Ch0 Total

Chapter 0
Theory of complex variables
Complex Numbers
A complex number z is an ordered pair ( x, y ) of real
numbers x and y, written
z = ( x, y ) .
x and y are respectively the real part and the imaginary

part of z, written
x = Re z ,
y = Im z.
By definition, two complex numbers are equal if and only
if their real parts are equal and their imaginary parts are
equal.
In particular, (0,1) is called the imaginary unit and is
denoted by i ,
i = (0,1).
The following abbreviations are frequently used:
( x,0) = x,
PHY 4602 Methods in Theoretical Physics II
(0, y ) = iy.
0- 1
Chapter 0
Addition of two complex numbers

z1 + z2 = ( x1 , y1 ) + ( x2 , y2 ) = ( x1 + x2 , y1 + y2 ).
Multiplication of two complex numbers

z1 z2 = ( x1 , y1 )( x2 , y2 ) = ( x1 x2 y1 y2 , x1 y2 + x2 y1 ).
In practice, complex numbers z = ( x, y ) are written as:

z = x + iy.
With this notation and the identity
i 2 (0,1)(0,1) = ( 1, 0) = 1,
the algebra of complex numbers is much simplified:
( x1 + iy1 ) + ( x2 + iy2 ) = ( x1 + x2 ) + i ( y1 + y2 ).
( x1 + iy1 )( x2 + iy2 ) = x1 x2 + ix1 y2 + iy1 x2 + i 2 y1 y2

= ( x1 x2 y1 y2 ) + i ( x1 y2 + x2 y1 ).
z1 z2 = ( x1 x2 ) + i ( y1 y2 ).
x1 + iy1 ( x1 + iy1 )( x2 iy2 )

z=
=
x2 + iy2 ( x2 + iy2 )( x2 iy2 )
=
x1 x2 + y1 y2
x2 y1 x1 y2
.
i
+
2
2
2
2
x2 + y2
x2 + y2
0- 2
Chapter 0
Complex Plane
Complex numbers can be represented by points in a
Cartesian coordinate system, whose x-axis and y-axis are
respectively called the real and imaginary axes.
The xy-plane in which the complex numbers are

represented is called the complex plane.
The addition and subtraction of complex numbers acquire

clear geometric picture in the complex plane notation.
0- 3
Chapter 0
Addition
of
complex
numbers
Subtraction of complex
numbers
Complex conjugate numbers

The complex conjugate z of a complex number z = x + iy
is defined by:
z = x iy.
Useful relations:
1
Re z = x = ( z + z ),
2
Im z = y =
1
( z z ).
2i
0- 4
Chapter 0
In the complex plane, z and z form a pair of mirror

images about the real axis.
Example:
Polar Form of Complex Numbers

A complex number can also be expressed in terms of its
polar coordinates, r and .
x = r cos ,
y = r sin .
z = r (cos + i sin ).
0- 5
Chapter 0
r is called the absolute value or modulus of z and is

denoted by z .
z = r = x 2 + y 2 = zz .
is called the argument of z and is denoted by arg z .

y
= arg z = arctan
x
Geometrically, is the directed angle measured from the
positive x-axis to the radius OP in the complex plane.
Note:
All angles are measured in radians and positive in the
counterclockwise sense.
For z = 0 this angle is undefined.
For given z 0 it is determined only up to integer
multiples of 2 .
The value of that lies in the interval < is

called the principal value of the argument of z ( 0)
and is denoted by Arg z , with capital A. That is,
< Arg z .
0- 6
Chapter 0
Triangle inequality
For any complex numbers we have the important triangle
inequality,
z1 + z2 z1 + z2 .
Proof:
The triangle inequality is a direct consequence of an
important theorem in geometry:
The
sum
of
the
lengths of any two

sides in a triangle is
larger than (or equal
to) the length of the
other side.
The generalized triangle inequality:

z1 + z2 + + zn z1 + z2 + + zn .
Proof:
z1 + z2 + + zn z1 + z2 + + zn
z1 + z2 + + zn
0- 7
Chapter 0
Multiplication and Division in Polar Form

Useful formulae:
z1 z2 = r1r2 [cos(1 + 2 ) + i sin(1 + 2 )].
z1 r1
= [cos(1 2 ) + i sin(1 2 )].

z2 r2
z1 z2 = z1 z2 .
z1
z1
=
.
z2
z2
arg( z1 z2 ) = arg z1 + arg z2 .

arg(
z1
) = arg z1 arg z2 .
z2
De Moivres formula
z n = r n (cos n + i sin n ).
(cos + i sin ) n = cos n + i sin n .
0- 8
Chapter 0
Roots
If z = wn (n = 1, 2, ) , each of these possible values of w
is called an n-th root of z, and written as
w = n z.
Obviously, there are n possible distinct values of w ,

namely:
n
+ 2 k
+ 2 k
z = n r cos
+ i sin
n
n
Hence the symbol
k = 0,1, n 1.
z is n-valued.
Example:
n
2k
2k
1 = cos
,
+ i sin
n
n
k = 0,1, n 1.
Some terminologies
Circle:
A circle of radius and centre a in the complex plane
is given by the solution of the equation:
za =
0- 9
Chapter 0
Open circular disk

An open circular disk is given by the interior enclosed
by a circle and satisfies the equation:
za < .
Closed circular disk
A closed circular disk is the interior of a circle, i.e.
z a < , plus the circle itself, given by the equation:
za .
Neighbourhood
An open circular disk z a < is called the

neighbourhood of a.
0- 10
Chapter 0
An open annulus (circular ring):

A closed annulus:
1 < z a < 2 .
1 z a 2
Half-planes.
(Open) upper half-plane:

The set of all points z = x + iy such that y > 0 .
Lower half-plane:
The set of all points z = x + iy such that y < 0 .
Right half-plane:
The set of all points z = x + iy such that x > 0 .
Left half-plane:
The set of all points z = x + iy such that x < 0 .
0- 11
Chapter 0
Open sets:
A set S is called open if every point of S has a
neighbourhood consisting entirely of points that
belong to S .
Domains:
An open set S is called connected if any two of its
points can be joined by a broken line of finitely many
straight-line segments all of whose points belong to
S . An open connected set is called a domain.
Complement:
The complement of a set S in the complex plane is
the set of all points of the complex plane that do not
belong to S .
Boundary points:
A boundary point of a set S is a point every
neighbourhood of which contains both points that
belong to S and points that do not belong to S .
0- 12
Chapter 0
Complex Functions
Let S be a set of complex numbers. A complex function f
defined on S is a rule assigning to every z = x + iy in S a
complex number w, called the value of the function f at z
and denoted as:
w = f ( z ).
z varies in S, a set of complex numbers, and is hence called
a complex variable. The set S is called the domain of
definition of f. The set consisting of all values of function
f is called the range of f.
A complex function f(z) is equivalent to a pair of real

function u ( x, y ) and ( x, y ) , each depending on the two
real variables x and y.
w = f ( z ) = u ( x, y ) + i ( x, y ).
Examples:
Consider w = f ( z ) = z 2 + 3 z .
f ( z ) = z 2 + 3 z = ( x + iy ) + 3 ( x + iy )
2
= x 2 + 2ixy y 2 + 3 x + 3iy
0- 13
Chapter 0
u = Re f ( z ) = x 2 y 2 + 3 x , and
= Im f ( z ) = 2 xy + 3 y .
Consider w = f ( z ) = 2iz + 6 z .
f ( z ) = 2i ( x + iy ) + 6( x iy )
u ( x, y ) = 6 x 2 y , and
( x, y ) = 2 x 6 y .
Limit
Consider a function f ( z ) defined in a neighbourhood of
z0 (expect perhaps at z0 itself). If the value of f is close to
l for all z close to z0 , f ( z ) is said to have the limit l as z
approaches z0 , written
lim f ( z ) = l ,
z z0
Note:
z may approach z0 from any direction in the complex
plane and the way
in which z approaches should not
affect the value of the limit. If a limit exists, it is then

unique.
0- 14
Chapter 0
z-plane
w-plane
In other words, if 0 as 0 , then

lim f ( z ) = l .
z z0
Continuity
A function f ( z ) is said to be continuous at z = z0 if f ( z0 )
is defined and
lim f ( z ) = f ( z0 ).
z z0
f ( z ) is said to be continuous in a domain if it is

continuous at every point of this domain.
0- 15
Chapter 0
Derivative
Definition:
The derivative of a complex function f at a point z0 is
written f ( z0 ) and is defined by
f ( z0 ) = lim
z z0
f ( z ) f ( z0 )
f ( z0 + z ) f ( z0 )
= lim
z 0
z
z z0
provided this limit exists. Then f is said to be differentiable

at z0 .
Note:
Differentiability of f at z0 means that the quotient in this
formula always approaches a common value no matter
along which path z approaches z0 .
Example:
The function f ( z ) = z 2 is differentiable for all z and has
the derivative f ( z ) = 2 z because
( z + z ) 2 z 2
f ( z ) = lim
z 0
z
z 2 + 2 z z + ( z ) 2 z 2
= lim
= lim (2 z + z ) = 2 z.
z 0
z 0
z
0- 16
Chapter 0
Differentiation rules:
(cf ) = cf ,
( f + g ) = f + g ,
( fg ) = f g + fg ,
f f g fg
.
=
2
g
g
They are the same as in real calculus.
Differentiability of a complex function is a rather severe

requirement.
Example:
Despite that z is obviously a differentiable function, its
complex conjugate, z , is not differentiable.
Proof:
f ( z + z ) f ( z ) ( z + z ) z z x iy
=
=
=
.
z
z
z x + iy
If y = 0 , the limit is +1.
If x = 0 , it is 1.
The limit is path-dependent and z is not differentiable.
0- 17
Chapter 0
Analytic Functions
Definition of Analytic Functions
A function f ( z ) is said to be analytic in a domain D (or,
equivalently, holomorphic in D) if f ( z ) is defined and
differentiable at all points of D. The function f ( z ) is said
to be analytic at a point z = z0 in D if f ( z ) is analytic in a
neighbourhood of z0 .
Analytic functions are functions that are differentiable in
some domain.
Examples:
The nonnegative integer powers 1, z , z 2 , and also
the polynomials,
f ( z ) = c0 + c1 z + c2 z 2 + + cn z n
where c0 , , cn are complex constants are analytic

functions.
The quotient of two polynomial g ( z ) and h( z ) ,
f ( z) =
g ( z)
h( z )
( a rational function),
is analytic expect at the point where h( z ) = 0 .

0- 18
Chapter 0
Cauchy-Riemann Equations
Let f ( z ) = u ( x, y ) + iv( x, y ) be defined and continuous in
some
neighbourhood
of
point
z = x + iy
and
differentiable at z itself. Then at that point, the first-order

partial derivatives of u and v exist and satisfy the CauchyRiemann equations:
u x = v y ,
u y = vx
Hence if f ( z ) is analytic in a domain D, those partial

derivatives exist and satisfy Cauchy-Riemann equations at
all points of D.
Proof:
By definition,
f ( z + z ) f ( z )
z 0
z
u ( x + x, y + y ) u ( x, y )
= lim
z 0
x + iy
v( x + x, y + y ) v( x, y )
.
+i lim
z 0
x + iy
f ( z ) = lim
0- 19
Chapter 0
The limit z 0 can proceed along different paths.

z = x (i.e. along a horizontal path).
u ( x + x, y ) u ( x, y )
x 0
x
v( x + x, y ) v( x, y )
.
+i lim
x 0
x
f ( z ) = lim
f ( z ) = u x + ivx .
z = iy (i.e. along a vertical path).
u ( x, y + y ) u ( x, y )
y 0
i y
v( x, y + y ) v( x, y )
.
+i lim
y 0
iy
f ( z ) = lim
f ( z ) = iu y + v y .
The two results should be the same, resulting in the
Cauchy-Riemann equations:
u x = v y ,
u y = vx
0- 20
Chapter 0
Examples:
Consider f ( z ) = z 2 = ( x + iy ) 2 = u + iv
u = x 2 y 2
v = 2 xy
Since f ( z ) = z 2 is analytic for all z, its CauchyRiemann equations are given by:
ux = 2 x = vy ;
and vx = 2 y = u y
Consider f ( z ) = z = x iy .
We have:
u = x, v = y .
One of the Cauchy-Riemann equations is not satisfied.
u x = 1 v y = 1.
Conclusion: f ( z ) = z is not analytic.
Condition for Analyticity

