understand them

2
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

better. Cauchy and Riemann further developed complex analysis, solidifying its place in
mathematics.

B. Definition of complex Numbers.

Definition 1:

a) A complex number z is an ordered pair (𝑥, 𝑦) of real numbers x and y, written

𝑧 = (𝑥, 𝑦)

b) If 𝑧 = (𝑥, 𝑦) is a complex number , then x is called the real part of z and y is called the
imaginary part of z, written

𝑥 = 𝑅𝑒(𝑧) and 𝑦 = 𝐼𝑚(𝑧)

c) The pair (1, 0) is called the imaginary unit and it is denoted by I and sometimes by j.
Thus
𝑖 = (0,1)
d) The pair (0, 𝑦) is called pure imaginary number.

Remarks:

a) Notice the following analogies between complex numbers and other number systems:
✓ Just as some real numbers are denoted by alphabets, for example e, so the complex
number (0,1) is denoted by a letter.
✓ Just as a rational number is constructed from a pair of integers (𝑎, 𝑏) (where 𝑏 ≠ 0), a
complex number is constructed from a pair of real numbers.
b) The set of all complex numbers is denoted by ℂ.
c) Complex numbers of the form (𝑥, 0) are considered the same as real numbers x because they behave
identically under addition and multiplication, and there exist an isomorphism between them. In
particular 0 = (0,0). Hence, ℝ ⊆ ℂ..
d) Geometry of Complex numbers: Geometrically, the complex number z can be thought as (or
represented by) the point corresponding to (𝑥, 𝑦) in xy-plane or the vector 𝑥𝑖 + 𝑦 𝑗, where 𝑥 = 𝑅𝑒(𝑧)
and 𝑦 = 𝐼𝑚(𝑧) Thus corresponding to each point P or (each vector v) on xy-plane there is a unique
complex number and conversely.

3
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

e) When the xy-plane is used to represent ℂ it is called the complex plane, the Argand plane or the z-
plane; the x-axis is called the real axis and the y-axis is called the imaginary axis.

Definition 2: Two complex numbers 𝒛 and 𝑤 are said to be equal, written 𝑧 = 𝑤, if and only if their
real parts are equal and their imaginary parts are equal. That is:

z = w ⇔ Re(z) = Re(w) and Im(z) = Im(w).

C. Operations with Complex Numbers: Basic Operations

The basic operations with complex numbers are addition, subtraction, multiplication and
division. There are also special operations such as conjugation, finding modulus (absolute
value), representing in polar form and finding root and power.

Definition 3(arithmetic operations on ℂ): Given two complex numbers 𝑧 = (𝑎, 𝑏) and 𝑤 =
𝑧
(𝑐, 𝑑), we define the sum 𝑧 + 𝑤, the difference z − w, the product 𝑧 × 𝑤 and the quotient 𝑤 (for

𝑤 ≠ (0,0)) as follows:

a) 𝑧 + 𝑤 ≔ (𝑎 + 𝑐, 𝑏 + 𝑑)
b) 𝑧 − 𝑤 ≔ (𝑎 − 𝑐, 𝑏 − 𝑑)
c) 𝑧 × 𝑤: = (𝑎𝑐 − 𝑏𝑑, 𝑎𝑑 + 𝑏𝑐) (we can also write 𝑧 × 𝑤 as 𝑧. 𝑤 or 𝑧𝑤.)
𝑎𝑐+𝑏𝑑 𝑏𝑐−𝑎𝑑
d) 𝑧 ÷ 𝑤 ≔ (𝑐 2 + 𝑑2 , 𝑐 2 +𝑑2 ) (provided that w ≠ 0)

Note:
i) Note that (𝑥, 𝑦) = 𝑥 + 𝑦𝑖 = 𝑥 + 𝑖𝑦 because (𝑥, 𝑦 ) = (𝑥, 0) + (0, 𝑦) = 𝑥 + (𝑦, 0)(0,1) =
𝑥 + 𝑦𝑖.
The expression "𝒙 + 𝒚𝒊" is the standard notation for complex numbers.
ii) 𝑧 – 𝑤 = 𝑔 if and only if 𝑧 = 𝑤 + 𝑔 and
iii) 𝑧 ÷ 𝑤 = 𝑔 if and only if 𝑤𝑔 = 𝑧.
𝑧
iv) Let 𝑧 = 𝑎 + 𝑏𝑖 and 𝑤 = 𝑐 + 𝑑𝑖,. Then 𝑤 can be computed as follows:
𝑧 𝑎+𝑏𝑖 𝑎+𝑏𝑖 𝑐−𝑑𝑖 𝑎𝑐+𝑏𝑑+(𝑏𝑐−𝑎𝑑)
= 𝑐+𝑑𝑖 = 𝑐+𝑑𝑖 . 𝑐−𝑑𝑖 = .
𝑤 𝑐 2 +𝑑 2

𝑧 𝑎𝑐+𝑏𝑑 𝑧 𝑏𝑐−𝑎𝑑
Hence 𝑅𝑒(𝑤) = and 𝐼𝑚(𝑤) = 𝑐 2 +𝑑2
𝑐 2 +𝑑 2

4
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

v) By the definition of multiplication, we have: 𝒊𝟐 = −𝟏 and (−𝒊)𝟐 = −𝟏. (That is i and –i are square
roots of -1).
vi) Let z and w be complex numbers. Then w is called an additive inverse of z if +𝑤 = 0 . Likewise, w
is called a multiplicative inverse of z if 𝑧 × 𝑤 = 1. Every complex number has a unique additive
inverse; every nonzero complex number has a unique multiplicative inverse. Justify!

Example 1:

a) Find real numbers 𝑥 and 𝑦 where:

3𝑥 + 2𝑦𝑖 − 𝑥𝑖 + 5𝑦 = 7 + 5𝑖

Solution: Applying the definition of addition and equality of complex numbers gives:
3𝑥 + 5𝑦 = 7
{ ⇒ 𝑦 = 2 and 𝑥 = −1. //
2𝑦 − 𝑥 = 5

2+3𝑖
b) Let 𝑧 = . Find 𝑅𝑒(𝑧)&𝐼𝑚((𝑧).
4+𝑖

2+3𝑖 2+3𝑖 4−𝑖 11 10 11 10


Solution: 4+𝑖
= . 4−𝑖 = 17 + 17 i . Thus 𝑅𝑒(𝑧) = 17 and 𝐼𝑚(𝑧) = 17 //
4+𝑖

c) Find the additive inverse and the multiplicative inverse of 𝑧 = 3 + 4𝑖.

1 1
Solution: The additive inverse of 𝑧 is -𝑧 = 3 − 4𝑖 and the multiplicative inverse is 𝑧 = 3+4𝑖 =
1 𝟑−𝟒𝒊 𝟑 𝟒
3+4𝑖
× 𝟑−𝟒𝒊
=
𝟐𝟓

𝟐𝟓
𝒊 //

5
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

Theorem 1 [Field Properties of complex number system): (ℂ, +, ×) is a field, i.e.,

a) ∀z1, z2 𝜖 ℂ , z1 + z2 𝜖 ℂ
b) ∀z1 , z2 , z3 𝜖 ℂ , (z1 + z2 ) + z3 = z1 + ( z3 + z3 )
c) ∀z1 , z2 𝜖 ℂ , z1 + z2= z2 + z1
d) ∀z 𝜖 ℂ , 0 + z = z
e) ∀z 𝜖 ℂ , there exists a unique w 𝜖 ℂ such that z + w = 0. (𝑤 is called the additive inverse of 𝑧
and it is denoted by −𝑧).
f) ∀z1 z2 𝜖 ℂ , z1 z2 𝜖 ℂ
g) ∀z1 , z2 , z3 𝜖 ℂ , (z1 z2 ) z3 = z1 ( z3 z3 )
h) ∀z1 , z2 𝜖 ℂ , z1 z2= z2 z1
i) ∀z 𝜖 ℂ , 1 z = z
j) ∀z 𝜖 ℂ\{0} , there exists a unique w 𝜖 ℂ such that z w = 1. (w is called the multiplicative inverse
of z .)
k) ∀z1 , z2 , z3 𝜖 ℂ , z1 (z2 z3 ) = z1 (z2 z3 ) and (z1 z2 ) z3 =( z1 z2) z3

D. Operations with Complex Numbers: conjugation and Finding Absolute Value

Definition 4: Given a complex number 𝑧 = 𝑥 + 𝑖𝑦,

a) the complex number 𝑧 = 𝑥 – 𝑦𝑖 𝑖𝑠 called the conjugate of z.


b) the number |𝑧| = √𝑥 2 + 𝑦 2 is called the absolute value (modulus) of z. The modulus of z is
also denoted by 𝑚𝑜𝑑(𝑧).

Note that geometrically 𝑧 is the reflection of z through real axis (x-axis) and |𝑧| is the distance between
origin and 𝑧. This interpretation of absolute value is consistent with that of absolute value of real
numbers.

Example 3: (a) ̅̅̅̅̅̅̅̅


1 + 3𝑖 = 1 – 3𝑖 (b) |3 + 4𝑖| = 5

6
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

Theorem 2(properties of conjugates and Absolute Value)

a) 𝑧̅̅ = z
b) 𝑧̅̅̅̅̅̅̅̅
+ 𝑤 = 𝑧̅ + 𝑤
̅
c) 𝑧𝑤
̅̅̅̅ = 𝑧̅𝑤
̅
d) |z| ≥ 0
e) | 𝑧⃗| = |𝑧|
f) 𝑧⃗ 𝑧 = |𝑧|2
g) |𝑧 + 𝑤| ≤ |𝑧| + |𝑤|
h) |𝑧𝑤| = |𝑧||𝑤|

E. Polar Forms of a Complex Number

Definition 5: Let 𝑧 = (𝑥, 𝑦) be a complex number. Let 𝑟 = √𝑥 2 + 𝑦 2 (the distance between z and the

origin and let 𝜃 be the angle between the positive real axis and the line joining the origin and z. Then we
can write 𝑧 = 𝑟 (𝑐𝑜𝑠 𝜃 + 𝑖 𝑠𝑖𝑛 𝜃) . This representation is called the polar form or trigonometric form of
𝑧 . Note that this representation is not unique. The real number 𝑟, is the modulus |𝑧| of z. The angle θ is
called an argument of z and it is written 𝑎𝑟𝑔(𝑧). Note that the value of 𝑎𝑟𝑔(𝑧) is not unique, in fact,
𝑎𝑟𝑔(𝑧) is multivalued. But the value of 𝑎𝑟𝑔(𝑧) in (-𝜋, 𝜋] is unique. The unique value of 𝑎𝑟𝑔(𝑧) that lies
in (-𝜋, 𝜋] is called the principal argument of 𝑧 and it is denoted by 𝐴𝑟𝑔(𝑧).

Remarks:

𝑦
- Algebraically, 𝜃 (that is arg(z) ) is given by: 𝜃 = arctan (𝑥 ). In using this formula, we must pay
attention to the quadrant in which z lies because the period of tan is 𝜋.
- a𝑟𝑔 0 is undefined;
- 𝑎𝑟𝑔( ⃗⃗⃗⃗
𝑧 ) = −arg (𝑧) (up to integer multiples of 2𝜋)
- 𝑎𝑟𝑔 (𝑧 ) is not unique. In fact it is multivalued..
- arg ( 𝑧 ) is unique up to integer multiples of 2𝜋.
- The value of 𝑎𝑟𝑔(𝑧) in (-𝜋, 𝜋] is unique.
- The unique value of 𝑎𝑟𝑔(𝑧) that lies in (-𝜋, 𝜋] is called the principal argument of z, 𝐴𝑟𝑔(𝑧).
- Here, as in calculus, all angles are measured in radians and positive in the counter clockwise sense.

7
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

- To write a polar form of a complex number 𝑧, we can use any value of the argument (not
necessarily the principal argument) of 𝑧

Example 4: Let 𝑧 = 1 − √3𝑖. find 𝑚𝑜𝑑( 𝑧), 𝑎𝑟𝑔 (𝑧) 𝑎𝑛𝑑 𝐴𝑟𝑔( 𝑧). Express 𝑧 in polar form.

Solution:

2
✓ mod(z) = √12 + (−√3) = 2

−√3 −𝜋
✓ 𝑎𝑟𝑔(𝑧) = 𝑎𝑟𝑐𝑡𝑎𝑛( )= 3 + 2𝑘𝜋, 𝑘 is any integer.
1
𝜋
✓ The principal argument is − .
3

✓ The polar form of 𝑧 = 1 − √3𝑖 is:

−𝜋 −𝜋
𝑧 = 2 (𝑐𝑜𝑠 ( ) + 𝑖 𝑠𝑖𝑛 ( ))
3 3

Proposition (Multiplication and Division in Polar form)

Let 𝑧1 = 𝑟1 (𝑐𝑜𝑠 𝜃1 + 𝑖 𝑠𝑖𝑛 𝜃1 ) and 𝑧2 = 𝑟2 (𝑐𝑜𝑠 𝜃2 + 𝑖 𝑠𝑖𝑛 𝜃2 ). Then

1. 𝑧1 𝑧2 = 𝑟1 𝑟2 [𝑐𝑜𝑠 (𝜃1 + 𝜃2 ) + 𝑖 𝑠𝑖𝑛( 𝜃1 + 𝜃2 )].


𝑧1 𝑟
2. = 𝑟1 [cos (− 𝜃2 ) + 𝑖 𝑠𝑖𝑛( 𝜃1 − 𝜃2 )]
𝑧2 2

In other words,

1. 𝑚𝑜𝑑(𝑧1 𝑧2 ) = 𝑚𝑜𝑑(𝑧1) 𝑚𝑜𝑑(𝑧2 ) (Modulus is multiplicative).


2. 𝑎𝑟𝑔(𝑧1 𝑧2 ) = 𝑎𝑟𝑔(𝑧1 ) + 𝑎𝑟𝑔( 𝑧2 ) (𝑎𝑟𝑔(𝑧) is additive up to integer multiples of 2𝜋)
𝑧
3. 𝑚𝑜𝑑(𝑧1 ) = 𝑚𝑜𝑑(𝑧1 ) ÷ 𝑚𝑜𝑑(𝑧2 ) and
2
𝑧
4. 𝑎𝑟𝑔 ( 1) = 𝑎𝑟𝑔(𝑧1 ) − 𝑎𝑟𝑔( 𝑧2 ) .
𝑧2

8
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

F. Powers and Roots

Definition 6: Let z ∈ ℂ and n be a natural number. We define

a) 𝑧 𝑛 = 𝑧. 𝑧 … . . 𝑧
1
b) 𝑧 −𝑛 = 𝑧𝑛

c) 𝑧 0 = 1

Theorem 3(De Moivre’s Theorem): Let 𝑧 = 𝑟(𝑐𝑜𝑠 𝜃 + 𝑖 𝑠𝑖𝑛𝜃). Then

𝑧 𝑛 = = 𝑟 𝑛 (𝑐𝑜𝑠 𝑛𝜃 + 𝑖 sin 𝑛𝜃)

Proof: We use the principle of mathematical induction.


Since 𝑧1 = 𝑧 = 𝑧 = 𝑟(𝑐𝑜𝑠 𝜃 + 𝑖 𝑠𝑖𝑛 𝜃 ) = 𝑟1 (𝑐𝑜𝑠 1. 𝜃 + 𝐼 𝑠𝑖𝑛 1. 𝜃), so the formula works for 𝑛 =
1. Assume that the result is true for the particular positive integer k, i.e., assume 𝑧 𝑘 =
𝑟 𝑘 (𝑐𝑜𝑠 𝑘𝜃 + 𝑖 sin 𝑘𝜃)
Then, 𝑧 𝑘+1 = 𝑧. 𝑧 𝑘 = 𝑟(𝑐𝑜𝑠 𝜃 + 𝑖 sin 𝜃) 𝑟𝑘(𝑐𝑜𝑠 𝜃 + 𝑖 𝑠𝑖𝑛𝜃) = 𝑟𝑘 + 1(𝑐𝑜𝑠 𝜃 + 𝑖 𝑠𝑖𝑛 𝜃). Thus,
if the result is true for 𝑘, then it is also true for 𝑘 + 1. By the PMI the formula is true for all
positive integers.

Definition 7: Let 𝑛 be a positive integer. A complex number 𝑤 is called an nth root of a


1
𝑛
complex number 𝑧 if 𝑤 𝑛 = 𝑧, and we write 𝑤 𝑛 = 𝑧 or w = √𝑧.

𝑛
Caution: The symbol 𝑤 = √𝑧 is multivalued in contrast to the case of real roots.

Theorem 4: Let 𝑧 = 𝑟(𝑐𝑜𝑠 𝜃 + 𝑖𝑠𝑖𝑛𝜃) ≠ 0. Then

a) 𝑧 has exactly 𝑛 distinct nth roots.


𝑛
b) The n distinct values of √𝑧 are given by
𝑛 𝑛 𝜃+2𝑘𝜋 𝜃+2𝑘𝜋
√𝑧= √𝑟[cos ( 𝑛
) + i sin (
𝑛
)] for 𝑘 = 0,1, … , 𝑛 − 1

9
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

Remark: . If we denote the n roots of z by 𝑤0 , 𝑤1,…,𝑤𝑛−1 , then :

𝑛 𝜃 𝜃
𝑤0 = √𝑟[cos (𝑛) + i sin (𝑛)],

𝑛 𝜃+2𝜋 𝜃+2𝜋
𝑤1 = √𝑟[cos ( 𝑛
) + i sin (
𝑛
)]

𝑛 𝜃+2𝑛𝜋 𝜃+2𝑛𝜋
𝑤𝑛−1 = √𝑟[cos ( 𝑛
) + i sin ( 𝑛
)]

Proof: Let 𝑤 = 𝑅(𝑐𝑜𝑠 ∅ + 𝑖𝑠𝑖𝑛∅) be an nth root of z.

⇒ wn = z.
⇒ [𝑅 cos ∅ + 𝑅𝑖sin]n = r(cos 𝜃 + 𝑖sin𝜃)
⇒ 𝑅 𝑛 [cos n∅ + 𝑅𝑖sin 𝑛∅)] = r(cos 𝜃 + 𝑖sin𝜃)
⇒ 𝑅 𝑛 = 𝑟 and n∅ =𝜃+2k𝜋
𝑛 𝜃+2𝑘𝜋
⇒ 𝑅 = √𝑟 and 𝜙 = for k=0,±1,±2, …
𝑛

But distinct roots are obtained when k = 0, 1,⋯ , 𝑛 − 1. ∎

Note:

1. The n nth roots of a nonzero complex number z form vertices of a regular polygon of n sides.
2. If 𝑤 ≠ 1 is an nth root of unity, then 1 + 𝑤 + 𝑤 2 + ⋯ + 𝑤 𝑛−1 = 0.
G. Topology of Complex Plane

Complex analysis (the study of derivatives and integrals of complex functions) relies heavily on the
concept of open sets and related concepts such as closed sets, connectedness, limit points, limits, etc.
Roughly speaking, open sets are generalizations of open intervals on the real line, open disks (or more
generally open regions) in the plane and open balls (or more generally open regions) in 3D Euclidean
space. The study of open sets and related issues is called topology.

We end this section with the definitions of some important subsets of the complex plane.

10
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

1. Neighbourhoods. A 𝛿- neighborhood of a point z0 is the set of all points z such that


|𝑧 – 𝑧0 | < 𝛿 where 𝛿 is any given positive number. A deleted 𝛿 neighborhood of 𝑧0 is a
neighbourhood of z0 in which the point z0 is omitted, i.e., 0 < |𝑧 − 𝑧0 | < 𝛿.
Notation: The standard notation for 𝛿- neighbourhood of a point z0 of a complex number

is typically written as 𝑵𝜹 (𝒛𝟎 ) or 𝐵𝛿 (𝑧0 ) or 𝐵(𝑧0 , 𝛿).


2. Interior points, Exterior points and Boundary Points. A point z0 is called an interior
point of a set S if we can find a 𝛿 neighborhood of z0 all of whose points belong to S. If
every 𝛿 neighborhood of z0 contains points belonging to S and also points not belonging
to S, then z0 is called a boundary point. If a point is not an interior or boundary point of a
set S, it is an exterior point of S.

3. Limit Points. A point z0 is called a limit point, cluster point, or accumulation point of a
point set S if every deleted neighbourhood of z0 contains points of S. Since 𝛿 can be any
positive number, it follows that S must have infinitely many points. Note that z0 may or
may not belong to the set S.
4. Closure Points: A point z0 is called a closure point of a set S if every neighbourhood of
z0 contains points of S. Note that all limit points are closure points.
5. 𝐎𝐩𝐞𝐧 𝐒𝐞𝐭𝐬: A set S is called open if and only if every point 𝑧0 of S has a
neighbourhood 𝑁𝛿 (𝑧0 ) such that 𝑁𝛿 (𝑧0 ) ⊆ 𝑆 .
For example, the set of points z such that |𝑧 – 𝑧0 | < 1 is an open set.
Note: A set is S open if and only if all elements of S are interior points of S.

6. Closed Sets. A set S is said to be closed if and only if ℂ ∖ 𝑆 is an open set.


For example, the set of all points z such that |𝑧 – 𝑧0 | ≤ 1 is a closed set.
Note: Every limit point of S belongs to S, i.e., if S contains all its limit points.
7. Bounded Sets. A set S is called bounded if we can find a constant M such that jzj , M for
every point z in S. An unbounded set is one which is not bounded. A set which is both
bounded and closed is called compact.
8. Connected Sets. An open set S is said to be connected if any two points of the set can be

11
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

joined by a path consisting of straight line segments (i.e., a polygonal path) all points of
which are in S.
9. Open Regions or Domains. An open connected set is called an open region or domain.
10. Closure of a Set. If to a set S we add all the limit points of S, the new set is called the
closure of S and is a closed set.
11. Closed Regions. The closure of an open region or domain is called a closed region.

Class work: From Kreyszig (Page 612 and 613: 2, 6, 9), page 618(Q13, 21, 31) and page 619 ( No.
33,34,35)

Exercise 5.1:

1
1. Let 𝑧 = 𝑥 + 𝑦𝑖. Then find real part of the complex number = 1−𝑧 .
5𝜋
2. Let z be a complex number such that |z| = 4 and 𝑎𝑟𝑔(𝑧) = 6
. Then express z in the form of 𝑥 +

𝑖𝑦.
4+2𝑖
3. Find the modulus of the complex number 1−2𝑖 .

