COMPLEX

VARIABLES
Complex functions
A function of a complex variable is a mapping f which
gives for each complex number z = x + i y another complex
number w = u + i v.
We write: w = f(z), u + i v = f(x + i y).
w is also called Image of z and z is called Pre-image.
EXAMPLES
If w = z2, then u = x2 - y2, v = 2xy.
If w = 1/z, where z is not 0, then
u = x/(x2 + y2), v = -y/(x2 + y2).
If w  z , then u = x, v = -y.
u  x 2  y 2 , v 0.
If w = | z |, then
Domain
Set D of complex numbers z for which the function f(z) is
defined is called the Domain of the function.
If domain of a function is not specified, we take Natural
Domain as the domain of definition of function.
Natural Domain is whole of complex plane (xy – plane) or
part thereof over which the function is defined.
Example:
Natural domain of the function w = f(z) = a z3+ b z+ c is
whole of the complex plane.
P( z )
Natural domain of rational function w  f ( z) 
Q( z )
is the complex plane excluding the points where Q(z) = 0.
 Range
Set R of complex numbers w = f(z) is called the Range of the
function.
In theory of complex variables we have some functions which
Take more than one value for each z.
1
The function wn  z or w z n, z | z | ei
where n is a positive integer, takes n values
   2k 
1   2k  
wk  z cos
n  i sin  , k 0,1, 2,........, ( n  1).
 n n 
The function w ln z
takes values
wk ln z  i (  2 k ), k 0,1, 2,3,...................
If for every value of z in the domain of definition, there exists
a unique image in the w-plane given by w = f(z), the function
f(z) is said to be single valued.
In case of multi-valued functions we restrict our study to the
part of domain in which the function behaves like a single
valued function.
Each of such single valued functions is known as Branch of
the function.
Many valued property of complex functions is useful in
solving some problems in simple manner.
Example: Find all roots of xn -1 = 0
where n is a positive integer.
Solution: We have xn = 1 = cos 2π + i
sin 2π
1
x n ei 2  
or x  ei 2 n

Therefore x takes n values

2  2k 2  2k
xk cos  i sin , k 0,1, 2,............., ( n  1).
n n
For n = 3 we have cube roots of unity
2 2 1 3
x0 cos  i sin   i
3 3 2 2
4 4 1 3
x1 cos  i sin   i
3 3 2 2

x2 cos 2  i sin 2 1
Remarks
1. It is not always possible to write a complex function as a
function of z = x + i y alone.
For example if u(x,y) = 2 x + y and v = 6 x y are two real
valued functions of real variables x and y, then we can
obtain a complex function
w = (2 x + y) + i (6 x y)
It is not possible to express this function as function z.
2. If z is taken in the polar form, then real and imaginary
parts of f(z) can be expressed as real valued functions of r
and .
Thus we can write w  f ( z )  f (r ei ) u (r ,  )  i v(r ,  )
 Example: Write the function w = z2 in the form and give its
graphical representation.

w u ( x, y )  iv( x, y ) and w u (r ,  )  i v(r ,  )

Solution:We have w = z2. Substituting z = x + i y, we get

w ( x  iy ) 2  x 2  y 2  2ixy

Therefore u ( x, y )  x 2  y 2 v( x, y ) 2 xy

Substituting z = r ei, we get

w (rei ) 2 r 2 e 2i r 2 [cos 2  i sin 2 ]

Therefore u (r ,  ) r 2 cos 2 v( r ,  ) r 2 sin 2

Writing w Rei  R[cos   i sin  ] we get

R r 2 and  2
Graphical representation
A circle r = a in z-plane is mapped onto circle R = a2 in w-
plane.
r  a , 0   
The region 2 in the z-plane is mapped onto the
region R a 2 , 0   in w-plane.
x2-y2 = c1 . (a 2
,)
. (a ,) v c2
2xy c2

