Unit_III

Complex Numbers:

In the system of real numbers R we can solve all quadratic equations of the
form
0 , 0 x
2
= + + a c bx a , and the discriminant 0 4
2
ac b . When the discriminant
0 4
2
< ac b , the solution of this quadratic equation do not belong to the
system of .
In fact , a simple quadratic equation of the form 0 1
2
= + x , does possesses
solution in real. This difficulty was overcame by introducing the imaginary
part unit i, where 1
2
= i .
Thus the set of complex numbers defined as . } { 1 , : ) ( = + = i and R y x iy x C .
Some Basic Results:
1. If z = x +iy is a complex number, then the complex number iy x z = is
called the complex conjugate of z , and
2 2
) )( ( z z y x iy x iy x + = + =
2.If z= x+iy is a complex number , then the modulus of z, denoted by
2 2
y z + = x
3. A complex number z = x + iy is represented by a point p(x ,y) in the
Cartesian plane with abscissa x and ordinate y. Then the x-axis is called real
axis and the y-axis is called the imaginary axis.The point p(x, y) is referred
to as the point z.

Let OP= r and . = XOP Then rsin y , cos r x = =
Every Complex number can z = x +iy be expressed in the form as given
below
) isin r(cos z + = polar form
i
re z = exponential form
We observe that
2 2
x r y + =
the modulus of z and it represents the
distance of the point z from the origin. Also
. arg tan
1 -
z of ument the called is angle the
x
y
|

\
|
=
4. Let
0 0 0
z iy x + = then ) ( ) ( z - z
0 0 0
y y i x x + =
( ) ( )
2
0
2
0
y y x x + = .
Now
0
z - z may be represented as
) sin (cos z - z
0
i R + = and R z z =
0
serves as the complex equation
of the circle C with ( )
0 0
, y x and radius R. In particular 1 z = represents the
circle with center at the origin and radius equal to 1.

Functions of a complex variable:
Let C be a set complex numbers. If to each complex number z in C there
corresponds a unique complex number w ., then w is called a complex
function of z defined on C, and we write w = f(z). Hence ,w has a real part ,
say u and an imaginary part , say v. Then , w has the representation
W= f(z) = u(x,y) + i v(x,y) (Cartesian form)
W = u(r,) + i v(r,) (polar form)
Continuity :
A complex valued function f(z) is said to be continuous at a point
0
z if f(z)
is defined at
0
z and
0
z z
z f(z) lim
0
=

.
Note:
1. If a complex valued function f(z) is differentiable at a point
0
z , then
it is continuous at
0
z .
2. The converse of the above result is not always true. The continuity of
a complex function need not imply its differentiability.
Derivative of a complex function:

The derivative of a complex function f at a point
0
z z = , denoted by ), (z f
0

is defined as
z
z
z

+
=

) f(z ) f(z
) ( f
0 0
0 z
0
1
lim
, provided this limit exists.
Substituting have we , z - z z
0
=

z
z

=

) f(z ) f(z
) ( f
0
0 z
0
1
lim

We should remember that by the definition of limit f(z) is defined in a
neighborhood of
0
z and z may approach
0
z from any direction in the
complex plane . The derivative of a function at a point is unique if it exists.
Analytic Functions. Cauchy-Riemann equations.
In complex analysis we are interested in the functions, which are
differentiable in some domain, called the analytic functions. A large variety
of functions of complex variables which are useful for applications purpose
are analytic.

A Function f(z) is said to be analytic at a point
0
z , if it is differentiable
at
0
z and, in addition , it is differentiable throughout some
neighborhood of
0
z .
Further a function f(z) is said to be analytic in a domain D if f(z) is defined
and differentiable at all points of D. In fact , analyticity is a global
property while differentiability is a local property.

The terms regular and holomorphic are also used in place of analytic.

