Unit 10 Functions of a Complex Variable – Analyticity

UNIT 10
FUNCTIONS OF A
COMPLEX VARIABLE –
ANALYTICITY
Structure

10.1 Introduction 10.4 Some Elementary Complex

Expected Learning Outcomes Functions
10.2 Complex Variables The Exponential Functions e z
10.3 Complex Analytic Functions Trigonometric Functions
Limit, Continuity and Derivative of a Hyperbolic Function
Complex Function The Logarithmic Function
Cauchy-Riemann Conditions Gamma Function
Analytic Functions 10.5 Singularities of a Complex Function
Harmonic Functions 10.6 Summary
10.7 Terminal Questions
10.8 Solutions and Answers

10.1 INTRODUCTION
From the basic calculus course in earlier classes, you are familiar with the
concepts of limit, continuity and differentiability of the functions of a real variable.
We now turn to a study of these aspects for functions of a complex variable. As
pointed out in the Block Introduction, complex analysis is a powerful tool as it
facilitates solutions of many problems in different areas of physics, engineering
and other allied sciences. In your earlier Physics course on Oscillations and
Waves, you learnt how conveniently we could obtain the resultant of n
superposed harmonic oscillations using complex numbers. Similarly, calculation
of the intensity distribution of Fraunhofer diffraction from N identical slits became
possible in a compact form using the complex numbers. While analysing a.c.
circuits, this theory simplifies derivations considerably. For such elementary
problems, a basic knowledge of complex numbers is sufficient. However, for
slightly involved but interesting practical problems in fluid dynamics, the theory of
heat, electrostatics and quantum mechanics, a detailed knowledge of the theory
of complex analytic functions is necessary.

We begin this unit by introducing complex numbers. You will then learn the
properties of functions of a complex variable; their limit and continuity. You will
note that differentiability of a function of complex variable needs careful attention
as the conditions for the existence of the derivative of a function of a complex
variable are different from those of a real variable. In this connection you will 7
Block 3 Complex Analysis
learn to derive the Cauchy-Riemann conditions and state the conditions under
which a function of a complex variable is analytic.

Some elementary functions of a complex variable and their properties are

discussed in Sec. 10.4. We have also discussed the Gamma function. And
singularities of a complex function form the subject matter of discussion of
Sec. 10.5.

Expected Learning Outcomes

After studying this unit, you should be able to:
 calculate the derivative of a function of a complex variable;

 identify complex analytic functions and domains of their analyticity;

 work with elementary complex functions; and

 locate the singularities of a complex function.

10.2 COMPLEX VARIABLES

You know that a complex number is an ordered pair of two real numbers a and
 
b written as a + ib, where i   1 . Similarly, a complex variable is an ordered
pair of two real variables

z = x + iy

The ordering is important as a + ib is not equal to b + ia and x + iy  y + ix, in

general.

Sometimes we use j instead of i. It is called the imaginary unit. You will recall
that x is real part of z and y is imaginary part of z:

x = Re (z)
y = Im (z)

It is useful to represent complex numbers and variables graphically. For this

we generally plot x, the real part of z as the abscissa, and y, the imaginary
part of z as the ordinate and have the complex xy-plane, as shown in
Fig. 10.1(a). We can plot a complex number as a point in the plane if we
assign specific values to x and y. The magnitude of the complex number is the
distance from the origin to the point representing the number.

In polar coordinates you can write

x = r cos

and
Fig. 10.1: a) A complex y = r sin
number in rectangular
and polar coordinate Then we get a useful polar representation of a complex number:
systems; b) complex
conjugate points. z = x + iy = r (cos + i sin)

8 = r ei
Unit 10 Functions of a Complex Variable – Analyticity
where

ei = cos + i sin

is called the Euler identity. In this representation, r is called the modulus of z

and the angle  is the argument or phase of z:

r  z  x2  y 2

  tan1( y / x )

Fig. 10.1(a) depicts a complex number in rectangular and polar coordinate

systems.

You should note here that the choice of coordinate system is dictated by
mathematical convenience and/or geometry. For instance, addition and
subtraction of complex numbers are easier in the Cartesian representation,
whereas multiplication, division, powers and roots are easier to handle in polar
form.

The complex conjugate z* of a complex number z is obtained by changing i to

i everywhere in its expression. Thus z*  x  i y. Notice that

zz*  ( x  i y )( x  i y )

 x2  y 2  r 2

or

r  zz *

Hence zz *  z is the magnitude of z.

You should note that a complex variable and its conjugate are mirror images
of each other reflected in the x-axis (Fig. 10.1b).

So far we have confined ourselves to complex variables. But complex analysis

concerns itself with functions of complex variables or complex function. In
particular we will focus on complex analytic function. You will learn about
these in the following section.

10.3 COMPLEX ANALYTIC FUNCTIONS

A function of a complex variable, f (z) is a prescription for assigning a unique
complex number to each z. The set of values for which the function is defined
is called the domain of z.

If you express f (z) in terms of x and y by putting x + iy for z, the complex

function can be expressed in terms of real and imaginary parts; each being a
function of x and y:

f (z) = f (x + i y) = u(x, y) + iv(x, y) (10.1)

where u(x, y) and v(x, y) are respectively the real and imaginary parts of f (z).

We illustrate this through an example. 9

Block 3 Complex Analysis
Example 10.1

Express f ( z )  z2 in the form u(x, y) + iv(x, y).

Solution : f (z)  z 2  ( x  iy ) 2  x 2  (iy ) 2  2x(iy )

 x 2  y 2  2i xy

Hence u( x, y )  x 2  y 2 and v ( x, y )  2 xy .

10.3.1 Limit, Continuity and Derivative of a Complex

Function
Neighbourhood

The set of all points z such that z  z0  , where  is any given positive
number, is called a -neighbourhood of the point z0 . When the point z0 is
omitted from the set, i.e., 0  z  z0  , the set of points is called a deleted
-neighbourhood. Geometrically, a -neighbourhood is represented by all
points lying inside a circle of radius  drawn around z0 in the complex plane.