z-plane
w-plane u = c1

The image of hyperbola x2 – y2 = c1 is the line u = c1

The image of rectangular hyperbola 2xy = c2 is the line v = c2.
Elementary functions
Exponential function w = ez
Taking z = x + i y , we get
x  iy x iy x
w e e .e  w  e [cos y  i sin y ]
Therefore the image of z = (x, y) in the z-plane is w = (u, v) in
the w-plane, where
u e x cos y and v e x sin y
ez is real if sin y = 0 or y = Im(z)= nπ , n = 0, ±1, ±2, ±3, ….
ez is imaginary if cos y = 0 or y = Im(z)= (2n+1)π/2 , n
= 0, ±1, ±2, ±3, ….
Properties w = ez
1. It is defined for all values of z. Therefore its domain of
definition is whole complex plane.
2. It is a single valued function.
3. |ez| = ex and Arg(ez) = y.
4. ez ≠ 0 for any z. Therefore its range is whole of complex
plane except origin.
5. ez = 1, if and only if z = 2nπi , n any integer.
z1 z2
6. e  e if and only if z1 =z2 + 2nπi , n any integer.
z1 z2 z1  z2
7. e e  e
z1 z2 z1  z2
8. e / e  e
 
n
9. ez enz , n any integer
 Trigonometric functions
We know that
eiz cos z  i sin z and e  iz cos z  i sin z
Solving these equations for cos z and sin z, we get
cos z  1 (eiz  e  iz ) and sin z  i (e  iz  eiz )
2 2
Substituting z = x + i y, we obtain
1 i ( x  iy )  i ( x  iy )  1 [eix  y  e  ix  y ]
cos z  [e e ] 2
2
 1 [e y (cos x  i sin x)  e y (cos x  i sin x)]
2
 1 [cos x(e y  e  y )  i sin x(e y  e  y )]
2
cos x cosh y  i sin x sinh y
Similarly we obtain sin z sin x cosh y  i cos x sinh y
Other trigonometric functions are defined as usual.
Complex trigonometric functions satisfy the same identities as
real trigonometric functions.
The functions sin z and cos z are periodic with period 2π and the
function tan z is periodic with period π.
*
In addition we have sin z  sin z *
We further have
2 2
sin z  sin x cosh y  i cos x sinh y
sin 2 x cosh 2 y  cos 2 x sinh 2 y
sin 2 x (1  sinh 2 y )  (1  sin 2 x) sinh 2 y
sin 2 x  sinh 2 y
Since sinh y is not bounded, therefore |sin z| and hence sin z is
not a bounded function.
Note that real function sin x is a bounded function as |sin x|  1.
Now sin z 0  sin z 0
2 2
Therefore sin x  sinh y 0 for all x and y.
This implies sin x 0 and sinh y 0
or x n , (n any integer) and y 0.
Hence, sin z = 0, only when z is real and z = n, n is any integer.
Similarly cos z = 0, only when z is real and z = (2n+1)/2, n is
any integer.
 Hyperbolic functions
z z
We have sinh z  12 [e  e ]
Substituting z = x + i y, we get
sinh z  1 [e x  iy  e ( x  iy ) ]  1 [e x eiy  e x e iy ]
2 2
 12 [e x (cos y  i sin y )  e  x (cos y  i sin y )]
 12 [(e x  e  x ) cos y  i (e x  e  x ) sin y )]
sinh x cos y  i cosh x sin y
Similarly we obtain cosh z cosh x cos y  i sinh x sin y
2
Now sinh z sinh 2 x cos 2 y  cosh 2 x sin 2 y
sinh 2 x (1  sin 2 y )  (1  sinh 2 x) sin 2 y
sinh 2 x  sin 2 y
2 2 2
Similarly cosh z sinh x  cos y
Now sinh z 0  sinh z 0
2 2
Therefore sinh x  sin y 0 for all x and y.
This implies sinh x 0 and sin y 0
or x 0 and y n , (n any integer).
Hence, sinh z = 0, only when z is pure imaginary and z = ni,
n is any integer.
Similarly cosh z = 0, only when z is pure imaginary and z =
(2n+1)i/2, n is any integer.