If two real-valued continuous functions u ( x, y ) and v( x, y )
of two real variables x and y have continuous first partial
derivatives satisfying the Cauchy-Riemann equations in
some
domain
D,
then
the
complex
function
f ( z ) = u ( x, y ) + iv( x, y ) is analytic in D.
0- 21
Chapter 0
Question:
Is f ( z ) = z 3 an analytic function?
Solution:
u = x 3 3 xy 2 and v = 3 x 2 y y 3 .
ux = 3x 2 3 y 2 ,
v y = 3x 2 3 y 2
u y = 6 xy,
vx = 6 xy.
The Cauchy-Riemann equations:

u x = v y ,
u y = vx
are satisfied and hence f ( z ) = z 3 is analytic for every z.
Classwork:
Express z in terms of the polar coordinates:
z = r (cos + i sin )
and consider an analytic function:
f ( z ) = u (r , ) + iv(r , ) .
Show that the Cauchy-Riemann equations in the polar
coordinates are:
1
u
=
r r v ,
v = 1 u .
r
r
0- 22
Chapter 0
Laplaces Equation
If f ( z ) = u ( x, y ) + iv( x, y ) is analytic in a domain D, then u
and v satisfy Laplaces equation
2 u = u xx + u yy = 0,
2 v = vxx + v yy = 0,
respectively, in D and have continuous second partial
derivatives.
Proof:
Differentiate the Cauchy-Riemann equations once:

=
u
x x x v y ,
u xx = v yx
u yy = vxy .
u y = ( vx )
y
y
One then obtains:
2 u = u xx + u yy = 0 .
The other Laplaces equation,
2 v = vxx + v yy = 0,
0- 23
Chapter 0
can be obtained from similar argument:

u
=
y x y v y ,
u xy = v yy
u yx = vxx .
u = ( v )
x
x y x
Harmonic Functions
Solutions of Laplaces equation having continuous second
partial derivatives are called harmonic functions and
relevant theory is called potential theory.
Note:
Real and imaginary parts of an analytic function are
harmonic functions.
If a pair of harmonic functions u and v satisfy the
Cauchy-Riemann equations:
u x = v y and u y = vx ,
in a domain D, they are the real and imaginary parts
of an analytic functions f in D. v is then said to be a
conjugate harmonic function of u in D.
0- 24
Chapter 0
Example:
Show that u = x 2 y 2 y is a harmonic function and find
a conjugate harmonic function v of u.
Solution:
It is straightforward to show that 2u = 0 and hence
u ( x, y ) is indeed a harmonic function.
To find a conjugate harmonic function of u, we write
down the Cauchy-Riemann equations:
v y = u x = 2 x,
vx = u y = 2 y + 1.
Integrate the first Cauchy-Riemann equation with respect

to y :
v = 2 xy + h( x),
where h( x) is a function of x and is independent of y .
Substitute this expression into the second CauchyRiemann equation.
vx = 2 y +
dh
dh
=1
= 2 y +1
dx
dx
h( x ) = x + c .
v = 2 xy + x + c
f ( z ) = u + iv = x 2 y 2 y + i (2 xy + x + c) = z 2 + iz + ic.
0- 25
Chapter 0
Note:
A conjugate of a given harmonic function is uniquely
determined up to an arbitrary real additive constant.
The curves u = constant (or, similarly, v = constant )

are called equipotential lines (or level curves).
Conformal Mapping
Mapping
A given function f
f ( z ) = u ( x, y ) + iv( x, y )
assigns to each point z in its domain of definition D the
corresponding point
w = f ( z)
in the w-plane. We say that f defines a mapping of D into
the w-plane.
For each point z in D , the corresponding point w = f ( z )
is called the image of z with respect to the function f.
0- 26
Chapter 0
Example: Consider the mapping w = u + iv = z 2 = ( x + iy ) 2

u = Re( w) = Re( z 2 ) = x 2 y 2
v = Im( w) = Im( z 2 ) = 2 xy
Vertical lines x = c = constant are mapped into
u = c 2 y 2 , v = 2cy . Eliminating y from these two
equations, we obtain:
v 2 = 4c 2 ( c 2 u ) ,
which represent parabolas opening to the left.
Similarly,
horizontal
lines
y = k = constant
are
mapped onto parabolas opening to the right:

v 2 = 4k 2 ( k 2 + u )
Mapping of
w = z 2 = u + iv
Red dashed lines:
x = c = constant
Blue solid lines:
y = k = constant
0- 27
Chapter 0
Conformal Mapping
A conformal mapping is a mapping which preserves
angles between any two oriented curves (i.e. the angle
between the oriented tangents at the intersection point)
both in magnitude and in sense.
the mapping w = f ( z )
z -plane
Theorem of conformality
The mapping defined by an analytic function f ( z ) is
conformal, except at critical points where the derivative
f ( z ) is zero.
Example:
The
mapping
w = z n , n = 1, 2,3,
conformal, except at z = 0 .
0- 28
is
Chapter 0
Proof:
the mapping w = f ( z )
z -plane
Let C1 : z = z1 (t ) and C2 : z = z2 (t ) be two oriented curves

in the z-plane. The images of these curves by the mapping
w = f ( z ) are respectively C1* and C2* . Besides,
z0 = z1 (t0 ) = z2 (t0 ) .
The orientation of the directed tangent of C1 at z0 :
1 = lim arg [ z1 (t ) z0 ]
t t0
The orientation of the directed tangent of C2 at z0 :
2 = lim arg [ z2 (t ) z0 ]
t t0
is the angle between the two curves.

0- 29
Chapter 0
= 1 2 = lim {arg [ z1 (t ) z0 ] arg [ z2 (t ) z0 ]}

t t0
z1 (t ) z0
= lim arg
t t0
z
t
z
(
)
2
0
( z1 (t ) z0 ) /(t t0 )
dz1 / dt
= lim arg
= arg
t t0
(
z
(
t
)
z
)
/(
t
t
)
dz
/
d
t
0
0
2
The orientation of C1* at w0 = f ( z0 ) :
1* = lim arg [ f ( z1 (t )) f ( z0 ) ]
t t0
The orientation of C2* at w0 = f ( z0 ) :
2* = lim arg [ f ( z2 (t )) f ( z0 ) ]
t t0
* = 1* *
2
= lim {arg [ f ( z1 (t )) f ( z0 ) ] arg [ f ( z2 (t )) f ( z0 ) ]}

t t0
f ( z1 (t )) f ( z0 )
= lim arg
t t0
f
(
z
(
t
))
f
(
z
)

2
0
( f ( z1 (t )) f ( z0 )) /(t t0 )
= lim arg
t t0
(
f
(
z
(
t
))
f
(
z
))
/(
t
t
)

2
0
0
df ( z1 ) / dt
= arg
d
f
(
z
)
/
dt
2
0- 30
Chapter 0
This expression can be simplified by the chain rule of

differentiation:
df ( z (t )) dw dw dz
.
=
=
dt
dt
dz dt
df ( z1 ) / dt
df
z
dt
(
)
/
2
* = arg
df
dz
= arg
df
dz
dz1
dt
dz2
dt
If f ( z ) 0 , then we have:
dz1
dt
*
= arg
dz
2
dt
It then proves the conformality of the mapping f ( z ) as

long as f ( z ) 0 .
0- 31
Chapter 0
Exponential Function
The complex exponential function is defined as:
e z = e x (cos y + i sin y ).
Properties:
e z is an entire function, that is, analytic for all z.
(e z ) = e z .
e z1 + z2 = e z1 e z2
Eulers formula:
eiy = cos y + i sin y.
ez 0
Periodicity:
e z + 2 i = e z
Parity:
e z + i = e z
The mapping w = e z is conformal everywhere in the

complex plane.
0- 32
Chapter 0
Trigonometric Functions
For real variable we have:
eix = cos x + i sin x,
e ix = cos x i sin x.
Hence,
1
cos x = (eix + e ix ),
2
sin x =
1 ix ix
(e e ).
2i
These two relations are generalized to define cosine and

sine functions for complex variables.
1
cos z = (eiz + e iz ),
2
1 iz iz
sin z = (e e ).
2i
Other trigonometric functions are defined by cosine and
sine functions in the usual manner.
sin z
,
cos z
cos z
cot z =
,
sin z
tan z =
1
,
cos z
1
csc z =
.
sin z
sec z =
0- 33
Chapter 0
Hyperbolic Functions
The hyperbolic cosine and sine functions for complex
variables are defined by:
1 z
(e + e z ),
2
1
sinh z = (e z e z ).
2
cosh z =
Similarly,
sinh z
,
cosh z
1
sech z =
,
cosh z
tanh z =
cosh z
,
sinh z
1
csc hz =
.
sinh z
coth z =
Differentiation:
(cosh z ) = sinh z ,
(sinh z ) = cosh z ,
Trigonometric and hyperbolic functions:

cosh iz = cos z ,
sinh iz = i sin z
cos iz = cosh z ,
sin iz = i sinh z.
0- 34
Chapter 0
Logarithm Function
The inverse function of the exponential function, e w = z , is
defined as the natural logarithm ln z , i.e.
w = ln z
Properties:
ln(e z ) = z
a = eln a
a z = e z ln a
ln( z1 z2 ) = ln z1 + ln z2
ln( z1 / z2 ) = ln z1 ln z2
(ln z ) =
1
z
ln( rei ) = ln r + i
The
complex
natural
logarithm ln z ( z 0) is
infinitely many-valued. Since

e w = z = e w 2 n i , n = 1, 2,
ln z = w i 2n
0- 35
Chapter 0
Principal value
Since
ln z = ln(| z | ei ) = ln | z | +i ,
where = arg z , one can define the principal value of ln z
by:
Ln z = ln z + iArg z .
Therefore,
ln z = Ln z 2n i .
Examples:
ln1 = 0, 2 i, 4 i,
Ln 1 = 0
ln i = i / 2, 3 i / 2,5 i / 2,
Ln i = i / 2
ln 4i = 1.386294 + i / 2 2n i
Ln 4i = 1.386294 + i / 2
ln(4i ) = 1.386294 i / 2 2n i
Ln (4i ) = 1.386294 i / 2
0- 36
Chapter0
Analyticity of the Logarithmic Function

We study here the continuity and analyticity of the singlevalued function Ln z , defined by:
Ln z = Ln r + i ,
r > 0,
< .
This function is discontinuous

(i)
at z = 0 since it is undefined there; and
(ii)
along the negative real axis where the argument

does not possess a unique limit at any point along
this axis.
According to the definition of Ln z , any point on the

negative real axis has an argument = . However,
points approaching the negative axis in the third quadrant
have an argument tending towards . The principal
0- 37
Chapter 0
argument of z goes through a discontinuity of 2 as a

point crosses the negative real axis.
However, Ln z is single-valued and continuous in the

domain D that consists of the z-plane with the points on
the negative real axis and origin removed.
Domain D
We note that continuity is a prerequisite for analyticity. In

the following we will show that Ln z is also analytic there.
0- 38
Chapter 0
Proof:
We write Ln z = u ( x, y ) + iv( x, y ) with
u = Ln r ,
v =,
and
and apply the Cauchy-Riemann equations in polar form:
u 1 v
=
, and
r r
1 u
v
=
,
r
r
to show the analyticity of Ln z in domain D . As
u 1
= ,
r r
u
and v
1 v 1
= ,
r r
obviously
v
1 u
= 0,
= 0,
r
r
satisfy
the
Cauchy-Riemann
equations. Moreover, the four partial derivatives u / r ,

v / , v / r and u / are continuous in domain D ,
where r 0 . Therefore, Ln z is analytical there.
Note:
ln z is a multi-valued function in the entire z -plane.
Ln z is a single-valued function in the entire z -plane.
Ln z is an analytic single-valued function in domain
D.
0- 39
Chapter 0
Branch, branch cut and branch point

Branch
A branch of a multi-valued function is a single-valued
function analytic in some domain. At every point of the
domain the single-valued function must assume exactly
one of the various possible values that the multi-valued
function can assume.
Example:
Ln z is called the principal branch of the logarithmic
function. This function is defined for all z except z = 0

and values of z on the negative real axis. The domain D is
formed by removing the semi-infinite line y = 0 , x 0
from the z -plane. This line is called a branch cut.
Branch cut
A line used to create a domain of analyticity is called a
branch line or branch cut.
0- 40
Chapter 0
Examples:
The principal value of the logarithmic function, Ln z ,
defined by:
Ln z = Ln r + i ,
r > 0,
< .
Domain D
The semi-infinite
line y = 0 , x 0
is the branch cut.
Another branch of ln z :
f ( z ) = Ln r + i ,
where 3 / 2 < / 2 .
is analytic in domain D1 .
The
infinite
semiline
x = 0 , y 0 is
the branch cut.
0- 41
Chapter 0
Branch point
When we are trying to establish branches of a multi-valued
function, we find that all the possible branch cuts share a
common point. We call that point a branch point of the
multi-valued function.
Examples:
In the case of ln z the origin z = 0 is a branch point.
Both of the two branch cuts that we investigated

previously for this function passed through z = 0 .
Find (i) the largest domain of analyticity; (ii) branch
cut; and (iii) branch point of f ( z ) = Ln z ( 3 + 4i ) .
branch point z = 3 + 4i
0- 42
Chapter 0
Power functions
In general, we can define z c , where c can be any real or
complex number, by:
z c = exp ( c ln z ) .
Since the logarithm of z is multi-valued, depending on the
value of c, z c may have more than one numerical value. If
we employ the principal value of the logarithm in the
above definition, we obtain the principal value of z c .
Examples:
Square-root function z1/ 2
1/ 2 )[ Ln z + i (2 k )]
z1/ 2 = e(
1/ 2 ) ln z
= e(
= e(
1/ 2 ) Ln z
= e(
1/ 2 ) Ln z + ik
[cos(k ) + i sin(k )]
= ( 1) e(
k
1/ 2 ) Ln z
where k = 0, 1, 2,..... The cut is along the negative

real axis with z = 0 being the branch point.
z c has an infinite set of possible values except if c is a
rational number and is a multi-valued function of z .
0- 43
Chapter 0
Consider the function ( z z0 ) , where z0 and c are

c
complex constants. If c is a non-integer, these

functions are multi-valued.
( z z0 )
=e
c ln ( z z0 )
Since ln ( z z0 ) has a branch point at z0 , so does
( z z0 ) .
c
Consider the function ( z 1) . The branch point is

1/ 3
z = 1 and we let a branch cut be constructed along the
line y = 0 , x 1.
r1 = z 1 ,
1 = arg ( z 1) ,
0 1 < 2 .
( z 1)
1/ 3
1/ 3
= r1ei ( + 2 k )
= 3 r1 e ( 1
i / 3 + 2 k / 3 )
where k = 0,1, 2 ,...

