1

Lecture 4

Properties of Logarithmic Function (Contd…)

Since, Log z  ln z  i Arg z
1
u  Re Log z  ln( x 2  y 2 )
2

y
v  Im Log z  tan 1  constant
x

x y
It follows that u x  2  v y , u y   v x
x y 2
x y
2 2

This shows that Re Log z and Im Log z are (i) continuous in
C {z : Re z  0, Im z  0} (ii) partially differentiable and first
order partial derivatives are continuous in
C {z : Re z  0, Im z  0} (iii) Cauchy‐Riemann equations hold.

Therefore,

 Log z is analytic in C {z : Re z  0, Im z  0} and
d x i y z 1
Log z  u x  i v x  2  2  .
dz x y 2
z z

 The branch of logarithm log0 z , with 0  arg z  0  2 , is
a single valued function, and its properties are similar to
the above properties of Log z .
 2

Exponential Function.

Define e z  e x (cos y  i sin y )

Note that e x   and e  x  0 as x  . But, lim eiy does not
y 
exist, since cos n takes the values 1 and ‐1 for even and odd n,
respectively. Consequently,

 lim e z does not exist.
z 
 It follows easily by using the truth of CR equations for e z
and the continuity of first order partial derivatives of real
and imaginary parts of e z , that e z is analytic for all z and
d z
e  e z .
dz
 Since, e z  e x  0 , it follows that e z  0 for any z .

 e z is a periodic function of complex period 2 i , since
e z  2 i  e x i ( y  2 )  e z .

 3
The diagram sketched below illustrates that the fundamental
period strip {z  x  iy :   x  ,    y   } is mapped one‐
one on‐to C  {0} by the function e z .


ez

0  z 
0
 w



The above diagram is explained by the following analytical
arguments:

w  e x (cos y  i sin y )  u  e x cos y , v  e x sin y ,    y  
 e 2 x  u 2  v 2 or x  ln w , w  0
and
v
1 1 v 1 v
y  Arg w  Tan ; Tan   or Tan 
u u u
( According as u  0; u  0, v  0 or u  0, v  0)

  a unique z  ( x, y ), with    x  ,    y   , such that
Log w  z  e z  w .


 4

 Log z is inverse function of e z , since e Log z  z and
Log e z  z .

In fact, that any branch of logarithm is inverse of
exponential function, can be seen as follows:

Let, z  r (cos   i sin  )  x  iy . Then,
log0 z  ln z  i arg z, where arg z   , with 0    0  2 .

log0 z ln z
e e .ei (arg z )
 e .ei
ln z

 z ei
 z .

Similarly,

log0 e z  ln e z  i arg e z ,  
where,   2k0  arg ( e z )    2k0 for some integer k0 .

 ln e x  i y
= z.

 5
Note: After introducing Power Series in the sequel we will be
able to prove that

(i)If f is differentiable in the entire complex plane C, f(0) = 1 and
f ( z )  f ( z ) for all z, then f(z) is the exponential function.

(ii) If f is differentiable in the entire complex plane C,
f(0) = f (0) =1 and f ( z1  z2 )  f ( z1 ). f ( z2 ) for all z1 and z2 , then
f(z) is an exponential function.

Another characterization of the exponential function can be
found in G. P. Kapoor, A new characterization of the exponential
function, Amer. Math. Monthly (1974).
 6


Trigonometric and Hyperbolic Functions. The definitions of
Trigonometric and Hyperbolic Functions of complex valued
functions of a complex variable as given below are analogous
to corresponding Trigonometric and Hyperbolic Functions of
real valued functions of a real variable.

However, some of properties of Trigonometric and Hyperbolic
functions of a complex variable as pointed out below are
drastically different from corresponding functions of a real
variable.

Definitions
eiz  e iz eiz  e iz sin z cos z
cos z  , sin z  , tan z  , cot z  ,
2 2i cos z sin z
1 1
cosec z  , sec z  .
sin z cos z

e z  e z e z  e z
cosh z   cos iz , sinh z   i sin iz ,
2 2
sinh z cosh z
tanh z   i tan iz , coth z   i cot iz ,
cosh z sinh z
1 1
cosech z  , sech z 
sinh z cosh z

 7
Properties: The following properties of Trigonometric and
Hyperbolic functions of a complex variable are drastically
different from corresponding functions of a real variable, rest
of standard properties are analogous:


1. sin z  sin x cosh y  i cos x sinh y ,
cos z  cos x cosh y  i sin x sinh y , for z  x  iy .
Similar identities for other Trigonometric and Hyperbolic
functions can also be easily derived.

