CHAPTER II

Analytic Functions

Functions of complex variable

Let S be a set of complex numbers. A function f defined on S is a rule that assigns to each
z  S , there is a complex number w. The number w is called the value of f at z , denoted by
f (z ) and is defined as w  f (z ). The set S is called the domain of f .

Example 1
1
Consider w  f ( z )  on the half plane Re z  0. The domain of f is the set of all nonzero
z
points in the complex plane.

Suppose that w  u  iv  f ( z )  f ( x  iy ). So, each of the real numbers u and v depends on the
real variables x and y. It follows that f (z ) can be expressed in terms of a pair of real-valued
functions u and v of real variables x and y.

Example 2
Consider w  f ( z )  z 2 .
We can write w  u ( x, y )  iv ( x, y )  z 2  ( x  iy ) 2  x 2  y 2  2ixy.
Therefore, we have u ( x, y )  x 2  y 2 and v( x, y )  2 xy.

Example 3
1
Consider w  f ( z )  z  .
z
1 1
We write f (re i )  re i  e i  r (cos   i sin  )  (cos   i sin  )
r r
1 1
 (r  ) cos   i (r  ) sin  .
r r
1 1
Hence, we have u (r ,  )  (r  ) cos  and v(r ,  )  (r  ) sin  .
r r
Therefore, we have u ( x, y )  x  y and v( x, y )  2 xy.
2 2

In polar coordinates r and  , we have u  iv  f (re i ), where w  u  iv and z  re i . We

write f ( z )  u (r ,  )  iv (r ,  ).

Example 4
Consider w  f ( z )  z .
2

We write w  z  x 2  y 2  i 0. So f (z ) is a real-valued function of a complex variable.

2

If n  0, n  N and if a 0 , a1 , a 2 ,  , a n are complex constants, then the function

p ( z )  a 0  a1 z  a 2 z 2    a n z n , ( a n  0 )
is a polynomial in z of degree n . The domain of p (z ) is the entire z-plane. The rational
p( z )
(quotient) function is r ( z )  if q ( z )  0.
q( z )
1
Polynomials and rational functions play a vital role in the theory of functions of complex
variables.
A generalization of the concept of function is a rule that assigns more than one value of z. These
multiple-valued functions occur in the theory of functions of complex variables.

Example 5
Consider z  re i , z  0.
1 
Then z 2   r e where   Arg z ,       . If we choose only the positive value of  r ,
i
2


then f ( z )  r e 2 , r  0,     is a single-valued function.
i

We also have f (0)  0. The function f (z ) is defined in the entire complex plane except    ,
the negative real axis.

Transformation or Mapping
If w is the value of f at z , then w is the image of z. We write w  Im z. Consider a subset
T  S . The set of images of all points in T is the image of T. If T  S , that is, if we consider the
entire domain, then Im f  Im T . Like elementary calculus, we can consider the inverse image of
a point w.

The terms translation, rotation and reflection are used for geometric characteristics of certain
mappings. The mapping w  z  c, where c is a complex constant, is a translation, the mapping
w  iz rotates each nonzero point z counterclockwise through a right angle about the origin and
the mapping w  z transforms each point z into its reflection in the real axis.

General linear transformation

The transformation w  Az  B, where A  0 and A and B are given complex constants, is called
a linear transformation.
(i) Translation ( w  z  B ):
Under this transformation, figures in the z-plane are transformed into w-plane in a direction
of the vector B through the distance equal to B .
(ii) Rotation ( w  e i z with B  0 and A  1 ):
Under this transformation, figures in the z-plane are rotated through an angle  about the
origin. If   0, then the rotation is counterclockwise and if   0, then the rotation is
clockwise.
(iii) Expansion or Contraction ( w  Az with A  1 ):
If 0  A  1, the figure is contracted and if A  1, the figure is expanded.
From the above discussion, we see that w  Az  B is evidently an expansion or contraction and
a rotation followed by a translation because w  re i z   .

1
(iv) Inversion: w  (4)
z

2
 1
This provides a mapping of the z -plane into the w -plane because w  f ( z )  is a single-
z
valued analytic function with only the simple pole at the origin z  0.
1
If we set z  re i  , then the mapping becomes w  e i  . This shows that points on the unit
r
circle z  1 are mapped into points on the unit circle w  1.
1
Because w  , it follows that points inside z  1 are mapped into points outside w  1, and
r
1
conversely. This can be confirmed by taking z  as a test point inside the unit circle z  1,
2
and that it transforms to the point w  2 which lies outside the circle w  1.

Problem 1
Find the image of the rectangle with vertices 1, 4, 4+6i and 1+6i in the z-plane under the
transformation w  (1  i ) z  2  i.
Solution. We can write w  (1  i )( x  iy )  2  i  ( x  y  2)  i ( x  y  1) so that u  x  y  2
and v  x  y  1.
Under this transformation, the points in the z -plane (1,0), (4,0), (4,6) and (1,6) are mapped into
the points (3, 0), (6, 3), (0, 9) and (-3, 6) in the w -plane, respectively.
Since B  2  i  5 , the mapping accomplishes a translation in the direction of 2  i through

the distance equal to 5. Again, since A  1  i   2  1, an anticlockwise rotation of the
i
2e 4


rectangle takes place through an angle about the origin and the figure is expanded by a factor
4
of 2.