0- 44
Chapter0
Complex Line Integrals

Definite integrals in the complex plane are called
(complex) line integrals:
f ( z )dz.
Here the integrand f ( z ) , a complex function, is integrated

over a given curve C in the complex plane, called the path
of integration.
Note: A curve C in the complex plane is usually

represented by a parametric equation:
z (t ) = x(t ) + iy (t ) ,
where t is a real parameter, and x(t ) and y (t ) are two real-
valued functions of t.
Definition of line integral:

Let C: z (t ) = x(t ) + iy (t ) be a smooth curve in the complex
plane, f ( z ) be a continuous function given (at least) at
each point of C.
0- 45
Chapter 0
The curve C
We partition the interval a t b by points

t0 = a, t1 , , tn 1 , tn = b
where t0 < t1 < tn . The values of z (t ) at these points are:

z0 = z (t0 ), z1 = z (t1 ), , zn 1 = zn 1 (tn 1 ), zn (tn ) = Z .
On each portion of subdivision of C we choose an

arbitrary point, say, a point 1 between z0 and z1 , a point
2 between z1 and z2 , and so on, and evaluate the sum:

n
S n = f ( m )zm
m =1
where
zm = zm zm1.
0- 46
Chapter 0
If the greatest value of tm = tm tm1
( m = 1, 2, , n )
approaches zero as n , the limit of the sequence of

complex numbers S 2 , S3 , obtained is called the line
integral of f ( z ) over the oriented curve C, denoted by:
f ( z )dz
, or by
f ( z ) dz
if C is a closed path.
Theorem of Existence of line integral

A complex line integral exists if
f ( z ) is continuous; and
C consists of finitely many smooth curves joined end
to end, that is, piecewise smooth.
Unless otherwise stated, we assume in the following

discussion that f ( z ) is continuous and all paths of
integration are piecewise smooth. Consequently, the
existence of complex line integrals is readily guaranteed.
0- 47
Chapter 0
Properties of line integrals:

Linearity
[k f ( z) + k
C
1 1
f ( z ) ] dz = k1 f1 ( z )dz + k2 f 2 ( z )dz.
2 2
Directionality
z0
z0
f ( z )dz = f ( z )dz.
Z
Additivity
f ( z )dz = f ( z )dz + f ( z )dz.

C1
C2
Evaluation of line integrals

Let C be a piecewise smooth path, represented by z = z (t ) ,
where a t b . Let f ( z ) be a continuous function on C.
Then
f ( z )dz = f [ z (t )]z (t )dt

a
0- 48
Chapter 0
Examples:
1. Integral of 1/ z around the unit circle C in the counterclockwise direction.
z (t ) = cos t + i sin t = eit
( 0 t 2 ) .
2
2
dz
it
it
=
e
(
ie
dt
)
=
i
C z 0
0 dt = 2 i.
2. Consider
dz
C z = 2 i.
( z z0 ) m dz .
Here m is an integer, z0 is a constant, and C is a circle of

radius and centered at z0 . The integration is taken
along the counter-clockwise direction.
z (t ) = z0 + (cos t + i sin t ) = z0 + eit
( z z0 ) m = m eimt ,
( z z0 ) dz = i
m +1
= i
m +1
( 0 t 2 ) .
dz = i eit dt
ei ( m +1) t dt
[ cos(m + 1)t + i sin(m + 1)t ] dt.

0
2 i, m = 1
C ( z z0 ) dz = 0, m 1
m
0- 49
Chapter 0
Path-dependence of line integral:

A complex line integral depends not only on the endpoints
of the path but in general also on the path itself.
Example:
Integral of a non-analytic function.
Integrate f ( z ) = Re z = x from 0 to 1 + 2i along (a) C * ,

and (b) C consisting of C1 and C2 .
C * : z (t ) = t + 2it ,
z (t ) = 1 + 2i,
f ( z (t )) = x(t ) = t
C1 : z (t ) = t ,
(0 t 1)
(0 t 1)
z (t ) = 1,
f ( z (t )) = x(t ) = t
C2 : z (t ) = 1 + it ,
(0 t 2)
z (t ) = i,
f ( z (t )) = x(t ) = 1 .
(a)
(b)
Re zdz = t (1 + 2i )dt =
*
0
1
1
(1 + 2i ) = + i.
2
2
1
xdz = xdz + xdz = tdt + 1 idt =

C1
C2
1
+ 2i.
2
0- 50
Chapter 0
The evaluation of a complex line integral resembles that of

a real line integral in the 2-dimensional space. However,
the integral of an analytic function in a simply connected
domain (to be defined later) can be evaluated by obtaining
an indefinite integral in a way analogous to integration
method in calculus of real variables.
Upper bound of integrals
The absolute value of a complex line integral is bounded
by the inequality:
f ( z )dz ML ,
where L is the length of the path C and M is a positive

constant such that f ( z ) M everywhere on C.
Simply connected domains

A simply closed path is a closed path that does not
intersect or touch itself.
0- 51
Chapter 0
A domain D is called simply connected if every simple

closed curve (closed curve without self-intersections) in D
encloses only points of D.
Examples:
An open annulus:
An open circle:
za <
is
connected domain.
simply 1 < z a < 2 is not a

simply connected domain.
Simply and multiply connected domains:
0- 52
Chapter 0
Cauchys integral theorem

Theorem
If f ( z ) is analytic in a simply connected domain D, then
for every simple closed path C in D,
f ( z ) dz = 0.
A simply closed path is

called a contour and an
integral over such a path is
called a contour integral.
Note:
f ( z ) is of course analytic in the domain enclosed by C.
Corollary:
If f ( z ) is analytic in the domain enclosed by a contour C,
f ( z ) dz = 0.
0- 53
Chapter 0
Examples:
Integrals of entire functions that are analytic for all z .
cos z dz = 0,
e z dz = 0,
z n dz = 0.
sec z dz = 0
, C is the unit circle.
sec z = 1/ cos z is not analytic at z = / 2, 3 / 2, .

However, all these points lie outside C.
dz
=0
2
z +4
, C is the unit circle.
The integrand ( z 2 + 4) 1 is singular at z = 2i , which

lie outside the contour C
Non-analytic functions:
(i)
(ii)
z dz = e it ieit dt = 2 i , C is the unit circle.

0
z 1 dz = 2 i
C lies in the annulus 0 < z < R where 1/ z is analytic,

but this domain is not simply connected. Thus, the
Cauchys integral theorem cannot be applied.
0- 54
Chapter 0
Proof of Cauchys integral theorem:

dz = dx + idy ;
f ( z )dz =
f ( z ) = u ( x, y ) + iv( x, y ) ,
and
(udx vdy ) + i (udy + vdx).

C
The complex line integral is expressed in terms of two real

line integrals, namely:
Ia =
(udx vdy ),
Ib =
(udy + vdx).
We will use the Greens theorem for real line integrals to

transform these two line integrals into double integrals.
Greens theorem
Let R be a closed bounded region in the xy-plane whose
boundary C is piecewise smooth. If F1 ( x, y ) and F2 ( x, y )
are two differentiable functions in the domain R, then:
F2 F1
+
=
(
F
dx
F
dy
)
2
C 1
R x y dxdy
v u
=
udx
v
dy
(
)
C
R x y dxdy
u v
C (vdx + udy) = R x y dxdy
0- 55
Chapter 0
From Cauchy-Riemann equations:
v u
=
x y
u v
=
x y
v u
=
udx
v
dy
(
)
C
R x y dxdy = 0
u v
C (vdx + udy) = R x y dxdy = 0
f ( z )dz =
(udx vdy ) + i (vdx + udy ) = 0

C
Path independence of contour integrals

We call an integral of f ( z ) independent of path in a
domain D if for every z1 , z2 in D its value depends only on
the initial point z1 and the terminal point z2 but not on the
choice of the path C in D.
Theorem
If f ( z ) is analytic in a simply connected domain D, then
the integral of f ( z ) is independent of path in D.
0- 56
Chapter 0
Proof:
Consider two directed paths C1 and C2 in D from z1 to z2 .
Reverse the direction of C2 to obtain the path C2* .
C2
fdz = * fdz.
Apply
Cauchys
integral
C2
theorem to this contour.
C1
fdz + * fdz = 0,
C2
Therefore,
C1
C1
fdz = * fdz.
C2
f ( z )dz = f ( z ) dz
C2
This shows that the integral of f ( z ) is independent of path

in an analytic domain D. Hence,
C1
f ( z )dz = f ( z )dz =
C2
z2
z1
f ( z ) dz
0- 57
Chapter 0
Deformation of path
If
path
deforms
C1
continuously to path C2
without crossing any point
at which
f ( z ) is non-
analytic, then
C1
f ( z )dz = f ( z )dz
C2
Line integrals of analytic functions

Theorem
Existence of an indefinite integral:
Let f ( z ) be analytic in a simply connected domain D.

Then there exists an indefinite integral of f ( z ) in the
domain D, that is, an analytic function F ( z ) such
that:
F ( z ) = f ( z ) for all z in D .
For all paths in D joining two points z0 and z1 in D,
we have:
z1
z0
f ( z )dz = F ( z1 ) F ( z0 ) .
0- 58
Chapter 0
Examples:
1+ i
1 3
z dz = z
3
2
1+ i
0
1
2 2
3
= (1 + i ) = + i
3
3 3
cos zdz = sin z
i
i
= 2sin( i ) = 23.097i
Here D is the complex plane excluding the point
z = 0 and the negative real axis, obviously a simply

connected domain. Ln z is then analytic in
dz
i
i
i
i
=
Ln
Ln
(
)
=
) = i .
i z
2
2
i
Proof:
Consider the line integral from z0 to z :
z
F ( z ) = f ( z * )dz *
z0
If the upper limit moves

from z to z + z , then
F ( z + z ) =
z +z
z0
f ( z * )dz *
z
F ( z + z ) F ( z ) 1 z + z
*
*
*
*
(
)
(
)
f
z
dz
f
z
dz
=
z0
z
z z0
1 z +z
*
*
(
)
.
=
f
z
dz
z
z
0- 59
D.
Chapter 0
f ( z )z ]
[
F ( z + z ) F ( z )
= lim
= f ( z ).
F ( z ) = lim
z 0
0
z
z
Therefore, F ( z ) is an indefinite integral of f ( z ) , written
F ( z ) = f ( z )dz.
There are many possible indefinite integrals for f ( z ) ,
corresponding to different choices of the starting point:
f ( z * )dz * = Fj ( z ) .
zj
The difference between any two indefinite integrals of a

function is a constant because:
z
Fj ( z ) Fk ( z ) = f ( z )dz f ( z * ) dz *
*
zj
zk
zk
zj
= f ( z * )dz * +
f ( z * )dz * =
zk
zj
f ( z * )dz *
Hence, if by definition
z
z1
f ( z )dz = G ( z )
and F ( z ) = G ( z ) + C is another indefinite integral of f ( z ) .