2 2
2. sin z  sin 2 x  sinh 2 y , cos z  cos2 x  sinh 2 y for
z  x  iy

3. sin z and cos z are unbounded functions in C (A complex
valued function f ( z ) is said to be bounded in a set A if
f ( z )  M for some M and all z  A , otherwise it is said to
be unbounded). Since, sin iy   and cos iy  
as y   , sin z and cos z are unbounded functions in C.
Recall that, since sin x  1 and cos x  1 for all x R ,
sin x and cos x are bounded functions in R.
 8

Harmonic Conjugate:

Let u :C  R be a harmonic function, i.e. the function u and its

partial derivatives up to the second order are continuous and
satisfy the Laplace’s equation
 2u  2u
 2  0 .
x 2
y

Definition: A function v : C  R is said to be Harmonic
Conjugate of harmonic function u :C  R, if

u x  v y and v x  u y .

The following Leibnitz Rule of differentiation under integral sign
is needed for proving the existence of harmonic conjugate:

Let  : [a , b]  [c, d ]  C be continuous. Define
b
g (t )    ( s, t ) ds .
a

If exists and is cont. on [a, b]  [c, d ], then g is diff. &
t
 ( s, t )
b
g (t )  
 ds
a
t




 9
Theorem 1 (Existence of Harmonic Conjugates).

Let G  B (0, R ), 0  R   and u : G  R be harmonic. Then, u
has a harmonic conjugate in G

Proof. Define v(x, y) by ( x, y )
y
L1
v ( x, y )   u x ( x, t ) dt   ( x )  R
0 0 L2
and determine  ( x ) such that v x  u y .
(Note that v y  u x is obviously satisfied)

The above integral is well defined since L1  B (0, R ) .

Using Leibnitz Rule,

y
v x ( x, y )   u xx ( x, t ) dt   ( x )
0
y
   u yy ( x, t ) dt   ( x )
0
= u y ( x, y )  u y ( x,0)   ( x )
  ( x )  u y ( x,0) (sin ce v x  u y )
y x
Consequently, v ( x, y )   u x ( x, t ) dt   u y ( s,0) ds
0 0
(since L2  B (0, R ) )
is the required harmonic conjugate.
 10

Notes.

1. The proof of above theorem gives a method to construct
harmonic conjugate of a given harmonic function in the
disk B(0, R ), 0  R  

2. The arguments of the proof and the result of the above
theorem are valid for any domain G that is convex both in
the direction of x and direction of y.

3. If G  B ( c, R ) , where c = (a, b), the above arguments could
be modified to give the harmonic conjugate as:
y x
v ( x, y )   u x ( x, t ) dt   u y ( s, b) ds
b a
 11

Remarks.

1. Harmonic conjugate of u is unique up to a constant (To see
this, let v and w be two harmonic conjugates of u. Then,

ux  v y  wy

 u y  v x  wx

Therefore,

v y  wy  v  w   ( x)

v x  wx  v  w   ( y )

so that  ( x )   ( y )  constt.

2. The function f, whose real part is u is determined uniquely
up to purely imaginary constant (Since harmonic
conjugate v of u is unique up to a real constant, f = u + i v is
uniquely determined up to a purely imaginary constant.)








 12


Examples.

1. u ( x, y )  x 2  y 2 .
Since, u x ( x, y )  2 x, u y ( x, y )  2 y , the harmonic conjugate is
y x
v ( x, y )   u x ( x, t ) dt   u y ( s,0) ds
0 0
y x
  2 x dt   0 ds  2 xy .
0 0

2. u( x, y )  2 xy .
Since u x ( x, y )  2 y , u y ( x, y )  2 x , the harmonic conjugate is

y x
v ( x, y )   u x ( x, t ) dt   u y ( s,0) ds
0 0
y x
  2t dt   2 s ds  y 2  x 2 .
0 0
The corresponding analytic function is
f ( z )  2 xy  i ( y 2  x 2 )
 i (( x 2  y 2 )  2ixy )

 iz 2

 13

Other Methods to find Harmonic Conjugates

Method 2. If u is harmonic in a region contained in {z: z > 0}
(i.e., x > 0, y > 0 or first quadrant) and homogenous of degree
1
m, m  0 , i.e. for any t > 0, u(tz )  t m u ( z ) , then v  ( yu x  xu y )
m
is a conjugate harmonic function of u.