Cauchy-Riemann Equations :
Cauchy Riemann equations provide a criterion for the analyticity of a
complex function W = f(z) = u(x,y) + i v(x,y) .
Statement: Necessary conditions for a function to be analytic.:
If f(z) = u(x,y) + i v(x,y) is continuous in some neighborhood of a point
z= x+ iy and is differentiable at z , then the first order partial derivatives of
u(x,y) and v(x,y) exist and satisfy the Cauchy-Riemann equations

x y y x
v u v u = = and . At the point z = x + iy.
Proof:
Since f(z) is differentiable at z, we have


z
z
z

+
=

) f(z ) f(z
) ( f
lim
0 z
1

{ } { }
y i
) y v(x, i (x.y) u y) y x, (x v ) y y , ( u
) ( f
lim
0 z
1
+
+ + + + + +
=

x
i x x
z
--------(I)
Let us assume z to wholly real and wholly imaginary.
Case I: When z wholly real, then y = 0 , so that z = x .The limit on
the right side of equation (I) becomes,

{ }
x
x x
z

+
=

y) u(x. - y) , ( u
) ( f
lim
0 x
1

{ }
x
x x

+
+

y) v(x. - y) , ( v
lim i
0 z

=
x
v
i
x
u

. --------------------------------------(II)
Case II: When z wholly imaginary, then x = 0 , so that z = y i .The
limit on the right side of equation (I) becomes,

{ }
y
x
z

+
=

i
y) u(x. - y) y , ( u
) ( f
lim
0 y
1
{ }
y
x

+
+

i
y) v(x. - y) y , ( v
lim i
0 y

=
y
v
y
u
i
1

y
u
i
y
v

= ---------------------------------------(III)
Since f(z) is differentiable the value of the limits obtained from (II) and (III)
must be equal.
x
v
i
x
u

y
u
i
y
v

=
Comparing the real and imaginary parts, we get
x
v
y
u
y
v
x

and
u
at the z = (x,y).
These are known as the Cauchy-Riemann equations. Satisfaction of these
equations is necessary for differentiability and analyticity of the function f(z)
at a given point. Thus, if a function f(z) does not satisfy the Cauchy-
Riemann equations at a point, it is not differentiable and hence not analytic
at that point.

Ex 1: If w= logz, find
dz
dw
, and determine where w is not analytic.
Let us consider z in exponential form , ) sin (cos re z
i

i r + = =
( )
|

\
|
= + =

x
y
tan , y x r
1 2 2

iy) log(x iv u w + = + =
( )
|

\
|
+ + =
x
y
tan i y x log
2
1
1 - 2 2

Equating real and imaginary parts
( )
|

\
|
= + =
x
y
tan v : y log
2
1
u
1 - 2 2
x
2 2 2 2
y x
y
x
v
:
y
x
x
u
+
=

+
=

x


2 2 2 2
y x
x
y
v
:
y
y
y
u
+
=

+
=

x


Now from- C-R equations

x
v
y
u
and
y
v
x

u

Thus w = logz the C-R equations holds good for ( ) 0 y x
2 2
+
Further ,
( )
2 2 2 2
y x
y
i
y x
x
dx
dv
i
x
u
dz
dw
+

+
= +

=


dz
dw

( )( ) z
1
z
z
y x
iy - x
2 2
= =
+
=
z

Thus every point other than origin ( ) 0 y x i.e.
2 2
+ w=logz is
differentiable and the function logz is analytic every where except at origin.