Note that the functions sinh z and cosh z are not bounded, since
2 2
2
sinh z sinh x  sin y and2
cosh z sinh 2 x  cos 2 y
Logarithm function
The real valued natural logarithm of a real variable x is defined as
y ln x, if x e y , x  0.
Similarly we define logarithm of a complex variable z = x + i y as
w ln z , if z e w , z 0.
i
Writing z = rei and w = u+i v, we get re e u iv
or re i ( 2 n ) e u iv , where n is any integer.
Therefore
e w e u r or u ln r
which is the logarithm of real variable |z|, and
arg(e w ) v   2n , n any integer.
Thus for any complex number z  0, we have
w ln z ln r  i (  2n ), n any integer. (1)
ln z  i arg( z ), z 0.
Since Im(ln z ) arg( z )   2n , n any integer, takes
infinite number of values, ln z is a multi-valued function.
For each value n we obtain a different branch of the multi-valued
function ln z.
If we restrict arg(z) to the interval (-, ], it is called Principal
argument of z and is denoted by Arg(z).
The corresponding Principal Branch of ln z is written as
Ln z ln | z | i Arg( z )
or Ln z ln | z | i  ,      (2)
or Ln z ln x 2  y 2  i tan  1 ( y / x)
The function Ln z is single valued and is defined for all z  0.
 Properties of Logarithm function
Verify the following properties
(1) e ln z  z.
( 2) ln (e z )  z  2n i, n any integer.
(3) ln ( z1 z 2 ) ln z1  ln z 2  2n i, n any integer.
z1
( 4) ln ( ) ln z1  ln z 2  2n i, n any integer.
z2

Note: ln z is also written as log z.

 Example 1. Show that the following limit does not exist
z
lim .
z  0 Re( z )

z x  iy
lim  lim
z  0 Re( z ) z  0 x

x  iy x  iy x
lim  lim lim  lim  lim 1 1
z 0 x x 0 y 0 x x 0 x x 0

x  iy x  iy lim   limit does not exist.

lim  lim lim 
z 0 x y 0 x 0 x y 0

Taking y = x (That is taking limit along line y = x)

x  iy x (1  i ) (1  i ) 1  i
lim  lim  xlim
0
z 0 x x 0 x
Example 2. Compute the following limits
z 3  iz 2  z  i Ans. 0
(1) lim .
z i z i
z2
(2) lim . Ans. 1
z i z

iz 3  iz  1 i
(3) lim . Ans.
z   (2 z  3i )( z  i ) 2 2
Continuity of a function
A function f(z) is continuous at z = z0 if the function satisfies
following three conditions:
1 f z0  exists
2 lim f z 
z  z0
exists

3 lim f z   f z0 

z  z0

 Im( z )
 z for z 0
Example 2. Show that the function f ( z ) 
is not continuous0at origin.
for z 0
 Differentiability of a function
A single valued function f(z) defined in a domain D is differentiable
at a point z = z0 of D if
f z  f z0 
lim exists
z  z0 z  z0
This limit is called derivative of f(z) at the point z = z0 and is
denoted by f ( z0 ).
Therefore we write
f ( z )  f ( z0 )
lim  f ( z0 )
z  z0 z  z0
If we write w  f ( z ) then
dw
f ( z ) 
dz
If the function f(z) is differentiable at z = z0, then it must be
continuous at z = z0.
From the identity
 f ( z )  f ( z0 ) 
f ( z )   ( z  z0 )  f ( z0 ), z  z0
 z  z0 
we obtain
  f ( z )  f ( z0 )  
lim f ( z )  lim    ( z  z0 )  f ( z0 ) 
z  z0 z  z0   z  z0  
 f ( z )  f ( z0 ) 
 lim   lim ( z  z0 )  lim f ( z0 )
z  z0  z  z0  z  z0 z  z0

Since the function is differentiable, therefore

f ( z )  f ( z0 )
lim  f ( z0 )
z  z0 z  z0
lim f ( z )  f ( z0 ) 0  f ( z0 )  f ( z0 )
z  z0
Remarks:

1. A function which is not continuous at a point can not be

differentiable at that point.
2. A function which is continuous at a point may not be
differentiable there.
Example: The function f ( z ) 
iszcontinuous at origin
but not differentiable there.
3. The rules of differentiation of function of real variable also hold
for function of complex variable.
 Analytic function
A function f(z) of a complex variable z is said to be analytic at a
point z = z0, if it is differentiable at z0 and also at each point in some
neighborhood of the point z0.
The function f(z) = |z|2 is differentiable at z = 0 only and nowhere
else. Therefore it is not analytic anywhere.
A function f(z) is analytic in a domain D if it is analytic at every
point of D.
f(z) = zn, n positive integer, is differentiable at every point,
therefore it is analytic at every point of the finite complex plane.
Notation:
For the function u(x, y) of two variables x and y, the partial
derivatives are denoted as
u u
u x , u y
x y
 Cauchy-Riemann Equations
Necessary conditions for a function to be analytic
Suppose that a function f(z)= u(x, y)+ i v(x, y) is continuous in
some neighborhood of z = x +i y and is differentiable at z. Then at
the point z, the first order partial derivatives of u(x, y) and v(x, y)
exist and satisfy the equations
u x v y and u y  vx . (1)

The equations (1) are called Cauchy-Riemann equations.