Proof. Since u ( x, y ) is a homogenous function, by Euler’s
Formula (see the derivation after this proof),
1
u( x, y )  ( xu x  yu y )
m
1
To show that v  ( yu x  xu y ) is the harmonic conjugate.
m
It is easily verified that
1
u x  (u x  xu xx  yu yx )
m

1
v y  (u x  yu xy  xu yy )
m
 u x  v y (since, u xy is continuous and u satisfies Laplace’s
equation.

The equation u y   v x is verified similarly.




 14

Example. Find an analytic function whose real part is
u( x, y )  x 2  y 2  xy .

The function u is homogenous of degree 2 and harmonic for all
z = x + i y. Therefore, by Method 1 above,

1
v ( x, y )  [ y (2 x  y )  x ( 2 y  x )]
2
1 2
 2 xy  ( y  x 2 )
2

The corresponding analytic function is therefore

1
f ( z )  u  iv  (1  i ) z 2 .
2



 15


Derivation of Euler’s Formula for Homogenous Functins:


d u x  u y 
(u(tz ))  mt m 1 (u ( z ))    mt m 1u( z ) ,
dt x  t y  t
where x   tx, y   ty.

Now, make t  1 and use continuity of u, to get

u( x   h )  u( x ) u( x  h )  u( x )
u x  u x and u y   u y (  )
h h
as x   x and y   y


 16



Method 3 (Milne‐Thompson Method: A completely informal
method). Let u ( x, y ) be a given Harmonic function.

z z
 In the given expression of u( x, y ) , put x  , y  (!) and
2 2i
z z _____
consider g ( z )  2 u ( , )  f (0) .
2 2i

 The imaginary part of g ( z ) is the desired harmonic
conjugate of u( x, y ) .

Example. u ( x, y )  x 2  y 2

Using Milne‐Thompson method,
z z
f ( z )  2u( , )  u (0,0)
2 2i
z z
 2[( )2  ( )2 ]  z 2
2 2i

Thus the desired harmonic conjugate is

v ( x, y )  Im f ( z )  2 xy .

 17
Informal Justification of Milne‐Thompson Method:

Let v ( x, y ) be a Harmonic conjugate of the given Harmonic
function u ( x, y ) and g  u  i v be the corresponding analytic
function,
 1  
Denote,  (  i ) . Then,
z 2 x y

 _____   
g ( z )  (u  iv )  (  i )(u  iv )
z z x y
1
 [u x  iv x  i (u y  iv y )]
2
1
 [u x  iv x  iu y  v y ]
2
1
 [u x  v y  i ( v x  u y )]
2
0 (1)

Informally assuming that z, z are independent variables (!),
_____
deduce from (1) that f ( z ) is independent of z, i.e. it is a
_____
function of z alone, i.e. g ( z )  g * ( z ) ( say ) .

1 _____
 u( x, y )  [ g ( z )  g ( z )]
2
1
 [ g ( z )  g * ( z )] (2)
2
 18

In (2), z = x + i y, where, x and y are real. Let us informally
assume that (2) holds as well with x and y complex (!) and put

z z
x  , y  .
2 2i

Then, (2) gives

z z 1 z z
u( , )  [ g ( z )  g (0)] *
( since z  x  iy   i  0)
2 2i 2 2 2i
1 _____
 [ g ( z )  g (0)] (3)
2
z z _____
Equation (3)  g ( z )  2u( , )  g (0)
2 2i
z z _____
 2u ( , )  u(0,0)  u (0,0)  g (0)
2 2i
_____
Since u(0,0)  f (0) is a purely imaginary constant, it can be
dropped from the above expression (since harmonic
conjugates are unique only up to an imaginary constant).

z z _____
 g ( z )  2 u( , )  f (0)
2 2i

is an analytic function whose real part is u( x, y ) .

The imaginary part of g ( z ) is therefore the desired harmonic
conjugate of u( x, y ) .