G ( z1 ) = 0
F ( z1 ) = +C
G ( z ) = F ( z ) F ( z1 )
z2
z1
f ( z )dz = F ( z2 ) F ( z1 )
0- 60
Chapter 0
Cauchys integral formula

Let f ( z ) be analytic in a simply connected domain D.
Then for any point z0 in D and any simple directed closed
path C in D that encloses z0 in counter-clockwise
direction, we have the Cauchys integral formula:
f ( z)
dz = 2 if ( z0 ) , or
z z0
1
f ( z0 ) =
2 i
f ( z)
dz
z z0
Proof:
f ( z)
dz
dz = f ( z0 )
+
C
z z0
z z0
f ( z0 )
f ( z ) f ( z0 )
dz.
z z0
dz
= f ( z0 ) 2 i .
z z0
f ( z ) f ( z0 )
dz = 0
z z0
0- 61
Chapter 0
Arguments:
(i)
Since f ( z ) is analytic at z = z0 ,
lim
z z0
f ( z ) f ( z0 )
= f ( z0 ) .
z z0
Hence
f ( z ) f ( z0 )
is bounded if z z0
z z0
is
sufficiently small. That is,

f ( z ) f ( z0 )
< M if z z0 < .
z z0
(ii)
We deform the contour C to a small circle

centered at z0 with a radius < .
(iii)
From Cauchys integral theorem we have:
f ( z ) f ( z0 )
dz =
z z0
z z0 =
f ( z ) f ( z0 )
dz .
z z0
0- 62
Chapter 0
(iv)
z z0 =
f ( z ) f ( z0 )
dz < M (2 ) < M (2 )
z z0
Taking the limit 0 , we show that
(v)
z z0 =
f ( z ) f ( z0 )
dz = 0 =
z z0
f ( z ) f ( z0 )
dz
z z0
The Cauchys integral formula is then established:
f ( z)
= 2 if ( z0 )
z z0
Examples:
ez
dz = 2 ie z
z2
z3 6
dz =
2z i
z =2
= 2 ie 2 46.4268i.
z3 3
dz
1
z 2i
1
2
C
= 2 i 12 z 3 3
z =i / 2
6 i.
Integrate g ( z ) = ( z 2 1) 1 tan z around the circle
C : z = 3/ 2 in counter-clockwise direction.
Use method of partial fractions to expand ( z 2 1) 1 .
0- 63
Chapter 0
1
1 1
1
=
.
2
z 1 2 z 1 z +1
tan z
1 tan z
tan z
dz
dz
dz
=
C
C
2 z 1
z +1
z 1
2 i
=
tan1 tan( 1)] = 2 i tan1 9.785i.
[
2
Cauchys formula for multiply connected domains

Consider the domain D
bounded by two simple
closed curves C1 and C2 . If
f ( z ) is analytic on C1 and
C2 and in D as well, for
any z0 in D the Cauchys

formula reads:
f ( z0 ) =
1
2 i
C1
f ( z)
1
dz +
2 i
z z0
C2
f ( z)
dz ,
z z0
where the outer integral (over C1 ) is taken along counterclockwise direction and the inner integral (over C2 ) is
taken along clockwise direction.
0- 64
Chapter 0
Proof:
By adding two equal but
opposite curves, C3 and C4 ,
we
can
join
the
two
contours C1 and C2 to form

a single contour C .
Since f ( z ) is analytic in the simply connected domain

enclosed by C , for any z0 in D the Cauchys formula
reads:
f ( z0 ) =
1
2 i
1
2 i
1
=
2 i
f ( z)
dz
z z0
C1 + C2 + C3 + C4
C1
f ( z)
dz
z z0
f ( z)
1
dz +
z z0
2 i
C2
f ( z)
dz.
z z0
This proves the Cauchys formula for doubly connected

domains. One can make use similar trick to obtain the
Cauchys formula for multiply connected domains.
0- 65
Chapter 0
Derivatives and contour integrals

Theorem:
If f ( z ) is analytic in a domain D, then it has derivatives of
all orders in D, which are then also analytic functions in D.
The values of these derivatives at a point z0 in D are given
by the formulas:
f ( z0 ) =
1
2 i
f ( z)
C ( z z0 )2 dz,
f ( z0 ) =
2!
2 i
f ( z)
C ( z z0 )3 dz.
f ( n ) ( z0 ) =
n!
2 i
f ( z)
C ( z z0 )n +1 dz.
Here C is any simple closed path in D that encloses z0 and

whose full interior belongs to D ; and we integrate
counter-clockwise around C.
These formulas are obtained formally by differentiating
the Cauchy formula, under the integral sign with respect to
z0 .
0- 66
Chapter 0
Proof:
f ( z0 ) = lim
z 0
f ( z0 + z ) f ( z0 )
.
z
Apply Cauchys integral formula to find f ( z0 + z ) and

f ( z0 ).
f ( z0 + z ) f ( z0 )
z
f ( z)
1
dz
=
2 iz C z ( z0 + z )
=
1
2 i
f ( z)
dz
z z0
f ( z)
C ( z z0 z )( z z0 ) dz
Taking the limit z 0 , we obtain the first Cauchys

formula:
f ( z0 ) =
1
2 i
f ( z)
C ( z z0 )2 dz
Other Cauchys formulas can be obtained in a similar way

using mathematical induction.
0- 67
Chapter 0
Examples:
Let C be a contour enclosing the point z = i
z 4 3z 2 + 6
2 i 4
2
dz
=
(
z
3
z
+ 6)
3
2!
( z + i)
= i 12 z 2 6
Find
z = i
z = i
= 18 i.
cos z
C ( z 1)3 ( z 5)2 dz , where C is the circle
z 4 = 2.
z 1 is nonzero both inside and on C.
z 5 becomes zero at z = 5 inside C.

cos z
C ( z 1)3 ( z 5)2 dz
1
=
C ( z 5) 2
= 2 i
cos z
( z 1)3 dz
64sin 5 48cos 5
d cos z
=
2
i
dz ( z 1)3 z = 5
46
0- 68
Chapter 0
Related mathematical theorems

Moreras theorem
If f ( z ) is continuous in a simply connected domain D and
if
f ( z )dz = 0
for every closed path in D, then f ( z ) is analytic in D. In

other words, this is just the converse of Cauchys integral
theorem.
Liouvilles theorem
If an entire function f ( z ) is bounded in absolute value for
all z, then f ( z ) must be a constant.
Gausss mean value theorem
Let f ( z ) be analytic in a simply connected domain.
Consider any circle lying in this domain. The value
assumed by f ( z ) at the center equals the average of the
values assumed by f ( z ) on its circumference. That is,
f ( z0 ) =
1
2
f (z0 + rei )d ,
where r is radius and z0 is the center of the circle.

0- 69
Chapter0
Complex Sequences
A complex sequence z1 , z2 , is a sequence of complex
numbers.
Convergent sequence
A complex convergent sequence z1 , z2 , is a sequence of
complex numbers that has a finite limit c, written
lim zn = c
n
In other words, as n , zn c.
Rigorous definition:
For every > 0 , no matter how small it is, we can always
find a sufficiently large N > 0 such that
zn c <
for all n > N .
A divergent sequence is one that does not converge.

_____________________________________________________________________
0-70
Chapter0
Examples:
The
sequence
{i / n} = {i, 1/ 2, i / 3,1/ 4, }
n
is
convergent with limit 0.

The sequence {i n } = {i, 1, i,1, i, } is divergent.
The sequence { zn } with
zn = xn + iyn = 1 1/ n 2 +i (2 + 4 / n)
converges with the limit c = 1 + 2i .
Theorem of convergence of complex sequences

A complex sequence z1 , z2 , , zn , with
zn = xn + iyn ,
where n = 1, 2, ,
converges to c = a + ib if and only if
the sequence of the real parts x1 , x2 , converges to a ;

and
the sequence of the imaginary parts y1 , y2 ,
converges to b .
_____________________________________________________________________
0-71
Chapter0
Complex Series
Given a complex sequence z1 , z2 , , zm , , we can form
another sequence of sums,
s1 = z1 ,
s2 = z1 + z2 ,
s3 = z1 + z2 + z3 ,
sn = z1 + z2 + + zn .
sn is called the n-th partial sum of the infinite series and
z1 , z2 , are called the terms of the series.
Convergent series
A (complex) convergent series is a series whose sequence
of partial sums converges,
lim sn = s.
n
s is called the sum (or value) of the series.
A series that is not convergent is called a divergent series.
_____________________________________________________________________
0-72
Chapter0
Theorem
A complex series with zm = xm + iym converges and has the
sum s = u + iv if and only if
x1 + x2 + converges and has the sum u ; and
y1 + y2 + converges and has the sum v.
Necessary condition for convergence

If a series z1 + z2 + converges, then lim zm = 0 . Hence if
m
lim zm 0 ,
the series diverges.
Absolute convergence
A series z1 + z2 + is called absolutely convergent if the
series of the absolute values of the terms
z
m =1
= z1 + z2 +
is convergent.
Theorem:
If a series is absolutely convergent, it is convergent.
_____________________________________________________________________
0-73
Chapter0
Conditional convergence
If the series z1 + z2 + converges while the series
z1 + z2 + diverges, then the series z1 + z2 + is called
conditionally convergent.
Example:
1 1 1
The series 1 + + converges, and is a
2 3 4
conditionally
convergent
z1 + z2 + = 1 +
series
since
the
series
1 1 1
+ + + diverges.
2 3 4
Test of convergence of infinite series

Comparison test
If a (complex) series z1 + z2 + is given and we can find a
converging series b1 + b2 + with non-negative real terms
such that
zn bn
for n = 1, 2, ,
then the given series converges absolutely.

_____________________________________________________________________
0-74
Chapter0
In this connection, the geometric series
= 1 + q + q 2 +
m =0
is often used for purpose of comparison. Since
1 q n +1
1
q n +1
sn = q =
=
.
1 q
1 q 1 q
m=0
n
the series converges (absolutely) with the sum 1/(1 q ) if
q < 1 and diverges if q 1.

From the comparison test and also the convergence of the
geometric series one can establish the ratio test and the
root test, to be detailed as follows.
Ratio test 1
If a series z1 + z2 + with zn 0 (n = 1, 2, ) has the
property that for every n > N ,
zn +1 / zn q < 1, where q is a fixed number,

this series converges absolutely.
If for every n > N ,
zn +1 / zn 1, the series diverges.
_____________________________________________________________________
0-75
Chapter0
Ratio test 2
If a series z1 + z2 + with zn 0 (n = 1, 2, ) is such that
lim
n
zn+1
= L , then:
zn
a) The series converges absolutely if L < 1.

b) The series diverges if L > 1.
c) If L = 1, no conclusion can be drawn from this test.
Example:
Consider the series
(100 + 75i ) n
1
=
1
+
(100
+
75
i
)
+
(100 + 75i ) 2 +
n!
2!
n =0
n +1
100 + 75i /(n + 1)! 100 + 75i 125

zn +1
=
=
=
.
n
zn
n +1
n +1
100 + 75i / n !
The limit L exists and L = lim
zn +1
=0 .
zn
By the ratio test 2 the series converges absolutely.
_____________________________________________________________________
0-76
Chapter0
Let
z n = i / 23 n / 2
, n = 0, 2, 4,
zn = 1/ 2(3n 1) / 2
, n = 1,3,5,
1 i 1
i
1
z0 + z1 + z2 + z3 + = i + + + + +
+
2 8 16 64 128
The ratios of the absolute values of successive terms

are 1/ 2,1/ 4,1/ 2,1/ 4, . Since the sequence of these
ratios has no limit, the ratio test 2 is inapplicable.
However, from the ratio test 1, the series converges
absolutely.
Root test 1
If a (complex) series z1 + z2 + is such that for every n
greater than some N ,
n
zn q < 1
(where q < 1 is fixed), this series converges absolutely. If ,

instead, for every n greater than some N ,
n
zn 1
the series diverges.

_____________________________________________________________________
0-77
Chapter0
Root test 2
If a series z1 + z2 + is such that lim n zn = L , then:
n
a) The series converges absolutely if L < 1.

b) The series diverges if L > 1.
c) If L = 1, the test fails.
Power series
A power series in powers of z z0 is a series of the form
a (z z )
n=0
= a0 + a1 ( z z0 ) + a2 ( z z0 ) 2 +
where z is a complex variable, a0 , a1 , are complex (or

real) constants, called the coefficients of the series, and z0
is a complex (or real) constant, called the center of the
series.
If z0 = 0 , we obtain as a particular case a power series in
powers of z :
n
2
a
z
=
a
+
a
z
+
a
z
+
0
0
1
2
n =0
_____________________________________________________________________
0-78
Chapter0
Convergence of power series

The convergence of a power series can be determined
from the tests mentioned previously. Since each term in a
power series is a power function of z , the convergence of
a power series depends on the value of z .
Examples:
Consider the geometric series
= 1 + z + z 2 +
n =0
The ratio between two successive terms is
z n +1
= z.
n
z
It converges absolutely if z < 1 and diverges if z 1.