Limit

A function f (z) is said to have the limit  as z approaches z0 , if for a positive

real number , (no matter how small, but not zero), there is a positive real
number  depending on  and z0 such that f (z)     for all z for which
0 < z  z0  . This means that f (z) can be made to approach  as close as
desired by bringing z close to z0 (Fig. 10.2). We write this condition as

lim f (z)  .
z z0

Fig. 10.2:

Continuity

In the definition of limit, you must have noted that z may approach z0 from
any direction in the complex plane. Moreover, nothing is said about the value
of f(z) at z0 . In particular, f ( z0 ) need not be equal to . But if f ( z0 ) = , the
function f (z) is said to be continuous at z0 :

lim f (z)  f (z0 )  

zz0

This definition implies that f (z) may be said to be continuous in some domain
if it is continuous at each point of that domain.
10
Unit 10 Functions of a Complex Variable – Analyticity
Having established functions of a complex variable and discussed their limit as
well as continuity, we now proceed to calculate the derivative of a complex
function.

Derivative of a complex function

A function f (z) is said to be differentiable at the point z = z0 if the limit

f ( z0  z )  f ( z0 )
lim
z z0 z

exists. Before we proceed further, let us pause for a moment and recall what is
implied by the definition of a limit. It implies that f (z) is defined in a
neighbourhood of z0 . Moreover, z may approach z0 from any direction.
Hence, differentiability at z0 means that f ( z0 ) has a unique value independent
of the path followed to approach this point.

The derivative of a function of a complex variable f (z) is defined in the same

way as that for a function of a real variable:
df ( z ) f ( z  z )  f ( z )
 f ( z )  lim (10.2)
dz z 0 z

If x and y are increments in the variables x and y, respectively, we can

write z = x + iy. Then z  0 means that the real part x and the
imaginary part y each must tend to zero separately. There are an infinite
number of ways of taking this limit. Refer to Fig. 10.3 and consider a point z at

Fig. 10.3: The point z is approached in many different ways as z  0.

which you are calculating the derivative of the function f(z). You take another
adjacent point z + z. Depending upon the arg z, the vector z will orient in
the complex plane. So when the limit z 0 is considered, you may think that
you can approach z from all possible directions. In other words, z  0, but
the arg z can have arbitrary values. Then according to Eq. (10.2), the limit
may assume an infinite number of values depending on the value of arg z
chosen (which is any value between 0 to 2). This is very unsatisfactory. It is
therefore important to know as to under what conditions the limit given by
Eq. (10.2) does not depend on the path chosen to approach z and is therefore
unique.

To enable you to appreciate the problem better, we consider a specific

example. Let f(z) = x + 3iy. Then 11
Block 3 Complex Analysis
f ( z  z)  f ( z ) x  x  3i ( y  y )  ( x  3iy )

z x  i y

x  3i y

x  i y

If we first put y = 0, and let x  0, the limit is 1. But if we let x = 0, and

y  0, the limit is +3. That is, we obtain a different value for each path and
the limit is not unique. Then we say that the derivative of f(z) does not exist.

This example clearly illustrates that differentiability of a complex function is a

rather stronger requirement than in real analysis.

SAQ 1

Show that the function f (z)  z 2 is differentiable only at z = 0.

10.3.2 Cauchy-Riemann Conditions

Consider a function f (z) whose first derivative f (z) exists. If we write from
Eq. (10.1) f ( z )  u( x, y )  i v ( x, y ) and note that

z  z  x  i y  ( x  i y )  ( x  x )  i ( y  y )

then we find that

f ( z  z )  u( x  x, y  y )  i v ( x  x, y  y )

df ( z ) f ( z  z )  f ( z )
  f ( z )  lim
dz z 0 z

u  x  x, y  y   iv  x  x, y  y   u  x, y   iv  x, y 
 lim x  i y
x 0
y 0

(10.3)

In evaluating this double limit, you can take x  0 and y  0 in any order.
But the derivative f (z ) should be independent of the path you take in making z
approach zero. Refer to Fig. 10.4. It shows two different paths. Let us now find
the conditions necessary for the existence of f (z ) .

Fig. 10.4: Two alternative approaches to the point z.

12
Unit 10 Functions of a Complex Variable – Analyticity
Path 1
Let y  0 so that z  0 parallel to the x-axis. With z  x , we can write
Eq. (10.3) as

u  x  x, y   i v  x  x, y   u  x, y   i v  x, y 
f ( z )  lim
z 0 x

Now let x  0. The expression for f (z ) takes the form

f (z)  lim  u( x  x, y )  u( x, y )   i  v ( x  x, y )  v ( x, y 

   
x 0 x   x 

u( x  x, y )  u( x, y )   v ( x  x, y )  v ( x, y ) 
 lim    i lim  
x 0  x  x 0  x 

Since f (z ) exists, we can safely assume that the real limit is non-zero. Hence,
we obtain

u v
f ( z )  i (10.4)
x x

u v
where and are the partial derivatives of u and v respectively with
x x
respect to x.

Path 2
Put x  0 so that z  0 parallel to the y-axis. Then z becomes iy and
let y  0 to obtain

 u  x, y  y   i v  x, y  y   u  x, y   i v  x, y  
f ( z )  lim  
y 0  i y 

 u  x, y  y   u  x, y    v  x, y  y   v  x, y  
 lim    lim  
y 0 i y  y 0  y 

1 u v u v
   i  (10.5)
i y y y y

You should note that the existence of the derivative f (z ) implies the existence
of four partial derivatives. Further, uniqueness of the derivative requires that
the expressions for f (z ) be equal:
u v u v
i  i 
x x y y

By equating the real and the imaginary parts, we get

u v
 (10.6a)
x y

and
u v
 (10.6b)
y x 13
Block 3 Complex Analysis
These basic relations are known as Cauchy-Riemann conditions. These
were derived by Cauchy and used extensively by Riemann. We may now
summarise the important result of this sub-section.