Problem 2
Let the rectangular region R in the z -plane be bounded by x  0, y  0, x  2, y  1. Determine
the region R ' of the w -plane into which R is mapped under the following
transformations:
(a) w  z  (1  2i ) ;
i
(b) w  2 e 4 z ;
i
(c) w  2 e 4 z  (1  2i ).
Solution. (a) We have w  z  (1  2i ) . We can write
u  iv  ( x  iy )  (1  2i )  ( x  1)  i ( y  2).
Therefore, u  x  1 and v  y  2.
The line x  0 is mapped into u  1
The line y  0 is mapped into v  2
The line x  2 is mapped into u  3
The line y  1 is mapped into v  1.
Therefore, the points in the z-plane (0,0), (0,1), (2,1) and (2,0) are mapped into the points (1,-2),
(1,-1), (3,-1) and (3,-2) in the w-plane, respectively.

3
Similarly, we can show that each point of R is mapped into one and only one point of R '
and conversely.
The transformation accomplishes a translation of the rectangle. In general, w  z  
accomplishes a translation of any region.
i
(b) We have w  2 e 4 z .
  1 1
i
Now 2 e 4  2 (cos
 i sin )  2 ( i )  1  i.
4 4 2 2
Therefore, u  iv  (1  i )( x  iy )  ( x  y )  i ( x  y ) and u  x  y , v  x  y.
The line x  0 is mapped into u   y, v  y  u  v
The line y  0 is mapped into u  x, v  x  u  v
The line x  2 is mapped into u  2  y, v  2  y  u  v  4
The line y  1 is mapped into u  x  1, v  x  1  u  v  2.

The mapping accomplishes a rotation of R through an angle and a stretching of length 2.
4
In general, the transformation w  z accomplishes a rotation and stretching of a region.

Exercise
Let the rectangular region R in the z -plane be bounded by x  0, y  0, x  2, y  1. Determine
the region R ' of the w -plane into which R is mapped under the transformation
i
w  2 e 4 z  (1  2i ).

Bilinear or fractional transformation

az  b
The transformation w  , ad  bc  0 and a, b, c, d are given complex constants, is called
cz  d
is called a bilinear or linear fractional transformation. This transformation can be considered as
a combination of the transformation of translation, rotation, stretching and inversion.
( z  z 2 )( z 3  z 4 )
If there are four points z1 , z 2 , z 3 , z 4 , taken in order, then the ratio 1 is called the
( z 2  z 3 )( z 4  z1 )
cross-ratio of z1 , z 2 , z 3 , z 4 .

Theorem
az  b
If under the bilinear transformation w  , (ad  bc  0), w1 , w2 , w3 , w4 are the images of
cz  d
z1 , z 2 , z 3 , z 4 , respectively, then the cross-ratio of w1 , w2 , w3 , w4 is equal to the cross-ratio of
z1 , z 2 , z 3 , z 4 .
Proof. Since w1 , w2 , w3 , w4 are the images of z1 , z 2 , z 3 , z 4 , respectively, we have
az1  b az  b az  b az  b (ad  bc)( z1  z 2 )
w1  , w2  2 , w3  3 and w4  4 . So w1  w2  .
cz1  d cz 2  d cz 3  d cz 4  d (cz1  d )(cz 2  d )
Similarly, we have
(ad  bc)( z 2  z 3 ) (ad  bc)( z 3  z 4 ) (ad  bc)( z 4  z1 )
w2  w3  , w3  w4  and w4  w1  .
(cz 2  d )(cz 3  d ) (cz 3  d )(cz 4  d ) (cz 4  d )(cz1  d )

4
 ( w1  w2 )( w3  w4 ) ( z1  z 2 )( z 3  z 4 )
We know  .
( w2  w3 )( w4  w1 ) ( z 2  z 3 )( z 4  z1 )
Hence ( w1 , w2 , w3 , w4 )  ( z1 , z 2 , z 3 , z 4 ).

Problem 1
Find the bilinear transformation which maps the points z1  0, z 2  i and z 3  1 into

w1  i, w2  1 and w3  0, respectively.

Solution. We know that the cross-ratio of w, w1 , w2 , w3 is equal to the cross-ratio of z , z1 , z 2 , z 3 .

( w  w1 )( w2  w3 ) ( z  z1 )( z 2  z 3 )
Therefore, 
( w1  w2 )( w3  w) ( z1  z 2 )( z 3  z )
( w  i )(1  0) ( z  0)(i  1)
 
(i  1)(0  w) (0  i )(1  z )
wi z (i  1)
 
(i  1)( w)  i (1  z )
 ( w  i )(i  iz )  (iz  z )(iw  w)

  iw  iwz  i 2  i 2 z  i 2 wz  iwz  iwz  wz

  iw  1  z  iwz
 iw( z  1)  1  z
z 1
 w  i
z 1
z 1
Hence w  i is the required transformation.
z 1
Problem 2
Find the bilinear transformation which maps the points z1  2, z 2  i and z 3  2 into

w1  1, w2  i and w3  1, respectively.