The power series
zn
z 2 z3
= 1 + z + + +
2! 3!
n =0 n !
is absolutely convergent for every z . By the ratio test,
z
z n+1 /(n + 1)!
=
0
n
z / n!
n +1
as
_____________________________________________________________________
0-79
Chapter0
Theorem
Consider a power series of the form:
n
2
a
(
z
z
)
=
a
+
a
(
z
z
)
+
a
(
z
z
)
+
0
0
0
1
0
2
0
n=0
a) It converges at the center z0 .

b) If it converges at a point z = z1 z0 , it converges
absolutely for every z closer to z0 than z1 , that is,
z z0 < z1 z0 .
c) If it diverges at a z = z2 , it diverges for every z farther
away from z0 than z2 .
_____________________________________________________________________
0-80
Chapter0
Radius of Convergence of a Power Series

Given a power series:
n
2
a
(
z
z
)
=
a
+
a
(
z
z
)
+
a
(
z
z
)
+
0
0
0
1
0
2
0
n=0
one can always find the smallest open disk,
z z0 < R
that is centered at center z0 and with radius R, and, in
addition, includes all the points at which the series
converges. The circle
z z0 = R
is called the circle of convergence and its radius R the

radius of convergence.
Note:
The series converges absolutely everywhere within the
circle of convergence, i.e. z z0 < R .

The series diverges for all z for which z z0 > R , i.e.
in the exterior of the circle.

No general statements can be made about the
convergence on the circle of convergence itself.

_____________________________________________________________________
0-81
Chapter0
R if the series converges for all z .
R = 0 if the series converges only at the center z = z0 .
Determination of the radius of convergence

Method 1 (Cauchy-Hadamard formula)
Consider a power series:
a (z z )
n=0
= a0 + a1 ( z z0 ) + a2 ( z z0 ) 2 +
Suppose that the sequence of the absolute value of the

ratio of two successive coefficients an +1 / an , (n = 1, 2, ) ,
converges with the limit L* , i.e. L* = lim an +1 / an .
n
a) If L* = 0 , then R = ; that is, the power series

converges absolutely for all z .
b) If L* 0 , then
R=
an
1
=
lim
.
*
n
L
an+1
(Cauchy-Hadamard formula).
c) If an +1 / an , then R = 0 (i.e. the series converges

only at the center z0 ).
_____________________________________________________________________
0-82
Chapter0
Method 2
Suppose that the sequence of
an , ( n = 1, 2, ) ,
converges with the limit L , i.e. L = lim n an .

n
a) If L = 0 , then R = ; that is, the power series

converges absolutely for all z .
b) If L 0 , then
R=
1
1
.
=
n
L lim an
n
c) If lim n an , then R = 0 , (i.e. the series

n
converges only at the center z0 ).
Example: Use the Cauchy-Hadamard formula to find the
radius of convergence for the series
R = lim
(2n)!
n
(
3
)
z
i
2
n = 0 ( n !)
an
(2n)! (2n + 2)!
= lim
an +1 n (n !) 2 ((n + 1)!) 2
(n + 1) 2
1
= lim
= .
n (2 n + 2)(2n + 1)
4
_____________________________________________________________________
0-83
Chapter0
Power Series Function

If a given power series has a radius of convergence R > 0 ,
its sum is a function of z and we write
f ( z ) = an ( z z0 ) n = a0 + a1 ( z z0 ) + a2 ( z z0 ) 2 +
n=0
for z z0 < R . We say that f ( z ) is represented by (or

developed in) the power series.
Example:
For z < 1, the power series 1 + z + z 2 +
converges and is
equal to (1 z ) 1 .
Theorem
Continuity of the sum of a power series
The function f ( z ) defined by
f ( z ) = an ( z z0 ) n = a0 + a1 ( z z0 ) + a2 ( z z0 ) 2 +
n=0
with the radius of convergence R > 0 is continuous at
z = z0 .
_____________________________________________________________________
0-84
Chapter0
Theorem
Uniqueness of representation
Suppose that the two power series
an ( z z0 )
n
b
(
z
z
)
n
0
and
n=0
n=0
converge and have the same sum for z z0 < R , where
R > 0 . Then these series are identical, that is,

an = bn
for all n = 0,1,
Corollary:
If a function f ( z ) has a power series representation with
any center z0 , this representation is unique.
Example:
The function (1 z ) 1 is represented by the power series
1+ z + z2 +
= (1 z ) 1
for z < 1 . However, it can also represented by power

series with other centers. For example,
(1 z ) 1 = [1 ( z ) ]
1
=
1
z z 2
1 +
+
+
1 1
The radius of convergence is now R = 1 .

_____________________________________________________________________
0-85
Chapter0
Termwise operations on power series

Consider the power series:
k
a
(
z
z
)
= a0 + a1 ( z z0 ) + = f ( z )
k
0
k =0
b (z z )
k =0
= b0 + b1 ( z z0 ) + = g ( z )
The radii of convergence of these two series are R1 and R2 ,

respectively.
Termwise addition (or subtraction):
k
(
a
b
)(
z
z
)
k k
0
k =0
Termwise addition (or subtraction) of the two power series

yields a power series with radius of convergence at least
equal to the smaller of R1 and R2 , and in addition:
k
(
a
b
)(
z
z
)
= f ( z) g ( z)
k k
0
k =0
In general, for any complex constants and , we have:
( a
k =0
bk )( z z0 ) k = f ( z ) g ( z )
_____________________________________________________________________
0-86
Chapter0
Termwise multiplication:
Termwise multiplication of the two power series means
the multiplication of each term of the first series by each
term of the second series and the collection of like powers
of z . That is,
s ( z ) = ak ( z z0 ) k bm ( z z0 ) m
k =0 m=0
= (a0 bn + a1bn 1 + + an b0 )( z z0 ) n .
n=0
This power series converges absolutely for each z within

the circle of convergence of each of the two given series
and has the sum
k
s ( z ) = ak ( z z0 ) bk ( z z0 ) k
k =0
k =0
.
= f ( z) g ( z)
_____________________________________________________________________
0-87
Chapter0
Termwise differentiation:
Termwise differentiation of a power series,
d
n
n 1
a
z
z
n
a
z
z
(
)
(
)
n
n
0
0
dz
n =1
n =1
= a1 + 2a2 ( z z0 ) + 3a3 ( z z0 ) 2 +
gives the derived series of the original one. It is
permissible in the sense that
d
d
n
n
(
)
(
)
a
z
z
a
z
z
n
0
0
n
dz n =1
n =1 dz
= nan ( z z0 ) n 1
n =1
In addition, the derived series of a power series has the

same radius of convergence as the original series, which
follows directly from an application of the ratio test:
lim
n an
(n + 1) an +1
a
n
lim n
n ( n + 1) n a
n +1
= lim
= lim
an
an +1
_____________________________________________________________________
0-88
Chapter0
Termwise integration:
Termwise integration of the series
a0 + a1 ( z z0 ) + a2 ( z z0 ) 2 + yields a new series:
an
a1
n +1
2
z
z
F
z
a
z
z
z
z
(
(
)
=
(
)
+
(
)
+ ,
0
0
0
0
2
n=0 n + 1
which has the same radius of convergence as the original
series, and F ( z ) = f ( z ).
Theorems Analyticity of power series function

A power series,
a (z z )
n=0
= f ( z ) , with a radius of
convergence R > 0 , represents an analytic function

everywhere inside its circle of convergence.
The derivatives of this function can be obtained by
differentiating the original series term by term, e.g.
df
= nan ( z z0 ) n 1 , etc.
dz n =1
The radius of convergence of the derived series are

also equal to R. Hence, by the first statement, each of
them represents an analytic function for z z0 < R .
_____________________________________________________________________
0-89
Chapter0
Taylor series of analytic functions

Theorem Taylor theorem
Let f ( z ) be analytic in a domain D, and let z = z0 be any
point in D. Then there exists a unique Taylor series with
center z0 that represents f ( z ) , namely
f ( z ) = an ( z z0 ) n
n =0
where an =
1 (n)
f ( z0 )
n!
This representation is valid in the largest open disk with

center z0 in which f ( z ) is analytic.
Note: From theory of contour integration, we know that:
(n)
n!
( z0 ) =
2 i
f ( z* )
*
dz
C ( z* z0 )n +1 .
The counterclockwise
integration is taken along
a simple closed curve C
that encloses z0 and f ( z )
is analytic in the interior
of C.
_____________________________________________________________________
0-90
Chapter0
Therefore, the coefficients of the Taylor expansion can be

expressed in the integral form:
1
an =
2 i
f ( z* )
*
dz
C ( z* z0 )n +1
The remainder of the expansion after the term an ( z z0 ) n ,

which is defined by:
z z0
( z z0 ) 2
f ( z ) = f ( z0 ) +
f ( z0 ) +
f ( z0 ) +
1!
2!
,
n
( z z0 ) ( n )
+
f ( z0 ) + Rn ( z )
n!
can also be expressed in the integral form:
( z z0 ) n +1
Rn ( z ) =
2 i
f ( z* )
*
dz
C ( z* z0 )n +1 ( z* z )
Note:
A Taylor series is also called a Maclaurin series if the
center z0 = 0 .
_____________________________________________________________________
0-91
Chapter0
Proof:
From theory of contour
integration, we know that:
1
f ( z) =
2 i
f ( z* ) *
dz .
z* z
Note:
z z0
< 1 for all z in the
*
z z0
interior of the circle C.
C is a circle centered at z0 with

a radius r .
Making use of the identity: 1 + r + r +

2
1 r n +1
+r =
,
1 r
n
1
:
we rewrite the quotient *
z z
1
z* z
=
1
=
*
z z0 ( z z0 )
1
( z * z0 )(1
z z0
)
*
z z0
n
n +1
z z0
z z0
1
1 z z0
1 + *
= *
+ + *
+ *
*
z z 0 z z0
z
z
z
z
z
z
0
0
_____________________________________________________________________
0-92
Chapter0
Hence,
1
f ( z) =
2 i
+
f ( z * ) * z z0
dz +
2 i
z * z0
( z z0 )
2 i
f ( z* )
*
dz
C ( z* z0 )2 +
f (z )
*
dz
C ( z* z0 )n +1 + Rn ( z )
where
( z z0 ) n +1
Rn ( z ) =
2 i
f ( z* )
*
dz
n
+
*
1
*
C ( z z0 ) ( z z ) .
This is the integral form of the Taylor series expansion,

which readily reduces to the differential form:
z z0
( z z0 ) 2
f ( z ) = f ( z0 ) +
f ( z0 ) +
f ( z0 ) +
1!
2!
( z z0 ) n ( n )
+
f ( z0 ) + Rn ( z )
n!
by evaluating the integrals with the theory of contour
integral. Furthermore, it can be shown that
lim Rn ( z ) = 0.
n
Therefore, the Taylor series converges with the limit f ( z ) :

( z z0 ) n ( n )
f ( z ) = f ( z0 ) +
f ( z0 )
n!
n =1
_____________________________________________________________________
0-93
Chapter0
Remarks:
The limit lim Rn ( z ) = 0 can be obtained as follows:
n
( z z0 ) n +1
Rn ( z ) =
2 i
f ( z* )
*
dz
C ( z* z0 )n +1 ( z* z )
( z z0 ) n +1
2 z z0
*
n +1
,
ML
f ( z* )
where M is the maximum value of *
on the
z z
circle C, and L = length of contour = 2 r .