If f ( z )  u( x, y )  i v ( x, y ) is defined and is differentiable in the neighbourhood

of a point z0 , then f (z ) exists at a point z0 , if Cauchy-Riemann conditions are
satisfied and the first partial derivatives of u and v are continuous at z0 .

For a polynomial function of degree n,

f ( z)  a0  a1z  a2z2  ... an zn

it follows that

f (z)  a1  2a2z  ... n an zn 1.

That is, f (z ) exists everywhere for a polynomial function of z.

Conversely, if the Cauchy-Riemann conditions are satisfied and the partial

derivatives of u(x, y) and v(x, y) are continuous, the derivative f (z ) exists.

10.3.3 Analytic Functions

A function f (z) of a complex variable z is said to be analytic at the point z0,

if it is differentiable at that point as well as at all points in its neighbourhood.
The function f (z) = z* is not an analytic function of z as its derivative exists
no-where in the complex plane. To enable you to appreciate this statement,
we note that by definition

df ( z ) f ( z  z )  f ( z )
 lim
dz z  0 z

If this limit exists independent of the manner in which z approaches zero and
is unique, we say that the function is analytic. Otherwise, the function will be
non-analytic. Here

dz* ( z  z )*  z *
 lim
dz z 0 z

( x  i y  x  i y ) * ( x  i y ) *
 lim
x 0 x  i  y
y 0

x  i y  x  i y  ( x  i y )
 lim
x 0 x  i y
y 0

x  i y
 lim
x 0 x  i y
y 0
14
Unit 10 Functions of a Complex Variable – Analyticity
dz * x
From this you will note that if we take y  0, then  lim  1. On the
dz x 0 x

other hand, if we put x  0, dz*  lim  i  y  1. Since the required limit is
dz y 0 y
not unique, i.e. it depends on the path in which z  0, the derivative does not
exist and f (z) = z* is non-analytic.

Next we consider the function f (z)  z 2 . While solving SAQ 1, you have
shown that it is non-analytic except at the origin. Now we use Cauchy-
Riemann equations to illustrate this point. From Sec. 10.2 you will recall that

z 2  zz*  x 2  y 2

so that

u( x, y )  x 2  y 2 and v(x, y) = 0

u v
  2x, 0
x y

and

u v
 2y, 0
y x

From these results, you will note that the Cauchy-Riemann equations will hold
only if x and y are zero. This is true only at the origin and not in the
neighbourhood, howsoever small, of the origin. That is, f (z)  z 2 is
differentiable only at z = 0 and not at any point in the neighbourhood of
z0  0. So z 2 is not an analytic function at z0  0.

Next, let us consider the function f (z)  z2. We can write it as

f (z)  z2  ( x  i y )2  x 2  y 2  i 2xy . Therefore, u( x, y )  x 2  y 2 and
v ( x, y )  2xy . The Cauchy-Riemann equations give

u v
 2x,  2x
x y

and

u v
 2y,   2y
y y

Since the Cauchy-Riemann equations are satisfied at all points in the xy-
plane, the function z 2 is analytic everywhere.

Analytic functions hold the centre stage in complex analysis. You will learn
more about these functions as you progress along this course. 15
Block 3 Complex Analysis
A function f(z) that is analytic at every point of the z-plane is called an entire
function. You can easily verify that every polynomial in z is an entire function.
You will come across other examples of entire functions in Sec. 10.4.

10.3.4 Harmonic Functions

You have studied the Laplace's equation in electrostatics in PHE-07 as well as

in the PHE-05 course on differential equations. Starting from Cauchy-Riemann
conditions, we now show that the function u(x,y) and v(x,y) separately satisfy
the Laplace's equation in x and y.

To proceed, we differentiate Eq. (10.6a) with respect to x, to obtain

 2u  2v

x 2 x y

provided these second derivatives exist. In fact, u(x,y) and v(x,y) have
continuous partial derivatives of all orders since the derivative of an analytic
function f(z) = u + iv is also analytic.

Similarly, by differentiating Eq. (10.6b) with respect to y, we get

 2u  2v

y 2 y x

Since the mixed second order derivatives of these functions must be equal,
we have

 2v  2v

x y y x

Using this result in the preceding two equations, we obtain the Laplace's
equation for u:

 2u  2u
 0 (10.7a)
x 2 y 2

By similarly differentiating Eq. (10.6a) with respect to y and Eq. (10.6b) with
respect to x, we get Laplace's equation for v:

 2v  2v
 0 (10.7b)
x 2 y 2

So we find that the real and imaginary parts of a complex function that is
analytic in a domain satisfy Laplace's equation. In fact, a function that has
continuous partial derivative of the second order and satisfies the Laplace's
equation is said to be a harmonic function in the variables x and y. For an
analytic function f ( z )  u( x, y )  iv ( x, y ), both u and v are harmonic functions;
one of these is said to be the conjugate harmonic of the other harmonic
function.
16
Unit 10 Functions of a Complex Variable – Analyticity
To illustrate these concepts, we will now like you to go through the following
example.

Example 10.2

Show that u = 2x (1y) is harmonic in some domain. Calculate its harmonic

conjugate v.
Solution : Since

u = 2x (1y)

u  2u
 2(1  y ) and 0
x x 2

Similarly

u  2u
 2x and 0
y y 2

On combining these results, we find that

 2u  2u
 0
x 2 y 2

So u is a harmonic function.