Proof:
Since f(z) is differentiable, we have
f ( z   z)  f ( z)
f ( z )  lim
 z 0 z
 u ( x   x, y   y )  iv( x   x, y   y )   u ( x, y )  iv( x, y )
 lim
 x  0,  x  i y
 y 0
Since the limit exists, it is independent of the path along which
δz → 0.
We consider two paths:
1. First Path: Let δy → 0 first. Then we have
 u ( x   x, y )  iv( x   x, y )  u ( x, y )  iv( x, y )
f ( z )  lim
 x 0 x
 u ( x   x, y )  u ( x, y )  i  v( x   x, y )  v( x, y )
 lim
 x 0 x
u ( x, y ) v( x, y )
 i (2)
x x

2. Second Path: Let δx → 0 first. Then we have

 u ( x, y   y )  iv( x, y   y )   u ( x, y )  iv( x, y )
f ( z )  lim
 y 0 i y
 1 u ( x, y ) v( x, y )
f ( z )  
i y y
v u
 i (3)
y y
The expressions for derivative of f(z) given by (2) and (3) must
be identical.
Therefore we have
u v v u
i  i
x x y y
u v u v
or  and  (4)
x y y x
or in short notation u x v y and u y  v x
Thus if the function f(z) is differentiable, then Cauchy-Riemann
equations (4) are satisfied.
The converse may not be true.
Example: Show that the function satisfies the Cauchy-Riemann
equations at z = 0 but the derivative does not exist.
 x3 (1  i )  y 3 (1  i )
 , z 0
f ( z )  2 2
x y

0 z 0
Solution: Writing f ( z ) u ( x, y )  iv( x, y ) we get

x3  y 3 x3  y 3
u ( x, y )  and v ( x, y )  , ( x, y ) (0, 0)
2 2 2 2
x y x y
f (0) 0  u (0, 0) 0 and v(0, 0) 0.
Now as z → 0, we obtain at z = 0
u ( x, 0)  u (0, 0) x
u x (0, 0)  lim  lim 1
x 0 x x 0 x
 u (0, y )  u (0, 0)  y
u y (0, 0)  lim lim  1
y 0 y y 0 y

v( x, 0)  v(0, 0) x
vx (0, 0)  lim  lim 1
x 0 x x 0 x

v(0, y )  v(0, 0) y
v y (0, 0)  lim lim 1
y 0 y y  0 y
Therefore at z = 0, u x v y and u y  vx
Thus the Cauchy-Riemann equations are satisfied at z = 0.
Now
f ( z )  f (0) x3 (1  i )  y 3 (1  i )
lim  lim
z 0 z z 0
3
x2  y 2 z 3
x (1  i )  y (1  i )
 lim
x  0,
y 0  
x 2  y 2 ( x  iy )
Choosing the path y = m x, we get
 f ( z )  f (0) x3 (1  i )  m3 x3 (1  i )
lim lim
z 0 z
 
x  0 x 2  m 2 x 2 ( x  imx )
3
(1  i )  m (1  i ) (1  i )  m3 (1  i )
 lim 
 
x  0 1  m 2 (1  im)  
1  m 2 (1  im)
Thus the limit value depends on m, hence the limit does not
exist and therefore f ' (0) does not exist.
Sufficient conditions for a function to be analytic
Suppose that real and imaginary parts u(x, y) and v(x, y) of the
function f(z) = u(x, y)+ i v(x, y) are continuous and have continuous
first order partial derivatives in a domain D. If u and v satisfy
Cauchy-Riemann equations at all points of D, then the function f(z)
is analytic in D and
f ( z ) u x  ivx v y  iu y
Proof:
We have
f  z   z  u  x   x, y   y   iv  x   x, y   y 
 u u 
u  x, y     x   y   ...
 x y 
  v v  
 i  v  x, y     x   y   ...
  x y  
 u v   u v 
or f  z   z   f  z   i  x  i  y
 x x   y y 