Ex2: Show that the function w= sinz is analytic and find the derivative.
iy) sin(x iv u w + = + =
siniy cosx cosiy sinx + = ------------------------(1)
Now
2
e e
cosx and
2i
e e
x sin
-ix ix ix

=
ix

coshx cosix : isinhx sinix = =
Using these in equation (1)
W ( ) sinhy icosx coshy sinx iv u + = + =
Equating real and imaginary parts, we get
u = sinx coshy : v= cosx sinhy
y sinh sin
x
v
, coshy cosx
x
u
x =

-------------------------(2)
coshy cosx
y
v
, sinhy sinx
y
u
=

The C-R equations are satisfied
x
v
y
u
and
y
v
x

u

f(z) = sinz is analytic.
( ) = z
1
f
dx
dv
i
x
u
+

( ) y sinh sinx - i coshy cosx + =

y i sin sinx cosiy cosx =
( ) cosz iy x cos = + =

Consequences of C-R Equations:
1). If f(z)= u + iv is an analytic function then u and v both satisfy the two
dimensional Laplace equation.
0
y x
2
2
2
2
=

This equation is also written as 0
2
= .
Here
2
is the two- dimensional Laplacian.
Since f(z) is analytic we have Cauchy-Riemann equations
) (
x
v
y
u
and ) (
y
v
x
II I
u

Differentiating (I) w,r,t. x and (II) w.r.t y partially we get
x y
v
y
u
and
y x
v
x
2
2
2 2
2
2

u


But
x y
v
y x
v
2 2

is always true and hence we have

or
y
u
x
2
2
2
2

u
0
y
u
x
u
2
2
2
2
=

,this implies u is harmonic.

Similarly Differentiating (I) w.r.t . y and (II) w.r.t.y partially we get
or
y
v
x
v
2
2
2
2

0
y
v
x
v
2
2
2
2
=

, this implies v harmonic.

If f(z) = u + iv is an analytic function, then u and v are harmonic
functions. Here , u and v are called harmonic conjugates of each other.

Consequence II:
If f(z) = u + iv is an analytic function, then the equations
1
c ) , ( = y x u
And
2
c ) , ( v = y x represent orthogonal families of curves.
Soln:
( )
1
c y x, u = -------------(i)

( )
2
c y x, v = --------------(ii)

Differentiating eqn (i) partially w.r.t x
( )
1
m
y
u
x
u
dx
dy
0
dx
dy
y
u
x
u
=
|

\
|

= =

or --------- (I)

Differentiating eqn (ii) partially w.r.t. x


( )
2
m
y
v
x
v
dx
dy
0
dx
dy
y
v
x
v
=
|

\
|

= =

or ---------(II)
The two families are orthogonal to each other , then 1 m m
2 1
= ,
And using C-R equations

( )( )
( )( ) y v y u
x v x u
m
2 1


= m =
( )( )
( )( )
1 =


y v y u
y u y v

Hence the curves intersect orthogonally at every point of intersection.

Note: The converse of the above result is not true. The following example
reveals the property.
2 2
2
y x v
y
x
u 2 : + = =

2
2 2
1
2
c y x c
y
x
= + = 2 :

( )
( )
( )
1
2 2
m
x
2y
y x -
y 2x
y
u
x
u
dx
dy
= = =
|

\
|

= for curve
1
c

( )
( )
( )
2
m
2y
x
4y
2x
y
v
x
v
dx
dy
= = =
|

\
|

= for curve
2
c
1 m m
2 1
= . They intersect orthogonally.
But C-R Equations are not satisfied
x
v
y
u
and
y
v
x

u
.

Some different forms of C-R Equations:

If w = f(z) = u+ iv , is analytic , then the following results follows.

1. ( ) = z
1
f
x
v
i
x
u

y
u
i
y
v

=
|
|

\
|

=
y
v
i
u
i
y


x
v
i
x
u

=
|
|

\
|

=
y
v
i
u
i
y


y
w
i
x
w

2. ( )
2 2
2
1
x
v
x
u
z f
|

\
|

+
|

\
|

= ( )
2 2
2
1
x
v
x
u
z f
|

\
|

+
|

\
|

=



2
2
u
x
u
|
|

\
|

+
|

\
|

=
y

2
2
x
v
|
|

\
|

+
|

\
|

=
y
u
using C-R Equations.
Based on the results above mentioned the following results are valid,

a) ( )
2
z f
x
(

( )
2
z f
y
(

+ ( )
2
1
f z =

b) is any differential function of x and y then


2
2
x
|
|

\
|

+
|

\
|

y


2 2
v u
|

\
|

+
|

\
|

=

( )
2
1
f z = .