4. Find the principal argument of the complex number −2√3 − 2𝑖 .


5. Find the conjugate of (1 + 𝑖)3 .
637 337
1 𝑖√3 1 𝑖√3
6. Find the value of the expression(− + ) + (− 2 − ) .
2 2 2

7. Express the multiplicative inverse of the complex number −1 − 𝑖 in the form 𝑥 + 𝑖𝑦.
8. Prove: If 𝑤 ≠ 1 is an nth root of unity, then 1 + 𝑤 + 𝑤 2 + ⋯ + 𝑤 𝑛−1 = 0.
2𝜋 2𝜋
9. If 𝑧 = 𝑐𝑜𝑠( ) + 𝑖𝑠𝑖𝑛( ), then show that
5 5

1 + 𝑧 + 𝑧 2 + 𝑧 3 + 5𝑧 4 + 4𝑧 5 + 4𝑧 6 + 4𝑧 7 + 4𝑧 8 + 5𝑧 9 = 5𝑧 4 .

𝑧−𝑧 𝜋
10. If 𝑧1 = 9 + 5𝑖, 𝑧2 = 3 + 5𝑖 and arg (𝑧 − 𝑧1) = 4 , then find the value of |𝑧 − 6 − 8𝑖|.
2

11. If a complex number 𝑧 lies in the interior or on the boundary of a circle of radius 3 units and
centre (– 4, 0), the greatest value of |𝑧 + 1|
12. Show that the n nth roots of a nonzero complex number z form vertices of a regular polygon of n
sides.

12
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

5.2 Complex Functions

A. Complex Function: Definition

Definition 1: Let 𝑆 ⊆ ℂ. A rule that assigns to every z in S a complex number w is called a


complex valued function (of a complex variable) or briefly a complex function. In this case, w is
called the value of 𝑓 at 𝑧. We write 𝑤 = 𝑓(𝑧).Here z varies in S and is called a complex variable.

Notes:

✓ Briefly, a complex function is any function 𝑓 of the form:


𝑓: 𝑆 ⊆ ℂ → ℂ
✓ In most cases the domain of definition S of 𝑓 will be an open and connected set, thus a
domain as defined just before (in section 5.1).
✓ A complex function is given usually by a formula of the form
𝑤 = 𝑓(𝑧)
✓ A complex function 𝑓 can be written as

𝑓(𝑧) = 𝑢(𝑥, 𝑦) + 𝑖𝑣(𝑥, 𝑦),

where u and v are real valued functions of two real variables 𝑥 and 𝑦. In this case 𝑢 is called
the real part of 𝑓 and 𝑣 is called the imaginary part of 𝑓.

✓ If not given, the domain of a complex function 𝑓 is:


𝐷𝑜𝑚𝑎𝑖𝑛(𝑓) = {𝑧: 𝑓(𝑧) ∈ ℂ} = {(𝑥, 𝑦) ∈ ℝ2 : 𝑢(𝑥, 𝑦)𝜖ℝ and 𝑣(𝑥, 𝑦)𝜖ℝ}

Example 1: Given the complex function 𝑓(𝑧) = 𝑧 2 + 3𝑧,

a) Evaluate 𝑓 (1 + 3𝑖).
b) Find the real part and imaginary part of 𝑓 and
c) Find the domain and range of 𝑓

Solution:

a) 𝑓(1 + 3𝑖) = −5 + 15𝑖.

13
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

b) Let 𝑧 = 𝑥 + 𝑦𝑖. Then, 𝑓(𝑧) = 𝑧 2 + 3𝑧 = (𝑥 + 𝑦𝑖)2 + 3(𝑥 + 𝑦𝑖) = (𝑥 2 −


𝑦 2 + 3𝑥) + (2𝑥𝑦 + 3𝑦)𝑖. Thus 𝑅𝑒(𝑓) = 𝑥 2 − 𝑦 2 + 3x and 𝐼𝑚(𝑓) = 2𝑥𝑦 + 3𝑦.
c) The domain of 𝑓 is ℂ.
To find the range of f:

3 2 9
Step 1: Rewrite the function by completing the square: 𝑓(𝑧) = (𝑧 + ) −
2 4

3 9
Step 2: Let = 𝑧 + . Then 𝑓(𝑧) becomes (𝑤) = 𝑤 2 − 4 . Since 𝑧 can be any complex number,
2

so 𝑤 can also be any complex number.

Step 3: The set of the square of all complex numbers is C and hence 𝑅𝑎𝑛𝑔𝑒(𝑓) = C.

Example 2: Find the domain of

1
a) 𝑓(𝑧) = 𝑧 2+1 Answer: ℂ \{𝑖, −𝑖}

b) 𝑓(𝑧) = 𝑦 ∫0 𝑒 −𝑥𝑡 𝑑𝑡 + 𝑖 ∑∞ 𝑛
𝑛=0 𝑦 . Answer: (0, ∞) × (−1,1)

Example 3: Find domain and range of

a) 𝑓(𝑧) = 𝑧 2
b) 𝑓(𝑧) = 𝑒 𝑥 (cos 𝑦 + 𝑖 sin 𝑦).

B. Graphing Complex Functions

A complex function cannot be graphically represented (in the usual sense) because it requires
four dimensions (two for the domain and two for the range) to do so. But some visual
information can be obtained by sketching various subsets S of the domain in a complex plane (z-
plane) and their images 𝑓(𝑆) on another complex plane (w-plane). A complex function 𝑓 can be
thought as a transformation of z-plane to 𝑓(𝑧) −plane. Especially S is a region or a curve.

Example 4: Let 𝑓(𝑧): = 𝑧 2 and let 𝑆 be the quarter unit disk {(𝑥, 𝑦): 𝑥 ≥ 0, 𝑦 ≥ 0 & 𝑦 ≤
√1 − 𝑥 2 }. Sketch 𝑆 and 𝑓(𝑆) in different complex planes.

𝜋
Solution: The upper quarter circle can be described by 0 ≤ |𝑧| ≤ 1 and 0 ≤ 𝑎𝑟𝑔(𝑧) ≤ .
2

14
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

Apply the given transformation 𝑓(𝑧) = 𝑧 2 to the polar form of a complex number and determine
the resulting region.

Step 1: Express 𝑧 in polar form. Let 𝑧 = 𝑟𝑒 𝑖𝜃 .

2
Step 2: Apply the transformation 𝑓(𝑧): = 𝑧 2 . 𝑓(𝑧) = (𝑟𝑒 𝑖𝜃 ) =𝑟 2 𝑒 𝑖2𝜃

𝜋
Step 3: Determine the range of 𝑟 2 and 2𝜃. Since 0 ≤ 𝑟 ≤ 1, then 0 ≤ 𝑟 2 ≤ 1. Since 0 ≤ 𝜃 ≤ ,
2

then 0 ≤ 2𝜃 ≤ 𝜋 .

Step 4: Determine the description of 𝑓(𝑆) (in polar) and describe the image. 𝑓(𝑆) = {(𝑟, 𝜃): 0 ≤
𝑅 ≤ 1 and 0 ≤ 𝜃 ≤ 𝜋}. The image of the upper quarter circle under the mapping 𝑓(𝑧): = 𝑧 2 is
the upper half of the unit disk, or a semicircle.

Step 5: Sketch 𝑆 and 𝑓(𝑆).

z -plane w- Plane

Example 5: Let 𝑓(𝑧): = 𝑧 2 and 𝑇: = {(𝑥, 𝑦): 𝑦 = 1}. Sketch 𝑇 and 𝑓(𝑇) on different complex
planes.

Step 1: Express T in rectangular form. Since 𝑇 = {(𝑥, 𝑦) ∶ 𝑦 = 1}, we can represent any 𝑧 ∈ 𝑇
as 𝑧 = 𝑥 + 𝑖(1) = 𝑥 + 𝑖.
Step 2: Apply 𝑓(𝑧) to T. Substitute 𝑧 = 𝑥 + 𝑖 in = 𝑓(𝑧) = 𝑧 2 . This gives 𝑤 = 𝑥 2 + 2𝑥𝑖 − 1.

Step 3: Separate real and imaginary parts of 𝑓(𝑇). If 𝑤 = 𝑢 + 𝑖𝑣 = 𝑥 2 + 2𝑥𝑖 − 1, then 𝑢 =


𝑥 2 − 1 and 𝑣 = 2𝑥.

15
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

Step 4: Find the equation of 𝑓(𝑇) and describe the image in rectangular (u and v). Substitute 𝑥 =
𝑣
into 𝑢 = 𝑥 2 − 1. Then 𝑣 2 = 4(𝑢 + 1). This is the equation of a parabola opening to the right
2

with vertex at (−1,0).

Step 5 . Sketch 𝑇 and 𝑓(𝑇).

z -plane w -plane

Example 6: Let 𝑓(𝑧): = √𝑥 2 + 𝑦 2 − 𝑖𝑦 and 𝑆: = {(𝑥, 𝑦): 𝑥 2 + 𝑦 2 = 𝑐 2 , 𝑐 > 0}. Sketch 𝑆 and
𝑓(𝑆) on different complex planes.

Solution: Put u = √𝑥 2 + 𝑦 2 and 𝑣 = −𝑦. Together with the equation describing S, these
equations imply that 𝑢 = 𝑐 and v can assume any value in [−𝑐, 𝑐]. Thus the image f(S) is the
line segment connecting −𝑐 + 𝑐𝑖 to 𝑐 + 𝑐𝑖. //

Exercises 5.2:

16
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

5.3 Limit and Continuity

A. Limit

Definition 1: Let 𝑧0 and 𝑤0 be complex numbers and 𝑓: 𝑆 ⊆ ℂ → ℂ be a complex function. We say


the limit of 𝒇 as 𝒛 approaches to 𝒛𝟎 is 𝒘𝟎 , written

lim 𝑓(𝑧) = 𝑤0
𝑧→𝑧0

if and only if

i) There exists a 𝑛𝑏ℎ𝑑 𝑁 of 𝑧0 such that 𝑓(𝑧) is defined on 𝑁 (except possibly at z0


itself ) and
ii) for every positive number 𝜀, there exists a positive number 𝛿 satisfying the following:
If 𝑧 ∈ 𝑆 and 𝑧0 < |𝑧 − 𝑧0 | < 𝛿 , then |𝑓(𝑧) – 𝑤0 | < 𝜀.

Remark 1: We can rewrite the definition briefly as follows:


lim 𝑓(𝑧) = 𝑤0 ⇔ (∀𝜀 > 0) (∃𝛿 > 0)(0 < |𝑧 − 𝑧0 | < 𝛿 ⇒ |𝑓(𝑧) − 𝑤0 | < 𝜀.
𝑧→𝑧0

Remark 2: Geometrically, we may put the definition as follows:

lim 𝑓(𝑧) = 𝑤0 ⇔ (∀𝜀 > 0) (∃𝛿 > 0)𝑠𝑢𝑐ℎ 𝑡ℎ𝑎𝑡 𝑧 ∈ 𝑁𝛿 (𝑧0 )\{𝑧0 } ⇒ 𝑓(𝑧) ∈ 𝑁𝜀 (𝑤0 ) .
𝑧→𝑧0

17
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

Remark 3: Consider the open proposition:

𝑝(𝜀, 𝛿): (∀𝜀 > 0)(∃𝛿 > 0)(0 < |𝑧 − 𝑧0 | < 𝛿 ⇒ |𝑓(𝑧) − 𝑤0 | < 𝜀

Then to show lim 𝑓(𝑧) = 𝑤0 means that the quantified proposition 𝑝(𝜀, 𝛿) is true. In turn, to
𝑧→𝑧0

show that "𝑝(𝜀, 𝛿) is true"; we can proceed as per the following steps:

Step 1: Assume that a positive number 𝜀 is given.

Step 2: Find a positive number 𝛿 so that 𝑝(𝜀, 𝛿) is true. The challenge is this.

Example 1. Prove that lim 𝑧 2 = 𝑧02


𝑧↦ 𝑧0

Assume that 𝜀 is given. We must find a 𝛿 > 0 such that

0 < |𝑧 − 𝑧0 | < 𝛿 ⇒ |𝑧 2 − 𝑧02 | < 𝜀.

Step 1 (Preliminary analysis). Guess 𝛿

Observe that: |𝑧 2 − 𝑧02 | = |𝑧 + 𝑧0 ||𝑧 − 𝑧0 | ≤ (|𝑧| + |𝑧0 |)|𝑧 − 𝑧0 |…….(*)

We find an upper bound for the factor on left side.

Assume that 0 < 𝛿 ≤ 1 and |𝑧 − 𝑧0 | < 𝛿

⇒ |𝑧 − 𝑧0 | < 1

⇒ |𝑧| − |𝑧0 | < 1

⇒ |𝑧| < 1 + |𝑧0 |. (**)

From (*) and (**) and the assumption 0 < 𝛿 ≤ 1 , we get that |𝑧 2 − 𝑧02 | = (1 + 2|𝑧0 |) |𝑧 − 𝑧0 |

Now, it suffices to choose a positive number 𝛿 such that 𝛿 ≤ 1 and (1 + 2|𝑧0 |) |𝑧 − 𝑧0 | < 𝜀.

𝜀
This suggests we can choose 𝛿 = min {1, 1+2|𝑧 |}
0

Step 2 (Formal Proof): Show that the chosen value of 𝛿 satisfies the requirement. (Exercise)

18
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

Example 2: Use the 𝜀 − 𝛿 definition of limit to show

a) lim 𝑅𝑒(𝑧) = 𝑅𝑒(𝑧0 )


𝑧→𝑧0

b) lim 𝑧̅ = 𝑧̅0
𝑧→𝑧0

Remark 4 [Two- Path Method to show Non-existence of Limit]: If 𝑓(𝑧) approaches two
complex numbers 𝐿1 ≠ 𝐿2 with along two different paths towards 𝑧0 , then lim f(z)does not
𝑧 ↦𝑧0

exist.

𝑧̅
Example 3: Show that lim 𝑧 does not exist.
𝑧 ↦𝑧0

Theorem 1 [Relation between limit of a complex function and limits of its real and imaginary
Parts]:

Let

𝑓(𝑧) = 𝑢(𝑥, 𝑦) + 𝑖𝑣(𝑥, 𝑦), 𝑧0 = 𝑥0 + 𝑖𝑦0 and , w0 = 𝑢0 + 𝑖𝑣0 .

Then

lim 𝑓(𝑧) = 𝑤0 ⇔ lim 𝑢(𝑥, 𝑦) = 𝑢0 and lim 𝑣(𝑥, 𝑦) = 𝑣0


𝑧→𝑧0 (𝑥,𝑦)→(𝑥0 𝑦0 ) (𝑥,𝑦)→(𝑥0 𝑦0 )

Note: If the limit on left or the two limits on right exist, then:

lim 𝑓(𝑧) = lim 𝑢(𝑥, 𝑦) + i lim 𝑣(𝑥, 𝑦)


𝑧→𝑧0 (𝑥,𝑦)→(𝑥0 ,𝑦0 ) (𝑥,𝑦)→(𝑥0 ,𝑦0 )

Remark 5 [Rules for Limits of Complex Functions]: The rules for limits of combination and
composition of complex functions are identical with those of real functions.

19
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

B. Continuity

Definition 1: A complex function 𝑓 (𝑧) is said to be continuous at a complex constant 𝑧0


if

lim 𝑓(𝑧) = 𝑓(𝑧0 ) ,


𝑧→𝑍0

Note

1. By definition of a limit, continuity of 𝑓(𝑧) at 𝑧0 implies that that 𝑓(𝑧) is defined in some
neighbourhood of 𝑧0 .
2. 𝑓(𝑧) is said to be continuous in a domain if it is continuous at each point of this domain.
3. 𝑓(𝑧) continuous ⇐⇒ 𝑢(𝑥, 𝑦) and 𝑣(𝑥, 𝑦) continuous.

Example 4: Show that the principal square root function is not continuous at 𝑧 = − 1.

Exercises 5.3:

1.Find 𝑅𝑒(𝑓) and 𝐼𝑚(𝑓) where


a) 𝑓(𝑧) = 5𝑧 2 − 12𝑧 + 3 + 2𝑖 at 4 − 3𝑖
𝑧−2
b) 𝑓(𝑧) = 𝑧+2 at 8𝑖

2.Use the formal definition of limit to prove that lim (2𝑥 + 𝑖𝑦 2 ) = 4𝑖.
𝑧 ↦𝑧0

3.If lim 𝑓(𝑧) exists, show that this limit is unique.


𝑧→𝑧0

4.Prove the sum , difference , product and quotient rules for limits of complex functions.

20
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

5.4 Derivatives, Analyticity and CR Conditions

A. Derivatives and Analyticity

Definition 1: Let 𝑧0 be a complex constant and let 𝑤 = 𝑓(𝑧) be a function of a complex


variable whose domain contains a neighbourhood of 𝑧0 . The derivative of f at 𝒛𝟎 is a
complex number denoted by 𝑓′(𝑧0 ) and it is defined by:

𝑓(𝑧0 +∆𝑧)−𝑓(𝑧0 )
f ′ (z0 ) = lim ,
∆𝑧→0 ∆𝑧

𝑓(𝑧)−𝑓(𝑧0 )
Provided that the limit exists. Equivalently, we can write f ′ (z0 ) = lim 𝑧−𝑧0
.
𝑧→𝑧0

Z-plane W-plane

∆𝑓 = 𝑓(𝑧) − 𝑓(𝑧)

𝑓(𝑧)
𝑧 + ∆𝑧
𝑓(𝑧 + ∆𝑧)
∆𝑧 = 𝑧 − 𝑧0

Remark 1 (Higher Order differentiation for complex functions): The existence of one
complex derivative automatically implies the existence of infinitely many!

Remark 2 (Differentiation Rules): The differentiation rules for combination and composition
of functions in for real functions also apply to complex functions.

Remark 3: Certain properties of derivatives in real calculus (eg. The Mean Value
Theorem)doesn't carry over directly to complex calculus .

21
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

Definition 2: Let 𝑧0 be a complex number and S be a subset of the complex plane.


A complex function 𝑓 is said to be

a) differentiable at 𝒛𝟎 if f ′ (z0 ) exist.


b) analytic(holomorphic) at z0 if 𝑓′(𝑧) exists at every point z in a neighbourhood of z0.
c) analytic on S if it is analytic at every point in some open set containing S.
d) an analytic(holomorphic) function if it is analytic at every point in its domain.
e) an entire function if it is analytic at every complex number.

Example1: Use the definition of complex derivative to find 𝑓 ′ (𝑧) where 𝑓(𝑧) = z2.

𝑓(𝑧+ℎ)−𝑓(𝑧) (𝑧+ℎ)2 −𝑧 2
Solution: f ′ (z) = lim = lim = lim (2𝑧 + ℎ) = 2𝑧 .
ℎ→0 ℎ ℎ→0 ℎ ℎ→0

Example 2: Show that the complex function 𝑓(𝑧) = 𝑅𝑒(𝑧) is nowhere differentiable.

𝑓(𝑧+ℎ)−𝑓(𝑧) 𝑅𝑒(𝑧+ℎ)−𝑅𝑒(𝑧)
Solution: lim ℎ
= lim ℎ
. If h is real , the limit is 1 ; if h is a pure
ℎ→0 ℎ→0

imaginary number, then the limit is 0. Since the limit values (1 and 0) are different depending on
how h approaches 0, the limit of the difference quotient does not exist, and therefore f(z) is not
differentiable. Later on, we use another approach (CR conditions).

Example 3: Show that the complex function 𝑓(𝑧) = |𝑧|2 is differentiable only at 𝑧 = 0.
𝑓(𝑧+ℎ)−𝑓(𝑧) |𝑧+ℎ|2 −|𝑧|2
Solution: f ′ (z) = lim = lim
ℎ→0 ℎ ℎ→0 ℎ

Using complex conjugates, we can express the last limit as:

|𝑧 + ℎ|2 − |𝑧|2 (𝑧 + ℎ)(𝑧 + ℎ) − 𝑧𝑧 𝑧ℎ + 𝑧ℎ + ℎℎ


lim = lim = lim
ℎ→0 ℎ ℎ→0 ℎ ℎ→0 ℎ

= +𝑧 + ℎ + 𝑧 lim
ℎ→0 ℎ

22
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)


The last limit exists only at 𝑧 = 0 ( because if ℎ = 𝑟𝑒 𝑖𝜃 , then ℎ = 𝑟𝑒 −𝑖𝜃 and = 𝑒−2𝑖𝜃 . Thus

ℎ ℎ
lim depends on 𝜃). If z=0 , however, lim 𝑧 =0).
ℎ→0 ℎ ℎ→0 ℎ

Example 4: Show that 𝑓(𝑧) = 𝑧̅ is nowhere differentiable.

𝑓(𝑧+ℎ)−𝑓(𝑧) (𝑧+ℎ−𝑧 ℎ
Solution: f ′ (z) = lim = lim = lim
ℎ→0 ℎ ℎ→0 ℎ ℎ→0 ℎ

The last limit DNE (as explained in Example 3).

Theorem 1: If 𝑓(𝑧) is differentiable at z0 , then 𝑓(𝑧) continuous at z0.

A point at which 𝑓(𝑧) fails to be analytic is called a singular point or a singularity of 𝒇(𝒛).
There are several types of singularities: poles, removable, branches, isolated, essential, etc.

B. Cauchy-Riemann Conditions

Theorem 2[CR Conditions as Necessary Conditions for Differentiability]

Let 𝑓(𝑧) = 𝑢(𝑥, 𝑦) + 𝑖 𝑣(𝑥, 𝑦) be continuous in some neighborhood of a point 𝑧0 =


𝑥0 + 𝑖𝑦0 and differentiable at 𝑧0 itself. Then

a) The partial derivatives 𝑢𝑥 , 𝑢𝑦 , 𝑣𝑥 and 𝑣𝑦 exist at 𝑧0 and satisfy the following


conditions there:
𝑢𝑥 = 𝑣𝑦 and 𝑢𝑦 = −𝑣𝑥
(These conditions are called Cauchy-Riemann conditions)
b) The derivative of 𝑓 at 𝑧0 is given by
𝑓 ′ (𝑧0 ) = 𝑢𝑥 (𝑥0 , 𝑦0 ) + 𝑖𝑣𝑥 (𝑥0 , 𝑦0 = 𝑣𝑦 (𝑥0 , 𝑦0 ) − 𝑖𝑢𝑦 (𝑥0 , 𝑦0 ) .

23
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

Remark (Contrapositive of (a)): In view of this theorem; if a complex function 𝑓(𝑧) =


𝑢(𝑥, 𝑦) + 𝑖𝑣(𝑥, 𝑦) does not satisfy the Cauchy Riemann conditions at some point (𝑥, 𝑦), then it
is not differentiable there.