Note:
z z0
< 1 for all z in
*
z z0
the interior of the circle
C.
Since
z z0
z z0
*
n +1
n +1
0 as n , and M is bounded for
z in the interior of circle, we then have lim Rn ( z ) = 0 .

n
_____________________________________________________________________
0-94
Chapter0
Similarly, the coefficients of the expansion satisfy the

inequality
1
an =
2 i
f ( z* )
M
*
dz
C ( z* z0 )n +1
rn
where M is the maximum of f ( z ) on the circle C.
The foregoing discussion is valid as long as f ( z ) is

analytic inside the circle C. Therefore, the Taylor
expansion holds in the largest open disk with center z0
in which f ( z ) is analytic.
Common Taylor Series

Geometric series
1
= z n = 1 + z + z 2 +
1 z n=0
( z < 1).
Note: (1 z ) 1 is non-analytic at z = 1 , which defines the

radius of convergence of this series.
_____________________________________________________________________
0-95
Chapter0
Exponential function
zn
z2
e = = 1 + z + + .
2!
n =0 n!
z
Trigonometric and hyperbolic functions
z 2n
z2 z4
cos z = (1)
= 1 + +
(2n)!
2! 4!
n =0
n
z (2 n +1)
z3 z5
sin z = (1)
= z + +
(2n + 1)!
3! 5!
n =0
z 2n
z2 z4
cosh z =
= 1 + +
2! 4!
n = 0 (2n )!
z (2 n +1)
z3 z5
sinh z =
= z + +
3! 5!
n = 0 (2n + 1)!
Logarithm
z 2 z3
Ln (1 + z ) = z + +
2 3
( z < 1).
Note: Ln (1 + z ) is non-analytic at z = 1 , which defines

the circle of convergence of the expansion.
_____________________________________________________________________
0-96
Chapter0
Expansion of non-analytic functions

It is also possible to expand a function f ( z ) around a
point where f ( z ) is non-analytic (i.e. singular) in terms of
a series consisting of both non-negative and negative
integer power of z z0 , which is called a Laurent series.
Theorem
Laurents theorem
If f ( z ) is analytic on two concentric circles C1 and C2

with center z0 and in the annulus between them, then f ( z )
can be represented by the Laurent series:
f ( z)
bn
n
n =1 ( z z0 )
= an ( z z0 ) +
n
n=0
= a0 + a1 ( z z0 ) + a2 ( z z0 ) 2 +
+
b1
b2
+
+
2
z z0 ( z z0 )
which consists of non-negative powers and the principal

part (the negative powers).
0- 97
Chapter 0
The coefficients of this Laurent series are given by the

integrals:
1
f ( z* )
*
an =
dz
,
n +1
*
v
C
2 i ( z z0 )
n = 0,1, 2,3,"
1
*
n 1
*
*
bn =
z
z
f
z
dz
(
)
(
)
,
0
C
2 i v
n = 1, 2,3,"
The integrals are taken counter-clockwise around any

simple closed path C that lies in the annulus and encircles
the inner circle.
This series converges and represents f ( z ) in the open
annulus where f ( z ) is analytic.
Note:
The domain of validity of this expansion can be
enlarged by increasing the outer circle C1 and

decreasing C2 until each of two circles reaches a point
where f ( z ) is singular.
0- 98
Chapter 0
If z0 is the only singular point of f ( z ) inside C2 , the
inner circle can be shrunk to the point z0 , giving

convergence of the Laurent series in a disk except at
the center z = z0 .
The Laurent series can also be written as:
f ( z) =
a (z z )
n =
1
f ( z* )
*
an =
dz
C ( z* z0 )n +1
2 i v
with
(n = 0, 1, 2, ).
Proof:
(a). Apply the Cauchys integral formula for multiply
connected domains to f ( z ) .
If f ( z ) is analytic on C1 and
C2 and in domain D bounded
by two simple closed curves

C1 and C2 , for any z in D the
Cauchys
integral
formula
reads:
0- 99
Chapter 0
f ( z) =
1
f ( z) *
1
f ( z) *
dz
+
dz ,
*
*
v
C
C
2 i 1 z z
2 i 2 z z
where the outer integral (over C1 ) is taken along counterclockwise direction and the inner integral (over C2 ) is
taken along clockwise direction.
In this case,
f ( z ) = g ( z ) + h( z )
1
f ( z* ) *
=
C1 z* z dz
2 i v
1
f ( z* ) *
dz
*
v
C
2
2 i
z z
Note: Both the integrals along C1 and C2 are taken in the
counter-clockwise direction.
(b). We can show that the integral along C1 yields:
1
f ( z* ) *
n
g ( z) =
dz
=
a
z
z
(
)
n
0
C1 ( z* z )
2 i v
n=0
where
1
f ( z* )
*
an =
dz
.
n
+
*
1
v
C
1
2 i
( z z0 )
0- 100
Chapter 0
The proof parallels that of the Taylors series.

1
f ( z* ) *
g ( z) =
dz .
*
v
C
2 i
z z
We note that
z z0
< 1 for all z in
*
z z0
the annulus.
Here C is the outer-circle C1 .
1
z* z
=
1
=
*
z z0 ( z z0 )
1
( z * z0 )(1
z z0
)
*
z z0
n
n +1
z z0
z z0
1
1 z z0
1 + *
= *
+ + *
*
+ *
z z 0 z z0
z
z
z
z
z
z
0
0
Therefore,
0- 101
Chapter 0
f ( z * ) * z z0
1
g ( z) =
dz +
*
v
C
2 i
2 i
z z0
+
( z z0 )
2 i
f ( z* )
*
dz
v C ( z* z0 )2 +
f (z )
*
dz
v C ( z* z0 )n +1 + Rn ( z )
where
( z z0 ) n +1
Rn ( z ) =
2 i
f ( z* )
*
dz
v C ( z* z0 )n +1 ( z* z ) .
Furthermore, it can be shown that

lim Rn ( z ) = 0.
n
Therefore, the function g ( z ) can be written as:
f ( z* ) *
1
n
g ( z) =
dz
=
a
z
z
with
(
)
n
0
*
v
C
1
2 i
( z z)
n=0
f ( z* )
1
*
an =
dz
C1 ( z* z0 )n +1 .
2 i v
(c). For the integral along the inner circle we have
z * z0
1
< 1. Therefore, we rewrite *
as:
z z0
z z
1
1
=
=
*
*
z z z z0 ( z z 0 )
1
z z0
( z z0 )(1
)
z z0
*
0- 102
Chapter 0
and expand it as follows:

2
n
*
*
*
z z0
1 z z 0 z z 0
1
=
+
1 +
+

*
z z0 z z0
z
z
z z z z0
0
1 z * z0
z z * z z0
n +1
f ( z* ) *
1
h( z ) =
dz
*
v
C
2 i 2 z z
1 1
*
*
=
f
z
dz
(
)
C2
2 i z z0 v
+
1
( z z0 ) 2
C2
( z * z0 ) f ( z * )dz *
1
+ +
( z z0 ) n +1
Since
+
f
z
dz
(
z
z
)
(
)
R
( z)
n
0
v C2
z * z0
< 1 , we can also show that the
z z0
remainder
1
Rn* ( z ) =
2 i ( z z0 ) n +1
( z * z0 ) n +1
*
*
(
)
f
z
dz
v C2 z z*
approaches zero for large n, i.e. lim Rn* ( z ) = 0.

n
0- 103
Chapter 0
This then completes the proof of Laurent series.

Uniqueness of Laurent series
The Laurent series of a given analytic function f ( z ) in its
annulus of convergence is unique. However, f ( z ) may
have different Laurent series in two different annuli with
the same center.
Examples:
Find the Laurent series of z 5 sin z with center z = 0 .
1 (1) n 2 n +1
z sin z = 5
z
z n = 0 (2n + 1)!
1
1
1
1 2
= 4 2+
z +
6 z 120 5040
z
5
The series converges in the whole complex plane

with the origin removed.
Find the Laurent series of z 2 e1/ z with center z = 0 .
1
1
z 2 e1/ z = z 2 1 +
+
+
2
1! z 2! z
1 1
1
= z2 + z + +
+
+
2
2 3! z 4! z
( z > 0)
0- 104
Chapter 0
The function (1 z ) 1 is singular at z = 1 . Find its

Laurent series with center z = 0 if (a) z < 1; and (b)
z > 1.
1
(a)
= zn
1 z n =0
(b)
if z < 1.
If z > 1, we have
1
1
=
1 z z (1 z 1 )
1 1
1 1
= n = 2
z n=0 z
z z
The function ( z 3 z 4 ) 1 is singular at z = 0,1. Find its

Laurent series with center z = 0 if (a) 1 > z > 0 ; and
(b) z > 1.
(a) If 0 < z < 1, we have:
0- 105
Chapter 0
1
1
1 n
= 3
= 3 z
3
4
z z
z (1 z ) z n = 0
1 1 1
= 3 + 2 + + 1 + z
z
z
z
(b) If z > 1
1
1
=
z 3 z 4 z 4 (1 z 1 )
1 1
1 1
= 4 n = 4 5
z n=0 z
z
z
2 z + 3
The function f ( z ) = 2
is singular at z = 1, 2 .
z 3z + 2
Find its Laurent series with center z = 0 if (a) 1 > z ;
(b) 2 > z > 1; and (c) z > 2 .
Note:
f ( z) =
1
1
.
z 1 z 2
1
= zn
1 z n =0
if z < 1.
1
1 1
= 2
1 z
z z
if z > 1.
0- 106
Chapter 0
1
1
1 n
=
=
z if z < 2 .
z 2 2(1 1 z ) n = 0 2n +1
2
1
1
2n
=
= n +1 if z > 2 .
2
z2
n=0 z
z (1 )
z
Region I: z < 1
Region II: 1 < z < 2
Region III: z > 2
(a) For z < 1,
1
3 5
9
f ( z ) = 1 + n+1 z n = + z + z 2 + .
2
2 4
8
n =0
(b) For 1 < z < 2 ,
0- 107
Chapter 0
1 n 1
f ( z ) = n +1 z n+1
n =0 2
n =0 z
1 1
1
1 1
= + z + z 2 + 2 .
2 4
8
z z
(c) For z > 2 ,
f ( z ) = (2n + 1)
n =0
1
z n+1
2 3 5 9
= 2 3 4 .
z z
z
z
Singular functions
Singularity
A function f ( z ) is singular or has a singularity at a point
z = z0 if f ( z ) is not analytic (or not even defined) at
z = z0 , but every neighbourhood of z = z0 contains points
at which f ( z ) is analytic.
Isolated singularity
z = z0 is called an isolated singularity of f ( z ) if there
exists a neighbourhood of z = z0 that does not contain

other singularity of f ( z ) besides z = z0 .
Examples:
0- 108
Chapter 0
tan z has isolated singularities at z =
(2n + 1)
,
2
n = 0,1, 2,".
2
1
, n = 0,1, 2,". The
tan( ) is singular at z =
(2n + 1)
z
singularities
densely
populate
around
z=0 .
Therefore, it has a non-isolated singularity at z = 0 .

Classification of isolated singularities
Isolated singularities of f ( z ) at z = z0 can be classified by
the Laurent series
bn
,
n
n =1 ( z z0 )
f ( z ) = an ( z z0 ) +
n
n=0
valid in the immediate neighbourhood of the singular point

z = z0 except at z0 itself, according to the number of terms
in the principal part of the series.
(a) If it has only finitely many terms and is of the form:

bm
b1
+ +
, m 1,
m
z z0
( z z0 )
0- 109
Chapter 0
the singularity of f ( z ) at z = z0 is called a pole, and m is

called its order. Poles of the first order are also known as
simple poles.
(b) If the principal part has infinitely many terms, we say
that f ( z ) has at z = z0 an isolated essential singularity.
Example:
1/ z
1
1
1
1
=
+
+
+
n
2
z 2! z
n=0 n! z
has an isolated essential singularity at z = 0 .

Removable singularities
We say that a function f ( z ) has a removable singularity at
z = z0 if f ( z ) is not analytic at z = z0 , but can be made
analytic there by assigning a suitable value f ( z0 ) .
Example: f ( z ) = (sin z ) / z becomes analytic at z = 0 if we
define f (0) = 1.
Entire function
An entire function is one that is analytic everywhere in the
(finite) complex plane.
Meromorphic function
0- 110
Chapter 0
An analytic function whose only singularities in the finite

plane are poles is called a meromorphic function.
Zeros of analytic functions

A zero of an analytic function f ( z ) in a domain D is a
z = z0 in D such that f ( z0 ) = 0 . A zero has order n if not

only f but also the derivatives f , f , , f ( n 1) are all 0 at
z = z0 but f ( n ) ( z0 ) 0 . A first-order zero is also called a

simple zero.
Examples:
The function 1 + z 2 = ( z + i )( z i ) has simple zeros at

i .
The function (1 z 4 ) 2 = ( z + 1) 2 ( z 1) 2 ( z + i ) 2 ( z i ) 2
has second-order zeros at 1 and i .
The function ( z a )3 has a third-order zero at z = a .

Taylor series at a zero
If z = z0 is a n-th order zero of f ( z ) , f ( z ) and the
derivatives
f , f , , f ( n 1) are all 0 at z = z0 but
0- 111
Chapter 0
f ( n ) ( z0 ) 0 . Therefore, the leading n terms of the Taylor

series of f ( z ) about z = z0 vanish.
f ( z ) = an ( z z0 ) n + an +1 ( z z0 ) n +1 +
= ( z z0 ) n an + an +1 ( z z0 ) + an + 2 ( z z0 ) 2 +
where an 0 .
Theorem
Let f ( z ) be analytic at z = z0 and has a zero of n-th order
at z = z0 . Then 1/ f ( z ) has a pole of n-th order at z = z0 .
Residue Integration Method

Consider an integral
f ( z )dz taken around a simple
closed path C in the counter-clockwise direction.

If f ( z ) is analytic everywhere on C and inside C, such
an integral is zero by Cauchys integral theorem.