c)
(

2
2
2
2
y x
( ) [ ]
2
z f Re ( )
2
1
f 2 z =

d)
(

2
2
2
2
y x
( )
2
f z ( )
2
1
f 4 z =


Construction of An Analytic Function When real or
Imaginary part is Given
(Putting in Exact differential M dx + N dy = 0)
The Cauchy-Riemann equations provide a method of constructing
an analytic function f(z) = u+iv when u or v or v u is given.
Suppose u is given, we determine the differential dv , since
v = v(x,y),

dy
y
v
dx
x
v
dv

=
using C-R equations ,this becomes
dy N dx M. dy
x
u
dx
y
u
dv + =

=

And it is clear that
y
M
x
N

= 0
y
u
x
u
2
2
2
2
=

Because u is harmonic. This shows that M dx + N dy is an exact differential.
Consequently , v can be obtained by integrating M w.r.t. x by treating y as a
constant and integrating w.r.t. y only those terms in N that do not contain x,
and adding the results.
Similarly, if v is given then by using


dy
y
u
dx
x
u
du

=

dy
x
v
dx
y
v

=
.
Following the procedure explained above we find u, and hence
f(z) u + iv can be obtained. Analogous procedure is adopted to find
u+iv when v u is given.

Milne-Thomson Method:
An alternative method of finding iv u when u or v or v u
is given.
Suppose we are required to find an analytic function f(z) = u+ iv
when u is given. We recall that
(

=
y
u
i
x
u
) ( f
1
z
------------------------(I)
Let us we set y) x, (
y
u
and y) x, (
x
u
2 1
=

-------(II)
Then y) x, ( i ) y x, ( (z) f
2 1
=

--------(III)

Replacing x by z and y by 0, this becomes

z,0) ( i ) z,0 ( (z) f
2 1
=

-------(IV)
From which the required analytic function f(z) can be got.
Similarly , if v is given we can find the analytic function f(z) = u+ iv by
starting with

(

=
x
v
i
y
v
) ( f
1
z
Analogous procedure is used when v u is given.






Applications to flow problems:
As the real and imaginary parts of an analytic function are the
solutions of the Laplaces equation in two variable. The conjugate functions
provide solutions to a number of field and flow problems.
Let v be the velocity of a two dimensional incompressible fluid with
irrigational motion, j
y
v
i
x
v
V

= ------------------------------(1)
Since the motion is irrotational curl V= 0.
Hence V can be written as
j
y
i
x

=

-----------------------------(II)
Therefore , is the velocity component which is called the velocity
potential. From (I) and (II) we have



y y
v
,
x x
v

------------------------(III)
Since the fluid is incompressible div V = 0.
0
y y x x
=
|
|

\
|

+
|

\
|

-----------------------(IV)
0
y x
2
2
2
2
=

This indicates that is harmonic.
The function ( ) y x, is called the velocity potential , and the curves
( ) c y x, = are known as equi -potential lines.

Note : The existence of conjugate harmonic function ( ) y x, so that

( ) ( ) y x, i y x, w(z) + = is Analytic.
The slope is Given by
x
y
v
v
x
y
y
x
dx
dy
=

=
|

\
|

\
|

This shows that the velocity of the fluid particle is along the tangent to the
curve ( )
1
c y x, = , the particle moves along the curve.

( )
1
c y x, = - is called stream lines ( ) c y x, = - called equipotential
lines. As the equipotential lines and stream lines cut orthogonally.

( ) ( ) y x, i y x, w(z) + =

y
i
x x
i
x dz
dw

=




y x
v v =
The magnitude of the fluid velocity ( )
dz
dw
v v
2
y
2
x
= +
The flow pattern is represented by function w(z) known as complex
potential.