To find its harmonic conjugate, we use the Cauchy-Riemann condition given

by Eq. (10.6a):

u v
 2(1  y ) 
x y

Integrating v with respect to y, we obtain

v  2y  y 2  g ( x ) (i)

Again

u v
 2x  
y x

Integrating v with respect to x, we get

v  x 2  h( y ) (ii)

By equating (i) and (ii), we get

2y  y 2  g( x )  x 2  h( y )

Comparison of terms containing x and y leads us to

h( y )  2y  y 2 and g( x )  x 2

 v ( x, y )  x 2  2y  y 2 17
Block 3 Complex Analysis
On combining this result with the given expression for u, you can construct the
complex function:

f (z)  u  i v  2x(1  y )  i ( x 2  2y  y 2 )

 2x  2xy  i ( x 2  y 2 )  2i y

 2( x  iy )  i ( x 2  (iy )2  2xi y )

 2z  i ( x  i y )2

 2z  iz 2

You may now solve an SAQ.

SAQ 2

Obtain the analytic function whose real part is u( x, y )  e x cos y.

10.4 SOME ELEMENTARY COMPLEX FUNCTIONS

You are familiar with many important elementary functions involving a real
variable e.g. polynomial, exponential, trigonometric, logarithmic etc. and their
properties. You may now like to learn about such functions when the
independent variable is extended to the complex domain. You will note that
some of the elementary functions have interesting properties, which find useful
applications in physical sciences and engineering.

10.4.1 The Exponential Function e z

The exponential function e z is defined by the same power series in z as e x is
represented by real x:
z z 2 z3
ez  1     ...
1! 2! 3!
zn
 1  n! (10.8)
n

From Cauchy-Riemann equations, you can show that e z is an analytic, in fact

an entire function in the z-plane.
The most important property of the exponential function is the functional
equation
e z1 e z2  e z1  z2

To establish this, we note that

z2 z13
e z1  1  z1  1   ...
2! 3!
and
z2 z23
e z2  1  z2  2   ...
18 2! 3!
Unit 10 Functions of a Complex Variable – Analyticity
On multiplying these expressions, we have
 z2 z2 
e z1e z2  1  ( z1  z2 )   1  z1 z2  2   ...
 2! 2! 

 z n z n 1z2 z1z2n 1 z2n 

  1  1  ...     ...
 n! (n  1)! (n  1)! n! 

(z1  z2 )2 (z  z )n
 1  (z1  z2 )   ...  1 2  ...
2! n!
 e z1  zz

If z1  z and z2  z, then it readily follows that e z . e z  e z  z  e0  1 holds

even for complex z.
1
 ez 
ez

 ez  0

Example 10.3

d z
Show that e  e z.
dz
Solution : ez  e x iy  e x eiy  e x (cos y  i sin y )

 u( x, y )  i v ( x, y )

where we have put u( x, y )  e x cos y and v ( x, y )  e x sin y. From Eq. (10.4) you
u v
will recall that f ( z )  i . Here
x x
u v
 e x cos y and  e x sin y
x x

d z u v
 f ( z )  e  i  e x (cos y  i sin y )
dz x x

 e x eiy  e x  iy  e z

10.4.2 Trigonometric Functions

The geometrical definition of sin  (opposite side/hypotenuse) in a right-angled
triangle loses its meaning, when angle  is replaced by a complex number.

We define sin z and cos z in terms of power series:

z3 z5 z7
sin z  z     ... (10.9)
3! 5! 7!
and

z2 z 4 z6
cos z  1     ... (10.10)
2! 4! 6!
19
Block 3 Complex Analysis
Let us replace z by iz in Eq. (10.8). Then we obtain

iz z 2 iz3 z 4 iz 5 z 6
eiz  1        ...
1! 2! 3! 4! 5! 6!

 z2 z4 z6   z3 z5 
 1     ...   i  z    ...
 2! 4! 6!   3! 5! 

= cos z  i sin z (10.11a)

That is, Euler’s formula is valid for complex numbers.

From Euler's formula, it follows that exp (2 i) = 1. Further, if we let z = z, we
can write

e iz  cos ( z )  i sin ( z)

You will recall that cos z is an even function of z whereas sin z is an odd
function of z. Therefore, we have cos (z) = cos z and sin (z) = sin z.
Hence, above equality takes the form

e iz  cos z  i sin z (10.11b)

On combining Eqs. (10.11a) and (10.11b), we find that

1 iz
cos z  (e  e iz )
2

and

1 iz
sin z  (e  e iz ) (10.12)
2i

Both sin z and cos z are entire functions as they are linear combinations of
eiz and eiz, which are entire functions.

You can easily show that trigonometric addition formulae also hold good with
complex arguments (SAQ 3):

i) sin ( z1  z2 )  sin z1 cos z2  cos z1 sin z2 (10.13a)

ii) cos ( z1  z2 )  cos z1 cos z2  sin z1 sin z2 (10.13b)

iii) sin ( z1  z2 )  sin z1 cos z2  cos z1 sin z2 (10.13c)

iv) cos ( z1  z2 )  cos z1 cos z2  sin z1 sin z2 (10.13d)

SAQ 3

Prove Eq. (10.13).

By Putting z1  z2  z in (iv), you will note that it reduces to

20 cos2 z  sin2 z  1 (10.13e)

Unit 10 Functions of a Complex Variable – Analyticity
It is important to note that unlike the inequalities sin x  1and cos x  1
(x real), a similar result does not follow for complex argument. To see this, let
 be any positive angle. Then

1 
cos (i )  (e  e  )  e 
2

as e   0. Thus cos (i) is real and can attain large values if  is not small.