Solution. We know that the cross-ratio of w, w1 , w2 , w3 is equal to the cross-ratio of z , z1 , z 2 , z 3 .

( w  w1 )( w2  w3 ) ( z  z1 )( z 2  z 3 )
Therefore, 
( w1  w2 )( w3  w) ( z1  z 2 )( z 3  z )
( w  1)(i  1) ( z  2)(i  2)
 
(1  i )(1  w) (2  i )(2  z )
 (2 w  wz  2  z )(3  i )  ( z  2  wz  2 w)(3  i )
  6 w  3wz  6  iz  2iw  iwz  2i  iz  3 z  6  3wz  6 w  iz  2i  iwz  2iw
 12 w  6 z  4i  2iwz  0
 w(iz  6)  3 z  2i

5
 3 z  2i
 w .
iz  6
3 z  2i
Hence w  is the required transformation.
iz  6
Problem 3
Find the bilinear transformation which maps the points z1  , z 2  i and z 3  0 into

w1  0, w2  i and w3  , respectively.

Solution. It is given that w1 , w2 , w3 in the uv-plane are the images of z1 , z 2 , z 3 in the xy-plane.
We know that the cross-ratio of w, w1 , w2 , w3 is equal to the cross-ratio of z , z1 , z 2 , z 3 .
1
Given z1  . Let z1  then z1 '  0.
z1 '
1
Also, given w3  . Let w3  then w3 '  0.
w3 '
We know that the cross-ratio of w, w1 , w2 , w3 is equal to the cross-ratio of z , z1 , z 2 , z 3 .
1 1
( w  w1 )( w2  ) (z  )( z 2  z 3 )
w3 ' z1 '
Therefore, 
1 1
( w1  w2 )(  w) (  z 2 )( z 3  z )
w3 ' z1 '

( w  w1 )( w2 w3 '1) ( zz1 '1)( z 2  z 3 )

 
( w1  w2 )(1  ww3 ' ) (1  z1 ' z 2 )( z 3  z )
w i
 
i  z
1
 w
z
1
Hence w   is the required transformation.
z
Problem 4
z 1
Find the image of Re z  0 under the bilinear transformation w  . Sketch the region
z 1
Re z  0 and its image.
z 1
Solution. Under the given bilinear transformation w  , the images of z-plane are given as
z 1
follows:
0 1
(i) The image of (0, 0) in the z-plane will be w   1, i.e., (-1, 0) in the w-plane.
0 1
i 1
(ii) The image of (0, 1) in the z-plane will be w   i, i.e., (0, 1) in the w-plane.
i 1
6
  i 1
(iii) The image of (0, -1) in the z-plane will be w   i, i.e., (0, -1) in the w-plane.
 i 1
1
1
z 1 z  1, i.e., (1, 0) in the w-plane.
(iv) The image of (0,  ) in the z-plane will be w  
z 1 1
1
z
1
1
z 1 z  1, i.e., (1, 0) in the w-plane.
(iv) The image of (0,   ) in the z-plane will be w  
z 1 1
1
z
Finally, any point z  a  ib with a  0 maps into
a  ib  1 (a  1)  ib
w 
a  ib  1 (a  1)  ib

a  ib  1 (a  1) 2  b 2
and so w    1 which represents the region within the circle w  1.
a  ib  1 (a  1) 2  b 2

Hence, under the given bilinear transformation, the image of Re z  0 in the z-plane is the region
within the circle w  1 in the w-plane.

Special case of bilinear transformation

1
(Inversion or Reciprocal mapping: w  )
z
1 1 x  iy x y
w   2  2 i 2
z x  iy x  y 2
x y 2
x  y2
x y
 u  iv  i 2
x y
2 2
x  y2
x y u v
Hence u  , v 2 and x  2 , y 2 .
x y
2 2
x y 2
u v 2
u  v2
1
The mapping w  transforms lines and circles into circles and lines.
z
To show this, consider the circle a ( x 2  y 2 )  bx  cy  d  0, b 2  c 2  4ad , a  0 (1)
u v
Substituting the values of x  , and y  2 in a ( x 2  y 2 )  bx  cy  d  0, we get
u v
2 2
u  v2
d (u 2  v 2 )  bu  cv  a  0 (2)

7
Setting, a  0, b  1, c  0, d  c1 in (1), we get the line x  c1 in the z-plane and its image will
u
be  c1 (u 2  v 2 )  u  0  u 2  v 2  0
c1

1 2 1
 (u  )  v2  2 ,
2c1 4c1
1 1
which represents a circle with centre at ( ,0) and radius .
2c1 2c1