Complex potential w(z) can be taken to represent other two-dimensional
problems. (steady flow)

1. In electrostatics ( ) c y x, = --- interpreted as equipotential lines.
( )
1
c y x, = --- interpreted as Lines of force




2. In heat flow problems:
( ) c y x, = --- Interpreted as Isothermal lines
( )
1
c y x, = --- interpreted as heat flow lines.

Cauchy Riemann equations in polar form:

Let ) iv(r, ) u(r, ) f(re f(z)
i

+ = = be analytic at a point z, then

there exists four continuous first order partial derivatives ,

v
,
r
v
,
u
,
r
u
and satisfy the equations
, .
u
r
1
r
v
:
v 1
r
u

r

Proof: The function is analytic at a point
i
re z = .

z
z
z

+
=

) f(z ) f(z
) ( f
lim
0 z
1
exists and it is unique.

Now ). iv(r, ) u(r, f(z) + =
Let z be the increment in z , corresponding increments are
. and r in , r

{ } { }
z
) v(r, i ) (r. u ) r, (r v ) , ( u
lim
) ( f
0 z
1

+ + + + + +
=

i r r
z


{ }
z
) u(r. - ) , r (r u
lim
) ( f
0 z
1

+ +
=


z

{ }
z
r

+ +
+

) . v(r - ) , (r v
lim i
0 z


------------------------(I)

Now
i
re z = and z is a function two variables r and , then we have

.
z
r
r
z
z

=


( ) ( )

=
i i
re r re
r
z

+ =
i i
e r i r e z
When z tends to zero, we have the two following possibilities.

(I). Let r e z that so , 0
i
= =



And 0 r implies , 0 Z

{ }
r e
) u(r. - ) , r (r u
lim
) ( f
i
0 r
1

+
=



z

{ }
r e
) v(r. - ) , (r v
lim i
i
0 r

+
+


r


The limit exists,

(

=

r
v
i
r
u
) ( f
1 i
e z
-----------------(I)
2. Let
i
e r i z that so , 0 r = =
And 0 imply 0, z

{ }

+
=

i
0
1
e r i
) u(r. - ) , (r u
lim
) ( f z

{ }

+
+

i
0 z
e ir
) v(r. - ) , (r v
lim i


=
(

v
i
u
re i
1
i (

v u
i
e r
1
i




(

u
r
i v
r
1
e z) ( f
i - 1
------------ (II)
From (I) and(II) we have

v
r
1
r
u

u
r
1
r
v
or

u rv , v ru
r r
= =
Which are the C-R Equations in polar form.

Harmonic Function:

A function -is said to be harmonic function if it satisfies Laplaces
equation 0
2
=

Let ) iv(r, ) u(r, ) f(re f(z)
i

+ = = be analytic. We shall show that

u and v satisfy Laplaces equation in polar form.

0
1
r r
1
r
2
2
2 2
2
=

r


The C-R equations in polar form are given by,

v
r
1
r
u
---------(I)

u
r
1
r
v
------------(II)

Differentiating (I) w.r.t r and (II) w.r. , partially , we get

r
u
r
u
r
2
2
2

v
r
:
2
2 2
u
r
v
r

And we have
r

v
r
v
2 2


2
2
2
2
u
r
1
r
u
r
u
r

Dividing by r we, get

0
u 1
r
u
r
1
r
u
2
2
2 2
2
=

r

Hence u-satisfies Laplaces equation in polar form.
The function is harmonic. Similarly v- is harmonic.

Orthogonal System:
Let ) f( r = and
=

,
dr
d
tan r
being the angle between
the radius vector and tangent. The angle between the tangents at
the point of intersection of the curves is
2 1
.
1 Tan Tan
2 1
=
,is
the condition for orthogonal.

Consider
1
c ) u(r, = .
Differentiating w.r.t , treating r as a function of .