SAQ 4
d sin z d cos z
Show that  cos z and   sin z.
dz dz

Using the definition of cos z, we can write

1 ix  y
cos z  cos ( x  i y )  cosi ix  y   (e  eix  y )
2

1 y 1
 e (cos x  i sin x )  e y (cos x  i sin x )
2 2

 cos x
e y  ey   i sin x e y  ey 
2 2
 cos x cosh y  i sin x sinh y (10.14a)

Similarly you can prove that

sin z  sin( x  iy )  sin x cosh y  i cos x sinh y (10.14b)

The hyperbolic sine and cosine functions are defined by the formulae

e y  ey
sinh y 
2

and

e y  ey
cosh y =
2

As these functions are connected with the hyperbola, the name hyperbolic has
been attributed to this class of functions. This follows from the formula

cosh2 x  sinh2 x  1

These relations are useful in numerical computation of cos z and sin z.

Zeros of sin z and cos z

If f (z ) is zero for a value of z  z 0 , i.e. f ( z0 )  0, then z0 is said to be a zero

of the function f (z). For sine function the condition for a zero is sin z = 0.

From Eq. (10.14b), this condition reduces to sin x cosh y = 0 and

cos x sinh y  0 to be satisfied simultaneously. For x and y real, cosh y > 1
and the first condition gives sin x = 0 which occurs for x = 0,  ,  2,  3, … 21
Block 3 Complex Analysis
For these values of x, cos x does not vanish. So it is also required that
sinh y = 0. From the definition of sinh y, you will note that sinh y = 0 is satisfied
only for y = 0. Therefore

sin z = 0 occurs for x = 0,  ,  2, ... and y = 0 or when

z = 0 or z = n; n = 1, 2, 3, ...

SAQ 5

Show that cos z = 0 occurs when z   (2n  1) ; n  1, 2, 3, ...
2

From the zeros of sin z and cos z you will note that the singularities of tan z

occur at z   (2n  1) , where cos z = 0, but sin z is not zero. tan z is
2
analytic at all other points.

10.4.3 Hyperbolic Functions

The hyperbolic sine and cosine functions with complex argument are defined
in the same way as they are defined with the real argument:
e z  e z
sinh z 
2
e z  e z
cosh z 
2
and
sinh z
tanh z 
cosh z
Other hyperbolic functions  coth z, sech z, cosech z  are the reciprocals of
tanh z, cosh z and sinh z respectively. As e z and e  z are entire functions, it
follows that sinh z and cosh z are also entire functions. tanh z is analytic
everywhere except at the zeros of cosh z. Derivatives of hyperbolic functions
can be easily calculated using their definitions. Thus
d sinh z
 cosh z
dz
d cosh z
 sinh z
dz
d tanh z
 sech2z
dz
d coth z
and   cosech2z
dz
You may work out the derivatives of other hyperbolic functions with little effort.
An important identity is
cosh2 z  sinh2 z  1

22 Verify it yourself using the definitions of cosh z and sinh z.

Unit 10 Functions of a Complex Variable – Analyticity

10.4.4 The Logarithmic Function

For a given real positive number x, logarithm to the base e, written as ln x, is a
number u such that

eu  x (10.15)

You should note that only one real number u satisfies this equation for a given
x. So ln x is a single-valued function of x.

By analogy with ln x, logarithm of a complex variable z, ln z is defined as a

function w such that

ew  z (10.16)

Writing w = u + i v in the above relation, we find that

ew  eu iv  eueiv  eu (cos v  i sin v ). So Eq. (10.16) takes the form

eu (cos v  i sin v )  z (10.17)

Remembering the polar form of z(  rei ), it follows that v is one of the values
of arg z and eu  z .

 u  ln z

A solution of Eq. (10.16) gives a value of logarithm of z:

w  ln z  ln z  i argz

You will recall that arg z has an infinite number of values,

arg z  p  2n; n  0,  1,  2, ... where p is the principal value of arg z
defined by    p  .

Since the function ln z has an infinite number of values differing by multiples of

2i for a given complex variable z, we say that natural logarithm is an infinitely
many-valued function.

If we restrict to a particular choice of arg z, we define a particular ln z function;

a branch of the logarithm. The most important branch is the principal value of
ln z and is denoted by ln z. This is obtained by giving arg z its principal
value    arg z    . Thus

Ln z  ln z  i,    

Note that ln z is identical with the ordinary logarithm when z is real and
positive.

Branch cut and branch point

Consider now the behaviour of ln z along the negative real axis. Take two
points in the complex z-plane: z1  rei ( ) and z2  rei (  ) which approach
the negative real axis from above and below respectively as   0. Writing

f (z) = ln z = ln |z| + i ,  <  <  23

Block 3 Complex Analysis
we find that

lim f ( z1)  ln r  i, r  z

 0

and

lim f z2   ln r  i
 0

Thus ln z has different limits as the negative real axis is approached from
above and from below. ln z is not continuous along the negative x-axis and so
is not differentiable at any point on it. Also ln z is not defined for z = 0. Thus all
the points on the negative x-axis including z = 0 are singular points of ln z
(Fig. 10.5).

Fig. 10.5: Singularities of ln z on negative x-axis.

Let us now introduce a cut extending from  to 0. Then it will not be possible
to cross this line. Next we define a bounded domain in the complex plane
excluding the cut and given by  <  < . Then ln z will be single-valued and
analytic in this domain. Negative x-axis representing the cut is called the
"branch cut" for the principal logarithmic function.

Depending on the choice of the value of arg z, we can think of other branches
of ln z. If we define ln z = ln r + i , where 0    0  2, it corresponds to a
different branch of logarithmic function and the straight line   0 from the
origin defines the branch cut in this case.

For all the branches of ln z, the point z = 0 is a common singular point which is
called the branch point of the logarithmic function.