Applying Cauchy-Riemann equations

u v u v
 and 
x y y x
in the coefficient of δy, we get
  u v   v u 
f  z   z   f  z    i   x    i   y
 x x   x x 
 u v   v u 
  i   x   i   i y
 x x   x x 
 u v 
  i   x  i y 
 x x 
f z   z   f z 
 f  z   lim
 z 0 z
u v
 i
x x
v u
Also f ( z )  i
y y
Real and imaginary parts of an analytic function satisfy
Laplace equation.
Real and imaginary parts of an analytic function are
harmonic functions.

Let the complex function f ( z ) u ( x, y )  iv( x, y ) be an analytic

function.
Then the real part u(x, y) and imaginary part v(x, y) satisfy the
Cauchy-Riemann equations
u x v y and u y  vx
Since the function f(z) is analytic, its derivative will also be
analytic and therefore u(x, y) and v(x, y) have continuous second
order derivatives.
Taking partial derivative of first equation with respect to x and
of second with respect to y, we get
 u xx v yx and u yy  vxy
Since the second order derivatives are continuous, therefore
v yx vxy
Hence u xx  u yy 0 or 2 u 0
Similarly, differentiating u x v y with respect to y, and u y  vx
with respect to x and subtracting we get
vxx  v yy 0 or 2 v 0
The function v is called Harmonic conjugate of u.

Cauchy-Riemann equations in polar form

u 1 v v 1 u
 and 
r r  r r 
Laplace equation in polar form
2 2 u1 u 1  2
u
 u   0
2 2
r r r  2
r
If the function f(z) = u(x, y) + i v(x, y) is analytic, then the
families of curves u(x, y) = c1 and v(x, y) = c2 are orthogonal.

Differentiating u(x, y) = c1 and v(x, y) = c2, we get

u u dy v v dy
 0  0
x y dx x y dx
Therefore the slopes are given by
dy u u
m1   
dx x y
dy v v u u
m2     
dx x y y x
 m1m2  1

Hence the families of curves are orthogonal.

 Given a harmonic function g(x, y), using Cauchy-Riemann
equations an analytic function can be constructed with g(x, y),
as real (or imaginary) part.
3 2
Example: Check whether the function u ( x, y ) 2 x  y  3x y
is harmonic. If so, find its conjugate function v(x, y) and the
corresponding analytic function f(z).
3 2
Solution: We have u ( x, y ) 2 x  y  3x y
Therefore u x 2  6 xy u xx  6 y
u y 3 y 2  3x 2 u yy 6 y

Since u xx  u yy  6 y  6 y 0, therefore u(x, y) is harmonic.

From Cauchy-Riemann equation u x v y , we get

v y u x 2  6 xy
Integrating partially with respect to y, we get
 v 2 y  3xy 2   ( x)
where (x) is arbitrary function of x. From the Cauchy-Riemann
equation vx = - uy, we get
 3 y 2   ( x)  (3 y 2  3 x 2 ) or  ( x) 3 x 2
  ( x )  x 3  c, where c is constant
Hence, we obtain v 2 y  3xy 2  x3  c
and f ( z ) u  iv
2 x  y 3  3 x 2 y  i (2 y  3xy 2  x3  c)
2( x  iy )  y 3  3 x 2 y  i ( 3 xy 2  x3 )  ic
2( x  iy )  i[ 3 xy 2  x3 )  i ( y 3  3x 2 y )]  ic
2( x  iy )  i[ x  iy ]3  ic
2z  iz 3  ic
Graphical representation of functions of Complex variable.

In complex domain the function w = f(z) i.e. u +iv =

f(x +iy) (1) involves four real variables x,
y, u, v.
Therefore to plot (1) in Cartesian form a four dimensional space
is required.
As it is not possible to get 4D graph, we make use of two complex
planes:
One for the variable z = x +iy and the other for variable w = u +iv.
For each point (x, y) in z-plane there is a corresponding point (u, v)
in w-plane.
Thus if the point z = x +iy moves along a curve C in z-plane, the
point w = u +iv will move along a corresponding curve C' in w-
plane.
We sum up:
The curve C in the z-plane is mapped into the corresponding
curve C' in the w-plane by the function w = f(z), which defines a
mapping or transformation of the z-plane into the w plane.
Example: For the function f(z) = z + ez, z = x + i y, y = x2

1.2 2
1
0.8 1
0.6
0.4 0
0.2 -1 0 1 2 3
0 -1
-2 -1 0 1 2
z-plane w-plane
Some Standard Transformations.