Similarly, if we set a  0, b  1, c  0, d  c1 in (1), we get the line x  c1 in the z-plane and its
1 1
image will be a circle having centre at ( ,0) and radius .
2c1 2c1

Again, setting, a  0, b  0, c  1, d  c 2 in (1), we get the line y  c 2 in the z-plane and its
v
image will be  c 2 (u 2  v 2 )  v  0  u 2  v 2  0
c2

1 2 1
 u 2  (v  )  v2  2 ,
2c 2 4c 2
1 1
which represents a circle with centre at (0, ) and radius .
2c 2 2c 2

Similarly, setting a  0, b  0, c  1, d  c 2 in (1), we get the line y  c 2 in the z-plane and its
1 1
image will be a circle having centre at (0, ) and radius .
2c 2 2c 2

Problem 5
1 1
Find the image of the infinite strip 0  y  under the bilinear transformation w  . Sketch
2c z
1
the strip 0  y  and its image.
2c
1
Solution. We know for a reciprocal mapping w  ,
z
x y u v
u , v 2 and x  2 , y 2 .
x y
2 2
x y 2
u v 2
u  v2
y 1
Since y  0, we get v   0 and since y  , we get
x y
2 2
2c
v 1 1 2
  v  (u  v 2 ).
u v22
2c 2c

8
Since c  0, we have

 2cv  u 2  v 2  u 2  v 2  2cv  u 2  (v  c) 2  c 2

Since u 2  (v  c) 2  c 2 represents a circle with centre at (0,c) and radius c, u 2  (v  c) 2  c 2

represents the region outside the circle u 2  (v  c) 2  c 2 .

For points in the z-plane, we should have the idea of an  -neighborhood z  z 0   of a given
point z 0 . It consists of all points z lying inside but not on a circle centered at z 0 and with a
specified positive radius  . A deleted neighborhood 0  z  z 0   consists of all points z in an
 -neighborhood of z 0 except for the point z 0 itself.
When working with functions of a complex variable, it is necessary to generalize the related
concepts of limit, continuity and differentiability by extending the corresponding definitions
from elementary calculus.

Limits
Let a function f be defined at all points z in some deleted neighborhood of z 0 . Then the limit of
f (z ) as z approaches z 0 is a number w0 , or lim f ( z )  w0 means that the point w  f (z ) can be
z  z0

made arbitrarily close to w0 if we choose the point z close enough to z 0 but distinct from it.
The statement lim f ( z )  w0 means: For every positive number  , there is a positive number
z  z0

 such that f ( z )  w0   whenever 0  z  z 0   .

Geometrically, for each  -neighborhood w  w0   of w0 , there is a deleted  -neighborhood
0  z  z 0   of z 0 such that for every point z it has an image w lying in the  -neighborhood.
The function w  f ( z )  u ( x, y )  iv ( x, y ) (1)
has the limit L which can be written in the form as
lim f ( z )  L  L1  iL2 (2)
z  z0

where L1 and L2 are real numbers if

lim f ( z)  L  0 (3)
z  z0 0

If z  x  iy and z 0  x0  iy 0 , then z  z 0 whenever ( x, y )  ( x0 , y 0 ). So, in this case,

lim f ( z)  L  0 (4)
( x , y )  ( x0 , y 0 )

However, by the triangle inequality,

f ( z )  L  u ( x, y )  iv( x, y )  ( L1  iL2 )
 u ( x, y )  L1  i (v( x, y )  L2 )
 u ( x, y )  L1  v( x, y )  L2
Therefore, lim f ( z )  L if
z  z0

lim u ( x, y )  L1 and lim v ( x , y )  L2 (5)

( x , y )  ( x0 , y 0 ) ( x , y )  ( x0 , y 0 )

These show the relationship between the limit of a function f (z ) for a complex variable z and
the limits of the real functions u ( x, y ) and v( x, y ) for real variables x, y. Therefore, all of the

9
properties of limits of functions for real variables hold for limits of functions for complex
variables. For example,
lim  f ( z )  g ( z )   lim f ( z )  lim g ( z ) (6)
z  z0 z  z0 z  z0

lim  f ( z ) g ( z )    lim f ( z )  lim g ( z )  (7)

z  z0  z  z0  z  z0 

lim f ( z )
f ( z ) z  z0
lim  ; if lim g ( z )  0 (8)
z  z0 g ( z ) lim g ( z ) z  z0
z  z0

Limit of a function is unique.