10.4.5 Gamma Function

z can be real or The Gamma function Γ(z) is a function of a real variable z defined by the
complex. integral formula


( z )  t z 1e  t dt (Re z  0) (10.18)
0

The integral converges absolutely for Re(z) > 0. It is called the second Euler
Integral and these improper integrals have several applications in various
branches of science.
24
Unit 10 Functions of a Complex Variable – Analyticity
For z > 0, the value of ( z  1) using Eq. (10.18) can be expressed as :

 a


( z  1)  t z e  t dt  lim
a   t ze t dt
0 0

Integrating by parts with the exponential function as first function, we can write

 a 
( z  1)  lim t z e  t   zt z 1( e  t )dt 

a
a   0 
 0

a
 lim  a z e  a   z lim  t z 1e t dt
a  a 
0

Since lim {e a }  0 and using the definition of Gamma function from

a 
[Eq. 10.18], we can write the second term as
( z  1)  z( z ) (10.19)

Now, on substituting z = 1 in Eq. (10.18), the value of (1) can be obtained as

a a
(1)   e  t dt  lim
a   e t dt
0 0

a
(1)  lim {e  t )  lim (1  e  a )
a  0 a 

Applying the limits, (1)  1. (10.20)

Using the value of (1) in Eq. (10.19), we can calculate the values of
(2), (3) etc. as:

(2)  1! , (3)  2(2)  2.1!  2! , (4)  3(3)  3.2!  3!

and so on.

Using Eq. (10.19), and taking z = x, where x is a natural number. Thus, for a
given x, we get :
( x  1)  x! (10.21)

Eq. (10.21) is a fundamental property of Gamma function.

Fractional values of (x)

Let us calculate the value of  . Putting the value of x  in Eq. (10.18),
1 1
2 2
we get

   t 1/ 2e t dt
1
2 
0

Substituting the values of t  y 2, dt  2y dy , the above equation can be

rewritten as 25
Block 3 Complex Analysis

 1

   2 e  y dy
2
(10.25a)
 
2
0

Since y is a random variable, it can be replaced by z in the Eq. (10.25a) to

obtain

   2 e  z dz
1

2
(10.25b)
2
0
Multiplying the above integrals [Eq. 10.25(a) and Eq. 10.25(b)], we get the
relation:

2 
  1 
 e ( y
2 z 2 )
   4 dy dz
  2 
0

Changing the above integral into polar coordinates (r, ) and

y  r cos , z  r sin , dydz  rdrd

2  2
  1 
   4  rer  d
2
dr
  2 
r 0  0

Substituting u  r 2, du  2r dr , the above equation can be rewritten as

 
2  2
 1   2     2(0  1)   
  
  2   eudu  d  2(eu )  
2 2
u 0  0
0

   
1
Thus (10.26)
2

SAQ 6

Obtain the values of Gamma functions   and  .

3 5
a)
2 2

1
b) Find the value of  .
 2 

10.5 SINGULARITIES OF A COMPLEX FUNCTION

In Unit 3 of the PHE-05 course on Mathematical Methods in Physics-II, you
learnt that a differential equation of the form

y   p( x )y   q( x )y  0

is said to have a singular point at x  x0 if p(x) or q(x) are not analytic at

x  x0 . And the singular point x  x0 is said to be regular when p(x) and/or
26
Unit 10 Functions of a Complex Variable – Analyticity

q(x) diverge but lim ( x  x0 ) p( x ) and lim ( x  x0 )2 q( x ) remain finite. If a

x  x0 x  x0
singular point is not regular, it is called irregular. You will now learn these
concepts in the context of a complex function.

Suppose that f (z) is a single-valued function of a complex variable. If f (z) is

not analytic at a point z  z0 , then z0 is called a singular point or singularity of
f (z). To enable you to grasp the idea of singularity, we take an example.
Consider a simple function f (z)  z2  1. Its derivative f (z ) at a point z is 2z
and exists for all finite values of z. So f (z ) is an entire function. Now consider
z2  1 2
the function g ( z)  . At z = 1, g(1) =   which is undefined and g(z)
z 1 0
does not exist at z = 1. That is, the function is not analytic at z = 1. This point
is a singularity of g(z). But it is an isolated singularity because there is no
other singularity around z = 1.

Definition

The point z  z0 is called an isolated singularity of f (z) if around z0 we can

always draw a circle of suitable radius  > 0, i.e. z  z0  , such that no
singular point other than z0 exists inside the circle. If no such  exists, then
z0 is called a non-isolated singularity.

For the function g(z) defined above, we note an interesting result. Though
z  1 is a singularity of the function, lim ( z  1) g ( z )  2  0 exists. Therefore,
z 0
z0  1 is called a simple pole of g (z ).

If z0 is a singularity of f (z ) but lim ( z  z0 )n f ( z )  A  0 exists (where n is a

z z0
positive integer), then we say that the singularity of f (z ) at z0 is a pole of
order n.

Removable singularity

If z0 is a singularity of f (z ) but lim f (z) exists, then z0 is called a removable

zz0
singularity of f (z ).

To explain this concept, we consider the function

sin z
f (z) 
z

sin z
You will recall that this function is undefined at z = 0 but lim  1. So we
z 0 z
sin z
can say that z = 0 is a removable singularity of .
z

Next consider the function f (z)  e1/ z . We can write an exponential function in
the form of a power series. For the given function, we have

1 1 1
e1/ z  1     ...
z 2! z 2 3! z3 27
Block 3 Complex Analysis
From this we note that z = 0 is a singularity. As it is an infinite series in 1/z, we
cannot associate a pole of definite order with the function. So the singularity at
z = 0 is not a pole; it is an isolated essential singularity. You will learn more
about it when we discuss Laurent series expansion of a function in the next
unit.

Meomorphic function

A function which is analytic at all the points of a finite region of the complex
plane, except at a finite number of singular points that are poles of the
function, is called a meomorphic function.

We now summarise what you have learnt in this unit.

10.6 SUMMARY
 A complex number has the form z  x  iy , where x and y are two real
numbers and i   1 .

 The complex conjugate of a complex number z  x  iy is z*  x  iy

and zz*  x 2  y 2 is always a positive real number.