Proof: By assumption, the derivative at z exists. It is given by

𝑓(𝑧 + ∆𝑧) − 𝑓(𝑧)


f ′ (z) = lim
∆𝑧→0 ∆𝑧

𝑓(𝑧+∆𝑧 )−𝑓(𝑧)
We write ∆𝑧 = ∆𝑥 + 𝑖∆𝑦. Since f ′ (z) exists, so the approaches the same limit as
∆𝑧

∆𝑧 approaches 0 (i.e., 𝑧 + ∆𝑧 approaches z) along all paths. Let us consider the following two
paths (path I and path II).

𝑢(𝑥0 + ∆𝑥, 𝑦0 + ∆𝑦) + 𝑖𝑣(𝑥0 + ∆𝑥, 𝑦 + ∆𝑦) − [𝑢(𝑥0 , 𝑦0 ) + 𝑖𝑣(𝑥0 , 𝑦0 )]


f ′ (z0 ) = lim
∆𝑥+𝑖∆𝑦→0 ∆𝑥 + 𝑖∆𝑦

𝑢(𝑥0 + ∆𝑥, 𝑦0 + ∆𝑦) − 𝑢(𝑥0 , 𝑦0 )


= lim
∆𝑥+𝑖∆𝑦→0 ∆𝑥 + 𝑖∆𝑦
𝑣(𝑥0 + ∆𝑥, 𝑦0 + ∆𝑦) − 𝑣(𝑥0 , 𝑦0 )
+𝑖[ lim ]
∆𝑥+𝑖∆𝑦→0 ∆𝑥 + 𝑖∆𝑦

Path I: Along the path 𝑧 = 𝑖𝑦0 (that is = 𝑦0 ) ∆𝑦 = 0 and we get

f ′ (z0 ) = 𝑢𝑥 (𝑥0 , 𝑦0 ) + 𝑖𝑣𝑥 (𝑥0 , 𝑦0 ) (*)

Path II: Along the path 𝑧 = 𝑥0 ; ∆𝑥 = 0 and hence, we get

f ′ (z0 ) = 𝑣𝑦 (𝑥0 , 𝑦0 ) − 𝑖𝑢𝑦 (𝑥0 , 𝑦0 ) (**)

From (*) and (**), we get that:

f ′ (z) = 𝑢𝑥 + 𝑖𝑣𝑥 = 𝑢𝑦 − 𝑖𝑣𝑦 .


This implies 𝑢𝑥 = 𝑣𝑦 and 𝑢𝑦 = −𝑣𝑥 ∎

24
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

Theorem 3 [CR Conditions as Sufficient Conditions for Differentiability]

If two real valued continuous functions 𝑢(𝑥, 𝑦) and 𝑣(𝑥, 𝑦) of two real variables 𝑥 and 𝑦 have
continuous first partial derivatives that satisfy the Cauchy-Riemann equations in some domain
𝐷, then the complex function 𝑓(𝑧) = 𝑢(𝑥, 𝑦) + 𝑖 𝑣(𝑥, 𝑦) is analytic in a domain 𝐷.

Proof (Read):

Example 5: Use CR conditions to show that the following complex functions are nowhere
differentiable.

a) 𝑓(𝑧) = 𝑧̅
b) 𝑓(𝑧) = 𝑅𝑒(𝑧)
c) 𝑓(𝑧) = 𝐼𝑚(𝑧)

Example 6: Use CR conditions to show that the complex function 𝑓(𝑧) = |𝑧|2 is differentiable
only at 𝑧 = 0

Example 7: Find 𝑓′(𝑧) for

a) 𝑓(𝑧) = 𝑒 𝑥 (cos(𝑦) + 𝑖𝑠𝑖𝑛(𝑦))


b) 𝑓(𝑧) = 𝑙𝑛|𝑧| + 𝑖𝐴𝑟𝑔(𝑧)

Solution (a): 𝑓 ′ (𝑧) = 𝑢𝑥 + 𝑖𝑣𝑥 = 𝑒 𝑥 cos 𝑦 + 𝑖𝑒 𝑥 𝑠𝑖𝑛 𝑦 = 𝑒 𝑧 //

𝑦
Solution (b): Note that 𝑓(𝑧) = 𝑙𝑛|√(𝑥 + 1)2 + 𝑦 2 | + 𝑖 𝑡𝑎𝑛−1 (𝑥+1)

𝑥+1 1 −𝑦 1
Therefore, 𝑓 ′ (𝑧) = 𝑢𝑥 + 𝑖𝑣𝑥 = (𝑥+1)2+𝑦2 + 𝑖 ( 𝑦 2
) ((𝑥+1)2) = 𝑧 . Later, we define this
( ) + 1
𝑥+1

function as one of the different complex version of natural logarithmic functions. //

Remark: There is a version of CR conditions in polar Form. (See Exercises No., ).

C. Analytic Functions and Harmonic Functions

Example 8

25
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

a) 𝑓(𝑧) = |𝑧|2 is differentiable at 0 but it is not analytic at 0. In fact it is nowhere analytic.


1
b) 𝑓(𝑧) = 𝑧 is analytic at every complex number different from 0.

c) Every polynomial function is an analytic function.

Definition 3 (Laplace equation and Harmonic Function): Let Φ be a function of two


variables.

a) The PDE Φxx + Φyy = 0 is called Laplace equation.


b) A function u of several variables is said to be a harmonic function if
(i) it is a solution of a Laplace’s differential equation
(ii) it has continuous second partial derivatives

The theory of solutions of Laplace equation is called potential theory.

The solutions of Laplace equations are closely related to complex analytic functions

. Theorem 4[Analyticity and Laplace Equation]

The real and imaginary parts of an analytic function are harmonic functions, i.e. if 𝑓 = 𝑢 + 𝑖𝑣
is analytic in a domain D, then

i) 𝑢𝑥𝑥 + 𝑢𝑦𝑦 = 0 and 𝑣𝑥𝑥 + 𝑣𝑦𝑦 = 0 in 𝐷 and

ii) 𝑢 and 𝑣 have continuous second partial derivatives in 𝐷

D Definition 4 (harmonic conjugates): Two functions 𝑢 (𝑥, 𝑦) and 𝑣(𝑥, 𝑦) are said to be harmonic
conjugates or conjugate harmonic functions if the complex function 𝑓 = 𝑢 + 𝑖𝑣 is analytic.

Note

26
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

1. Every harmonic function has a harmonic conjugate. (The proof is given in the section on
Complex Line Integrals).
2. Given a harmonic function, CR conditions can be sued to find a harmonic function conjugate to
it.

Example 9: Let 𝑢(𝑥, 𝑦) = 𝑦 3 − 3𝑥 2 𝑦 .


a) Show that 𝑢 is harmonic.
b) Find a harmonic conjugate of 𝑢.

Exercises 5.4

1. If 𝑓(𝑧) is differentiable at 𝑧0 , show that 𝑓(𝑧) is continuous at 𝑧0 .

2. Derive the CR-equations in polar form.

3. Use polar form of CR-equations to show that the function


1
𝜃 𝜃
𝑓(𝑧) = 𝑧 2 =√𝑟[cos (2 ) + 𝑖𝑠𝑖𝑛 (2 )] (where 0 < 𝜃 < 2𝜋) is differentiable except at non-negative real

numbers.

4. Show that if

5.5 Elementary Complex Functions

In mathematics, functions that have established names are broadly classified into two types:
elementary functions and special functions. The term special function lacks a general formal
definition. Elementary functions encompass power functions, trigonometric functions, and
exponential functions, as well as those functions created from them through combinations,
compositions, and inverses. Recall that a power function is a function that can be expressed in
the form 𝑓(𝑥) = 𝑥 𝑎 , where 'a' is a real number. In particular the constant function and identity
function are power functions. The domain of a power function depends on its exponent. Other
elementary functions include polynomial functions, rational functions, logarithmic functions,
hyperbolic functions, etc.

5.5.1 The Complex Polynomial and Rational Functions

27
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

Definition 1: A function of the form

𝑃(𝑧) : = 𝑐𝑛 𝑧 𝑛 + … + 𝑐1 𝑧 + 𝑐0 ; (2)

where cn , …, c0 are complex constants and 𝑐𝑛 ≠ 0 is called a polynomial function in 𝑧 with


degree 𝑛. A function of the form

𝑝(𝑧)
𝑅(𝑧) = (3)
𝑞(𝑧)

where 𝑝, 𝑞 are polynomials in z and 𝑞(𝑧) ≠ 0 , is called a rational function.

Properties

1. Every complex polynomial function 𝑝𝑛 (𝑧) is an entire function.


2. The complex rational function 𝑅(𝑧) is analytic at every z in its domain.

5.5.2 The Complex Natural Exponential Functions

Definition 2: Let 𝑧 = 𝑥 + 𝑦𝑖. We denote and define the natural exponential


function of z by:

exp (𝑧) = 𝑒 𝑥 (𝑐𝑜𝑠 𝑦 + 𝑖 𝑠𝑖𝑛𝑦)

Notes:

1. Some books use the notation 𝑒 𝑧 as an alternative for 𝑒𝑥𝑝(𝑧). But there is a slight
difference of meaning between these two notations. See Definition 7. For now let as
accept the convention that they are identical.
2. This definition is inspired by the requirement that the infinite series expansion of
elementary calculus holds for e𝑖𝑦 .

Example-1 Find the value of e2+5i

Example-2 Find 𝑅𝑒 𝑤 and 𝐼𝑚 𝑤 where 𝑤 = 𝑒 −3𝑧 and 𝑧 = 𝑥 + 𝑦𝑖

28
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

Properties I
1. 𝑒 𝑧 ≠ 0 for every z.
2. 𝑒 𝑧1 𝑒 𝑧2 = 𝑒 𝑧1+𝑧2
3. 𝑒 𝑧1 ÷ 𝑒 𝑧2 = 𝑒 𝑧1− 𝑧2
4. (𝑒 𝑧 )𝑛 =𝑒 𝑧 𝑛
5. 𝑒 𝑧̅ = ̅̅̅
𝑒𝑧
6. 𝑒 𝑧1 = 𝑒 𝑧2 if and only if 𝑧2 = 𝑧1 + 2𝑘𝜋𝑖, 𝑘 = 0, ±1, ±2,…
7. 𝑒 𝑧 is a periodic function with period 2πi.

Proof: Let z1 = x1 + y1i and z2 = x2 + y2i

𝑒 𝑧1 𝑒 𝑧2 = 𝑒 𝑥1 + 𝑖𝑦1 𝑒 𝑥2 +𝑖𝑦2
= 𝑒 𝑥1 [𝑐𝑜𝑠𝑥1 +𝑠𝑖𝑛𝑦1 ] 𝑒 𝑥2 [𝑐𝑜𝑠𝑥2 +𝑠𝑖𝑛𝑦2 ]
= 𝑒 𝑥1 𝑒 𝑥2 [𝑐𝑜𝑠(𝑥1 +𝑥2 )+ i𝑠𝑖𝑛(𝑦1 + 𝑦2 )]
= 𝑒 𝑥1+𝑖𝑦1+𝑥2+𝑖𝑦2
= 𝑒 𝑧1+ 𝑧2

Properties II /Analyticity /

𝑑
1. (𝑒 𝑧 ) = 𝑒 𝑧
𝑑𝑧

2. 𝑒 𝑧 is an entire function.

5.5.3 The Complex Trigonometric Functions

29
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

Definition 3:

𝑒 𝑖𝑧 +𝑒 −𝑖𝑧
i) 𝑐𝑜𝑠 𝑧 =
2
𝑒 𝑖𝑧 − 𝑒 −𝑖𝑧
ii) 𝑠𝑖𝑛 𝑧 = 2𝑖
sin 𝑧
iii) 𝑡𝑎𝑛 𝑧 = cos 𝑧
1
iv) 𝑐𝑠𝑐 𝑧 = sin 𝑧
1
v) 𝑠𝑒𝑐 𝑧 =
cos 𝑧
1
vi) 𝑐𝑜𝑡 𝑧 =
tan 𝑧

Caution: The denominator in the definition of 𝑠𝑖𝑛 𝑧 is 2𝑖 ; not just 2.

Example-3

a) Evaluate 𝑐𝑜𝑠( 𝑖) and 𝑐𝑜𝑠 (𝜋).


b) Find the real part and the imaginary part of 𝑡𝑎𝑛 𝑧.

Properties I [Domain, Range, Periodicity and Boundedness)

1. Domain of Cos(z)=Domain of sin (z) =ℂ


2. Range of 𝑐𝑜𝑠(𝑧) = range of 𝑠𝑖𝑛(𝑧) = ℂ
3. in z is periodic with period 2π.
4. 𝑐𝑜𝑠 𝑧 is periodic with period 2π.
5. tan z is periodic with period π.
6. sin z and 𝑐𝑜𝑠 𝑧 are unbounded.

30
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

Properties II [Some Trig Identities]

1. sin 2 z + cos2 z = 1
2. sin (z1± z2) = sin z1 cos z2 ± sin z2 𝑐𝑜𝑠 z1
3. 𝑐𝑜𝑠 (z1± z2) = cos z1 cos z2 ∓ sin z1 sin z2
𝑡𝑎𝑛 𝑧 ±tan 𝑧2
4. t𝑎𝑛(z1± z2) = 1∓tan1𝑧
1 tan 𝑧2

5. 𝑠𝑖𝑛 (−𝑧) = − 𝑠𝑖𝑛 𝑧


6. 𝑐𝑜𝑠 (−𝑧) = 𝑐𝑜𝑠 𝑧
7. 𝑡𝑎𝑛 (−𝑧) = − 𝑡𝑎𝑛 𝑧

Proof (Exercise)

Properties III/Differentiability, analyticity, entirety/

1. sin 𝑧 and 𝑐𝑜𝑠 𝑧 are entire functions. The rest are analytic at all points in their domain.
𝑑
2. (𝑠𝑖𝑛 𝑧) = 𝑐𝑜𝑠 𝑧
𝑑𝑧
𝑑
3. 𝑑𝑧
(𝑐𝑜𝑠 𝑧) = − 𝑠𝑖𝑛 𝑧
𝑑
4. (tan z) = sec2 z
𝑑𝑧
𝑑
5. (csc z) = - csc2 z
𝑑𝑧
𝑑
6. 𝑑𝑧
(𝑠𝑒𝑐 𝑧) = 𝑠𝑒𝑐 𝑧 𝑡𝑎𝑛 𝑧
𝑑
7. (𝑐𝑜𝑡 𝑧) = −𝑐𝑠𝑐 𝑧 𝑐𝑜𝑡 𝑧
𝑑𝑧

Question: Where does the analogy with the real case fails?

31
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

5.5.4 The Complex Hyperbolic Functions

Definition 4: Let 𝑧 be a complex variable. We define

𝑒 𝑧 +𝑒 −𝑧
a) 𝑐𝑜𝑠ℎ 𝑧 =
2
𝑒 𝑧 − 𝑒 −𝑧
b) 𝑠𝑖𝑛ℎ 𝑧 =
2
sinh 𝑧
c) 𝑡𝑎𝑛ℎ 𝑧 =
cosh 𝑧
1
d) 𝑐𝑠ℎ𝑐 𝑧 = sinh 𝑧
1
e) 𝑠𝑒𝑐ℎ 𝑧 = cosh 𝑧
1
f) 𝑐𝑜𝑡ℎ 𝑧 = tanh 𝑧

Example 4: Evaluate 𝑐𝑜𝑠ℎ 𝑖 and 𝑐𝑜𝑠ℎ(1)

Properties I (Domain, Range, Periodicity)

1. Domain of Cosh(z)=Domain of 𝑠𝑖𝑛ℎ (𝑧) =ℂ


2. Range of 𝑐𝑜𝑠ℎ(𝑧) = ℂ=Range ℎ(𝑧) . Thus they are unbounded.
3. 𝑠𝑖𝑛ℎ 𝑧 is periodic with period 2π.
4. 𝑠𝑖𝑛ℎ 𝑧 is periodic with period 2𝜋𝑖.
5. 𝑜𝑠ℎ 𝑧 is periodic with period 2𝜋𝑖.
6. 𝑡𝑎𝑛ℎ 𝑧 is periodic with period 𝜋𝑖.

Properties II: Some Hyperbolic identities

1. cosh2 z- 𝑠𝑖𝑛ℎ 2 z = 1
2. 𝑠𝑖𝑛ℎ (z1± z2) = 𝑠𝑖𝑛ℎ z1 cosh z2 ± sinh z2 cosh z1
3. cosh (z1± z2) = cosh z1 cosh z2 ± sinh z1 sinh z2
𝑡𝑎𝑛ℎ 𝑧 ±tanh 𝑧2
4. 𝑡𝑎𝑛ℎ(z1± z2) = 1±tanh1𝑧
1 tanh 𝑧2

5. 𝑐𝑜𝑠ℎ (−𝑧) = 𝑐𝑜𝑠ℎ 𝑧


6. 𝑡𝑎𝑛ℎ (−𝑧) = − 𝑡𝑎𝑛ℎ 𝑧
7. 𝑠𝑖𝑛(𝑖𝑧) = 𝑖 𝑠𝑖𝑛ℎ 𝑧
8. 𝑐𝑜𝑠(𝑖𝑧) = 𝑖 𝑐𝑜𝑠ℎ 𝑧

32
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

Properties III: Differentiability, analyticity, entirety/

𝑑
1. 𝑑𝑧
(𝑠𝑖𝑛ℎ 𝑧) = 𝑐𝑜𝑠ℎ 𝑧
𝑑
2. (𝑐𝑜𝑠ℎ 𝑧) = − 𝑠𝑖𝑛ℎ 𝑧
𝑑𝑧
𝑑
3. (tanh z) = sech2 z
𝑑𝑧
𝑑
4. (csch z) = − csch2 z
𝑑𝑧
𝑑
5. 𝑑𝑧
(sech z) = 𝑠𝑒𝑐ℎ 𝑧 𝑡𝑎𝑛ℎ 𝑧
𝑑
6. (𝑐𝑜𝑡ℎ 𝑧) = −𝑐𝑠𝑐ℎ 𝑧 𝑐𝑜𝑡ℎ 𝑧
𝑑𝑧

7. 𝑠𝑖𝑛ℎ 𝑧 and 𝑐𝑜𝑠ℎ 𝑧 are entire functions. The rest are analytic at all points in their domain.

5.5.5 Complex Natural Logarithmic "Function"

Definition 5: Let 𝑧𝜖ℂ \ {0}. The natural logarithm of 𝑧 is denoted by 𝑙𝑛 𝑧 or 𝑙𝑜𝑔 𝑧 (in Churchill) and it is
defined as:

𝑙𝑛 𝑧=w ⟺ 𝑒𝑤 = z

Caution: If 𝑟 is a positive real number the real natural logarithm of 𝑟 and the complex natural logarithm
of 𝑟 are different. To avoid confusion, we denote the real natural logarithm of 𝑟 by Ln r or 𝐿𝑜𝑔 ( 𝑟 ).

Proposition [Re(ln z) and Im(ln z)]: Let 𝑧 = 𝑟(𝑐𝑜𝑠 𝜃 + 𝑖𝑠𝑖𝑛𝜃) 𝜖 ℂ.

Then, 𝑙𝑛 𝑧 = 𝐿𝑛|𝑧| + 𝑖(𝜃 + 2𝑛𝜋) , 𝑛 𝜖 ℤ.

Proof: Let 𝑤 = 𝑙𝑛 𝑧, 𝑧 = 𝑟[cos(𝜃) + 𝑖𝑠𝑖𝑛(𝜃)]𝑎𝑛𝑑 𝑤 = 𝑢 + 𝑖𝑣.

⇒ 𝑟[cos(𝜃) + 𝑖𝑠𝑖𝑛(𝜃)] = 𝑧 = 𝑒 𝑤 = 𝑒 𝑢+𝑖𝑣 = 𝑒 𝑢 [cos 𝑣 + 𝑖𝑠𝑖𝑛 𝑣].

⇒ 𝑟 = 𝑒 𝑢 𝑎𝑛𝑑 𝑣 = 𝜃 + 2𝑛𝜋 where n is any integer.

33
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

⇒ 𝑢 = 𝐿𝑛(𝑧) 𝑎𝑛𝑑 𝑣 = 𝜃 + 2𝑛𝜋 where n is any integer.

Therefore. 𝑙𝑛 𝑧 = 𝐿𝑛|𝑧| + 𝑖(𝜃 + 2𝑛𝜋) , 𝑛 𝜖 ℤ. ∎

Remark 1 (multivalued-ness of 𝒍𝒏(𝒛)) : The complex logarithm 𝑓(𝑧) = 𝑙𝑛(𝑧) is a multi-


valued function, which can be formally written as: 𝑙𝑛(𝑧) = {𝑤 ∈ 𝐶: 𝑒 𝑤 = 𝑧}. When z is

nonzero, it can be written as 𝑙𝑛(𝑧) = 𝐿𝑛|𝑧| + 𝑖𝑎𝑟𝑔(𝑧) where 𝑎𝑟𝑔(𝑧): = {𝜃 ∈ 𝑅: 𝑐𝑜𝑠𝜃 +


𝑧
𝑖𝑠𝑖𝑛𝜃 = | |}.
𝑧

Remark 2(Analogies with Certain Inverse Functions and Implicit Functions): Recall that
functions such as 𝑦 = 𝑥 2 and 𝑦 = 𝑠𝑖𝑛(𝑥) do not have inverse functions; because they are not
one to one. But we can define their inverse functions by restricting their domain or range. On the
other hand, an implicit equation 𝐹(𝑥, 𝑦) = 0 can define multiple functions, but can be made to
define a single function by restricting its x-values or y-values.

Similarly the multivalued function 𝑦 = ln (𝑧) can be made single valued by restricting its
domain/range (eg. by assuming 𝛼 < 𝜃 ≤ 𝛼 + 2𝜋, where 𝛼 is any fixed angular measure in
radians). See the next definition and remark. In particular if 𝛼 = 𝜋, we get the following:

Definition 6: The function 𝑓: ℂ ∖ (−∞, 0] → ℂ given by 𝑓 (𝑧) = 𝑙𝑛|𝑧| + 𝑖𝐴𝑟𝑔(𝑧) is called the principal
value or principal branch of 𝑙𝑛(𝑧). We denote this function by 𝐿𝑛(𝑧).

Thus 𝐿𝑛(𝑧) = 𝑙𝑛|𝑧| + 𝑖𝜃, −𝜋 < 𝜃 < 𝜋

Caution: Notice the difference between 𝑙𝑛(𝑧) (complex natural log), 𝐿𝑛(𝑧) (principal complex natural
log) and Ln|𝑧| (real natural log).