1. Translation: w = z + c, where c is a complex constant.

The transformation can be written as
u+iv=x+iy+a+ib where u, v, x, y, a,
b are real and u = x + a and v = y + b.
Thus a point P(x, y) in z-plane is mapped onto the point
P'(x +a, y +b) in w-plane.
Every point in the z-plane is mapped onto a point in w-plane in
the same manner. Thus if w-plane is superposed on the z-plane,
we observe that every point is shifted, the shift represented by
the vector c = (a, b).
Thus a figure in the z-plane is mapped as an identical figure in
w-plane. In particular a circle transformed to a circle of same
radius.
Example:
Find the image of the triangle with vertices at i, 1 + i, 1 – i in
the z-plane under the transformation w = z + (4 – 2 i).

Solution:
We rewrite w = z + (4 – 2 i) as u + iv = (x
+ iy) + (4 – 2 i)
or u + iv = (x + 4) + i (y – 2)
Therefore u = x+ 4 and v = y – 2

z ↔ (x, y) (u,v) ↔ w
i ↔ (0, 1) (4, -1) ↔ 4 - i
1 + i ↔ (1, 1) (5, -1) ↔ 5 - i
1 - i ↔ (1, -1) (5, - 3) ↔ 5 - 3i
 z ↔ (x, y) (u,v) ↔ w v
z - p la n e

i ↔ (0, 1) (4, -1) ↔4 - i 2

1 0, 1 1, 1 u

1 + i ↔ (1, 1) (5, -1) ↔ 5 - i 0

-1 0 1 1 , -1 2 3 4 5 6

1 - i ↔ (1, -1) (5, -3) ↔ 5 - 3i -2

w - p la n e
v
0
0 1 2 3 4 4 , -1 5 5 , -1 6u
-2
5 , -3
-4

2
1 0, 1 1, 1
0
z
-1 0 1 , -1 2 4 4 , -1 5 , -1 6
w
-2
-3 5 , -3
-4
2. Magnification and Rotation:
w = c z where c is a complex constant.
Taking c = a ei , z = r ei , w = R ei  the transformation
can be written as
R ei  =(a ei ) (r ei ) or R ei  =ar ei ( + )
Therefore R = ar and  =  +
Thus the transformation w = c z represents the rotation through
angle  = arg(c) and magnification by a = |c|.

The transformation
w=cz+d
represents translation, magnification and rotation.
Example:
Find the image of the triangle with vertices at i, 1 + i, 1 – i in
the z-plane under the transformation w = (3+i)z + 4 – 2 i.

Solution:
We rewrite w = (3+i) z + 4 – 2 i as u + iv
= (3+i) (x + iy) + 4 – 2 i
or u + iv = (3x - y + 4) + i (x + 3y – 2)
Therefore u = 3x - y + 4 and v = x + 3y – 2

z ↔ (x, y) (u,v) ↔ w
i ↔ (0, 1) (3, 1) ↔ 3 + i
1 + i ↔ (1, 1) (6, 2) ↔ 6 +2 i
1 - i ↔ (1, -1) (8, -4) ↔ 8 - 4i
 z-Plane

z ↔ (x, y) (u,v) ↔ w 2
y

i ↔ (0, 1) (3, 1) ↔ 3 + i 1 0, 1 1, 1

1 + i ↔ (1, 1) (6, 2) ↔ 6 +2 i 0
-1 0 0.5 1 1, -1
x
1.5
1 - i ↔ (1, -1) (8, -4) ↔ 8 - 4i -2

w-plane
3
v
5 2 6, 2
6, 2
u 1 0, 1 1, 1 3, 1
3, 1
0 0
z
0 2 4 6 8 10 -1 0 1, -1
2 4 6 8 10
8, -4 w
-5 -2
-3
-4 8, -4
-5
 1
3. Inversion and Reflection: w (1)
z
Taking z = r ei  and w = R ei  the transformation can be
written as 1  i
1
i
Re   e . P(z)
rei r
 R
1
and    1 P1 .
r O . P'(w)