If f ( z ) has a singularity at a point z = z0 inside C, but
is otherwise analytic on C and inside C, then f ( z ) has

a Laurent series
0- 112
Chapter 0
f ( z ) = an ( z z0 ) n +
n =0
b1
b2
+
+
2
z z0 ( z z0 )
The coefficient of the term ( z z0 ) is given by:

1
b1 =
f ( z )dz ,
v
C
2 i
Since we can obtain Laurent series (and hence b1 ) by
other methods, we can in fact use the above formula
to evaluate the integral, namely:
f ( z )dz = 2 ib1 .
The coefficient b1 is called the residue of f ( z ) at

z = z0 and is denoted it by
b1 = Res f ( z ).
z = z0
Examples:
sin z
Integrate f ( z ) = 4 counter-clockwise around the
z
contour C : z = 1.
Expand the function about z = 0 :
0- 113
Chapter 0
sin z 1
1
z z3
+ +
f ( z) = 4 = 3
z
z 3! z 5! 7!
b1 = 1/ 3!.
sin z
i
dz
2
ib
.
=
1
v C z 4
3
Integrate
f ( z ) = 1 ( z 3 z 4 ) clockwise around the
circle C: z = 1/ 2 .
f ( z ) is singular at z = 0 and z = 1 . We have to
seek the Laurent series that converges for 0 < z < 1.
From previous calculation we have:
1
1 1 1
=
+ 2 + + 1 +
3
4
3
z
z z
z
z
for 0 < z < 1.
dz
= 2 i Res [ f ( z = 0)] = 2 i
3
4
z z
There is an extra minus sign in the above equation

because the integration is taken along the clockwise
direction.
Evaluation of residues
0- 114
Chapter 0
(a) Simple poles:

From the Laurent series about a simple pole z = z0 :
f ( z) =
b1
+ a0 + a1 ( z z0 ) + a2 ( z z0 ) 2 + ,
z z0
we have
( z z0 ) f ( z ) = b1 + ( z z0 ) [ a0 + a1 ( z z0 ) + ] ,
and hence
Res f ( z ) = b1 = lim( z z0 ) f ( z ).
z = z0
z z0
Example:
Re s
z =i
9z + i
9z + i
=
lim
z
i
(
)
z ( z + i )( z i )
z ( z 2 + 1) z i
9z + i
10i
=
= 2 = 5i.
z
z
i
(
+
)
z =i
If the function is expressed in the form of a quotient,

f ( z) =
p( z )
, where p ( z0 ) 0 and q ( z ) has a simple
q( z )
zero at z = z0 , the residue there is given by:
0- 115
Chapter 0
Re s f ( z )
z = z0
= lim ( z z0 )
z z0
p( z )
q( z )
( z z0 ) p ( z )
z z0 ( z z ) q ( z ) + ( z z ) q ( z ) / 2 +"
]
0 [
0
0
0
= lim
Re s f ( z ) = Re s
z = z0
z = z0
p ( z ) p ( z0 )
=
.
q ( z ) q ( z0 )
Example:
9z + i 9z + i
10i
9z + i
=
=
=
= 5i.
Re s 2
3
z =i
z ( z + 1) ( z + z ) z =i 3 z + 1 z =i 2
(b) Higher-order poles:

If f ( z ) has a pole of order m > 1 at z = z0 , its Laurent
series is given by:
f ( z) =
bm
bm 1
b2
+
+
+
( z z0 ) m ( z z0 ) m 1
( z z0 ) 2
b1
+ a0 + a1 ( z z0 ) +
z z0
0- 116
Chapter 0
( z z0 ) m f ( z )
g ( z)
= bm + bm 1 ( z z0 ) + + b2 ( z z0 ) m 2
+b1 ( z z0 ) m 1 + a0 ( z z0 ) m + a1 ( z z0 ) m +1 +
This is just the Taylor series of g ( z ) . Therefore,
b1 =
1
g ( m 1) ( z0 )
(m 1)!
d m 1
1
m
Re s f ( z ) =
lim
( z z0 ) f ( z ) .
z = z0
( m 1)! z z0 dz m 1
f ( z) =
Example:
50 z
( z + 4)( z 1) 2
d
d 50 z
( z 1) 2 f ( z ) = lim
= 8.
z 1 dz
z 1 dz
z+4
Re s f ( z ) = lim
z =1
Residue integration
Residue theorem
Let f ( z ) be analytic inside a simple closed curve C and
on C, except for finitely many singular points z1 , z2 , , zk
inside C. Then the integral of f ( z ) taken counter-
0- 117
Chapter 0
clockwise around C equal 2 i times the sum of the

residues of f ( z ) at z1 , z2 , , zk :
f ( z )dz = 2 i Res f ( z ).
j =1
z=zj
Proof:
Let
Cn (n = 1, 2,.., k )
be a tiny circle
enclosing
the
singularity zn .
f ( z ) is analytic in the multiply connected domain D

bounded by C and C1 , , Ck and on entire boundary of D.
Application of the Cauchy integral formula yields:
f ( z )dz + v
f ( z )dz + v f ( z )dz + + v
C1
C2
Ck
f ( z )dz = 0,
with the integral along C being taken counter-clockwise

and the other integrals clockwise. Therefore,
0- 118
Chapter 0
f ( z )dz = v
f ( z )dz + v f ( z )dz + + v
C1
C2
Ck
f ( z )dz
where all the integrals are now taken counterclockwise.

From
Cj
f ( z )dz = 2 i Res f ( z ),
z=zj
k
f ( z )dz = 2 i Res f ( z ).
j =1
z=zj
Examples:
Evaluate the integral
4 3z
v C z 2 z dz where C is a simple
closed contour such that (a) 0 and 1 are inside C, (b) 0

is inside, 1 outside, (c) 1 is inside, 0 outside, (d) 0 and
1 are outside.
Res
z =0
4 3z 4 3z
=
= 4,
z ( z 1) z 1 z =0
Res
z =1
4 3z 4 3z
=
= 1.
z ( z 1) z z =1
0- 119
Chapter 0
Answers: (a) 6 i , (b) 8 i , (c) 2 i , (d) 0.
Evaluate the integration
tan z
dz around the circle
2
z 1
C: z = 3/ 2 .
The circle C encircles two singularities of the
integrand, namely, z = 1.
tan z
v C z 2 1dz
tan z
tan z
= 2 i Res 2
+ Res 2
z =1 z 1 z = 1 z 1
tan z
= 2 i
2z
tan z
+
2z
z =1
z = 1
= 2 i tan1.
ze z
/z
v C z 4 16 + ze dz ,
where C is the ellipse 9 x 2 + y 2 = 9 in the counterclockwise direction.
0- 120
Chapter 0
We note that z 4 16 = 0 at 2i and 2 . The first term

of the integrand has simple poles at 2i inside C.
ze z
ze
1
= 3
= ,
Res 4
z = 2 i z 16
16
4 z z =2i
ze z
ze z
1
Res 4
= 3
= .
z =2 i z 16
16
4 z z =2i
The second term of the integrand has an essential

singularity at z = 0 . The residue there equals 2 / 2
since
/z
ze
2
3
= z 1 + +
+
+
z 2! z 2 3! z 3
2 1
= z + +
+
2 z
ze z
/z
2
1
+
=
)i
ze
dz
4
v C z 4 16
0- 121
Chapter0
Applications of residue integration

Integrals of trigonometric functions
Consider integrals of type
2
I = F (cos ,sin )d .
0
Let z = ei be a complex number on the unit circle z = 1.
Since
1 i
1
1
(e + e i ) = ( z + ) ,
2
2
z
1
1
1
sin = (ei e i ) = ( z ) ,
2i
2i
z
cos =
dz / d = iei = iz ,
d = dz / iz ,
the real integral can be reduced to a complex contour

integral:
I=
dz
f ( z)
iz
where C is the unit circle z = 1 and the integral is taken in

the counter-clockwise direction. The value of the complex
contour integral can then be obtained by evaluating the
residues of the singularities inside C.
0- 121
Chapter 0
Example:
I =
d
2 cos
From cos = 12 ( z + 1/ z ) and d = dz / iz we have:
dz / iz
=
1
1
2 (z + )
2
z
dz
i
( z 2 2 2 z + 1)
2
2
dz
.
=
C
i
( z 2 1)( z 2 + 1)
C
There are two simple poles in the integrand. The first one,
z1 = 2 + 1, lies outside the unit circle C. The second one,
z2 = 2 1, is inside C with the residue given by:
Res
z = z2
1
1
=
( z 2 1)( z 2 + 1) z 2 + 1 z =
1
= .
2
2 1
1
C ( z 2 1)( z 2 + 1) = 2 i 2
dz
I = ( 2 / i ) [ 2 i ( 1/ 2) ] = 2
0- 122
Chapter 0
Evaluation of improper integrals

An improper integral of the form
f ( x)dx = lim
a a
If both lim
f ( x) dx is defined as:
b
f ( x)dx + lim f ( x)dx.

b 0
f ( x)dx and lim f ( x)dx independently
a a
b 0
exist, then it can be shown that
The limit lim
R R
principal value of
f ( x)dx = lim
R R
f ( x)dx.
f ( x)dx is usually called the Cauchy
f ( x)dx .
In the following we assume that the function f ( x) is a

rational function whose denominator is different from zero
for all real x and is of degree at least two units higher than
the degree of the numerator. In order to evaluate
f ( x)dx , we consider the corresponding contour
integral
f ( z )dz
0- 123
Chapter 0
S is a large semi-circle with a radius R.
f ( z )dz = f ( z )dz +
S
R
R
f ( x)dx = 2 i Re s f ( z )
On the large semi-circle S ( z = R > R0 ) , we have:

f ( z) <
k
z
Hence,
f ( z )dz <
k
k
R
=
.
2
R
R
Therefore, in the limit R ,
f ( x)dx = 2 i Res f ( z ) .
Note: All residues in upper half-plane have to be summed.

0- 124
Chapter 0
Example:
Evaluate the improper integral I =
dx
.
4
1+ x
Solution:
By symmetry we have I =
dx
1 dx
=
.
4
4
2 1+ x
1+ x
The integrand f ( z ) = 1/(1 + z 4 ) has four simple poles at the

points:
z1 = e i / 4 ,
z2 = e3 i / 4 ,
z3 = e 3 i / 4 ,
z 4 = e i / 4 .
Two of these poles lie in the upper half-plane.
1
1 3 i / 4
1 i / 4
1
= 3
= e
= e ,
Res f ( z ) =
4
z = z1
4
(1 + z ) z = z1 4 z z = z1 4
1
1 9 i / 4 1 i / 4
1
=
=
= e
Res f ( z ) =
.
e
4
3
z = z2
4
(1 + z ) z = z2 4 z z = z2 4
dx
2 i i / 4 i / 4
=
e e
(
)
1 + x 4
4
2 i
=
2i sin = sin =
.
4
4
4
2
1 dx
dx
=
=
.
4
4
1+ x
2 1+ x
2 2
0- 125
Chapter 0
Evaluation of Fourier Integrals

Consider Fourier integrals of the form:
f ( x) cos x dx
and
f ( x) sin x dx .
Here is a real parameter and is assumed to be positive in

the following discussion.
Jordans Lemma.
If as R , f ( z ) 0 uniformly in for 0 < < 2 ,
lim
R SU
f ( z )ei z dz = 0
lim f ( z )e i z dz = 0
R S L
Here SU is the semicircle in the upper-half plane, and S L

denotes the semicircle in the lower-half plane.
Examples:
f ( z ) = z 1 approaches zero uniformly as z .