 The polar form of complex number is z  r (cos   i sin ), where r is the
modulus and  is the argument of z. Using Euler's formula
e  cos   i sin , a complex number can be represented as z  rei .

 A function f (z) is said to have a limit  for z  z0 if for a positive real

number  how-so-ever small, there is a positive real number  such that
f z      for all z such that 0  z  z0  .

 A function f (z ) is said to be differentiable if the limit

f ( z0  z )  f ( z0 )
lim
 z 0 z

exists.

 A function f (z ) of a complex variable can be expressed as

f ( z )  u( x, y )  iv ( x, y )

where u( x, y ) is the real part of f (z ) and v ( x, y ) is the imaginary part of

f (z ).

 Derivative of a function exists at a point, if it is continuous at that

point and satisfies the Cauchy-Riemann conditions:

u v u v
 and 
x y y x

 A function of a complex variable is said to be analytic at a point z0 if it

is differentiable at that point as well as in some domain about it.
28
Unit 10 Functions of a Complex Variable – Analyticity
 The real and imaginary parts u( x, y ) and v ( x, y ) of an analytic function
f (z ) , separately satisfy the two-dimensional Laplace equation and are
said to be harmonic functions.

 Logarithmic function ln z is an infinitely many valued function.

 The Gamma function (z ) is defined by the integral formula


( z )  t z 1e  t dt [R(z) > 0]
0

 The fundamental property of a Gamma function is

( x  1)  x!

where x is a natural number.

 If a single-valued function of a complex variable is not analytic a point

z  z0 , then z0 is singularity of the function.

 If z0 is a singularity of f (z ) but lim ( z  z0 )n f ( z ) exists, then we say

z  z0
that the singularity of f (z ) is a pole of order n.

10.7 TERMINAL QUESTIONS

1. Show that whenever f ( z0 ) exists, f is necessarily continuous at z0 .

2. Show that in polar form, the Cauchy-Riemann equations can be written as

u 1 v v 1 u
 and 
r r  r r 

3. Discuss the analyticity of the function

f ( z)  e y (cos x  i sin x )

4. For each of the following functions of a complex variable, locate and name
the singularities in the finite z-plane.

z2  2z
a)
z2  2z  2
ln ( z  3i )
b)
z2

sin z 2
c)
z

5. a) Show that      21/ 3 

1 5
3 6

1 n 1
1

b) Using Gamma function, solve the integral   log  dx
 x
0
29
Block 3 Complex Analysis
10.8 SOLUTIONS AND ANSWERS
Self-Assessment Questions
1. f ( z )  zz *

Derivative at z0 :

( z0  z ) ( z0  z )*  z0 z0*
f ( z0 )  lim
z 0 z

z0 z *  z0* z  z z *
 lim
z 0 z

 z *   z * 
 lim  z0  z0*  z *   lim  z0  z0* 
z 0 z  z 0 z 

If z0 is approached along the real axis, i.e. y = 0, x 0, we have

 x  iy 
f ( z0 )  lim z0    z0*  z0  z0* (i)
x 0  x  iy 
y 0

If z0 is approached along the imaginary axis, i.e. x  0, and y  0,

we get

  iy 
f ( z0 )  lim z0    z0*  z0  z0* (ii)
x 0  iy 
y 0

From (i) and (ii) you will note that the two expression agree only at z0  0.
So f ( z0 ) does not exist except at z0  0.

2. Let the analytic function f ( z )  u( x, y )  i v ( x, y ). The real part is

u( x, y )  e x cos y. To find the imaginary part v, we use the Cauchy-
Riemann condition (10.16a):
u v
 e x cos y 
x y

Integrating v with respect to y, we obtain

v  e x sin y  g ( x ) (i)

From the second Cauchy-Riemann condition, we have

u v
 e x sin y  
y x

Integrating v with respect to x, we get

v  e x sin y  h( y ) (ii)

On equating (i) and (ii), we find that g(x) = h(y) = 0 so that

f (z)  u  i v  e x (cos y  i sin y )  e x iy

30  f (z)  ez is the required analytic function.

Unit 10 Functions of a Complex Variable – Analyticity
 e iz1  e iz1   e iz2  e iz2 
3. i) 2 sin z1 cos z2  2    
 2i  2 

 e 1 2  e i z1  z2   ei z1  z2   e i z1  z2  

1 i z  z 
2i
 sin ( z1  z2 )  sin ( z1  z2 ) (a)

On interchanging z1 and z2, we get

2 sin z 2 cos z1  sin( z 2  z1 )  sin( z 2  z1 )

 sin ( z1  z2 )  sin ( z1  z2 ) (b)

Adding (a) and (b) we get the required result.

  
When z2  , we see that sin   z   cos z
2 2 

Similarly it follows that

sin (   z )   sin z

and
cos (   z )   cos z

 e iz1  e iz1   e iz 2  e iz 2 

ii) 2 cos z1 cos z 2  2   
 2  2 

e i ( z1  z2 )  e i ( z1  z2 )  e i ( z1  z2 )  e i ( z1  z2 )

2

 cos ( z1  z2 )  cos ( z1  z2 ) (c)

 e iz1  e iz1   e iz2  e iz2 

2 sin z1 sin z2  2    
 2i   2i 

 e
1 i z1  z2 
 e i z1  z2   e i z1  z2   e  i z1  z2  
2i 2

  cos ( z1  z2 )  cos ( z1  z2 ) (d)

On subtracting (d) from (c), we get

cos ( z1  z2 )  cos z1 cos z2  sin z1 sin z2

iii) We know that

sin ( z1  z2 )  sin z1 cos z2  sin z1 sin z2

Using  z2 in place of z 2 we find that

sin ( z1  z2 )  sin z1 cos ( z2 )  cos z1 sin ( z2 )

Since sin( z2 )   sin z2 and cos(  z2 )  cos z2 , we get

sin ( z1  z2 )  sin z1 cos z2  cos z1 sin z2 31

Block 3 Complex Analysis
iv) If you replace z 2 by  z2 in the addition formula for cos(z1  z2 ), you
will obtain the required relation.