Definition: The inverse of a point P with respect to a circle

with centre O and radius a is a point P1 on the
line OP such that OP × OP1 = a2.
Thus if P is the point (r, ) and P1 is (1/r, ), then P1 is inverse of
P with respect to unit circle.
The point P' (1/r, -) is reflection of P1 with respect to real axis.
Hence the transformation represents inversion with respect to
Unit circle and reflection with respect to real axis.
 4. Bilinear Transformation:
az  b
The transformation w  (1)
wherecz 
a,db, c, d are arbitrary complex
constants and ad – bc ≠ 0, is known as the bilinear or Möbius
transformation.
Differentiating (1), we get
dw a (cz  d )  (az  b)c ad  bc
 
2
dz (cz  d ) (cz  d )2
Since ad – bc ≠ 0, therefore there is no critical point.
 dw  b
Inverse transformation of (1) is z  (2)
cw  a
For each point in z- plane, except when cz +d = 0, there is a
unique point in finite w-plane. Also for each point in w-plane,
except when cw = a, there is a unique point in z- plane.
Excluding the two exceptional points, we see that there is one
to one correspondence between the points of extended planes.
Invariant Points of Bilinear Transformation
These are the points for which w = z.
For such a point we have
az  b 2
z or cz  ( d  a ) z  b 0
cz  d
The roots of this equation ( say z1 and z2) are defined as the
invariant or fixed points of the bilinear transformation.
If two roots are equal, then bilinear transformation is parabolic.
Properties of Bilinear Transformation
1. A bilinear transformation maps circles into circles.
2. A bilinear transformation preserves cross-ratio of four
points.
( w1  w2 )( w3  w4 )
( w1  w4 )( w3  w2 )
 A bilinear transformation maps circles into circles.
az  b
Consider the bilinear transformation w  , ad  bc 0.
cz  d
a bc  ad 1
Performing the division process, we get w  
c c2 z  d
We define the transformations c
d 1 bc  ad a
w1  z  , w2  , w3  2
w2 , w   w3
c w1 c c
By these transformations, we go from z-plane to w1-plane, from
w1-plane to w2-plane, from w2-plane to w3-plane and finally
from w3-plane to w-plane.
Each one of these transformation is one of the standard
transform w = z +c or w = cz or w = 1/z and each one of these
maps a circle onto a circle.
Hence, A bilinear transformation maps circles into circles.
A bilinear transformation preserves cross-ratio of four points.
az  b
Consider the bilinear transformation w  , ad  bc 0.
cz  d
Let the points z1 , z2 , z3 , of
z4 the z-plane map onto points
w
of1 ,the
w2 w-plane.
, w3 , w4
If these points are finite, we have
azn  b azm  b (ad  bc)( zn  zm )
wn  wm   
czn  d czm  d (czn  d )(czm  d )
Use this relation for n, m = 1, 2, 3, 4, in
( w1  w2 )( w3  w4 )
( w1  w4 )( w3  w2 )
Simplify to obtain
( w1  w2 )( w3  w4 ) ( z1  z2 )( z3  z4 )

( w1  w4 )( w3  w2 ) ( z1  z4 )( z3  z2 )
If one of the points → ∞, then factors containing that point are
replaced by 1 and we obtain
Thus if z1 → ∞, we have

( w1  w2 )( w3  w4 ) z3  z4

( w1  w4 )( w3  w2 ) z3  z2

and if w4 → ∞, we have
w1  w2 ( z1  z2 )( z3  z4 )

w3  w2 ( z1  z4 )( z3  z2 )
Example:
Find the bilinear transformation which maps the points -1, i, 1
of the z-plane to the points 1, i, -1 of the w-plane.