Remark 3 (Branches, Principal Branch, Branch Cuts and Branch Point of 𝒇(𝒛) = 𝒍𝒏(𝒛))

The principal argument 𝐴𝑟𝑔(𝑧), is discontinuous along the negative real axis. If we approach a
point on the negative real axis from above (the upper half-plane), 𝐴𝑟𝑔(𝑧) approaches π, while
if we approach it from below (the lower half-plane), 𝐴𝑟𝑔(𝑧) approaches -π. That is why it is

34
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

excluded from the domain of 𝐿𝑛(𝑧). The non-positive real axis is called the branch cut of
𝐿𝑛(𝑧).

Definition 7:

a) A branch of a multivalued function 𝑓 is any single valued function 𝐹 that is analytic in


some domain (connected open region) at which the value 𝐹(𝑧) is one of the values of 𝑓.
The assumption of analyticity prevents us to make random selection of 𝐹.
b) A curve along which a branch of a multivalued function f is discontinuous is called a
branch cut of 𝑓. Usually, the cuts are rays or lines.
c) Any point common to all branch cuts of 𝑓 is called a branch point.

Example 5:

1. 𝐿𝑛 1 = 0
2. 𝑙𝑛 1 = 2𝑛𝜋𝑖
3. 𝑙𝑛(−1) = (2𝑛 + 1) 𝜋𝑖
4. 𝐿𝑛(−1) = 𝜋𝑖
𝜋
5. 𝐿𝑛(𝑖) = 2 𝑖
(2𝑛+1)𝜋
6. 𝑙𝑛(𝑖) = 𝑖
2
𝜋
7. 𝐿𝑛 − (𝑖) = − 𝑖
2
1 𝜋
8. 𝑙𝑛 (1 + 𝑖 ) =2 𝐿𝑛2 + 𝑖 ( 4 + 2𝑛𝜋), 𝑛 ∈ ℤ.

9. Determine the values of 𝑙𝑛(1 − 𝑖). What is the principal value?


10. Find all values of 𝑧 for which (a) e3z = 1, (b) e4z= i.

35
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

Properties

II. Domain and Range

a) The domain of 𝑙𝑛(𝑧) is ℂ \{0}.

b) The range of 𝑙𝑛(𝑧) is ℂ .

II. Laws of Logarithm (Some Logarithmic Identities)

c) 𝑙𝑛(𝑧1 𝑧2 ) = 𝑙𝑛 𝑧1 + 𝑙𝑛 𝑧2 (that is, the set of values on left and right are equal).
d) 𝑙𝑛(𝑧1 ÷ 𝑧2 ) = 𝑙𝑛 𝑧1 − 𝑙𝑛 𝑧2 (the sets of values on left and right are the same.)
e) 𝑙𝑛 (𝑧 𝑚 ) = 𝑚 𝑙𝑛 𝑧
f) 𝑒𝑥𝑝(𝑙𝑛 𝑧) = 𝑧
g) 𝑙𝑛(exp 𝑧) = 𝑧 + 2𝜋𝑘𝑖, 𝑘 ∈ ℤ

III. Differentiability and Analyticity(Holomorphicity)

𝑑 1
𝑎) 𝑑𝑧
(Ln z) = 𝑧

c) b) For every fixed integer n, the single valued function 𝑙𝑛 𝑧: = 𝐿𝑛 𝑧 + 2𝑛𝜋𝑖 defines a function and
each function is analytic except at z = 0 and at each point on negative real axis

Note : 𝐿𝑛(𝑧1 𝑧2 ) ≠ 𝐿𝑛(𝑧1 ) + 𝐿𝑛(𝑧2 ).

5.5.6 General Power, General Exponential and General Logarithmic Functions

In school mathematics, the expression “𝑏 𝑥 ” was defined sequentially in cases: first when x is a
positive integer, then when 𝑥 is any integer, then when x is a special fraction, then when x is any
fraction and finally when 𝑥 is an irrational number. We now generalize it to the case of complex
exponents.

36
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

Definition 8 (complex exponent):

Let 𝑧, 𝑤 ∈ ℂ and 𝑧 ≠ 0. We define:

𝑧 𝑐 = 𝑒 𝑐 𝑙𝑛 𝑧

Definition 9 (general complex power function: Let 𝑐 be a complex constant and 𝑧 be a


complex variable. The function 𝑓(𝑧) = 𝑧 𝑐 (𝑧 ≠ 0) is called a complex power function.

Since 𝑙𝑛 𝑧 is multivalued so 𝑧 𝑐 is multivalued, in general. In particular, ecLnz is called the


principal value of 𝑧 𝑐 .

If 𝑐 is an integer, “𝑧 𝑐 ” is single valued. If 𝑐 is a rational number, it is multivalued. If 𝑐 is an


irrational number or a genuinely complex number, it is infinitely many valued.

Definition 10: The number 𝑃𝑉(𝑧 𝑐 ) ≔ 𝑒 𝑐 𝐿𝑛(𝑧) is called the principal value of 𝑧 𝑐 .

Note: Since 𝑙𝑛(𝑧) is multivalued, so is 𝑧 𝑐 .

Example 6: Find

a) 𝒊−𝟐𝒊 = 𝑒𝑥𝑝(−𝜋𝑖 + 4𝑛𝜋), where 𝑛 is an integer.


𝜋
b) Find the principal value of −𝑖 𝑖 . Answer: 𝑒𝑥𝑝( 2 )
2
2
c) Find the principal value of 𝑧 3 . Answer: 𝑒𝑥𝑝(3 𝐿𝑛(𝑧))

Properties

1. When 𝑐 is rational, the above definition reduces to the definition for rational exponents (given
in 5.1).
1
2. 𝑧 −𝑐 = 𝑧 𝑐 (the set of values on left and right are equal.)

𝑑 𝑐
(𝑧 ) = 𝑐𝑧 𝑐−1
𝑑𝑧

37
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

Definition 11(General complex exponential function):

Let 𝑐 be a complex constant with 𝑐 ≠ 0 and z be a complex variable. The function 𝑓(𝑧) = 𝑐 𝑧 .is
called a complex exponential function.

Remark (Inconsistency between the definition of 𝒆𝒛 via Definition 2 and Definition 7): If we
put 𝑧 = 𝑒 in Definition 7, it reduces to 𝑒 𝑧 . So it becomes multivalued; where as in Definition 2 it
was single valued. Paradox!!! To correct this confusion, we use 𝑒𝑥𝑝(𝑧) to denote the single
valued variant and 𝑒 𝑧 the multivalued variant of natural exponential function.

Example 7:

1. The principal value of (2i) 1/2 is 1 + i.


−𝜋
2. The principal value of 𝑖 𝑖 is 𝑒 2 .
3. Find the principal value of (−5)2−4𝑖
4. Find (a) Re(1 − 𝑖)1+𝑖 ) (b) |(−𝑖)−𝑖

Definition 12(General complex logarithmic function): For complex numbers 𝑏 and 𝑧


with 𝑏 ≠ 0,1 , we define: log b 𝑧 by:

ln (𝑧)
log b 𝑧 =
ln (𝑏)

The multivalued function 𝑓(𝑧) = 𝑙𝑜𝑔𝑐 𝑧 is called a complex logarithmic function.

Example 8:

38
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

5.5.7 Inverse Trigonometric Functions and Inverse Hyperbolic Functions

(See Exercises 5.4)

Exercises 5.5

39
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

CHAPTER 6: COMPLEX INTEGRALS

Complex integration is important for several reasons. Complex integrations allow us to


evaluate certain real integrals that are difficult or impossible to evaluate using the usual
integration techniques of calculus. Secondly, they allow us to understand the behaviour of
analytic functions. Thirdly, they have numerous applications in various fields, including: Fluid
dynamics, Electromagnetic field theory, Aerodynamics and Quantum mechanics: (Source:
Kreyszig and AI)

6.1 Derivatives and Integrals of Complex Valued Functions of Real Variables


(Read: from Churchill).

We can define the limit, continuity, derivative and integral of a complex valued function of a
complex variable (a function of the form 𝑤(𝑡) = 𝑥(𝑡) + 𝑖𝑦(𝑡) (𝑎 ≤ 𝑡 ≤ 𝑏) ) in the same
manner as those of the vector functions 𝑟(𝑡) = 𝑥(𝑡)𝑖 + 𝑦(𝑡)𝑗 (𝑎 ≤ 𝑡 ≤ 𝑏).

Definition 1: Let 𝑤(𝑡) = 𝑥(𝑡) + 𝑖𝑦(𝑡) (𝑎 ≤ 𝑡 ≤ 𝑏). The derivative of w is defined


by:𝑤 ′ (𝑡) = 𝑥 ′ (𝑡) + 𝑖𝑦′(𝑡).

Figure 1

Example 1:

40
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

Many of the ordinary derivative rules carry over to this derivative (Eg, sum rule, difference rule,
product rule and Quotient Rule and Chain Rule).. But some properties fail to hold .

Example 2: Show that the MVT fails for the derivative of w(t).

Solution: Consider the function 𝑤(𝑡) = 𝑒 𝑖𝑡 , 0 ≤ 𝑡 ≤ 2𝜋 . Suppose there exist a number


𝑐 between 0 and 2𝜋 such that

𝑤(2𝜋) − 𝑤(0)
𝑤 ′ (𝑐) =
2𝜋 − 0

This implies – 𝑠𝑖𝑛 ( 𝑐 ) + 𝑖 𝑐𝑜𝑠 ( 𝑐 ) = 0. As the LHS can never be zero, the last statement is a
contradiction.

Definition 2: Let 𝑤(𝑡) = 𝑥(𝑡) + 𝑖𝑦(𝑡) (𝑎 ≤ 𝑡 ≤ 𝑏) where 𝑢 and 𝑣 are real valued
functions. We define:

𝑏 𝑏 𝑏
∫ 𝑤(𝑡)𝑑𝑡 = ∫ 𝑢(𝑡)𝑑𝑡 + 𝑖 ∫ 𝑓(𝑡)𝑑𝑡
𝑎 𝑎 𝑎

provided that the real integrals on RHS exist.

The definition can be extended to improper integrals.

Example 3:

Many of the ordinary integration rules carry over to the above integral, but some properties fail

Example 4: Show that the Mean Value Theorem of Integrals fails for the definite integral of
𝒘(𝒕) = 𝒙(𝒕) + 𝒊𝒚(𝒕).

6.2 Line Integral in the Complex Plane; Basic Properties

Complex line integrals extend the concepts of line integrals from vector calculus to complex
functions. Complex line integrals share many concepts and properties with and multivariable
calculus line integrals but they are not exactly the same. The results of complex line integrals
are complex numbers, while those of multivariable line integrals are real numbers.

41
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

Recall that a plane curve 𝒞 is a the trace of a continuous vector function 𝑟(𝑡) =
〈𝑥(𝑡), 𝑦(𝑡)〉, 𝑎 ≤ 𝑡 ≤ 𝑏. We may parameterize a curve in the complex plane by an equation of
the form:

𝑧(𝑡) = 𝑥(𝑡) + 𝑖𝑦(𝑡) (𝑎 ≤ 𝑡 ≤ 𝑏)

Recall also the definitions of the different types of curves.

A curve C is said to be a :

a) closed curve if it starts and ends at the same point and if it encloses a region (an area).
b) Simple curve if it does not intersect (touch or cross) itself, except possibly at its
endpoints.
c) simple closed curve if it is a closed curve and does not intersect itself except at end points.
More precisely if it is closed and its parameterization 𝑟(𝑡) = 〈𝑥(𝑡), 𝑦(𝑡)〉, 𝑎 ≤ 𝑡 ≤ 𝑏 is one to
one on [𝑎, 𝑏).
d) Smooth curve if it (its parameterization) has a continuous nonzero derivative.
e) piecewise smooth curve if it consist of finitely many smooth curves that are joined end to end.

Definition 1: Let 𝒞 be a piecewise smooth curve in a complex plane with orientation from 𝑧1 =
𝑓(𝑎) to 𝑧2 = 𝑓(𝑏) . Let 𝑓(𝑧) = 𝑢(𝑧) + 𝑖𝑦(𝑧) be a complex function defined on 𝒞. Partition the
curve as shown in the figure and for each k, choose a point 𝑧𝑘∗ on C between 𝑧𝑘−1 and 𝑧𝑘 and let
∆𝑧𝑘 ≔ 𝑧𝑘 − 𝑧𝑘−1. Then we denote and define the line integral of 𝑓 from 𝑧1 𝑡𝑜 𝑧2 𝑎𝑙𝑜𝑛𝑔 𝒞 by:

∫ 𝑓(𝑧)𝑑𝑧 = lim ∑ 𝑓(𝑧𝑘∗ )∆𝑧𝑘


𝒞 𝑛→∞
𝑘=1

Provided that |∆𝑧| ≔ max {|∆𝑧1 |, … , |∆𝑧𝑛 |} approaches zero when n approaches infinity.

If C is closed we denote the integral by ∮𝐶 𝑓(𝑧)𝑑𝑧.

42
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

Figure 2

Evaluation I: Let 𝒞 be parameterized by 𝑧(𝑡) = 𝑥(𝑡) + 𝑖𝑦(𝑡) (𝑎 ≤ 𝑡 ≤ 𝑏). Then

𝑏
∫ 𝑓(𝑧)𝑑𝑧 = ∫ 𝑓(𝑧(𝑡))𝑧 ′ (𝑡)𝑑𝑡
𝒞 𝑎

Evaluation II (Conversion to Vector Calculus Line Integrals) : If 𝑓 = 𝑢 + 𝑖𝑣 and 𝑑𝑧 = 𝑑𝑥 + 𝑖𝑑𝑦,


then

∫𝒞 𝑓(𝑧)𝑑𝑧 = ∫𝒞 𝑢𝑑𝑥 − 𝑣𝑑𝑦 + 𝑖 ∫𝒞 𝑢𝑑𝑦 + 𝑣𝑑𝑥

The following are basic properties of complex line integrals that directly follow from the
definition.
Thus the complex line integral is equivalent to two real line integrals.
Properties:
1. Linearity: ∫𝒞[𝑘1 𝑓1 (𝑧) + 𝑘2 𝑓2 (𝑧)]𝑑𝑧 = ∫𝒞 𝑘1 𝑓1 (𝑧)𝑑𝑧 + ∫𝒞 𝑘2 𝑓2 (𝑧)𝑑𝑧
2. Sense reversal (Orientation Dependence): ∫− 𝒞 𝑓(𝑧)𝑑𝑧 = − ∫𝒞 𝑓(𝑧)𝑑𝑧 ( C & -C are same
curve with opposite sense)
3. Path Partitioning: ∫𝒞[𝑘1 𝑓1 (𝑧) + 𝑘2 𝑓2 (𝑧)]𝑑𝑧 = ∫𝒞 𝑓(𝑧)𝑑𝑧 + ∫𝒞 𝑓(𝑧)𝑑𝑧. (𝒞 = 𝒞1 ∪ 𝒞2 )
1 2

4. Parameter Independence: The value of ∫𝒞 𝑓(𝑧)𝑑𝑧 is invariant under change of


parameterization of 𝒞.

43
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

5. Path Dependence: In general ∫𝒞 𝑓(𝑧)𝑑𝑧 depends not only on endpoints of 𝒞 but on 𝒞.

1
Example 1: Evaluate ∫𝒞 𝑑𝑧, where 𝒞 is:
𝑧

a) the upper half of the circle |𝑧| = 1 with anticlockwise orientation. Ans. 𝜋𝑖
b) the lower half of the circle |𝑧| = 1 with anticlockwise orientation. Ans. 𝜋𝑖
c) the circle |𝑧| = 1 with anticlockwise orientation. Ans. 2𝜋𝑖

Example 2: Show that for any two points z_1 and z_2 and any path C joining z_1 to z_2,

𝑧22 𝑧12
∫ 𝑧𝑑𝑧 = −
𝒞 2 2

Example 3

44
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

(non-analytic integrand & simple closed path): Integrate 𝑓(𝑧) = 𝑅𝑒(𝑧) from 0 to 1 + 2𝑖 along
(a) 𝐶1 union 𝐶2 (b) 𝐶 ∗ .

Figure 3

Theorem 1 [ML-Inequality]: If length(𝐶) = 𝐿 and |𝑓(𝑧)| ≤ 𝑀, then

|∫𝐶 𝑓(𝑧)𝑑𝑧| ≤ 𝑀𝐿.

Proof:

Method I (via the definition of Line Integrals): Partition C as in Definition 1. Then, by the so
called generalized triangle inequality (for complex numbers) and the assumption we get that
|𝑓(𝑧)| ≤ 𝑀 , we have:

|∑𝑛𝑘=1 𝑓(𝑧𝑘∗ )∆𝑧𝑘 | ≤ ∑𝑛𝑘=1|𝑓(𝑧𝑘∗ )∆𝑧𝑘 | = ∑𝑛𝑘=1|𝑓(𝑧𝑘∗ )||∆𝑧𝑘 | ≤ 𝑀 ∑𝑛𝑘=1|∆𝑧𝑘 |.

⟹ |∑𝑛𝑘=1 𝑓(𝑧𝑘∗ )∆𝑧𝑘 | ≤ 𝑀 ∑𝑛𝑘=1|∆𝑧𝑘 |. The sum on the RHS represents the length 𝐿𝑛 of the line
segments formed by joining the points 𝑧0 , 𝑧1 , …, 𝑧𝑛 . When 𝑛 approaches infinity, 𝐿𝑛
approaches 𝐿. Thus by taking limits of both sides we get |∫𝐶 𝑓(𝑧)𝑑𝑧| ≤ 𝑀𝐿.

Method II (Read from Churchill). ∎

𝑧−2 4𝜋
Example 5 (An Illustration of ML-inequality): Show that |∫𝐶 𝑧 4+1 𝑑𝑧| ≤ 15
.

45
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

6.3 Cauchy Integral Theorem and Cauchy-Goursat Theorem


6.3.1 Cauchy’s Integral Theorem

In Section 6.2 we defined certain types of curves. We begin this section by defining some
additional point sets. A point set is a set where each element of the set is a geometric point.

Some Point Sets: in the complex pane A point set D in the complex plane is said to be

- open if for every point P of D there exists some positive number R such that the open
disk centred at P and having radius R is contained in D. .
- connected if any two of its points can be joined by a chain of finitely many straight-line
segments all of whose points belong to D.
- simply connected if it is connected & every simple closed path in D encloses only points of D.
- multiply connected if it is connected but not simply connected.
- doubly connected if its boundary consists two disjoint closed curves(if it has one hole).
- Triply connected if its boundary consists three disjoint closed curves(if it has two holes).
- Quadruply connected if its boundary contains four disjoint closed curves (if it has three holes).
- P-fold connected if its boundary consists p closed curves with no common points.
- a domain if it is open and connected set .( caution: domain of definition of a function is
not necessarily domain in the sense defined above.)
- a region if it the union of a domain and some/all of its end points.

46
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

Figure 4

Theorem 1 [Cauchy’s Integral Theorem]: Let 𝐶 be a simple closed contour.

If

(i) 𝑓 (𝑧) is analytic at all points interior to and on C and


(ii) 𝑓’(𝑧) is continuous at all points on C and interior to C,

then ∮𝐶 𝑓(𝑧)𝑑𝑧 = 0.

In short, Cauchy’s Theorem states that the integral of the function around a closed path is zero
if the function is analytic (holomorphic) and continuously differentiable within and on the
path.
A simple closed path is sometimes called a contour and an integral over such a path a contour integral.

Proof: We use Green’s Theorem. Recall that if 𝑀(𝑥, 𝑦) and 𝑁(𝑥, 𝑦) have continuous partial
derivatives on a closed bounded plane region R, then:

∫𝑀𝑑𝑥 + 𝑁𝑑𝑦 = ∬ (𝑁𝑥 − 𝑀𝑦 )𝑑𝐴


𝐶 𝑅

47
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

In our case, if 𝑓(𝑥, 𝑦) = 𝑢(𝑥, 𝑦) + 𝑖𝑣(𝑥, 𝑦), then ∫𝒞 𝑓(𝑧)𝑑𝑧 = ∫𝒞 𝑢𝑑𝑥 − 𝑣𝑑𝑦 + 𝑖 ∫𝒞 𝑢𝑑𝑦 + 𝑣𝑑𝑥.

Since 𝑓 ′ = 𝑢𝑥 + 𝑖𝑢𝑦 = 𝑣𝑦 − 𝑖𝑣𝑥 is continuous, so the partial derivatives of u and v are


continuous. By Green’s Theorem applied to the two integrals (since f’ is continuous on RHS we
get that:

∫𝒞 𝑢𝑑𝑥 − 𝑣𝑑𝑦 = ∬𝑅[−𝑣𝑥 − 𝑢𝑦 ]𝑑𝐴 and ∫𝒞 𝑢𝑑𝑦 + 𝑣𝑑𝑥 = ∬𝑅[𝑢𝑥 − 𝑣𝑦 ]𝑑𝐴. But by CR-
conditions, 𝑢𝑥 = 𝑣𝑦 and 𝑢𝑦 = −𝑣𝑥 . Hence, ∫𝒞 𝑢𝑑𝑥 − 𝑣𝑑𝑦 + 𝑖 ∫𝒞 𝑢𝑑𝑦 + 𝑣𝑑𝑥 = 0. ∎

Example 1 (entire functions,): Let C be any simple closed curve and n be any natural number.
Then ∮𝐶 𝑧 𝑛 𝑑𝑧 = 0, ∮𝐶 𝑒 𝑧 𝑑𝑧 = 0, ∮𝐶 cos 𝑧 𝑑𝑧 = 0.

Example 2 (analytic but not entire functions,): Let C be the unit circle. Then ∮𝐶 sec 𝑧 𝑑𝑧 = 0
1
and ∮𝐶 𝑧 4+4 𝑑𝑧 = 0 . (The integrands are not analytic outside C but analytic in and on C).

Example 3 (not analytic functions,): Let C be the unit circle. Then ∮𝐶 𝑧𝑑𝑧 = 2𝜋𝑖 .

Example 4 (Analyticity is Sufficient, Not necessary): Let C be the unit circle. Show that
1 1
𝑓(𝑧) = 𝑧 2 is not analytic on C but ∮𝐶 𝑧 2 𝑑𝑧 = 0. Does this contradict Cauchy’s Theorem?

6.3.2 Cauchy-Goursat Theorem: Various Versions

Goursat proved Cauchy’s integral theorem without the assumption of continuity of 𝑓′. There are
several versions of this theorem, like:

✓ Cauchy-Goursat Theorem for simple closed contour


✓ Cauchy Goursat Theorem for simply connected domain
✓ Cauchuy Goursat Theorem for multiply connected domain

Theorem 2A [Cauchy-Goursat Integral Theorem for simple closed curve]: If a function


𝑓(𝑧) analytic inside and on a simple closed contour C, then ∮𝐶 𝑓(𝑧)𝑑𝑧 = 0.