4. We know that

eiz  eiz eiz  eiz

sin z  and cos z 
2i 2
d d  e iz  e  iz  i e iz  e iz 
 sin z    
dz dz  2i  2i

eiz  e iz

2
 cos z

Similarly, we can write

d d  e iz  e iz   e iz  ie iz 
cos z     i  
dz dz  2   2 

e iz  e iz
   sin z
2i
5. cos z  cos x cosh y  i sin x sinh y

We know that cos z = 0 will occur when both the real and the imaginary
parts vanish. The real part vanishes when cos x cosh y  0. But
cosh y  0. Therefore it is required that cos x = 0. This happens when
x   (2n  1)  / 2; n  1, 2, 3, ...

For the imaginary part to vanish, we note that sin x  0 when

x  (2n  1)  / 2. So we need sinh y = 0. This occurs only when y = 0. So
the zeros of cos z will occur for
y = 0 and x  (2n  1)  / 2

which means that z  (2n  1)  / 2; n  1, 2, 3,...

in recurrence relation ( x  1)  x! and    

3 1
6. a) Putting x 
2 2


       
3 1 1
2 2 2 2

5
Similarly, for x  the recurrence relation yields
2

 5 3 1  1 3 
          
2 22 2 4

1
b) Substituting z  x  in Eq. (iii) of Example 10.4, we get
2

  1  3  
     .
 2   2  sin   
 
32  2
Unit 10 Functions of a Complex Variable – Analyticity

  3 
Since sin     sin   1 and the value of    , then the
 2 2 2 2
above equation can be rewritten as

1 
    
 2 2

 1
which on rearranging the terms yields    2 .
 2 

Terminal Questions
1. Derivative of f (z ) at z0 is defined by

f ( z0  z )  f ( z0 )
f ( z0 )  lim
z 0 z

f ( z0 ) exists when the right hand side limit exists. To see whether or not it
is so, we write

f ( z0  z)  f (z0 ) 

lim f (z0  z)  f (z0 )  lim  z
z 0 z 0 z 
f ( z0  z )  f ( z0 )
 lim lim z
z 0 z z 0

The first term exists but lim z  0.

z 0

 lim f ( z0  z )  f ( z0 )  0
z 0
or
lim f ( z0  z )  f ( z0 )
z 0

or
lim f ( z )  f ( z0)
z 0

Therefore, f (z ) is necessarily continuous if the derivative f ( z0 ) exists.

y
2. We have x = r cos, y = r sin, r = x 2  y 2 and   tan1 .
x

u u r u  u  x  u 
  y 

   
x 
r x  x r  x 2  y 2    
  x2  y 2 

u 1 u
 cos   sin 
r r 

Similarly
u u r u  u 1 u
   sin   cos 
y r y  y r r 

v v 1 v
 cos   sin 
x r r  33
Block 3 Complex Analysis
and

v v 1 v
 sin   cos 
y r r 

From the Cauchy-Riemann condition

u v
 .
x y

it readily follows that

 u  1 v  cos    v  1 u  sin   0
    (i)
 r r    r r  

Similarly, from the Cauchy-Riemann condition

u v

y x

we obtain

 u  1 v  sin    v  1 u  cos   0
    (ii)
 r r    r r  

On combining (i) and (ii), we obtain the polar form of the Cauchy-
Riemann equations:

u 1 v

r r 

and

v 1 u

r r 

3. f ( z)  e y (cos x  i sin x )  e y cos x  i e y sin x

If we write f (z) = u + i v, we find that

u( x, y )  e y cos x and v ( x, y )  e y sin x

u
   e y sin x
x

u
 e y cos x
y

v
 e y cos x
x

and

v
 e y sin x
34 y
Unit 10 Functions of a Complex Variable – Analyticity
From the Cauchy Riemann condition (10.13a) it readily follows that

 e y sin x  e y sin x

or

e y sin x  0 (i)

Similarly, from Eq. (10.13b) we obtain

e y cos x  e y cos x

or

e y cos x  0 (ii)

You will recognize that (i) and (ii) are not simultaneously satisfied at any
point in the complex plane. So f (z ) does not exist and f (z ) is no-where
analytic.

z 2  2z
4. a) f (z) 
z 2  2z  2

We note that

z2  2z  2  0 for z  z1  1 i and z  z2  1 i

z 2  2z
 f (z) 
( z  z1 ) ( z  z 2 )

From this expression it is clear that singularities occur at z1  1  i

and z 2  1  i . Moreover, each singularity is a simple pole.

ln ( z  2i )
b) f (z) 
z2
At z = 0, ln (2i)  0. Therefore f (z) has a pole of order 2 at z = 0.
z  2i  0 when z  2i . So logarithmic singularity occurs at z  2i ,
i.e. z  2i is a branch point.

sin z 2 sin z 2
c) f (z)  z
z z2
At z = 0, f (z ) is undefined. But

 sin z 2   sin z 2 
lim f ( z )  lim  z   lim z . lim  
z 0 z 0 z 2  z 0 z 0 z 2 


sin z 2
f (z)  has a removable singularity at z = 0.
z
1
5. a) Substituting x  in the duplication formula Eq. (10.31), we get
3


        
1 1 1 2
3 3 2 2 ( 2 / 3 1) 3
35
Block 3 Complex Analysis
Hence

   
1 5
 3   6   21/ 3 
 
2
3

1
b) Put log  t or x  e t so that dx  e t dt . It may be noted that for
x
x = 0, t = and for x = 1, t = 0. Now the integral can be rewritten as
0 

 t n 1( e t )dt 
 t n 1e t dt  (n )
 0

36