Problems
1. Using the definition of limit show that
lim [ x  i (2 x  y )]  1  i where z  x  iy.
z 1i

Solution. By definition, for each   0, however small, we can find a   0 such that
[ x  i (2 x  y )]  (1  i )   whenever z  (1  i )   (1)
That means ( x  1)  i (2 x  y  1)   whenever ( x  iy )  (1  i )  
i.e., ( x  1)  i (2 x  y  1)   whenever ( x  1)  i ( y  1)   (2)
Since ( x  1)  i (2 x  y  1)  x  1  2 x  y  1  x  1  2( x  1)  ( y  1)
 x 1  2 x 1  y 1  3 x 1  y 1 (3)
 
we see that if x  1  and y  1  then the inequality in (3) is satisfied. Now
4 4
z  (1  i )  ( x  1)  i ( y  1)  x  1  y  1 (4)
  
Hence z  (1  i )     .
4 4 2
Re z
2. Find lim f ( z ), where f ( z )  ; z  0.
z0 z
Solution. If z  0 along the positive real axis, then
Re z x x
lim f ( z )  lim  lim  lim 1
z 0 z 0 z y  0 , x  0  x  iy x  0 x
and if z  0 along the negative real axis, then
Re z x
lim f ( z )  lim  lim   1.
z 0 z 0 z y  0, x 0  x

Since the limiting values are not unique, lim f ( z ) does not exist.
z0

z
2
3. If f ( z )  ; z  0 find lim f ( z ).
( z )2 z0

z2 x 2  y 2  2ixy
Solution. We can write f ( z )   2
( z ) 2 x  y  2ixy
2

If z  0 along the Re z-axis, then

10
 x 2  y 2  2ixy
lim f ( z )  lim 1
z 0 y  0, x  0 x 2  y 2  2ixy
If z  0 along the Im z-axis, then
x 2  y 2  2ixy
lim f ( z )  lim 1
z 0 x  0, y  0 x 2  y 2  2ixy

If z  0 along the line y  mx , then

x 2  y 2  2ixy x 2  m 2 x 2  2ix (mx )
lim f ( z )  lim  lim
z 0 y  mx , x  0 x 2  y 2  2ixy x 0 x 2  m 2 x 2  2ix(mx)
1  m 2  2im
 lim  1 if m  0.
x 0 1  m 2  2im

Since the limiting values of f (z ) are not unique, lim f ( z ) does not exist.
z0

Limits involving the points at infinity

The statement lim f ( z )  w0 means: for each   0, we can find a   0 such that
z 0

1
f ( z)  whenever 0  z  z 0  

1 1
  0  whenever 0  z  z 0   .
f ( z) 

Consider the following properties:

1 iz  3 z 1
(i) lim f ( z )    lim  0, e.g., lim   because lim  0.
z  z0 z  z0 f ( z ) z  1 z  1 z  1 iz  3

1
(ii) lim f ( z )  w0  lim f    w0
z  z 0
z
2z  i (2 / z )  i 2  iz
e.g., lim  2 because lim  lim  2.
z  z  1 z 0 (1 / z )  1 z 0 1  z

1
(iii) lim f ( z )    lim 0
z  z 0 f (1 / z )

z2 1 (1 / z )  1 z (1  z )
e.g., lim   because lim  lim  0.
z  z 1 z 0 (1 / z )  1
2 z 0 1  z 2

Continuous Functions
A function f is said to be continuous at a point z 0 if the following properties are satisfied:
(i) f ( z 0 ) is defined;
(ii) lim f ( z ) exists;
z z 0

(iii) lim f ( z )  f ( z 0 ).
z  z0

If one or more of the conditions in this definition fails to hold, then f is called discontinuous at
z 0 and z 0 is called a point of discontinuity of f .
In terms of real functions, f ( z )  u  iv is continuous at z 0  x0  iy 0 if
lim u ( x, y )  u ( x0 , y 0 ) and lim v( x, y )  v( x0 , y 0 ).
( x , y )  ( x0 , y 0 ) ( x , y )  ( x0 , y 0 )

11
Theorem: For a complex function f (z ), the following conditions hold:
(i) If f (z ) and g (z ) are continuous, then so are f ( z )  g ( z ) and f ( z )  g ( z ).
f ( z)
(ii) If f (z ) and g (z ) are continuous at a point z 0 , then is also continuous provided that
g ( z)
g ( z 0 )  0.

Differentiable functions
A function f is said to be differentiable at a point z 0 if
f ( z)  f ( z0 )
f ' ( z 0 )  lim (1)
z  z0 z  z0
provided that limit exists. If f ' ( z 0 ) exists, we say that f is differentiable at z0 or f has a
derivative at z 0 .
Let z  z 0  z. Then z  z 0  z and then (1) takes the form
f ( z 0  z )  f ( z 0 )
f ' ( z 0 )  lim (2)
z  0 z
By dropping the subscript, we have the general formula for derivative of a complex function
f (z ) given by
f ( z  z )  f ( z )
f ' ( z )  lim (3)
z 0 z
is known as the differential coefficient of f (z ).
Let w  f (z ), then w  w  f ( z  z ). So f ( z  z )  f ( z )  w and then (3) becomes
w dw d
f ' ( z )  lim   f ( z ).
z  0 z dz dz
Example 1
Show that f ( z )  z is differentiable only at z  0.
2

f ( z  z )  f ( z ) z  z  z
2 2

Solution. We have f ' ( z )  lim  lim

z 0 z z  0 z
( z  z )( z  z )  z z z z  z z  z z  z z  z z
 lim  lim
z  0 z  z  0 z
 z 
Hence f ' ( z )  lim  z  z  z 
 (1)
z 0
  z 
If z  0 along the Re z -axis, then y  0 . So z  x  iy  x  z  z.
 