Proof: Read from Churchill. NB. There are various alternative proofs of it in literature.

48
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

Theorem 2B [Cauchy-Goursat Integral Theorem for any Closed Contour in a simply


connected domain]: If a function 𝑓(𝑧) is analytic on a simply connected domain D and C is a
closed contour (not necessarily simple) within D, then ∮𝐶 𝑓(𝑧)𝑑𝑧 = 0.

Proof (Read from Churchill.)

Note: If f is analytic in a multiply connected domain D and C is a closed curve in D, then we cannot , in
general, conclude that ∮𝐶 𝑓(𝑧)𝑑𝑧 = 0. But we have the following result.

Theorem 2C [Cauchy-Goursat Integral Theorem for multiply connected domain]: Let 𝐶


be the outer contour of a multiply connected domain D and 𝐶1 , … , 𝐶𝑛 be the inner contours . If
𝑓(𝑧) is analytic on the domain and on the contours , then

∫𝐶 𝑓(𝑧)𝑑𝑧 + ∑𝑛𝑘=1 ∫𝐶 𝑓(𝑧)𝑑𝑧 = 0;.


𝑘

Where C is oriented anticlockwise and each 𝐶𝑘 clockwise.

Note: If all contours are taken in counter clockwise sense, the conclusion becomes, clearly

∫𝐶 𝑓(𝑧)𝑑𝑧 = ∑𝑛𝑘=1 ∫𝐶 𝑓(𝑧)𝑑𝑧.


𝑘

Proof :

49
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

̃1 and 𝐶
Case 1: If D is doubly connected: Connect the outer and inner contours by two cuts 𝐶 ̃2
as shown. Then D is divided into two simply connected domains, say 𝐷1 and 𝐷2 . By Cauchy’s
integral theorem the integral over the entire boundary of D_1 (in anticlockwise) is zero, and so is the
integral over the boundary of 𝐷2, and thus their sum is zero. In this sum the integrals over the two cuts
cancel because we integrate over them in both directions and we are left with the integrals over 𝐶1 and
𝐶2 . Hence, ∫𝐶 𝑓(𝑧)𝑑𝑧 + ∫𝐶 𝑓(𝑧)𝑑𝑧 = 0.
1 2

Case 2: If D is n-fold connected (Read!).

6.3.3 Path Independence Complex Line Integrals and Related Issues

Recall that in vector integral calculus, we have the following equivalencies (under appropriate
conditions on a vector filed F and a path C in a domain D).

∫𝐶 𝐹. 𝑑𝑟 is path independent in D ⇔ 𝐹. 𝑑𝑟 is exact differential⇔ F is conservative⇔ ∮𝐶 𝐹. 𝑑𝑟 = 0 ⟺F has an


"𝑎𝑛𝑡𝑖𝑑𝑒𝑟𝑖𝑣𝑎𝑡𝑖𝑣𝑒"(potential function).
This sub-section attempts to consider analogous results for complex line integrals.

A. Path Independence of Complex Line Integrals

Definition: A complex line integral of a function 𝑓(𝑧)is called independent of path in a


domain𝐷 if for every 𝑧1 and 𝑧2 in D and every two paths 𝐶1 and 𝐶2 in D, we have
∫𝐶1 𝑓(𝑧)𝑑𝑧 = ∫𝐶2 𝑓(𝑧)𝑑𝑧.

For example, the integrals in Example 2 and Example 3 of section 6.2 are respectively path
independent and path dependent.

Theorem 3[Path Independence]: If 𝑓(𝑧) is analytic in a simply connected domain D,


then the line integral of f(z) is independent of path in D.

Proof:

Case 1: For paths that have only the end points in common

Case 2: For paths that have finitely many further common points
50
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

Case 3: For paths with infinitely many common points. (We illustrate by an example).

B. Principle of Deformation of Path (Principle of Homotopy invariance)

Theorem 4[Principe of Deformation of Path]: Let 𝐶1 and 𝐶2 be two positively oriented simple
closed contours, where 𝐶1 is inferior to 𝐶2 . If 𝑓 is an analytic function in the region between the
contours and on the contours, then

∫ 𝑓(𝑧)𝑑𝑧 = ∫ 𝑓(𝑧)𝑑𝑧
𝐶1 𝐶1

Proof (Exercise): It follows easily from Theorem 6.2C. ∎

Example 5: Use the principle of deformation of path to show that for counter clock wise integration
around any simple closed path C enclosing 𝑧0 ,
2𝜋𝑖 𝑖𝑓 𝑚 = −1
∮(𝑧 − 𝑧0 )𝑚 = {
0 𝑖𝑓 𝑚 ≠ −1
𝐶

Solution: First show for the special case when C is a circle and then use principle of deformation of path
for arbitrary C.

51
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

C. Existence of Indefinite Integrals

Theorem 5: Let be analytic in a simply connected domain D. Then there exists an analytic
function F such 𝐹 ′ (𝑧) = 𝑓(𝑧) for every z in D. (F is called an indefinite integral of 𝑓 on D)

Proof (Text book-Kreyszig-Page 656 and 657).

D. The Fundamental Theorem of Complex Line Integrals

Theorem 6: Let 𝑓 be analytic in a simply connected domain 𝐷 and 𝐹 be an indefinite integral


of 𝑓 on 𝐷. If 𝑧1 and 𝑧2 are any two points in 𝐷 and 𝐶 is any path in 𝐷 joining 𝑧1 to 𝑧2 , then

2 𝑧
∫𝑪 𝒇(z)dz = f(z2 ) − 𝑓(𝑧1 ). (In this case we also write: ∫𝑪 𝒇(z)dz = ∫𝑧1 𝑓(𝑧)𝑑𝑧.)

Remark: In view of this theorem, we now have two methods for evaluating the line integral of
analytic functions.

𝜋+2𝑖 𝑧
Example 6: Compute: (a) ∫0 cos (2) 𝑑𝑧. Answer: 0 (b).

Exercises 6.3:

52
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

6.4 Cauchy’s Integral Formula

Cauchy’s integral formula is a consequence of Cauchy’s integral theorem. In turn, it has


several important consequences. It is also useful in evaluating certain complex line integrals.
It has two main versions:

✓ The Basic Cauchy Integral formula


✓ The Extended Cauchy Integral formula. (This is also called the Cauchy Integral
formula for derivatives or the Generalized Cauchy Integral Formula).

A Visual Representation of Hierarchical Relationship among Some Theorems of this


Chapter

𝐶𝑎𝑢𝑐ℎ𝑦 ′ 𝑠𝑇ℎ𝑒𝑜𝑟𝑒𝑚 (Basic and Cauchy Goursat) → Cauchy Integral formula (basic+ extended)→
Lowville Theorem → Fundamental Theorem of Algebra
two branches{
𝐶𝑎𝑢𝑐ℎ𝑦 ′ 𝑠 𝑅𝑒𝑠𝑖𝑑𝑢𝑒 𝑇ℎ𝑒𝑜𝑟𝑒𝑚

Theorem 1 (Cauchy’s Integral formula): Let 𝑓(𝑧) be analytic inside and on a simple
closed contour C taken in the positive sense. If 𝑧0 is any point interior to C, then that is
oriented counter clock wise

1 𝑓(𝑧)
𝑓(𝑧0 ) = 2𝜋𝑖 ∮𝐶 𝑧−𝑧 𝑑𝑧 =
0

Note that the integrand in Cauchy integral formula is analytic except at 𝑧0 .

53
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

Remarks:

1. CIF can be used to evaluate certain complex line integrals:

𝑓(𝑧)
∮ 𝑑𝑧 = 2𝜋𝑖𝑓(𝑧0 )
𝐶 𝑧 − 𝑧0

2. CIF can be generalized to closed curves in, multiply connected domains. For instance if
𝑓(𝑧) is analytic on two closed contours 𝐶1 and 𝐶2 and the ring-shaped region between
them as shown in the fig, then
1 𝑓(𝑧) 1 𝑓(𝑧)
𝑓(𝑧0 ) = ∮ 𝑑𝑧 + ∮ 𝑑𝑧
2𝜋𝑖 𝐶1 𝑧 − 𝑧0 2𝜋𝑖 𝐶2 𝑧 − 𝑧0

Proof: By sum and constant multiple rules for complex contour integrals, we get:

𝑓(𝑧) 𝑓(𝑧)+𝑓(𝑧0 )−𝑓(𝑧0 ) 𝑓(𝑧 ) 𝑓(𝑧)−𝑓(𝑧0 )


∮𝐶 𝑧−𝑧 𝑑𝑧 = ∮𝐶 𝑧−𝑧0
= ∮𝐶 𝑧−𝑧0 𝑑𝑧 + ∮𝐶 𝑧−𝑧0
𝑑𝑧
0 0

𝑓(𝑧0 ) 𝑓(𝑧)−𝑓(𝑧0 )
=𝑓(𝑧0 ) ∮𝐶 𝑑𝑧 + ∮𝐶 𝑑𝑧…….(*)
𝑧−𝑧0 𝑧−𝑧0

By Example (5) of Section 6.3, we have:

𝑓(𝑧 )
∮𝐶 𝑧−𝑧0 𝑑𝑧 = 2𝜋𝑖𝑓(𝑧0 ) ………….(**)
0

We now use the ML inequality, Principle of Deformation of Path and the fact that analyticity
𝑓(𝑧)−𝑓(𝑧 )
implies continuity to prove that ∮𝐶 𝑧−𝑧 0 𝑑𝑧 = 0.
0

54
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

Let K be a circle of radius 𝜌 and centre 𝑧0 . Hence its length is L=2𝜋𝜌. By the Principle of
Deformation of path

𝑓(𝑧) − 𝑓(𝑧0 ) 𝑓(𝑧) − 𝑓(𝑧0 )


∮ 𝑑𝑧 = ∮ 𝑑𝑧
𝐶 𝑧 − 𝑧0 𝐾 𝑧 − 𝑧0

Let 𝜀 be a given positive number. Since 𝑓(𝑧) is continuous (Why?), so there exists a positive
number 𝛿 such that |𝑓(𝑧) − 𝑓(𝑧0 )| < 𝜀 when ever |𝑧 − 𝑧0 | < 𝛿. We can choose 𝛿 so small that
|𝑓(𝑧) − 𝑓(𝑧0 )| < 𝜀 when ever |𝑧 − 𝑧0 | < 𝜌.. By ML inequality:

𝑓(𝑧)−𝑓(𝑧0 ) 𝜀 𝑓(𝑧)−𝑓(𝑧0 )
|∮𝐾 𝑑𝑧| < 𝜌 (2𝜋𝜌) = 2𝜋𝜀. By taking 𝜀 → ∞, we get ∮𝐾 𝑑𝑧. ∎
𝑧−𝑧0 𝑧−𝑧0

cos(𝑧)
Example 1 : Evaluate ∮
|𝑧|=1
𝑑𝑧 (where the curve is positively oriented.)
𝑧

cos(𝑧)
Solution: By CIF, ∮|𝑧|=1 𝑑𝑧 = 2𝜋𝑖 cos(0) = 2𝜋𝑖. //.
𝑧−0

cos(𝑧)
Example 2: Let C: |𝑧| = 1 with positive orientation. Compute ∮
𝐶 𝑧(𝑧 2 +9)
𝑑𝑧 .

1
cos(𝑧) cos (𝑧)( 2 ) 2
Solution: ∮𝐶 𝑧(𝑧 2+9) 𝑑𝑧 = ∮𝐶 𝑧 +9
𝑧−0
=.9 𝜋𝑖. //

𝜋𝑤𝑖
e 2
Example 3 (two singularities): Compute ∮ 2
𝐶 𝑤 −1
𝑑𝑧. Answer: -2𝜋

55
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

Ln(z+1)
Example 4: Compute ∮
𝐶 𝑧 2 +1
𝑑𝑧 where 𝐶: |𝑧 − 𝑖| = 1.4

Answer: The singularities are 𝑖, −𝑖 and 1. But only 𝑖 lies inside the contour. .Therefore
Ln(z+1) 𝜋 𝑖𝜋2
∮𝐶 𝑧 2 +1 𝑑𝑧 = 2 𝑙𝑛(2) + 4 //

Theorem 2 (Extended Cauchy’s Integral formula): Let n be a natural number. Let 𝑓(𝑧)
be analytic inside and on a simple closed contour C oriented in positive sense. If 𝑧0 is any
point interior to C, then the nth order derivative of 𝑓 at 𝑧0 , 𝑓 [𝑛] (𝑧0 ), exists and it is given
by

𝑛! 𝑓(𝑧)
𝑓 [𝑛] (𝑧0 ) = ∮ 𝑑𝑧
2𝜋𝑖 𝐶 (𝑧 − 𝑧0 )𝑛+1

Remarks:

1. ECIF includes the CIF as a special case.


2. Note that the formula can be obtained by differentiating the corresponding formula for
CIF n times with respect to 𝑧0 .
3. ECIF can be generalized to closed curves in multiply connected domains.

4. ECIF can be used to evaluate certain complex line integrals:

𝑓(𝑧)
∮ 𝑑𝑧 = 2𝜋𝑖 𝑓 [𝑛] (𝑧0 )
𝐶 (𝑧 − 𝑧0 ) 𝑛+1

56
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

Proof (by PMI)

1 𝑓(𝑧)
Step 1 (For n=1): We want to show𝑓 ′ (𝑧0 ) = ∮
2𝜋𝑖 𝐶 (𝑧−𝑧 0)
2

Using the Cauchy integral formula (for f(z) and f(z+∆𝑧) ) we get:

𝑓(𝑧0 + ∆𝑧) − 𝑓(𝑧0 ) 1 𝑓(𝑧) 𝑓 (𝑧)


= [∮ −∮ ]
∆𝑧 2𝜋𝑖∆𝑧 𝐶 𝑧 − (𝑧0 + ∆𝑧) 𝐶 𝑧 − 𝑧0

1 𝑓(𝑧)
= [∮𝐶 ( 𝑧−𝑧0 −∆𝑧)(𝑧−𝑧0 )
]
2𝜋𝑖

Taking the limit of both sides of the last equation as ∆𝑧 approaches to 0 we get

1 𝑓(𝑧)
𝑓 ′ (𝑧0 ) = ∮ .
2𝜋𝑖 𝐶 (𝑧−𝑧 0)
2

(The LHS follows from the definition of derivative. The RHs follows from Analyticity (hence,
continuity) of 𝑓(𝑧) and ML-inequality. By continuity there exist K such that |𝑓(𝑧)| ≤ 𝐾 . Let d
𝑑
be the smallest distance from 𝑧0 to C and let |∆𝑧| ≤ 2. Then

𝑓(𝑧) 𝑓(𝑧)
|∮ 𝑑𝑧 − ∮ 𝑑𝑧|
𝐶 ( 𝑧 − 𝑧0 − ∆𝑧)(𝑧 − 𝑧0 ) 𝐶 (𝑧 − 𝑧0 )
2

𝑓(𝑧)∆𝑧 2𝐾𝐿|∆𝑧|
= |∮ |≤
2 𝑑2
𝐶 (𝑧 − 𝑧0 − ∆𝑧)(𝑧 − 𝑧0 )

This approaches zero as |∆𝑧| approaches zero.

Step 2 and Step 3 (Exercise).

57
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

exp (2𝑧)
Example 5: Evaluate ∮𝐶 where C is the standard unit circle with positive orientation.
𝑧4

Answer: 8𝜋𝑖

sin6 𝑧
Example 6 (Cauchy ‘s Theorem, CIF and ECIF): Compute ∮𝐶 𝜋 3
𝑑𝑧 where
(𝑧− )
6

a) C is the circle |𝑧 − 𝑖10| = 1 oriented in positive sense.


b) C is the circle |𝑧| = 1 oriented in positive sense.

Answers:

a) 0 by Cauchy Goursat Theorem


21𝜋𝑖
b) by the extended Cauchy integral formula.
16

Exercise 6.4

6.5 Consequences of Cauchy Integral Formulas

Cauchy integral formula is related to the theorems listed below and some of these are its
consequences:

✓ All analytic functions have derivatives of all orders,


✓ All analytic functions have Taylor series representations.
✓ Moreira’s Theorem
✓ Cauchy’s Inequality
✓ Lowville’s Theorem
✓ The Fundamental Theorem of Algebra
✓ Gauss’s MVT
✓ Maximum Modulus Theorem
✓ Minimum Modulus Theorem
✓ The Argument Theorem
✓ Rouche’s Theorem

58
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

✓ Poisson’s Integral formulas (for circle, for half plane)


✓ Weirstrass factorization theorem
✓ Miitag Leffler Expansion

Theorem 1 (Analyticity of derivatives): If 𝑓 (𝑧) is analytic in a domain 𝐷, then it has


derivatives of all orders in 𝐷, which are then also analytic functions in 𝐷.

Proof:

Theorem 2 [Morera’s Theorem (Converse of CGT)]: Let 𝑓 be continuous on a domain D.


If ∮𝐶 𝑓(𝑧)𝑑𝑧 = 0 for every closed contour C in D , then f is analytic throughout D.

Proof:

Theorem 3 [Cauchy’s Inequality]: Suppose that f(z) is analytic inside and on a


positively orinetd circle 𝐶𝑅 , centred at 𝑧0 and with radius R. If 𝑀𝑅 denotes the maximum
value of |𝑓(𝑧)| on 𝐶𝑅 , then

n!MR
|f n (z0 )| ≤ (𝑛 = 1,2, … )
Rn

Proof:

Theorem 4 [Liouville’s Theorem]: If an entire function is bounded in absolute value in the


whole complex plane, then this function must be a constant.

Proof:

59
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

CHAPTER 7: TAYLOR SERIES AND LAURENT SERIES

This Chapter extends the concepts of real sequences and real series, which are fundamental in
real calculus to the realm of complex calculus. One of the main results of this chapter is that
complex power series represent analytic functions and that; conversely, every analytic function
can be represented by power series. Here is a more detailed breakdown of the chapter.

7.1 Basic Definitions and Properties on Complex Sequences and Complex Series
7.2 Power series
7.3 Taylor Series and Taylor Theorem
7.4 Laurent Series and Laurent theorem
(Texts: Kreyszig, Churchill, Schaum)
7.1 Basic Definitions and Properties on Complex Sequences and Complex Series

The basic concepts for complex sequences and series and tests for convergence and
divergence are very similar to those concepts in (real) calculus. Thus if the reader feels the
next notes trivial, he/she can skip this section

7.1.1 Sequences and Series of Complex Constants


A. Definitions

Definition 1 (complex sequence and complex series):

a) An infinite sequence or just a sequence of complex numbers is an ordered and unending list of
complex numbers of the form
𝑧1 , 𝑧2 , 𝑧3 , ..
We can denote it by using various notations like {𝑧1 , 𝑧2 , . . } or {𝑧𝑛 }∞
𝑛=1 . Here 𝑛 is called the

running index, 1 is called the initial index and 𝑧𝑛 is called the nth term. The initial index can
also be any integer other than 1.
b) An infinite series or just a series of complex numbers is a formal sum of the form
𝑧1 + 𝑧2 + ⋯
𝑛
We denote it by ∑∞
1 𝑧𝑘 , The finite sum 𝑆𝑛 ≔ ∑1 𝑧𝑘 is called the nth partial sum of the series.

60
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

Definition 2 (Convergence and Divergence of complex sequence and complex series):

a) An infinite sequence {𝑧𝑛 } of complex numbers is said to converge to a complex number 𝑐,


written 𝑧𝑛 → 𝑐 ( or equivalently c is said to be the limit of {𝑧𝑛 }, written c = lim 𝑧𝑛 ) if for
𝑛→∞

every 𝜀 > 0 there exists a natural number 𝑁 such that


|𝑧𝑛 − 𝑐| < 𝜀 for all 𝑛 > 𝑁

Geometrically, 𝑧𝑛 → 𝑐 if and only if for every positive number 𝜀 there exists a natural number N
such that all of the terms 𝑧𝑁+1 , 𝑧𝑁+2 , … lie with in the open disk of radius 𝜀 and centre c.

A sequence {𝑧𝑛 } of complex numbers is said to be convergent of there exists a complex number c
such that 𝑧𝑛 → 𝑐; otherwise, it is called divergent.

b) An infinite series ∑∞
1 𝑧𝑘 of complex numbers is said to converge to a complex number 𝑠 if its

sequence of partials sums converge to s(i.e., 𝑠𝑛 → 𝑠). In analogy with real series, we can define
convergent series, divergent series, absolutely convergent series and conditionally convergent
series of complex constants.

Example 1: Show that

1
lim (𝑛2 + 𝑖) = 𝑖.
𝑛→∞

61
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

Example 2: Show that the sequence {𝑖 𝑛 } is divergent.

B. Properties of Complex Sequences and Series (of Constants)

The convergence of a complex sequence can be verified by verifying the convergence of the two
real sequences associated with it (that is; the sequences of the real parts and the imaginary parts).

Theorem 1[Convergence of Real and the Imaginary Parts of a Complex Sequence]: A


sequence {𝑧𝑛 } = {𝑥𝑛 + 𝑖 𝑦𝑛 } of complex numbers converges to a complex number 𝑐 = 𝑎 + 𝑖𝑏
if and only if the sequence of the real parts {𝑥𝑛 } converges to "𝑎" and the sequence of the
imaginary parts {𝑦𝑛 }converges to "𝑏".

Proof (Exercise):

The next theorem relates the convergence of a complex series to that of the two series of its real
parts and of its imaginary parts:

Theorem 2[Convergences of Real and the Imaginary Parts of a complex series]: Let 𝑥𝑘 +
𝑖𝑦𝑘 be a complex sequence and 𝑢 + 𝑖𝑣 be a complex number. Then

∑∞ ∞ ∞
𝑘=1(𝑥𝑘 + 𝑖𝑦𝑘 ) = 𝑢 + 𝑖𝑣 if and only if ∑𝑘=1 𝑥𝑘 = 𝑢 and ∑𝑘=1 𝑦𝑘 = 𝑣 .

Remark (Convergence Tests for Complex Sequences and Series): In view of Theorems 1 and
2, the convergence tests in complex calculus are practically the same as in real calculus. We
discuss only few convergence tests.

Theorem 3(Divergence Test): If ∑∞


1 𝑧𝑘 converges, then lim 𝑧𝑘 = 0.
𝑛→∞

62
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

C. Cauchy convergence Principle

In order to prove the convergence of a sequence(via its definition) , its limit must be known. This
is a practical difficulty, because the limit is not known in most cases. Cauchy provided a way to
test convergence of a sequence without any reference to its limit.