0
u
d
dr
r
u
=

\
|

\
|

=
r
u
u
d
dr

Thus
dr
d
r Tan
1

=
=
( )
( )

u
r
u
r
=
( )
( )

u
r
u
r -
---------(I)

Similarly for the curve
2
c ) v(r, =


=
2
Tan
( )
( )

v
r
v
r -
-----------(II)


( ) ( )
( ) ( )

=
v u
r
v
r
r
u
r
an an
2 1
T T

By C-R Equations

u rv , v ru
r r
= =


The equation reduces to

( ) ( )
( ) ( )
1
u v
u
-
v
an an
2 1
=

=


T T


Hence the polar family of curves
1
c ) u(r, = and
2
c ) v(r, = ,
intersect orthogonally.

Construction of An Analytic Function When real or
Imaginary part is Given(Polar form.)
The method due to Exact differential and Milne-Thomson is
explained in earlier section .

Ex: Verify that ( ) cos2
r
1
u
2
= is harmonic . find also an analytic
function.

Soln:
cos2
r
2
r
u
2
|

\
|
=

sin2
r
2 u
2
|

\
|
=

cos2
r
6
r
u
4 2
2
=

cos2
r
4 u
2 2
2
=

.
Then the Laplace equation in polar form is given by,


=

2
2
2 2
2
u 1
r
u
r
1
r
u
r
cos2
r
6
4

cos2
r
2
4
|

\
|

0 cos2
r
4
2
=


Hence u-satisfies the laplace equation and hence is harmonic.

Let us find required analytic function f(z) = u+iv.
We note that from the theory of differentials,

d
v
dr
r
v
dv

=

Using C-R equations

u rv , v ru
r r
= =

d
r
u
r dr
u
r
1
-
|

\
|

+
|

\
|

=


d cos2
r
2
d sin2
r
2
2 3
|

\
|

\
|
= r


|

\
|
= sin2
r
1
- d
2


From this
c sin2
r
1
- v
2
+ =


c +
|

\
|
+
|

\
|
= + = sin2
r
1
- i cos2
r
1
iv u f(z)
2 2



[ ] ic isin2 - cos2
r
1
2
+ =



( )
ic
e r
1
e
r
1
2
i
2i -
2
+ = + =

ic


ic.
z
1
f(z)
2
+ =

Ex 2:Find an analytic function f(z)= u+iv given that

sin
r
1
- r v
|

\
|
=
0 r
Soln:
sin
r
1
r
r
v
2
|

\
|
+ =

cos
r
1
r
v
|

\
|
=

To find u using the differentials

d
u
dr
r
u
du

=

Using C-R equations

u rv , v ru
r r
= =

d
r
v
- dr
v
r
1
|

\
|

+
|

\
|

=



d sin
r
1
1 r - dr cos
r
1
- r
r
1
2
|

\
|
+
|

\
|
=


=
d sin
r
1
r - dr cos
r
1
1
2
|

\
|
+
|

\
|




(

\
|
+ = cos
r
1
r d

c cos
r
1
r u +
|

\
|
+ =
f(z) = u+iv

sin
r
1
- r i cos
r
1
r
|

\
|
+ +
|

\
|
+ = c
( ) c isin - cos
r
1
isin r(cos + + + =
c
z
1
z
r
1
e r f(z)
i - i
+ + = + =

e
Ex: Construction an analytic function given cos2 r u
2
= .
(Milne Thomson Method)

cos2 r u
2
= -----------(I)
2rcos2
r
u
=

sin2 2r
u
2
=

=

r
v
i
r
u
) ( f
1 i
e z

Using C-R equations

u r v , v u r
r r
= =

( ) ( )
(

\
|
+ =

sin2 2r -
r
1 -
i rcos2 2 e z f
2 i - 1


( ) [ ]

sin2 2r i rcos2 2 e
-i
+ =

[ ]

isin2 cos2 e r 2
-i
+ =

Now put r = z , and 0 = .
( ) 2z z f
1
= on integrating
( ) c z z f
2
+ = .

COMPLETION OF UNIT-I