In this case, f ' ( z )  lim z  z  z  z  z
z 0
(2)
If z  0 along the Im z -axis, then x  0 and z   z for z  x  iy  iy.
 
In this case, f ' ( z )  lim z  z  z  z  z
z 0
(3)
Therefore, f ' ( z ) exists if
z  z  z  z  2z  0  z  0
showing that f ( z )  z is differentiable only at z  0.
2

12
Example 2
Test the function
 ( z) 2
 , z0
f ( z)   z
0, z0


for continuity and differentiability at z  0.

( x  iy ) 2 x 2  y 2  2ixy
Solution: We have z  x  iy and z  x  iy. So that f ( z )  
x  iy x  iy
x 2  y 2  2ixy
Now lim f ( z )  lim
z 0 ( x , y ) ( 0, 0 ) x  iy
If z  0 along the Re z -axis, then
x 2  y 2  2ixy
lim f ( z )  lim  lim x  0.
z 0 y  0, x 0 x  iy x 0

If z  0 along the Im z -axis, then

x 2  y 2  2ixy
lim f ( z )  lim  lim ( y )  0.
z 0 x  0 , y 0 x  iy y 0

If z  0 along the line y  mx, then

x 2  y 2  2ixy x 2  m 2 x 2  2ix(mx)
lim f ( z )  lim  lim
z 0 y  mx , x 0 x  iy x 0 x  i (mx)
(1  m 2  2im) x 2
 lim  0,
x 0 (1  im) x
showing that lim f ( z )  0  f (0).
z 0

Hence f (z ) is continuous at z  0.

f ( z  z )  f ( z )
Again, we have f ' ( z )  lim so that
z 0 z
( z ) 2
0
f (z )  f (0) ( z ) 2 ( x  i y ) 2
f ' (0)  lim  lim z  lim  lim
z 0 z z 0 z z 0 ( z ) 2 ( x , y )( 0 , 0 ) ( x  iy ) 2

If z  0 along the Re z -axis, then

( x  i y ) 2
f ' (0)  lim 1
y  0 , x  0 ( x  iy ) 2

If z  0 along the Im z -axis, then

( x  i y ) 2
f ' (0)  lim 1
x  0 , y  0 ( x  iy ) 2

If z  0 along the line y  mx, then

(x  iy ) 2 (x  imx) 2 (1  mi )2
f '(0)  lim  lim  lim 1
y  m x , x 0 ( x  iy ) 2 x  0 ( x  imx ) 2 x  0 (1  mi ) 2

for any values of m showing that f ' (0) does not exist.
Hence f (z ) is not differentiable at z  0.

13
Cauchy-Riemann equations
A single-valued function f is said to be analytic in a domain D of the finite complex plane if
f ' ( z ) exists at all points z in D. In particular, f (z ) is analytic at a point z 0 if it is analytic in a
neighborhood of z 0 .

Examples
1
(i) At z  0, the function f ( z )  is not analytic.
z
(ii) Since the function f ( z )  z is differentiable only at z  0, so f (z ) is not an analytic
2

function.
A function that is analytic everywhere in the finite complex plane is said to be an entire function.
Since a polynomial function is differentiable everywhere in the finite complex plane, it follows
that every polynomial function is an entire function.
The point at which a function f fails to be analytic is known as a singular point. In particular, if
a function f is not analytic at z 0 but is analytic at some points in every neighborhood of z 0 ,
then z 0 is a singular point of f (z ).

Example
1
The function f ( z )  has singularity at z  0 but the function f ( z )  z has no singularity.
2

z
f ( z 0  z )  f ( z 0 )
Let f ( z )  u ( x, y )  iv ( x, y ) and suppose that f ' ( z 0 )  lim exists.
z  0 z
If z  x  iy and z 0  x0  iy 0 , then
u ( x0  x, y 0  y )  iv ( x 0  x, y 0  y )  u ( x0 , y 0 )  iv ( x0 , y 0 )
f ' ( z 0 )  lim
z 0 x  i y