Theorem 4 A(Cauchy’s Convergence Principle for Sequence): A complex sequence {𝑧𝑘 } is


convergent if and only if for every given positive number ε (however small) we can find a
natural number N (which depends on ε) such that |𝑧𝑛 − 𝑧𝑚 | < 𝜀 for all 𝑛. 𝑚 > 𝑁.

Remarks

1. The condition |𝑧𝑛 − 𝑧𝑚 | < 𝜀 for all 𝑛, 𝑚 > 𝑁, in the theorem, can be replaced by other
equivalent conditions such as.|𝑧𝑛 − 𝑧𝑚 | → ∞ as . 𝑚 → ∞.
2. Intuitively, a Cauchy sequence is a sequence whose terms (not just consecutive terms)
become arbitrarily close to each other as the sequence progresses. In order for a sequence
{𝑧𝑘 } to converge, it is not sufficient for each term, after 𝑧𝑁 , to become arbitrarily close
to the preceding term. But every two terms after 𝑧𝑁 must become arbitrary close.
3. Cauchy convergence principle is useful in setting termination criteria for sequence of
iterates of algorithms.

Theorem 4 B(Cauchy’s Convergence Principle for Series): A complex series ∑∞


𝑘=1 𝑧𝑘 is

convergent if and only if for every given positive number 𝜀 (however small) we can find a
natural number N (which depends on 𝜀) such that |𝑆𝑛 − 𝑆𝑚 | < 𝜀 for all 𝑛. 𝑚 > 𝑁. (Here 𝑆𝑛 and
𝑆𝑚 are the 𝑛𝑡ℎ and 𝑚𝑡ℎ partial sums of the series.)

63
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

Definition 3(absolute and conditional convergence): A complex series 𝑧1 + 𝑧2 + ⋯ is called:

a) absolutely convergent if the series of the absolute values of the terms, i.e., |𝑧1 | +
|𝑧2 | + ⋯, is convergent.
b) conditionally convergent if 𝑧1 + 𝑧2 + ⋯ is convergent but |𝑧1 | + |𝑧2 | + ⋯ is divergent

Note: If a complex series is absolutely convergent, then it is convergent.

absolutely
convergent
convergent
conditionally
series

convergent

oscilatating
divergent
divergent to ±∞

Theorem 5[Convergence Test for Geometric Series]: The geometric complex series
1
∑∞ 𝑚
𝑚=0 𝑞 converges with the sum if |𝑞| < 1 and diverges if |𝑞| ≥ 1.
1−𝑞

Proof:

Case 1: If |𝑞| ≥ 1, then lim |𝑞|𝑚 ≥ 1 and by divergence test it implies divergence of
𝑚→∞

∑∞ 𝑚
𝑚=0 𝑞 .

1−𝑞 𝑛+1
Case 2: If |𝑞| < 1, the nth partial sum becomes 𝑆𝑛 = 1−𝑞
and hence the sum is S is given by:
1
S = lim 𝑠𝑛 = 1−𝑞. ∎
𝑛→∞

The next test is one form of Ratio test. We get it by combining Comparison test and Geometric
series Test:

64
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

Theorem 6[Ratio Test-Version I]: Let ∑∞


𝑘=1 𝑧𝑘 be a complex series with nonzero terms and q

be a fixed positive real number. If there exists 𝑁 such that:

𝑧𝑘+1
a) | 𝑧𝑘
| ≤ 𝑞 < 1 for every 𝑘 > 𝑁, then ∑∞
𝑘=1 𝑧𝑘 converges.

𝑧𝑘+1
b) | | ≥ 1 for every 𝑘 > 𝑁, then ∑∞
𝑘=1 𝑧𝑘 diverges.
𝑧𝑘

𝑧𝑛+1
Caution: The inequality in (a) cannot be relaxed to | 𝑧𝑛
| < 1. (As counter example, take the

harmonic series.).

Theorem 7[Ratio Test-Version II (The usual ratio test)]: Let ∑∞


𝑘=1 𝑧𝑘 be a complex series
𝑧𝑘+1
with nonzero terms and let lim | 𝑧𝑘
| = 𝐿. If
𝑘→∞

a) 𝐿 < 1, then ∑∞
𝑘=1 𝑧𝑘 converges.

b) 𝐿 > 1, then ∑∞
𝑘=1 𝑧𝑘 diverges.

c) 𝐿 = 1, then this test does not apply.

Example: Apply a ratio test to decide if the series 𝑎0 + 𝑏0 +1 + 𝑏1 + 𝑎2 + 𝑏2 + ⋯ is convergent


𝑖 𝑖
or divergent where 𝑎𝑛 = 23𝑛 and 𝑏𝑛 = 3𝑛+1.
2

𝑧𝑘+1 1 1 1 1 1 1
Solution: The first few terms of the sequence | | are: , , , , , ,…
𝑧𝑘 2 4 2 4 2 4

𝑧𝑘+1
The usual ratio test cannot be applied because lim | 𝑧𝑘
| DNE.
𝑘→∞

1
From Ratio Test I, we can conclude that the series is convergent. (Use Ratio Test I ,with 𝑞 = 2).

Note: We can also state two Root Tests (Root test I and II in the same manner as Theorem 6 & 7.

7.1.2 Sequences and Series of Complex Functions

65
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

Let 𝑓1 (𝑧), 𝑓2 (𝑧), 𝑓3 (𝑧), … be a sequence whose terms are complex functions having a common
domain D. If z is replaced by a constant complex number, then the sequence {𝑓𝑘 (𝑧)} reduces to a
sequence of constants. Similarly we can consider a series with variable terms.

Definition 4(convergence of complex sequence with variable terms): Let 𝑔(𝑧) be a


complex function and let {𝑓𝑘 (𝑧)} = {𝑓1 (𝑧), 𝑓2 (𝑧), … } (where 𝑧 ∈ 𝐷) be a sequence of
complex functions. The sequence {𝑓𝑘 (𝑧)} is said to

a) converge point wise to g on D if for every element 𝑧 ∗ in D and for every positive
number 𝜀, there exist a natural number N (N depends on 𝑧 ∗ and 𝜀 ) such that
|𝑓𝑘 (𝑧) − 𝑔(𝑧 ∗ )| < 𝜀 whenever 𝑛 > 𝑁.
a) converge uniformly to g on D if for every positive number 𝜀, there exist a natural
number N (N depends on 𝜀 ) such that for every two elements 𝑧, 𝑧 ∗ in D
|𝑓𝑘 (𝑧) − 𝑔(𝑧 ∗ )| < 𝜀 whenever 𝑛 > 𝑁.

Similarly we can define point wise and uniform convergence for complex series with variable
terms.

Note: A sequences/series with variable terms falls to one of these three categories:

a) It converges at every z in D.
b) It diverges at some value of z and diverges at other values of z.
c) It diverges at all values of z.

Exercise 7.1:

66
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

7.2 Power Series


A. Absolute Convergence and Uniform Convergence

Power series is a special type of series of functions.

Definition 1 (convergence of complex sequence with variable terms): Let


𝑧0 , 𝑐0 , 𝑐1 , 𝑐2 , …be complex constants and z be a complex variable. A series of the form

∑∞ 𝑘
𝑘=0 𝑐𝑘 (𝑧 − 𝑧0 ) ,

is called a power series centred at 𝒛𝟎 or a power series in 𝒛 − 𝒛𝟎 . In this case 𝑧0 is called


the centre of the power series.

Remark: The power series is always convergent when 𝑧 = 𝑧0 (because 𝑧 = 𝑧0 implies


∑∞ 𝑘 ∞ 𝑘
𝑘=0 𝑐𝑘 (𝑧 − 𝑧0 ) = ∑𝑘=0 𝑐𝑘 (𝑧0 − 𝑧0 ) = 0 + 0 + ⋯ = 0 . )

The next theorem implies that there are three main types of power series for complex variables,
based on convergence behaviour: A power series can be:

i) Convergent only at one point (centre), ( The useless case)


ii) everywhere convergent, or (The nicest case)
iii) Convergent on a disk of positive radius. ( Convergence on a Disk)

Theorem 1(Absolute Convergence of a Power Series):

a) If a power series ∑∞ 𝑘
𝑘=0 𝑐𝑘 (𝑧 − 𝑧0 ) converges at 𝑧 = 𝑧1 (𝑧1 ≠ 𝑧0 ) and if 𝑧2 is closer to 𝑧0

than 𝑧1 , then it converges absolutely at 𝑧2 .


b) If it diverges at 𝑧 = 𝑧1 and if 𝑧2 is further from 𝑧0 than 𝑧1 , then it diverges at 𝑧2 , too.

67
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

Remark:

Proof:

a) Assume that ∑∞ 𝑘
𝑘=0 𝑐𝑘 (𝑧1 − 𝑧0 ) is convergent and |𝑧2 − 𝑧0 | < |𝑧1 − 𝑧0 |. By

Divergence Theorem, 𝑐𝑘 (𝑧1 − 𝑧0 )𝑘 → 0 as 𝑘 → ∞. This implies boundedness in absolute


value, i.e., there exists a positive number M such that

|𝑐𝑘 (𝑧1 − 𝑧0 )𝑘 | ≤ 𝑀 for every 𝑘 = 0,1,2, …..

(𝑧 −𝑧0 )𝑘
Then multiplying both sides of the inequality by |(𝑧2 𝑘 |, we get
1 −𝑧0 )

(𝑧 −𝑧0 )𝑘
|𝑐𝑘 (𝑧2 − 𝑧0 )𝑘 | ≤ 𝑀 |(𝑧2 𝑘 |,
1 −𝑧0 )

Summation over k gives:

∞ ∞
(𝑧2 − 𝑧0 )𝑘
∑|𝑐𝑘 (𝑧2 − 𝑧0 )𝑘 | ≤ ∑𝑀| |
(𝑧1 − 𝑧0 )𝑘
𝑘=0 𝑘=0

The right hand side is a real Geometric series with common ratio less than 1. Hence, it is
convergent. By comparison test the left hand side converges absolutely.

b) Exercise ∎

Remark(three types of a power series.): The power series ∑∞ 𝑘


𝑘=0 𝑐𝑘 (𝑧 − 𝑧0 ) can :

a) Converge at 𝑧0 only (useless case)


b) Converge at every 𝑧 (Nicest case)
c) Converge at all points inside some circle C and diverges at all points outside
C(while it may converge at all/some /no point of C itself). This follows from
Theorem 1.

68
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

Definition 2(Circle of Convergence, Radius of Convergence & Disk of Convergence): Let


the power ∑∞ 𝑘
𝑘=0 𝑐𝑘 (𝑧 − 𝑧0 ) converges at some points other than 𝑧0 and diverges at some

other points. The smallest circle with centre 𝑧0 that includes all the points at which the power
series converges is called the circle of convergence and its radius 𝑅 is called the radius of
convergence. The set of all points at which a complex power series converges is called the
region of convergence or sometimes the disk of convergence. It follows that D is the union of
points interior to C and {z ∈ 𝐶: ∑∞ 𝑘
𝑘=0 𝑐𝑘 (𝑧 − 𝑧0 ) converges }.

Convention: If the power series converges for all 𝑧, we write 𝑅 = ∞; if it converges for 𝑧0 only
we write 𝑅 = 0.

Remark: A power series is absolutely convergent within its region (disk) of convergence, but it
may be absolutely convergent, conditionally convergent or divergent at the boundary points of
the region. (See also Uniform convergence of power series-theorem 4). The region of
convergence of a power series is always a disk (possibly infinitely large or point).

Theorem 2 (Determination of the Radius of Convergence from the Coefficients-by Ratio Test]):
𝑐𝑘+1
Consider the power series ∑∞ 𝑘
𝑘=1 𝑐𝑘 (𝑧 − 𝑧0 ) and let L: = lim | |. If
𝑘→∞ 𝑐𝑘

a) 𝐿 = 0, then it converges for every z; hence 𝑅 = ∞


b) 𝐿 = ∞, then it converges only at z_0; hence 𝑅 = 0
1 1 1
c) 𝐿 > 0 and 𝐿 ≠ ∞, then it converges for |𝑧 − 𝑧0 | < 𝐿 and diverges for |𝑧 − 𝑧0 | > 𝐿, hence 𝑅 = 𝐿.

69
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

Remarks:

1. Other methods to determine 𝑅 are by using the usual Root test or by using the so called
Cauchy-Hadamard Theorem (a specialized application of the root test tailored for power
series.).
2. Determination of the circle of convergence requires usually additional effort, after finding
𝑅.

Theorem 3[Cauchy, 1821] The radius of convergence of the power series ∑∞


𝑘=0 𝑐𝑘 (𝑧 − 𝑧0 )
𝑘

1
is 𝑅 = 1 .
limsup[𝑎𝑘 ]𝑘
𝑘→∞

Theorem 4(Uniform Convergence of Power series): A power series with centre 𝑧0 and
radius of convergence 𝑅 > 0 converges uniformly on every circular disk |𝑧 − 𝑧0 | ≤ 𝑟 with
radius 𝑟 < 𝑅.

Definition 3 (Representation of a Function by a Power Series): Let 𝑓(𝑧) be a function and


∑∞ 𝑘
𝑘=0 𝑐𝑘 (𝑧 − 𝑧0 ) be a power series with disk of convergence D such that

𝑓(𝑧) = ∑∞ 𝑘
𝑘=0 𝑐𝑘 (𝑧 − 𝑧0 ) .for all z∈ D

Then we say f(z) is represented by the power series on D.

𝑛
Note: Every power series of the form ∑∞ 𝑘 ∞
𝑘=0 𝑐𝑘 (𝑧 − 𝑧0 ) can be reduced to the form ∑𝑛=0 𝑐𝑛 𝑧̃

(Just let 𝑧̃ : = 𝑧 − 𝑧0 ).

70
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

B. Operations With Power Series

We can create new power series from other Power series by using the operstions of addition,
subtraction, multiplication, division, differentiation and integration

Theorem 5[Continuity of the Sum of a Power Series]: If a function 𝑓(𝑧) can be represented
by a power series ∑∞ 𝑘
𝑘=0 𝑐𝑘 𝑧 with radius of convergence R, then it is continuous at 𝑧 = 0.

Theorem 6 (Uniqueness of a Power Series Representation): A function cannot be represented


by two different power series with the same centre, i.e., if 𝑓(𝑥) = ∑∞ 𝑘
𝑘=0 𝑏𝑘 (𝑧 − 𝑧0 ) and 𝑓(𝑥) =

∑∞ 𝑘
𝑘=0 𝑐𝑘 (𝑧 − 𝑧0 ) , then 𝑏𝑘 = 𝑐𝑘 for all 𝑘 = 0,1,2, …

Theorem 7 [Term wise addition or subtraction of two power series]: Let ∑∞ 𝑘


𝑘=0 𝑎𝑘 𝑧 and

∑∞ 𝑘
𝑘=0 𝑏𝑘 𝑧 be two convergent power series with sums 𝑓(𝑧) and 𝑔(𝑧) and radius of convergence

𝑅1 &𝑅2 respectively. Then the power series ∑∞ 𝑘 ∞ 𝑘


𝑘=0(𝑎𝑘 + 𝑏𝑘 )𝑧 and ∑𝑘=0(𝑎𝑘−𝑏𝑘 ) 𝑧 have with

radius of convergence at least equal to the smaller of 𝑅1 and 𝑅2 .

Theorem 8 [Cauchy Product of two power series]: Let ∑∞ 𝑘 ∞ 𝑘


𝑘=0 𝑎𝑘 𝑧 and ∑𝑘=0 𝑏𝑘 𝑧 be two

absolutely convergent power series with sums 𝑓(𝑧) and 𝑔(𝑧) respectively. Then the power series
𝑛
∑∞ 𝑘
𝑘=0 𝑐𝑛 𝑧 (where 𝑐𝑛 ≔ ∑𝑘=0 𝑎𝑘 𝑏𝑛−𝑘 ) is called the Cauchy product of the two series and it

converges absolutely to 𝑓(𝑧)𝑔(𝑧).

Note: Term by term multiplication of two power series does not behave nicely as Cauchy
product of the two series.

71
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

Theorem 9(Term wise Differentiation of a Power Series): The power series. ∑∞


𝑘=0 𝑏𝑘𝑘 𝑧
𝑘−1

obtained by differentiating the power series ∑∞ 𝑘


𝑘=0 𝑏𝑘 𝑧 term by term , called a derived series,

has the same radius of convergence as the power series ∑∞ 𝑘


𝑘=0 𝑏𝑘 𝑧 .

𝑏𝑘
Theorem 10 (Term wise Integration of Power Series): The power series ∑∞
𝑘=0 𝑧 𝑘+1
𝑘+1

obtained by integrating the series ∑∞ 𝑘


𝑘=0 𝑏𝑘 𝑧 term by term has the same radius of

convergence as the power series ∑∞ 𝑘


𝑘=0 𝑏𝑘 𝑧 .

C. Power Series Representation of Analytic Functions

Let us first elaborate the difference between real analyticity and complex analyticity and the
equivalence of holomorphicity and analyticity for complex functions. Originally, the term
"analytic function" described any function (real or complex) that could be represented by a
convergent power series in a neighbourhood of each point in its domain. Originally, the term
“holomorphic function" meant a complex-valued function of a complex variable that is complex
differentiable in a neighbourhood of every point in its domain. The notion is not applicable to
real functions.

While initially distinct, it is a fundamental result in complex analysis that a complex


function is analytic (expressible as a power series) if and only if it is holomorphic (complex
differentiable), meaning the two concepts are equivalent. But in this course we used the term
analytic function as a function that is differentiable at every nbd of each point in its domain
and the term holomorphic function as its synonym. In this section we prove that power series
represent analytic functions (complex differentiable functions). In the next section, we prove the
converse of this; a property which does not hold for real analytic functions.)

72
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

Theorem 11 [Functions Represented by Power Series are Analytic (Differentiable )]:. The sum of
a power series (with a positive radius of convergence) is an analytic function and has derivatives of all
orders, which thus in turn are analytic functions.

Remark: In the next section we show that, conversely, every given analytic function can be
represented by power series, called Taylor series .

Proof : Let ∑∞ 𝑘
𝑘=0 𝑏𝑘 𝑧 be a power series with a nonzero (positive) radius of convergence 𝑅 and

sum 𝑓(𝑧). Let 𝑓1 (𝑧) be the sum of the derived series ∑∞ 𝑘 ′


𝑘=1 𝑘𝑏𝑘 𝑧 . We prove that 𝑓 (𝑧) =

𝑓1 ′(𝑧). The proof is involved; read it from Kreyszig).

Exercises 7:2:

73
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

7.3 Complex Taylor Series and Taylor Theorem


A. Taylor Series

Definition 1: Let 𝑓(𝑧) be a complex function which is infinitely differentiable at 𝑧0 . Then the
power series:

∑∞
𝑘=0 𝑐𝑘 (𝑧 − 𝑧0 )
𝑘

𝑓 [𝑘] (𝑧0 ) 1 𝑓(𝑠)


where 𝑐𝑘 ≔ ( or equivalently 𝑐𝑘 ≔ ∮ 𝑑𝑠 )., is called
𝑘! 2𝜋𝑖 𝐶 (𝑠−𝑧0 }𝑘+1

a) the Taylor series of 𝒇(𝒛) with centre 𝑧0 , at 𝑧0 or about 𝑧0 .


b) the Maclaurin series of 𝒇(𝒛) if 𝑧0 = 0.

Remarks

1. The Taylor series of 𝑓(𝑧) at 𝑧0 may not converge at certain elements 𝑧 in the Dom(𝑓).
But it always converges.at 𝑧 = 𝑧0.
2. Even if the Taylor series of 𝑓(𝑧)at 𝑧0 converges at 𝑧, it may not converge to 𝑓(𝑧).
3. If 𝑓(𝑧) is analytic and its Taylor series converges at 𝑧, then it converges to 𝑓(𝑧).
4. Every Taylor series is a power series. (by definition). Conversely, a power series with
positive radius of convergence is the Taylor series of its sum. (This is not valid for real
analytic functions.).
5. The polynomial 𝑇𝑛 (𝑧) ≔ ∑𝑛𝑘=0 𝑐𝑘 (𝑧 − 𝑧0 )𝑘 is called the nth Taylor polynomial of 𝑓(𝑧).
6. The difference 𝑅𝑛 (𝑧) ≔ 𝑓(𝑧) − 𝑇𝑛 (𝑧) is called the nth Taylor remainder of 𝑓(𝑧). Thus
𝑅𝑛 (𝑧), is the error introduced when approximating the function 𝑓(𝑧) using its Taylor
polynomial of degree 𝑛.

74
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

B. Taylor’s Theorem

In the previous section we have learnt that power series represents analytic(holomorphic)
functions. We now show that every analytic function can be represented by a power series,
namely, by Taylor series (with various centres).

Theorem 1 [Taylor’s theorem]: Let 𝑓(𝑧) be analytic in a domain 𝐷, and let 𝑧0 be any point in 𝐷.
Then the Taylor series of 𝑓(𝑧) with centre 𝑧0 represents 𝑓(𝑧) at every z inside its circle of
𝑓 [𝑘] (𝑧0 )
convergence i.e., ∑∞
𝑘=0 𝑘!
(𝑧 − 𝑧0 )𝑘 = 𝑓(𝑧) (for every z in the largest open disk with centre

𝑧0 in which 𝑓 is analytic).