Solution:
It is given that the points z1 = - 1, z2 = i, z3 = 1 in the z-plane are
mapped to w1 = 1, w2 = i, w3 = -1 of the w-plane.
Let the general point z in the z-plane be mapped to the point w, in
the w-plane.
Since the cross-ratio remains unchanged under bilinear
transformation, taking z4 = z and w4 = w in the identity
( w1  w2 )( w3  w4 ) ( z1  z2 )( z3  z4 )

( w1  w4 )( w3  w2 ) ( z1  z4 )( z3  z2 )
(1  i )( 1  w) ( 1  i )(1  z )
we get 
(1  w)( 1  i ) ( 1  z )(1  i)
  1  w  1  i ( 1  i )(1  z )

1 w 1  i ( 1  z )(1  i )

1  w (1  i )2 1  z 2i 1 z
   
w  1 (1  i ) 2  1 z  2i  1  z

1 w 1 z
 
w  1 1 z
Applying componendo dividendo, we get

2w 2

2  2z
1
 w 
z
 Example:
Find the bilinear transformation which maps the points 1, i, -1
of the z-plane to the points 0, 1, ∞ of the w-plane.

Solution:
It is given that the points z1 = 1, z2 = i, z3 = -1 in the z-plane are
mapped to w1 = 0, w2 = 1, w3 → ∞ of the w-plane.
Let the general point z in the z-plane be mapped to the point w, in
the w-plane.
Since the cross-ratio remains unchanged under bilinear
transformation, taking z4 = z and w4 = w in the identity
( w1  w2 )( w3  w4 ) ( z1  z2 )( z3  z4 )

( w1  w4 )( w3  w2 ) ( z1  z4 )( z3  z2 )
Since w3 → ∞, omitting the factors containing w3, we get and
w1  w2 ( z1  z2 )( z3  z4 )

w1  w4 ( z1  z4 )( z3  z2 )
0  1 (1  i )( 1  z )

0  w (1  z )( 1  i )

(1  z )( 1  i )
w
(1  i )( 1  z )
(1  i )(1  z )
w
(1  i )(1  z )

i (1  z )
w
1 z
5. Conformal Mapping or Transformation
Definition
Let two curves C , in
C1the z-plane intersect at the point P and their
image curves , C1
inCw-plane intersect at P' as shown in figure.

Y C1 C'
C V C' 1
 

P(z) P'(z)
O X U
O'
If the angle of intersection of the image curves at P' is same as
the angle of intersection of curves at P in magnitude and sense,
then the transformation is said to be conformal.
 Theorem:
The transformation given by analytic function w = f(z) is
conformal at all those point of the z-plane where f ( z ) 0.
The point for which f ( z ) is
0called critical point of the mapping.
Effects of Conformal Mapping
Under conformal transformation w  f ( z ),
1. Angle between the tangents to the curves C and C' at P and P'
arg[ f ( z )].
is given by
That is tangent is rotated through an angle arg[ f ( z )].
2. Length of any line segment passing through P in any direction
is changed in the ratio  :1, where  | f ( z ) |
Thus any infinitesimal length in z-plane is magnified by the
factor | f ( zand
) | hence any infinitesimal area in z-plane is
 2
magnified by the factor | f ( z ) | .
 Since the function w  f ( z ) is
u analytic,
iv its real and imaginary
parts satisfy Cauchy- Riemann equations.
Therefore
 u, v  u x u y u x  vx 2 2 2
J     u x  v x  f ( z )
 x , y  v x v y
v x u x

Hence under a conformal mapping w  f ( z ) u  iv

 u, v 
• Infinitesimal areas are magnified by the factor J  
 x , y 
• Condition for an analytic function to be a conformal
mapping is
 u, v 
J  0
 x, y 
Example: Discuss the transformation
1
w z  Joukowski Transformation
Answer: z
The function is differentiable at all points of z-plane and hence
is analytic.
The transformation is conformal at those points where its
derivative does not vanish.
dw 1 z2  1
1  
2
dz z z2
Therefore mapping is conformal except for z = 1, -1 which
correspond to w = 2, -2.
Writing the transformation in polar coordinates (r, ), we have
i 1  i  1  1
u  iv  w re  e  r   cos   i  r   sin 
r  r  r
  1  1
 u  r   cos  , v  r   sin  (1)
 r  r
From (1) we obtain
u2 v2
 1 (2)
2 2
 1  1
r   r  
 r  r
u2 v2
and  1 (3)
2 2
4 cos  4sin 
z-plane w-plane
Circle r = 1 Line segment v = 0 , -2 ≤ u ≤ 2
Circle r = ρ ≠ 1 Ellipse with foci (-2, 0) & (2, 0)
Radial line  =  Hyperbola (3)
Complete solution: BS Grewal : Sec 20.10 Complex Integration