If z  0 along the Re z -axis in the finite complex z -plane, then

u ( x0  x, y 0 )  iv ( x0  x, y 0 )  u ( x0 , y 0 )  iv ( x0 , y 0 )
f ' ( z 0 )  lim
x 0 x
u ( x 0  x , y 0 )  u ( x 0 , y 0 ) v ( x 0  x , y 0 )  v ( x 0 , y 0 )
 lim  i lim
x  0 x x 0 x
 u x ( x0 , y 0 )  iv x ( x0 , y 0 ).
Again, if z  0 along the Im z -axis in the finite complex z -plane, then
u ( x0 , y 0  y )  iv ( x0 , y 0  y )  u ( x0 , y 0 )  iv ( x0 , y 0 )
f ' ( z 0 )  lim
y 0 iy
u ( x 0 , y 0  y )  u ( x 0 , y 0 ) v( x0 , y 0  y )  v( x0 , y 0 )
 i lim  lim
y 0 y y 0 y
 iu y ( x0 , y 0 )  v y ( x0 , y 0 ).
If f ' ( z 0 ) exists, then u x ( x0 , y 0 )  iv x ( x 0 , y 0 )  iu y ( x 0 , y 0 )  v y ( x0 , y 0 ) implying that
u x ( x0 , y 0 )  v y ( x0 , y 0 ) and u y ( x0 , y 0 )  v x ( x0 , y 0 )
which are called the Cauchy-Riemann equations, named in honor of the French Mathematician
A. L. Cauchy (1789-1857) and in honor of the German Mathematician G. F. B. Riemann (1826-
1866).
14
Theorem 1
The necessary conditions for a function f ( z )  u ( x, y )  iv ( x, y ) to be analytic in a domain D is
that u ( x, y ) and v( x, y ) satisfy the Cauchy-Riemann equations
u v u v
 and 
x y y x
u u v v
provided that , , , exist.
x y x y
u u v v
If , , , are continuous in a domain D, then the Cauchy-Riemann equations are
x y x y
sufficient conditions for f (z ) to be analytic in D.
The functions u ( x, y ) and v( x, y ) are called conjugate functions.

Harmonic Functions
A real valued function h of two real variables x and y is said to be harmonic in a domain D of the
z-plane if throughout D it has continuous partial derivatives of the first and second order and
satisfies the Laplace’s equation hxx ( x, y )  h yy ( x, y )  0.

Theorem 2
If a function f ( z )  u ( x, y )  iv ( x, y ) is analytic in a domain D, then u ( x, y ) and v( x, y ) are
harmonic in D.
In this case, the first order partial derivatives of u ( x, y ) and v( x, y ) satisfy the Cauchy-Riemann
equations throughout D and v is called the harmonic conjugate of u.

Theorem 3
A function f ( z )  u ( x, y )  iv ( x, y ) is analytic in a domain D if and only if v is a harmonic
conjugate of u.

Example 1
Prove that the function u ( x, y )  x 2  y 2  2 xy  2 x  3 y is harmonic. Hence find a function
v( x, y ) such that f ( z )  u ( x, y )  i v( x, y ) is analytic.

Solution. We have u  x 2  y 2  2 xy  2 x  3 y
u  2u
So  2 x  2 y  2 and so 2
x x 2
u  2u
Again,  2 y  2 x  3 and so  2
y y 2
 2u  2u
Thus   0, i.e., u is harmonic.
x 2 y 2
u v
From the Cauchy-Riemann equations, we know   2x  2 y  2
x y

15
 v
Therefore,  2x  2 y  2 (1)
y
Integrating with respect to y, keeping x as constant, we get
v  2 xy  y 2  2 y  F ( x) (2)
where F (x) is an arbitrary real function of x.
Differentiating with respect to x, we get
v
 2 y  F ' ( x) (3)
x
u v
We also have     2 y  2 x  3.
y x
Therefore,  2 y  2 x  3  2 y  F ' ( x)
 F ' ( x)  2 x  3
Integrating with respect to x, we get F ( x)  x 2  3 x  c
So from (2), v  2 xy  y 2  2 y  x 2  3 x  c implying that
v( x, y )  x 2  y 2  2 xy  3 x  2 y  c.

Example 2
Prove that the function u ( x, y )  e  x ( x sin y  y cos y ) is harmonic. Find a function v( x, y ) such
that f ( z )  u ( x, y )  i v( x, y ) is analytic and write f (z ) in terms of z.
Solution. (First Part)
We have u ( x, y )  e  x ( x sin y  y cos y ) so that
u
 e  x (sin y )  e  x ( x sin y  y cos y )  e  x (sin y  x sin y  y cos y ),
x
 u
2
 e  x ( sin y )  e  x (sin y  x sin y  y cos y )  e  x (2 sin y  x sin y  y cos y ) (1)
x 2

u
Again,  e  x ( x cos y  cos y  y sin y )  e  x ( x cos y  cos y  y sin y ),
y
 2u
 e  x ( x sin y  sin y  y cos y  sin y )  e  x (2 sin y  x sin y  y cos y ) (2)
y 2
 2u  2u
Adding (1) and (2),   0 and hence u ( x, y ) is harmonic.
x 2 y 2

Second Part: From the Cauchy-Riemann equations, we have

u v
  e  x (sin y  x sin y  y cos y ) (1)
x y
v u
and   e  x ( x cos y  cos y  y sin y ).
x y
Integrating (1) with respect to y by keeping x constant, we get
v  e  x ( cos y  x cos y  y sin y  cos y )  F ( x),
where F (x) is an arbitrary real function of x,
 v  e  x ( x cos y  y sin y )  F ( x) (2)

16
 v
So  e  x cos y  e  x x cos y  e  x y sin y  F ' ( x).
x
v u
But we have  .
x y
Therefore, e cos y  e  x x cos y  e  x y sin y  F ' ( x)  e  x ( x cos y  cos y  y sin y )
x

 F ' ( x)  0 and so F ( x)  c where c is a constant.