Proof: The key tool is Cauchy’s integral formula. Let C be the circle of convergence, z be inside
C and s be on C (In the fig below s is erroneously written as z*). Then by Cauchy integral
formula,

1 𝑓(𝑠)
𝒇(𝒛) = ∮ 𝑑𝑠 (*)
2𝜋𝑖 𝐶 𝑠−𝑧

1
We first express 𝑠−𝑧 in powers of (𝑧 − 𝑧0 ). By algebraic manipulation, we obtain

1 1 1 1 1
= = 𝑧_𝑧 = .
𝑠 − 𝑧 𝑠 − 𝑧0 − (𝑧 − 𝑧0 ) (𝑠 − 𝑧0 )(1 − 0 (𝑠 − 𝑧0 ) 1 − 𝑧_𝑧0
𝑠 − 𝑧0 ) 𝑠 − 𝑧0

𝑧_𝑧
By Geometric progression formula (since |𝑠−𝑧0 |<1), we get
0

75
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

1 1 1
= .
𝑠 − 𝑧 (𝑠 − 𝑧0 ) 1 − 𝑧_𝑧0
𝑠−𝑧 0
1 𝑧_𝑧0 𝑧_𝑧0 2 𝑧 − 𝑧0 𝑛 1 𝑧 − 𝑧0 𝑛+1
= . [1 + +( ) + ⋯+ ( ) ]+ ( )
(𝑠 − 𝑧0 ) 𝑠 − 𝑧0 𝑠 − 𝑧0 𝑠 − 𝑧0 𝑠 − 𝑧 𝑠 − 𝑧0

Multiplying both sides by f(s) gives:

𝑓(𝑠) 𝑓(𝑠) 𝑧_𝑧0 𝑧_𝑧0 2 𝑧_𝑧0 𝑛+1 𝑓(𝑠) 𝑧_𝑧0 𝑛+1


= (𝑠−𝑧 ) . [1 + +( ) +⋯+( ) ]+ ( ) (**)
𝑠−𝑧 0 𝑠−𝑧0 𝑠−𝑧0 𝑠−𝑧0 𝑠−𝑧 𝑠−𝑧0

Inserting (**) into (*) gives us:

1 𝑓(𝑠) 1 𝑧−𝑧0 (𝑧−𝑧0 )𝑛


𝒇(𝒛) = 2𝜋𝑖 ∮𝐶 𝑠−𝑧 𝑑𝑠 = 2𝜋𝑖 𝑓(𝑧0 ) + 2𝜋𝑖
𝑓′(𝑧0 ) + ⋯ + 2𝜋𝑖
𝑓 [𝑛] (𝑧0 ) + 𝑅𝑛 (𝑧)

(𝑧−𝑧0 )𝑛+1 𝑓(𝑠)


Where 𝑅𝑛 (𝑧) ≔ ∮𝐶 (𝑠−𝑧)(𝑠−𝑧 𝑛+1
𝑑𝑠
2𝜋𝑖 0)

Claim: lim 𝑅𝑛 (𝑧) = 0.


𝑛→∞

This completes the proof.∎

C. Finding Taylor Series Expansions: Practical Methods


- There are two common Methods of obtaining Taylor series expansions of 𝑓(𝑧):
(i) By Taylor’s Theorem; if 𝑅𝑛 → 0, then 𝑓(𝑧) = 𝑓(𝑧0 ) + 𝑓 ′ (𝑧0 )(𝑧 − 𝑧0 ) +
𝑓 ′′ (𝑧0 )
(𝑧 − 𝑧0 )2+…
2!

(ii) By algebraically manipulating known expansion formulas: Using addition,


subtraction, Cauchy product, division, differentiation, integration, substitution , etc.

- Regardless of the method used, the result will be the same. This follows from the
uniqueness Theorem.

Example 1: Prove that

1 1
1.
1−z
= ∑∞
k=0 zk if (|𝑧| < 1)

76
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

zk
2. 𝑒 𝑧 = ∑∞
k=0 if (|𝑧| < ∞)
k!
z2k−1
3. sin(𝑧) = ∑∞
k=0(−1)
𝑘+1
if (|𝑧| < ∞)
(2k−1)!

z2k
4. cos (z) = ∑∞
k=0(−1)
𝑘
if (|𝑧| < ∞)
(2k)!

z2k+1
5. sinh(𝑧) = ∑∞
k=0 if (|𝑧| < ∞)
(2k+1)!

z2k
6. cos h(z) = ∑∞
k=0 if (|𝑧| < ∞)
(2k)!

7. Ln(1+z)=

Proof: They follow from analyticity of the functions and Taylor’s theorem.

Example 2: Use algebraic manipulations of known Taylor series expansions to find the
Maclaurin series expansion of:

𝑧
a) 𝑓(𝑧) = (𝑧 4+4).
1−𝑧
b) 𝑓(𝑧) = 𝐿𝑜𝑔(1+𝑧)

Solution: Rewrite the function in the form of a geometric series and then expand it.

n+1
Answer (a): ∑∞
n=0 (−1)
𝑛
4𝑛+1
//

Exercises 7.3:

77
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

7.4 Laurent Series and Laurent Theorem


A. Laurent Series

If 𝑓(𝑧) is not analytic at 𝑧0 , we cannot expand 𝑓(𝑧) in powers of 𝑧 − 𝑧0 as a Taylor series.


Instead we can represent 𝑓(𝑧) by a series consisting of both non-negative and negative powers of
(𝑧 − 𝑧0). Such a series is called a Laurent series. . Laurent series are also used for classifying
singularities and in a powerful integration method (called residue integration, Chapter 8). A
Laurent series of 𝑓(𝑧) converges in an annulus (in the “hole” of which 𝑓(𝑧) may have
singularities).

1+2𝑧 2
Example 1: Express 𝑓(𝑧) = 𝑧 3+𝑧 5 by a series in powers of z.

Solution: Since 𝑓 is not analytic at 0, the series as powers of z cannot be Taylor series. The

1
⏟ + 𝒛𝟑 − 𝒛𝟓 + ⋯ ) + ( ⏟
answer is the Laurent series (−𝒛 𝒛−𝟏 + 𝒛−𝟑 ).To get this, factorize 𝑧 3
𝒑𝒐𝒔𝒊𝒕𝒊𝒗𝒆 𝒑𝒂𝒓𝒕 𝒏𝒂𝒈𝒆𝒕𝒊𝒗𝒆 𝒑𝒂𝒓𝒕

and expand the other factor and multiply) //

𝑒𝑧
Example 2: Expand 𝑓(𝑧) = (𝑧+1)2 around -1 ; that is as powers of (𝑧 + 1).

Solution: Since 𝑓 is not analytic at -1, the series as powers of z cannot be Taylor series. The
(𝑧+1)𝑛−2
answer is the Laurent series ∑∞
𝑛=0 𝑒.𝑛!
. //

78
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

B. Laurent Theorem

Theorem 1 [Laurent’s theorem]: Let 𝑓(𝑧) be analytic in a domain 𝐷 containing two concentric circles
𝐶1 and 𝐶2 with centre 𝑧0 and the annulus between them. Then 𝑓(𝑧) can be represented by the Laurent
series.

∞ ∞
𝑏𝑛
𝑓(𝑧) = ∑ 𝑎𝑛 (𝑧 − 𝑧0 )𝑛 +∑
(𝑧 − 𝑧0 )𝑛
𝑛=0 𝑛=1

𝑏1 𝑏
= 𝑎0 + 𝑎1 (𝑧 − 𝑧0 ) + 𝑎2 (𝑧 − 𝑧0 )2 + … + + (𝑧−𝑧2 )2 + ⋯ (1)
𝑧−𝑧0 0

Where the coefficients 𝑎𝑛 and 𝑏𝑛 are given by the integrals,

1 𝑓(𝑠) 1
𝑎𝑛 ≔ 2𝜋𝑖 ∮𝐶 (𝑠−𝑧)𝑛+1 𝑑𝑠 , 𝑏𝑛 ≔ 2𝜋𝑖 ∮𝐶 𝑓(𝑠)(𝑠 − 𝑧0 }𝑛−1 𝑑𝑠 (2)

taken counter clockwise around any simple closed path C that lies in the annulus and encircles the inner
circle. [The variable of integration is denoted by s, not z, because z is used in (1).] This series converges
and represents in the enlarged open annulus obtained from the given annulus by continuously increasing
the outer circle 𝐶1 and decreasing the inner circle 𝐶2 until each of the two circles reaches a point where
𝑓(𝑧) is singular.

79
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

Remarks:

1. If 𝑧0 is the only singular point of 𝑓(𝑧) inside the circle 𝐶2 , the series (or finite sum) of the
negative powers of (1) is called the principal part of 𝑓(𝑧) at 𝑧0 [or of that of (1)].
2. Denoting 𝑏𝑛 by 𝑎−𝑛 , instead of formula (1) of the theorem: we may write

∑ 𝑎𝑛 (𝑧 − 𝑧0 )𝑛
−∞

And instead of formula (2) we may write

1 𝑓(𝑠)
: 𝑎𝑛 ≔ ∮ 𝑑𝑠 , 𝑛 = 0, ±1, ±2, … . .
2𝜋𝑖 𝐶 (𝑠−𝑧)𝑛+1

Proof (Laurent Theorem): We use Cauchy’s integral formula for multiply connected domain and
the principle of deformation of path.

Recall that if is analytic on 𝐶1 and 𝐶2 and in the ring-shaped domain bounded by 𝐶1 and 𝐶2 and z is
any point in that domain, then:

1 𝑓(𝑠) 1 𝑓(𝑠)
𝑓(𝑧) = ∮
2𝜋𝑖 𝐶1 𝑠−𝑧
𝑑𝑠 − ∮
2𝜋𝑖 𝐶2 𝑠−𝑧
𝑑𝑠 (*)

Where both integrals are is taken counter-clockwise.

1 𝑓(𝑠) 1 𝑓(𝑠)
Let 𝑔(𝑧) = 2𝜋𝑖 ∮𝐶 𝑠−𝑧
𝑑𝑠 & ℎ(𝑧) = − 2𝜋𝑖 ∮𝐶 𝑠−𝑧
𝑑𝑠.
1 2

Then 𝑓(𝑧) = 𝑔(𝑧) + ℎ(𝑧).

Step 1: For the nonnegative Powers

The first integra in (*) l is precisely as in the proof of Taylor’s theorem . Hence we get exactly
the same result, namely, the Taylor series of 𝑔(𝑧):


1 𝑓(𝑠)
𝑔(𝑧) = ∮ 𝑑𝑠 = ∑ 𝑎𝑛 (𝑧 − 𝑧0 )𝑛 ,
2𝜋𝑖 𝑠 − 𝑧
𝐶1 𝑛=0

With coefficients

80
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

1 𝑓(𝑠)
𝑎𝑛 ≔ ∮ (𝑠 − 𝑧)𝑛+1 𝑑𝑠 , 𝑛 = 0,1,2, , … .
2𝜋𝑖 .
𝐶1

By the principle of deformation of path we can replace 𝐶1 by any simple closed curve 𝐶 in the
annulus that encircles 𝑧0 .

Step 2: For the negative Powers

The formula for 𝑏𝑛 in formula (2) of the theorem is obtained if we consider ℎ(𝑧).

Since z lies in the annulus and s lies on C__2, so in this case |𝑧 − 𝑧0 | > |𝑠 − 𝑧0 | and
|𝑠−𝑧0 | |𝑠−𝑧 |
hence, |
|𝑧−𝑧0 |
| <1. Unlike the proof of Taylor series, we develop a power series in |𝑧−𝑧0 |,not
0
|𝑧−𝑧0 |
|𝑠−𝑧0 |
, Continuing as in the proof of Taylor series , we get:

1 𝑓(𝑠) 1 1 1
𝒉(𝒛) = − ∮ 𝑑𝑠 = [ ∮ 𝑓(𝑠)𝑑𝑠 + ∮ (𝑠 − 𝑧0 )𝑓(𝑠)𝑑𝑠 +
2𝜋𝑖 𝐶2 𝑠−𝑧 2𝜋𝑖 (𝑧−𝑧0 ) 𝐶 2 (𝑧−𝑧0 )2 𝐶2

1
⋯+ ∮ (𝑠 − 𝑧0 )𝑛 𝑓(𝑠)𝑑𝑠 ] +𝑅𝑛∗ (𝑧)
(𝑧−𝑧0 )𝑛+1 𝐶2

1 (𝑠−𝑧0 )𝑛+1
With the last term given by: 𝑅𝑛∗ (𝑧) = ∮ 𝑓(𝑠)𝑑𝑠. As before, we can
(𝑧−𝑧0 )𝑛+1 𝐶2 𝑧−𝑠

integrate over 𝐶 instead of in the integrals on the right.

Claim: 𝑅𝑛∗ → 0 𝑎𝑠 𝑛 → ∞. This completes the proof. ∎

81
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

Theorem 2(Uniqueness of Laurent Series): The Laurent series of a given analytic function in its
annulus of convergence is unique. However, it may have different Laurent series in two annuli with
the same centre.

C. Finding Laurent Series Expansions: Two Methods

Methods to find Laurent series expansion of f(𝒛):

1. Direct method: by using Laurent series Theorem


2. Indirect method: By combining known Taylor expansions.
- Because the Laurent expansion of a function is unique whenever it exists, any
expression of this form that equals the given function in some annulus must
actually be the Laurent expansion of .
- The coefficients in a Laurent series are generally found by means other than
appealing directly to their integral representations.

Procedures For Finding Laurent Series Directly by Laurent Theorem around a point 𝒛𝟎 :

1. Identify the Singularities: Determine where the function 𝑓(𝑧) has singularities in a region
containing the point 𝑧0 about which you want the Laurent series.

2. Choose a Contour: Select a contour \(a closed path) that encloses 𝒛𝟎 and avoids any other
singularities of f(z) within the region of interest.

3. Apply the Integral Formula

4. Construct the Laurent Series

5. Determine the Validity of the Series: Specify the annulus where the Laurent series
converges and represents the function 𝒇(𝒛). This is often determined by the distances between
𝒛𝟎 and the closest singularities.

82
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

1
Example 3: Use Laurent Theorem to find Laurent series of 𝑓(𝑧) = (𝑧−1)(𝑧−2) valid in 1 <
|𝑧| < 2).

1. Singularities: 𝑓(𝑧) has singularities at 𝑧 = 1 and 𝑧 = 2.


2. Contour: We can choose a contour : We can choose a circle with radius R such that 1 <
𝑅 < 2.

3. Apply the Integral Formula:

4. Laurent Series: Combining the two series, we get:

5. Validity: This series converges on the annulus 1<|𝑧| <2.

Example 4: Use algebraic manipulations of known series expansions to find the Laurent
1
expansion of 𝑓(𝑧) = (𝑧−1)(𝑧−2) in the region 1 < |𝑧| < 2?.

Solution: Using partial fraction decompositions we get:

1 −1 1
= +
(𝑧 − 1)(𝑧 − 2) 𝑧 − 1 𝑧 − 2

−1 1 1 1 −1 −1 1 𝑛 1 1
Then 𝑧−1 = 1−𝑧 = 1 = ( 1 )= ∑∞ ∞
𝑛=0 (𝑧 ) = − ∑𝑛=0 𝑧 𝑛+1 (for |𝑧 | < 1)
𝑧( −1) 𝑧 1− 𝑧
𝑧 𝑧

1 1 𝑧 𝑛 𝑧𝑛 𝑧
Similarly 𝑧−2 = − ∑∞ ( ) = − ∑∞
𝑛=0 2𝑛+1 (for |2| < 1).
2 𝑛=0 2

Combining the two series we obtain:

1 𝑧𝑛
𝑓(𝑧) = − ∑∞ ∞
𝑛=0 𝑧 𝑛+1 − ∑𝑛=0 2𝑛+1 //

83
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

Summary on Chain of Key Theorems:

• Cauchy Integral theorems


1

• Cauchy Integral Formulas


2

• Taylor Theorem and Laurent Theorem


3

• Cauchy Residue Theorem


4

Exercises 7.4:

1. Use algebraic manipulations of known Taylor/Laurent series expansions to find Laurent


series of 𝑓 in the indicated powers (at the indicated centres)
a) 𝑓(𝑧) = 𝑧 −5 sin 𝑧 with centre 0.
1
b) 𝑓(𝑧) = 𝑧 2 𝑒 𝑧 with centre 0.
1
c) 𝑓(𝑧) = 1−𝑧 in nonegative powers of z
1
d) 𝑓(𝑧) = 1−𝑧 in negative powers of z
1
e) 𝑓(𝑧) = (𝑧−1)(𝑧−2) with centre 0.
2. M

84
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

CHAPTER 8: INTEGRATION BY THE METHOD OF RESIDUES

Residue calculus is a core part of complex analysis. Residue theory is a theory for
computing integrals by looking at certain terms in the Laurent series of the integrated function.
Recall that if a function 𝑓 is analytic inside and on a closed curve C, then ∮𝐶 𝑓(𝑧)𝑑𝑧 = 0 ; but if
𝑓 is not analytic at certain points, we need other approaches.

Residue calculus involves using a key concept called residues and a key theorem called
residue theorem (also called Cauchy residue theorem). A residue is a complex number
associated with a singularity (pole) of a complex function 𝑓. It essentially measures the
"contribution" of that singularity to the integral of 𝑓 around a closed curve encircling the
singularity. Residue theorem generalizes the Cauchy integral theorem and Cauchy's integral
formula

Residue calculus provides a powerful method for evaluating integrals, particularly those
real integrals that are difficult or impossible to solve using standard real calculus
techniques. While primarily focused on integration, the residue theorem and the concept of
residues as well as related theorems and concepts such as Argument principle, Rouche’s
Theorem and winding number do also have applications in other areas of mathematics, physics
and engineering (eg. the Nyquist stability criterion which is fundamental in electrical
engineering). For instance, it leads to formulae for the number of solutions of an equation
𝑓 (𝑧) = 0 inside a simple closed loop by just evaluating an integral whose value must be an
integer.

This chapter is presented in the following THREE sections:

8.1 Singularities and Residues


8.2 The Residue Theorem
8.3 Applications of Residues
8.4 Further Topics: Zeros, Winding Number, Rouche’s theorem and Argument Principle

85
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

8.1 Singularities, Zeroes and Residues


A. Singularities
Recall that a function 𝑓 is said to be differentiable at a point 𝑧0 if 𝑓 ′ (𝑧0 ) exist. It is said to be
analytic(holomorphic) at 𝑧0 if it has a derivative at each point in some neighbourhood 𝑁∈ (𝑧0 ) of
𝑧0 .

Question: When do we say a function 𝑓 has a singularity at 𝑧0 ?

Definition 1: We say that a function f(z) is singular or has a singularity at a point 𝑧0 if

i) It is not analytic (perhaps not even defined) at 𝑧0


ii) every neighbourhood of 𝑧0 contains some point at which 𝑓(𝑧) is analytic.
Notice the following equivalence:

𝑓(𝑧) is singular at 𝑧0 ⟺ 𝑓(𝑧) has a singularity at 𝑧0 ⟺ 𝑧0 𝑖𝑠 𝑎 𝑠𝑖𝑛𝑔𝑢𝑙𝑎𝑟 𝑝𝑜𝑖𝑛𝑡 𝑜𝑓 𝑓(𝑧)

Types (Classifications) of Singularities

Singularities can be classified as isolated singularities and non-isolated singularities. Isolated


singularities can be further classified as removable, pole and essential singularities.

86
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

Definition 2: A point 𝑧0 is said to be an isolated singular point of 𝑓(𝑧) if :

i) it is a singular point of 𝑓
ii) there exists a positive number 𝜀 such that f(z) is analytic on 𝑁𝜀 (𝑧0 ) ∖ {𝑧0 } , i.e., if 𝑓(𝑧) is
analytic at every point in some deleted neighbourhood of 𝑧0 . We call 𝑧 = 𝑧0 an isolated
singularity of 𝑓(𝑧) if has a neighbourhood without further singularities of 𝑓(𝑧).
Caution: Contrast how the words some and every are used in Definition 1 and Definition 2.

Definition 3: Let 𝑧0 be an isolated singular point of 𝑓(𝑧) and let


𝑏 𝑏
𝑓(𝑧) = ∑∞ 𝑛 1 2
𝑛=1 𝑎𝑛 (𝑧 − 𝑧0 ) + 𝑧−𝑧 + (𝑧−𝑧 )2 + ⋯ ,
0 0

be the Laurent series expansion of 𝑓(𝑧) at 𝑧0 . Then 𝑧0 is said to be :

(a) a removable singular point if 𝑏𝑛 ≠ 0 for no n.


(b) a pole if 𝑏𝑛 ≠ 0 for finitely many n.
(c) an essential singular point if 𝑏𝑛 ≠ 0 for infinitely many n.
If 𝑧0 is a pole of 𝑓(𝑧) there exists a positive integer 𝑚 such that 𝑏𝑛 = 0 for all 𝑛 > 𝑚. A
removable singularity occurs when a function can be made analytic at a point by defining a
value at that point. This happens when the Laurent series of the function at the point has no
negative powers.

Definition 4 (pole of order m and simple pole): A pole 𝑧0 of 𝑓(𝑧) is said to be

a) a pole of order m if 𝑏𝑚 ≠ 0 and 𝑏𝑛 = 0 for all 𝑛 > 𝑚. (In other words, if 𝑏𝑚 ≠ 0 and
the principal part of the Laurent series (the part with negative powers of (z - z₀)) has the
form:
𝑏1 𝑏 𝑏
𝑧−𝑧0
+ (𝑧−𝑧2 )2 + ⋯ +(𝑧−𝑧𝑚 )𝑚
0 0

b) Simple pole if it is a pole of order 1.


Remark:

1. If 𝑓(𝑧) is analytic and has a pole at 𝑧0 , then |𝑓(𝑧)| → ∞ as 𝑧 → 𝑧0.


2. A function can have an infinite number of poles, but not uncountable poles.
3. The behavior of an analytic function in a neighbourhood of an essential singularity is
entirely different from that in the neighbourhood of a pole.
4. An essential singularity is sometimes called a pole of infinite order.
5. Removable singularities are of no interest since they can be removed.

87
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

Definition 5: We say that a function, f(z), has an isolated singularity at infinity if it is analytic
𝟏
outside some bounded set. Equivalently, 𝐟(𝐳) is analytic at infinity if 𝐟( 𝐳 ) is analytic at zero.

Definition 6 (meromorphic function): A meromorphic function is a complex function with poles


but no other singularities.

B. Zeros
A zero of a function in a domain D is a 𝑧0 in D such that 𝑓(𝑧0 ) = 0. A zero has order n if not
only 𝑓 but also the derivatives 𝑓 ′ , 𝑓 ′′ , ⋯ , 𝑓 [𝑛−1] of 𝑓 are all 0 at 𝑧0 but 𝑓 [𝑛] (𝑧0 ) ≠ 0. A first-
order zero is also called a simple zero. If 𝑓(𝑧)has such a Taylor series, it has an nth-order zero at
𝑧 = 𝑧0 .

Proposition: The zeros of an analytic function are isolated; that is each of them has a
neighbourhood that contains no further zeros of 𝑧0 .
C. Residues
Recall that If 𝑓(𝑧) is analytic everywhere on C and inside C, then the integral ∮𝐶 𝑓(𝑧) 𝑑𝑧 is zero
by Cauchy’s integral theorem. On the other hand if 𝑓(𝑧) is analytic everywhere on C and inside

88
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

𝑓(𝑧) 2𝜋𝑖𝑓 [𝑛] (𝑧)


C and if C encircles 𝑧0 , then for every positive integer n the integral ∮𝐶 (𝑧−𝑧 𝑛+1
𝑑𝑧 =
0) 𝑛!
By Cauchy integral formula. From the formula for the coefficient 𝑏1 of the first negative power
of Laurent series , we obtain

∮ 𝑓(𝑧) 𝑑𝑧 = 2𝜋𝑖𝑏1
𝐶

Here 𝑧0 is a singular point of 𝑓 and it is inside C but C contains no other singular points of 𝑓.
Since we can obtain Laurent series of 𝑓 by various methods without using the integral formulas
for the coefficients , so this is a useful integration method.

Definition 7 The coefficient 𝑏1 of the first negative power of the Laurent series of 𝑓(𝑧) at 𝑧0
𝑅𝑒𝑠 𝑓(𝑧)
is called the residue of 𝒇(𝒛). It is denoted by or 𝑅𝑒𝑠(𝑓, 𝑧0 ).
𝑧 = 𝑧_0

Thus we can write:

𝑅𝑒𝑠 𝑓(𝑧)
𝑏1 = .
𝑧 = 𝑧0

Alternative Methods for Calculating Residues


To calculate a residue at a pole, we need not produce a whole Laurent series, but, more
economically, we can use the following formulas.
1. Formula for formula for the residue at a simple pole
a) Formula 1:
𝑅𝑒𝑠 𝑓(𝑧)
= 𝑏1 = lim [𝑧 − 𝑧0 ] 𝑓(𝑧)
𝑧 = 𝑧0 𝑧→𝑧0

𝑝(𝑧)
b) Formula 2: If 𝑓(𝑧) = 𝑞(𝑧) with
𝑝′(𝑧0 ) ≠ 0 and if 𝑞 has a simple zero at 𝑧0 , then
𝑅𝑒𝑠 𝑓(𝑧) 𝑝(𝑧) 𝑝(𝑧0 )
= Ref =
𝑧 = 𝑧0 𝑧=𝑧0 𝑞(𝑧) 𝑞′(𝑧0 )

Proof (Kreyszig-Page 721).


2. Residue at a Pole of Higher Order: The residue of 𝑓(𝑧) at an 𝑚𝑡ℎ order pole at 𝑧0 is

89
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

𝑅𝑒𝑠 𝑓(𝑧) 1 𝑑𝑚−1


= lim {[𝑧 − 𝑧0 ]𝑚 𝑓(𝑧)}.
𝑧 = 𝑧0 (𝑚−1)! 𝑑 𝑧 𝑚−1
𝑧→𝑧0

8.2 The Residue Theorem


Residue integration can be extended from the case of a single singularity to the case of several
singularities within the contour C via the so called the residue theorem.

Theorem [Residue Theorem]: Let 𝑓(𝑧) be analytic inside a simple closed path C and on
C, except for finitely many singular points 𝑧1 , 𝑧2 , ⋯ , 𝑧𝑘 inside C. Then the integral of
𝑓(𝑧) taken counter clockwise around C equals 2𝜋𝑖 times the sum of the residues of 𝑓(𝑧)
at 𝑧1 , ⋯ , 𝑧𝑘 :
𝑘 𝑅𝑒𝑠 𝑓(𝑧)
∮ 𝑓(𝑧) 𝑑𝑧 = 2𝜋𝑖 ∑
𝑗=1 𝑧 = 𝑧𝑗
𝐶

Proof: Let each 𝑧𝑘 be the centre of a positively oriented circle 𝐶𝑘 such that each 𝐶𝑘 is interior to
𝐶 and the 𝐶𝑘 ‘s do not intersect. Then by Cauchy Goursat theorem for multiply connected
domain,
𝑘
∮ 𝑓(𝑧) 𝑑𝑧 = ∑ ∮ 𝑓(𝑧) 𝑑𝑧
𝑗=1
𝐶 𝐶𝑘
𝑅𝑒𝑠 𝑓(𝑧)
But for each k, ∮𝐶 𝑓(𝑧) 𝑑𝑧 = 2𝜋𝑖 𝑧 = 𝑧 . Hence, the result follows. ∎
𝑘 𝑗

Example 1: Use the-Residue theorem to evaluate.


4𝑧 − 5
∮ 𝑑𝑧
𝑧(𝑧 − 1)
𝐶:|𝑧|=2
Solution:

Step 1: Identify singular points and compute their residues

The only singular points of the integrand are 0 and 1 and both are isolated and they lie in the
circle |𝑧| = 2.

Step 2: compute the residues of the singular points

Let 𝐵1 : the residue of 𝑓(𝑧) at 1

90
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

𝐵2 : the residue of 𝑓(𝑧) at 0.

To find B1 and B2 , we can use Limits or Laurent series expansions.

Method I (Using Limits)

4𝑧 − 5
𝐁𝟏 = 𝐥𝐢𝐦(𝒛 − 𝟏) ( ) = −𝟏
𝒛 →𝟏 𝑧(𝑧 − 1)

And

4𝑧 − 5
𝐁𝟐 = 𝐥𝐢𝐦(𝒛 − 𝟎) ( )=𝟓
𝒛 →𝟏 𝑧(𝑧 − 1)

Method II (Using Laurent Series expansion)


1
To find the Laurent series of 𝑓(𝑧) at 0 , start with 𝑧−1 = −(1 + 𝑧 + 𝑧 2 + ⋯ )

4𝑧−5 1 5
Then 𝑧(𝑧−1)
= 𝑧 (4𝑧 − 5)[−1 − 𝑧 − 𝑧 2 + ⋯ ] = + 1 + 𝑧 + 𝑧 2 + ⋯
𝑧

1 1
To find the Laurent series of 𝑓(𝑧) at 1 , start with 𝑧 = 1−(−(𝑧−1)) = 1 − (𝑧 − 1) + (𝑧 − 1)2 −
(𝑧 − 1)3 + ⋯ )
4𝑧−5 1 1 1
Then = (4 − 𝑧−1) = (4 − 𝑧−1)( 1 − (𝑧 − 1) + (𝑧 − 1)2 − (𝑧 − 1)3 ) for |𝑧 − 1| < 1.
𝑧(𝑧−1) 𝑧

1
The coefficient of 𝑧−1 is -1 and so 𝐵2 = −1.

Step 3: Apply the residue theorem.


4𝑧−5
∮𝐶:|𝑧|=2 𝑧(𝑧−1) 𝑑𝑧 = 2𝜋𝑖(𝐵1 + 𝐵2 ) = 2𝜋𝑖(5 − 1) = 82𝜋𝑖 //

Example 2: Use CRT to evaluate the following integral counter clockwise around any simple
closed path such that (a) 0 and 1 are inside C, (b) 0 is inside, 1 outside, (c) 1 is inside, 0 outside,
(d) 0 and 1 are outside.

4 − 3𝑧
∮ 𝑑𝑧
𝑧(𝑧 − 1)
𝐶
Answer፡ a)−6𝜋𝑖 (b)−6𝜋𝑖 (c) 2𝜋𝑖 (d)0.

91
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

Summary of Methods s to compute a contour integral ∮𝑪 𝒇(𝒛) 𝒅𝒛


- parameterize the path (feasible when C is easily described)
- Basic Rules (sum, difference, constant multiple, addition of paths)
- Fundamental Theorem of Line Integrals for complex functions
- Cauchy Integral Theorem
- The Principle of Deformation of Path
- the Cauchy integral formula
- the Cauchy’s residue theorem
- theorem of the residue at infinity : Read
Exercises 8.2:

8.3 Applications of The Residue Theorem


The Cauchy Residue theorem is useful in/for:

a. Evaluating complex line integrals


b. Evaluating certain Real Integrals (definite and improper integrals)
c. Summing certain series
d. Analysing(Locating) the Zeros of Functions (using the so called Argument Principle)
e. Proving various theorems (eg. Jordan decomposition and Cayley-Hamilton theorems).
f. Others (in physics and engineering): Read!

92
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

8.3.1 Cauchy Residue Theorem for Real Integrals


A. Definite Integrals of Rational Functions of 𝒄𝒐𝒔 𝜽 and 𝒔𝒊𝒏 𝜽
Consider definite integrals of the form
2𝜋
𝐽 = ∫0 𝐹(𝒄𝒐𝒔 𝜽 , 𝒔𝒊𝒏 𝜽) 𝑑𝜃, (1)

Where 𝐹 is a real rational function of two variables which is bounded on the given interval of
integration..

Integral (1) can be viewed as a complex line integral over the unit circle with parameterization
𝑧 = 𝑒 𝑖𝜃 , 0 ≤ 𝜃 ≤ 2𝜋. Thus The substitution 𝑧 = 𝑒 𝑖𝜃 can be used to reduce these integrals to
complex line integrals.

(Recall: In section 6.2 we evaluated complex line integrals by converting them to real integrals;
here we are reversing this procedure). (Research Question: Can we use multiple integrals, real
line integrals, surface integrals to compute real integrals in the same way?).

Setting 𝑒 𝑖𝜃 = 𝑧, we can write:

𝑒 𝑖𝜃 +𝑒 −𝜃 1 1 𝑒 𝑖𝜃 −𝑒 −𝜃 1 1
cos 𝜃 = 2
= 2 (𝑧 + 𝑧) and sin 𝜃 = 2𝑖
= 2𝑖 (𝑧 − 𝑧)

𝑑𝑧
Since 𝑑𝜃 = 𝑖𝑧, so

𝑑𝑧
𝑑𝜃 =
𝑖𝑧

𝐹 is now a rational function of z, say, 𝑓(𝑧). On the other hand as 𝜃 ranges from 0 to 2𝜋, the
variable z ranges counter clockwise once around the unit circle. Thus the given integral takes the
form:

𝐽= ∫ 𝑓(𝑧)𝑑𝑧
𝐶:|𝑧|=1

93
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

2𝜋 𝒄𝒐𝒔 𝟐𝜽
Example 1: Compute ∫0 𝑑𝜃,
𝟓−𝒄𝒐𝒔 𝜽

Solution: For the values of z lying on the unit circle : |𝑧| = 1 , we have 𝑧 = 𝑒 𝑖𝜃 , (0 ≤ 𝑧 ≤
1
𝑧 2 +𝑧 −2 𝑧+ 2𝜋 𝒄𝒐𝒔 𝟐𝜽
2𝜋) and we can write : cos 2𝜃 = 2
and cos 𝜃 = 𝑧
. It follows that ∫0 𝑑𝜃 =
2 𝟓−𝒄𝒐𝒔 𝜽
𝑧2 +𝑧−2
𝑑𝑧
∫𝐶:|𝑧|=1 . 2
𝑧+
1
𝑖𝑧
5− 𝑧
2

The singularities of 𝑓(𝑧) lying inside C are 0 (pole of order 2) and ½ ( simple pole). By the
𝜋
Residue theorem, the result is . //
6

Example 2:

B. Improper Integrals With Infinite Limits of Integration


Recall the definition of the improper integral of a real function 𝑓(𝑥) over the infinite interval
(-∞, ∞) is given by
∞ 𝑡 0
∫−∞ 𝑓(𝑥)𝑑𝑥 = lim ∫0 𝑓(𝑥)𝑑𝑥 + lim ∫𝑡 𝑓(𝑥)𝑑𝑥 ----- (1)
𝑡→∞ 𝑡→−∞
(if both limits exist)

we may couple the two independent passages to and , and write


∞ 𝑅
∫−∞ 𝑓(𝑥)𝑑𝑥 = lim ∫−𝑅 𝑓(𝑥)𝑑𝑥 ---(2)
𝑅→∞


The limit on the right is called the Cauchy principal value of the integral. ∫−∞ 𝑓(𝑥)𝑑𝑥.

Definition 1 (The Cauchy Principal Value of Improper Integral):

The number
∞ 𝑅
𝑃. 𝑉 ∫−∞ 𝑓(𝑥)𝑑𝑥 = lim ∫−𝑅 𝑓(𝑥)𝑑𝑥 --- (3)
𝑅→∞


is called the Cauchy principal value of the improper integral ∫−∞ 𝑓(𝑥)𝑑𝑥

Remarks:

94
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

∞ ∞
1. If ∫−∞ 𝑓(𝑥)𝑑𝑥 exists, so does P.V ∫−∞ 𝑓(𝑥)𝑑𝑥 and they are equal.
∞ ∞
2. The converse of remark (1) is false. (eg.: 𝑃. 𝑉 ∫−∞ 𝑥𝑑𝑥 = 0 but ∫−∞ 𝑓(𝑥)𝑑𝑥 DNE,)
∞ ∞
3. If 𝑓 is an even function, then ∫−∞ 𝑓(𝑥)𝑑𝑥 = P.V ∫−∞ 𝑓(𝑥)𝑑𝑥 .
4. If 𝑓(𝑥) is a real rational function whose denominator is different from zero for all real 𝑥
and is of degree at least two units higher than the degree of the numerator, then
∞ ∞
𝑃. 𝑉 ∫−∞ 𝑥𝑑𝑥 exists and hence (by remark 1) ∫−∞ 𝑥𝑑𝑥
The residue theorem is a powerful tool for evaluating certain types of improper integrals, but
it's not applicable to all improper integrals.

Procedures for Evaluating the Improper Integral ∫−∞ 𝒇(𝒙)𝒅𝒙 , where 𝒇(𝒙) is a certain type
of Real Rational Function
𝑝(𝑥)
1. Make sure that 𝑓(𝑥) = 𝑞(𝑥) is a real rational function such that 𝑞(𝑥) ≠ 0 for all real 𝑥 and is of
degree at least two units higher than the degree of 𝑝(𝑥).
∞ 𝒑(𝒙) 𝑅
2. Rewrite it as the limit of Cauchy principal Value: ∫−∞ 𝒒(𝒙) 𝒅𝒙 = lim ∫−𝑅 𝑓(𝑥)𝑑𝑥 . Step 1
𝑅→∞
guarantees that this is possible.
𝑝(𝑧)
3. Extend to the complex plane: Consider the corresponding complex rational function (𝑧) = 𝑞(𝑧).

4. Find the zeros of 𝑞(𝑧). Since 𝑓(𝑧) is rational, so the line integral ∫𝐶 𝑓(𝑧)𝑑𝑧 can have at most
finitely many poles in the upper half-plane. Let 𝑧0 ,…,𝑧𝑛 be the poles in the upper half plane.
5. Choose a positively oriented semi-circular contour C of radius R that lie in the upper half-plane,
that encloses all the poles and that has its centre at (0,0), as shown. Let S be the circular portion
of C.

95
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

𝑅
6. Then ∫−𝑅 𝑓(𝑥)𝑑𝑥 = ∫𝐶 𝑓(𝑧)𝑑𝑥 − ∫𝑆 𝑓(𝑧)𝑑𝑥 = 2𝜋𝑖 ∑𝑛𝑘=0 𝑅𝑒𝑠(𝑓, 𝑧𝑘 ) − ∫𝑆 𝑓(𝑧)𝑑𝑧 (By Residue
Theorem & a property of line integrals).
7. If the integral over the semicircle S goes to 0 as R→∞,
then

∫−∞ 𝑓(𝑥)𝑑𝑥 = 2𝜋𝑖 ∑𝑛𝑘=0 𝑅𝑒𝑠(𝑓, 𝑧𝑘 ) --- (4)

(Use ML-inequality to show that the limit of the integral over S is 0!).

NOTE:
1. If 𝑓is even, then


∫𝟎 𝒇(𝒙)𝒅𝒙 = 𝜋𝑖 ∑𝑛𝑘=0 𝑅𝑒𝑠(𝑓, 𝑧𝑘 ) ---(5)

2. when the residue theorem is applied to real improper integrals, the result can have an
imaginary part. Forget the imaginary part. (See Eg. 5).
∞ 𝟏 𝝅√𝟐
Example 3: ∫𝟎 𝟏+𝒙𝟒
𝒅𝒙 = 𝟒
//

∞ 𝟏 𝝅
Example 4: ∫𝟎 𝒅𝒙 = //
𝟏+𝒙𝟔 𝟑

∞ 𝒙
Example 5: Evaluate ∫−∞ 𝟖−𝒙𝟑 𝒅𝒙 Answer:

Solution:
𝑖 2𝜋 𝑖 4𝜋
𝑧
Step 1: Find the poles of 𝑓(𝑧) = (8−𝑧 3). They are 2, 𝑒 3 and 𝑒 3

Step 2: Identify poles in the upper half-plane:


𝑖 2𝜋
The poles in the upper half-plane are 𝑧1 = 2 𝑎𝑛𝑑 𝑧2 = 𝑒 3 .
Step 3: Calculate the residues: The residues are

For 𝑧1 :
𝑧 1
𝑅𝑒𝑠(𝑓, 2) = lim[𝑧 − 2] (8−𝑧 3) = − 6
𝑧→2

For 𝑧2 :

96
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

𝑖 2𝜋 𝑖 2𝜋 𝑧 1 √3
𝑅𝑒𝑠 (𝑓, 𝑒 3 ) = lim [𝑧 − 𝑒 3 ] = +𝑖
𝑖 2𝜋 (8 − 𝑧 ) 12
3 12
𝑧→𝑒 3

𝜋√3
Step 4: Apply the Residue Theorem. The value of the integral is 36
//

C. Improper Integrals With Infinite Limits of Integration (Read!)


Definition 2( Cauchy Integral value of Improper Integral with Infinite Integrand):
Let the function 𝑓(𝑥) be infinite (that is has infinite discontinuity) at 𝑥 = 𝑎 and 𝑓 be
continuous on [𝐴, 𝑎) ∪ (𝑎 , 𝐵]. Then the Cauchy principal value of the improper
𝐵
integral ∫𝐴 𝑓(𝑥)𝑑𝑥 is denoted and defined by

𝑩 𝒂−∈ 𝑩

𝑷. 𝑽 ∫ 𝒇(𝒙)𝒅𝒙 = 𝐥𝐢𝐦 ( ∫ 𝒇(𝒙)𝒅𝒙 + ∫ 𝒇(𝒙)𝒅𝒙)


∈→𝟎
𝑨 𝑨 𝒂+∈

∞ 𝒙
Example 6: Evaluate ∫−∞ 𝟖−𝒙𝟑 𝒅𝒙

∞ ∞
D. Fourier Integrals ∫−∞ 𝒇(𝒙)𝒄𝒐𝒔 𝒙 𝒅𝒙 and ∫−∞ 𝒇(𝒙)𝒔𝒊𝒏 𝒙 𝒅𝒙
The method discussed in Subsection "B" extends to integrals of the following type; they
are called Fourier integrals.

∫−∞ 𝒇(𝒙)𝒄𝒐𝒔 𝒙 𝒅𝒙 = −2𝜋 ∑𝑛𝑘=0 𝐼𝑚 𝑅𝑒𝑠 [𝑓(𝑧)𝑒 𝑖𝑠𝑧 ]

∫−∞ 𝒇(𝒙)𝒔𝒊𝒏 𝒙 𝒅𝒙 = −2𝜋 ∑𝑛𝑘=0 𝑅𝑒 𝑅𝑒𝑠 [𝑓(𝑧)𝑒 𝑖𝑠𝑧 ] (s>0) ---(6)

Where we sum the residues of 𝑓(𝑧)𝑒 𝑖𝑠𝑧 at its poles in the upper half plane.

∞ 𝒄𝒐𝒔 𝒔𝒙
Example 7: Evaluate ∫−∞ 𝒌𝟐 +𝒙𝟐 𝒅𝒙 where (𝑠 > 0, 𝑘 > 0)

𝜋
Answer: 𝑘
𝑒 −𝑘𝑠

8.3.2 Cauchy Residue Theorem for Evaluating Sums of Series

The residue method can, also be used to find sums of certain series. This is evident from the fact
that the theorem relates contour integral and sum. This requires interpreting the terms of the
series whose sum we find as residues of some "nice function".

97
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

Procedures (To find the sum of a series using the Residue Theorem)

1. Choose a complex function 𝑓(𝑧) whose poles correspond to the terms of the series.
2. Choose a contour 𝐶 that encircles the poles of 𝑓(𝑧).
3. Calculate the residues of the function at each pole inside the chosen contour.
4. Use the Residue Theorem to evaluate the contour integral. U
5. Solve for the series sum.
(Usually we begin by choosing a contour 𝐶𝑁 that encircles only N poles for an arbitrary natural
number N, then we find the sum of the residues of the N poles. This gives us the Nth partial sum
of the given series. We proceed to finding the contour integral over N and take its limit as N
approaches to infinity. At the end we equate this limit with the sum of residues and figure out the
answer. ).
1 𝜋2
Example 8: Use the method of residues to prove that the sum of the series ∑∞
𝑘=−∞ is .
𝑛2 6

Solution:
πcot (𝜋𝑧)
Step 1: Choose 𝑓(𝑧) = 𝑧2
. (Why this choice?)

Step 2: Identify the poles of f (z) . It has a pole of order 3 at 0 and simple poles at all nonzero
integers.
1
Step 3; Calculate the residues of 𝑓(𝑧) at all poles. The residue at 𝑧 = 𝑛 is 𝑛2 for every nonzero
𝜋2
integer n. The residue of 𝑓(𝑧) at 0 is − 3
.

Step 4: Choose as a contour 𝐶𝑁 a square of side length 2N+1 as shown in the figure. It encloses
2N+1 poles. Then find ∫𝐶 𝑓(𝑧)𝑑𝑧 by parameterizations. We get lim ∫𝐶 𝑓(𝑧)𝑑𝑧 = 0.
𝑁 𝑁→∞ 𝑁

1 𝜋2
Step 5: Apply the residue theorem. It gives ∑∞
𝑘=−∞.𝑘≠0 𝑛2
− 3
= 0. The desired result follows
from this.

98
AAU Dep. Of Math. Applied Math. III Part III: Complex Analysis by Tadesse B. (PhD)

1
Example 9: Use the method of residues to prove that the sum of the series ∑∞
𝑘=1 𝑛2 +𝑎2 is
𝜋 1
(2𝑎) 𝑐𝑜𝑡ℎ(𝜋𝑎) − 2𝑎2 . Verify the result by using Fourier expansions.

1
Solution: Choose 𝑓(𝑧) = 𝜋 cot(𝑎𝜋𝑧) 𝑧 2 +𝑎2

The poles of 𝑓(𝑧) are the poles of the first function (all integers k) and the second function
(z=±𝑖𝑎 ).

The residues of the poles are:


1 −𝜋 −𝜋
𝑅𝑒𝑠(𝑓, 𝑘) = 𝑘 2 +𝑎2 , 𝑅𝑒𝑠(𝑓, 𝑖𝑎 ) = 2𝑎
𝑐𝑜𝑡ℎ(𝑎𝜋) and 𝑅𝑒𝑠(𝑓, −𝑖𝑎 ) = 2𝑎
𝑐𝑜𝑡ℎ(𝑎𝜋).

Exercises 8.3:

99

You might also like