Thus, (2) becomes
v( x, y )  e  x ( x cos y  y sin y )  c.

Last Part: We have u ( x, y )  e  x ( x sin y  y cos y ) and v( x, y )  e  x ( x cos y  y sin y )  c

So that f ( z )  u ( x, y )  i v( x, y )  e  x ( x sin y  y cos y )  e  x ( x cos y  y sin y )  c
 e  x ( x sin y  y cos y )  ie  x ( x cos y  y sin y )  ic
 e  x ( x sin y  y cos y  ix cos y  iy sin y )  ic
 e  x {( x  iy ) sin y  i ( x  iy ) cos y}  ic
 e  x ( x  iy )(sin y  i cos y )  ic
 e  x ( x  iy )ie  iy  ic
 i ( x  iy )e  ( x  iy )  ic
Hence f ( z )  ize  z  ic.

Example 3
1
(a) Prove that u ( x, y )  ln( x 2  y 2 ) is a harmonic function.
2
(b) Find v( x, y ) such that f ( z )  u ( x, y )  i v( x, y ) is an analytic function.
(c) Write f (z ) in terms of z.
Solution
1
(a) We have u ( x, y )  ln( x 2  y 2 )
2
x y2  x2 y x2  y2
So, u x  2 , u  , u  , u 
x  y2 (x 2  y 2 )2 x2  y2 (x2  y 2 )2
xx y yy

Since u xx  u yy  0, so that u ( x, y ) is harmonic.

(b) From the Cauchy-Riemann equations, we have

u v x
  2 (1)
x y x  y 2
u v y
and    2 (2)
y x x  y2
Integrating (1) with respect to y by keeping x constant, we get
1 y y
v  x tan 1  F ( x)  tan 1  F ( x),
x x x
where F (x) is an arbitrary real function of x.

17
 y 1  y  y
So from (2),   2 
 2   F ' ( x)   2  F ' ( x)
x  y2
2
y  x  x  y2
1 2
x
 F ' ( x)  0 and so F ( x)  c, where c is an arbitrary constant.
y
Hence, v( x, y )  tan 1  c.
x
1  y 
(c) Finally, f ( z )  u ( x, y )  iv ( x, y )  ln( x 2  y 2 )  i tan 1  c 
2  x 
z z
 log z  ln  ic  ln( z  )  ic  ln z  ic
z z
y
Note: We have x 2  y 2  z  z z , r  z  x 2  y 2 and   tan 1 .
2

x
z z y z
Also, z  re i  ln z  ln r  i  i  ln  i  ln  i tan 1  ln .
r z x z

Example 4

If f ( z )  e z , then f ' ( z ) does not exist at any point.

Solution. We can write
f ( z )  u  iv  e z  e x iy  e x (cos y  i sin y )
 u ( x, y )  e x cos y and v( x, y )  e x sin y
So that u x ( x, y )  e x cos y, u y ( x, y )  e x sin y, v x ( x, y )  e x sin y and v y ( x, y )  e x cos y.
For f (z ) to be differentiable at any point, Cauchy-Riemann equations must be satisfied at that
point. So u x  v y and u y  v x .
Hence e x cos y  e x cos y  2e x cos y  0  cos y  0 since e x  0
and  e x sin y  e x sin y  2e x sin y  0  sin y  0 since e x  0.
Since there is no y for which cos y and sin y are simultaneously zero, we can conclude that
Cauchy-Riemann equations are not satisfied at any point. Hence f ' ( z ) does not exist at any
point.

Example 5
If f ( z )  x 3  i (1  y ) 3 , show that f ' ( z ) exists only when z  i.
Solution. Given f ( z )  x 3  i (1  y ) 3
 u ( x, y )  x 3 and v( x, y )  (1  y ) 3 .
So that u x ( x, y )  3 x 2 , u y ( x, y )  0, v x ( x, y )  0 and v y ( x, y )  3(1  y ) 2 .
u u v v
Since f (z ) is defined and is continuous everywhere in the finite z-plane and , , ,
x y x y
exist and are continuous everywhere in the finite z-plane, so for the existence of f ' ( z ), Cauchy-
Riemann equations must be satisfied, i.e., u x  v y and u y  v x .
So 3x 2  3(1  y ) 2  x 2  (1  y ) 2  0 .
Since x and y are real, the above equation is satisfied only when x  0 and 1  y  0.
Hence f ' ( z ) exists only when z  0  1i  i.
18
Exercises
Prove that the following functions are harmonic:
(i) u ( x, y )  y 3  3 x 2 y
(ii) u ( x, y )  3 x 2 y  2 x 2  y 3  2 y 2
(iii) u ( x, y )  e x ( x cos y  y sin y )
Hence find a function v( x, y ) for each of the functions u ( x, y ) such that f ( z )  u ( x, y )  i v( x, y )
is analytic.

19