Consider the function f ( z ) = exp(iz ) . As z ,
0, for 0 < <
f ( z ) 1,
for = 0
, for < < 2
0- 126
Chapter 0
Applications:
If as R , f ( z ) 0 uniformly in for 0 < < 2 ,
then
By closing the contour in the upper-half plane,
where
f ( x)ei x dx = 2 i Res f ( z )ei z ,

upper
f ( z )e
i z
upper
is the sum of residues in the
upper-half plane.
By closing the contour in the lower-half plane,
where
f ( x)e i x dx = 2 i Res f ( z )e i z ,
Res f ( z )e
lower
lower
i z
is the sum of residues in the
lower-half plane.
0- 127
Chapter 0
Therefore, if f ( z ) is a real-valued function on the real

axis, its Fourier cosine and sine transforms can be
obtained by separating the real and imaginary parts of
the above equations:
i z
Res f ( z )e ,
f ( x) cos xdx = 2 Im upper
i z
Res f ( z )e ,
f ( x) sin xdx = 2 Re upper
f ( x) cos x dx = 2 Im Res f ( z )e i z ,
lower
i z
f ( x) sin x dx = 2 Re Res f ( z )e .
lower
Proof:
Let S be a large
semi-circle with a
radius R. Consider
the
contour
consisting of the
semicircle S and
the diameter of S.
0- 128
Chapter 0
From the residue theorem we have
f ( z )e dz = f ( z )e dz +
i z
i z
f ( x)ei z dz
= 2 i Re s f ( z )ei z
On the large semi-circle S ( z = R > R0 ) , we have:
f ( z ) M ( R) ,
where M ( R ) is the maximum value of f ( z ) on S and
lim M ( R ) = 0 . Besides, for / 2 0 we have:
ei z = ei x e y = e R sin < exp(2 R / ) ,

following directly from the inequality sin 2 / for
/ 2 0 . Hence,
f ( z )ei z dz
f ( z )ei z dz
M ( R) e
R sin
2M ( R)
/2
/2
Rd = 2 M ( R )
/2
e R sin Rd
exp( 2 R / )Rd
exp( 2 R / ) Rd =
[1 exp( R)]
2
0- 129
Chapter 0
[1 exp( R)]
f ( z )e dz M ( R )
i z
Since lim M ( R ) = 0 ,
R
lim f ( z )ei z dz = 0 , ( > 0) .
f ( x)ei x dx = 2 i Res f ( x)ei z ,
Taking the real and imaginary parts of this equation, we

find that:
i z
,
f
(
x
)
cos
xdx
2
Im
Res
f
(
z
)
e
=
upper
i z
Res f ( z )e ,
f ( x) sin xdx = 2 Re upper
Similarly, if S is a large semicircle in the lower-half plane,

we then have:
lim f ( z )e i z dz = 0 , ( > 0) , and
i z
f ( x) cos x dx = 2 Im Res f ( z )e ,
lower
f ( x) sin x dx = 2 Re Res f ( z )e i z .
lower
0- 130
Chapter 0
Example:
Consider a case with f ( x) = (k 2 + x 2 ) 1 .
Assume for the moment the parameter s > 0 .
eisz
eisx
s 2
2
k 2 + x 2 dx = 2 i Re
z = ik
+
k
z
eisz
= 2 i
= e ks .
2 z z =ik k
cos sx
ks
=
dx
e ,
k 2 + x 2
k
sin sx
k 2 + x 2 dx = 0
If, instead, the parameter s < 0 , we can still make use of

the above results to show that:
cos s x
k 2 + x2
dx =
k s
sin s x
cos sx
dx ,
k 2 + x 2
sin sx
k 2 + x 2 dx = 0 = k 2 + x 2 dx
Therefore, for any real parameter s, we have:

cos sx
k s
=
dx
e ,
k 2 + x 2
k
sin sx
k 2 + x 2 dx = 0
0- 131
Chapter 0
Evaluation of integrals with singular integrands

Consider an improper integral
B
A
f ( x)dx
whose integrand becomes infinite at a point A < a < B , i.e.

lim f ( x) = .
x a
The value of the improper integral is defined as:
B
A
f ( x)dx = lim
0
a
A
f ( x)dx + lim
0 a +
f ( x)dx
On the other hand, the limit

a
B
B
lim f ( x)dx + f ( x) dx P f ( x) dx
a +
A
0
A
is called the Cauchy principal value of the integral.
Example:
The improper integral
dx
1 dx
dx
1
1
+ lim 3 = lim 2
1 x3 = lim
0 1 x 3
0 x
2 0 x
1
1
1
+ lim 2
0 x
does not exist. However, its principal value is given by:

1 dx
dx
dx
P 3 = lim
+ 3 = 0.
1 x
1 x 3
x
0
0- 132
Chapter 0
Theorem
If f ( z ) has a simple pole at z = a on the real axis, then
lim f ( z )dz = i Res f ( z ) ,
r 0
z =a
C2
where C2 is a semicircle with center z = a and radius r.
Proof:
From the Laurent expansion we have:
f ( z) =
b1
+ g ( z) ,
za
where g ( z ) is analytic at z = a . Let z a = rei on C2 .
C2
f ( z )dz =
b1
i
ire
d + g ( z )dz.
i
C2
re
Taking the limit r 0 , we show that:

lim f ( z )dz = ib1 = i Res f ( z )
r 0
C2
z =a
0- 133
Chapter 0
Application:
If f ( z ) has several simple poles on real axis and the
integral
along
an
infinitely
large
semicircle
lim f ( z )dz = 0 , the principal value integral of f ( z ) ,
R S
f ( x)dx , is given by:

P
f ( x)dx = 2 i Res f ( z ) + i Res f ( z ).
Here the sum 2 i Res f ( z ) extends over poles in the

UHP, whereas the sum i Res f ( z ) only includes those
poles on the real axis.
Proof:
lim f ( z )dz = 0
R S
lim f ( z )dz
r 0
C2
= i Res f ( z )
z =a
ar
lim f ( x)dx + f ( x)dx P f ( x)dx
a+r
r 0

0- 134
Chapter 0

ar
f ( z )dz + f ( z ) dz +
a+r
C2
f ( z ) dz + f ( z )dz
S
= 2 i Res f ( z )
upper
f ( x)dx = 2 i Res f ( z ) + i
UHP
Res f ( z ).
x axis
Examples:
dx
.
( x 1)( x 2)( x 2 + 1)
Evaluate P
z = 1,
1
1
Res f ( z ) =
= 2,
2
z =1
+
(
2)(
1)
z
z
z =1
z = 2,
1
1
=
Res f ( z ) =
,
2
z =2
( z 1)( z + 1) z =2 5
z = i,
1
Res f ( z ) =
z =i
(
1)(
2)(
)
z
z
z
i
z =i
1
3i
,
=
=
6 + 2i
20
dx
=
.
( x 1)( x 2)( x 2 + 1)
10
0- 135
Chapter 0
Evaluate the improper integral
sin x
dx.
I =
Note:
The integral
sin z
dz does not vanish as R and
z
hence we cannot close the contour from above (or

below) to apply the residue theorem.
eiz
Instead, we consider the improper integral
dz .
z
The integrand has a simple pole at z = 0 . Besides, the

integral
eiz
dz does vanish as R . Therefore,
z
eiz
eix
P
dx = i Res
x
z =0
z
= i.
cos x + i sin x
eix
P
dx = P
dx = i.
x
cos x
dx = 0,
sin x
dx = .
0- 136
Chapter 0
Evaluation of integrals on the range ( 0, )

To evaluate an integral of the form
dx f ( x) ,
where the function f ( x) is not an even function, we can

introduce Ln z .
Ln z = ln r + i ,
0 < 2
The discontinuity of Ln z across the cut is

Ln z1 Ln z2 = 2 i.
C1
dz f ( z )Ln z = dz f ( z )Ln z1 dz f ( z )Ln z2
0- 137
Chapter 0
dx f ( x) =
1
dz f ( z )Ln z.
C
2 i 1
If furthermore f ( z ) 0 for z as fast as z
> 0 , we can complete the contour by adding in a

vanishing contribution from C2 . That is,
1
dz f ( z )Ln z
+
C
C
1
2
2 i
= Res [ f ( z )Ln z ]
dx f ( x) =
If, then, f ( z ) has only simple poles of order one at points

zk , with residues Res f ( zk ) , we find that:
dx f ( x) = Res f ( zk ) Ln zk .
k
Example: For n = 2,3, 4,5,
dx
Ln zk
1
=
n
n 1
1+ x
k nz k
zk = n th roots of ( -1) .
0- 138
Chapter 0
For the case n = 3, z1 = e i / 3 , z2 = e i , z3 = e5 i / 3 , we have:
dx
Ln zk
1
=
k nz n 1
1 + x3
k
i
5 i / 3
i /3
= 2 i / 3 + 2 i + 10 i / 3
3e
3e
3e
i 1 1
3
5 1
3
= i
+1+ + i
3 3 2
2
3 2
2
2 3
=
.
9
Evaluation of integrals with branch cuts
Consider integrals of the form
x 1 R ( x)dx ,
where R ( z ) is rational, analytic at z = 0 , and has no poles

on the positive real axis, and where z R( z ) 0
uniformly as z 0 and as z . The power function
z 1 has a branch point at the origin.
0- 139
Chapter 0
We will choose the principal branch of z 1 . That is,

z 1 = exp [ ( 1)Ln z ] = exp [ ( 1)(Ln r + i ) ] ,
where 0 < 2 . Consequently,
z 1 = r 1ei ( 1) ,
0 < 2 ,
z 1 = r 1 = x 1
if 0+ .
Consider the contour C as shown below. Since
z R( z ) 0 uniformly as z 0 and as z , both the

integrals along the large circle at infinity and the tiny
circle around the origin vanish.
z 1 R ( z )dz
= e 2 i ( 1) x 1 R ( x)dx
0
+ x 1 R ( x)dx
0
= 2 i
inside C
1 e
2 i ( 1)
Res z 1 R( z )
x
0
2i sin 1
R ( x) dx =
0 x R ( x)dx.
i
e
0- 140
Chapter 0
Therefore,
e i
R ( x)dx =
sin
inside C
Res z 1 R ( z )
(1) 1
Res z 1 R ( z ) .
=
sin inside C
Example:
I =
x 1
dx
1+ x
for 0 < < 1.

z
Note:
0 uniformly as z 0 and as z for
1+ z
0 < < 1. Therefore, we can apply the method introduced

above to evaluate the integral.
I =
z -1
x 1
(1) 1
dx =
Res
1+ x
sin
1 + z z =1
.
sin
0- 141
Chapter 0
Transformation of integrals
An integral can be transformed to a more convenient form
by using Cauchys theorem to deform the contour. As an
example, we consider the integrals
F1 = dx cos ax 2 ,
0
F2 = dx sinax 2 ,
0
F = F1 + iF2
= dx e
iax 2
F = dz e
iaz 2
C1
= dz e
iaz 2
C2
+ dz e
C3
iaz 2
Along the contour C3 at large R we have

eiaz = exp iaR 2 ( cos 2 sin 2 ) aR 2 sin 2 .
2
This goes exponentially to zero as R except in a

small
angular
1/ R 2
range
near = 0 .
Since
dz = iRei d , the contour C3 vanishes in the limit R .

F1 + iF2 = dz e
C2
iaz 2
i / 4
=e
dr e
ar 2
1+ i 1
,
2 2 a
dx cos ax = dx sin ax 2 = / 8a .
2
0- 142
Chapter 0
Evaluation of infinite series

Theorem
Let f ( z ) be a meromorphic function and let C be a
contour that encloses the zeros of sin z , located at
z = p, p + 1, , n . If we assume that the poles of f ( z ) and
sin z are distinct,
n
m= p
f ( m) =
1
2 i
cot z f ( z )dz
Res [ cot( z ) f ( z )].
poles of f ( z )
inside C
Proof: cot z is singular at z = 0, 1, 2,... .

Res [ cot z f ( z )] z = m =
cos z f ( z )
(sin z )
= f ( m)
z =m
cot z f ( z )dz
= 2 i
Res [ cot z f ( z )]
all the poles of the integrand
= 2 i f (m) + Res [ cot( z ) f ( z )] .

m= p
poles of f ( z )
inside C
0- 143
Chapter 0
If the value of the contour integral can be obtained from

other methods, then we can make use of this theorem to
evaluate the value of a series.
Example: Evaluate the value of the series
2x
.
2
2 2
x
n
+
n =1
Solution:
(a) We consider the value of a related series:
N
m= N
2x
x 2 + m 2 2
(b) We apply the above theorem to this series.

N
m= N
2x
1
=
x 2 + m 2 2 2 i
cot zf ( z )dz
Res [ cot( z ) f ( z )],
C
poles of f ( z )
inside C
where the contour C is a closed rectangle enclosing

the points z = 0, 1, 2,..., N .
0- 144
Chapter 0
(c) Let the length and width of the contour C approach .

1
2 i
cot zf ( z )dz
1
2
cot z
2x
dz .
2
2 2
x +z
We can show that cot z is bounded if the contour C is

suitably chosen. Thus the integrand goes as 1/ z 2 as
z , and the integral vanishes.
1
2 i
cot zf ( z )dz 0.
(d) The value of the infinite series can then be obtained:
2x
2 x( cot z )
Res
=
2
2 2
x 2 + z 2 2
m = x + m
z = ix /
2 x cot (ix / ) cot (ix / )
=
+
2ix /
2ix /
= 2i cot ix = 2 coth x.
2x
2
+
= 2coth x
2
2 2
x
m =1 x + m
1
2x
coth x = 2
,
2 2
x m=1 x + m
0- 145
Chapter 0
(e) Appropriate choice of contour C:
cos z
1/ 2
cos 2 x + sinh 2 y
= 2
cot z =
sin z sin x + sinh 2 y
Let the vertical sides of contour C cross the x-axis at a

large enough half-integer.
1/ 2
sinh y
cot z =
2
1 + sinh y
2
= tanh y 1.
Over the horizontal sides of the rectangle, we have:

1/ 2
cos x + sinh y
lim cot z = lim 2
y
y sin x + sinh 2 y
= 1.
Therefore, cot z is bounded on C.
Evaluation of series with alternating terms

Application of the residue theorem also leads to a method
for calculating infinite sums of the form:
n
(
1)
f ( n)
Note that the function sin z has an infinite series of zeros:

sin z = 0 at z = 0, 1, 2, .
0- 146
Chapter 0
Hence,
1
has first-order poles at
sin z
z = 0, 1, 2, n, .
The residues are given by

1
Res
sin z z = n
( 1)
1
1
=
=
=
.
(sin z ) z = n cos n
n
Following arguments similar to those for the previous

proof, we can establish the following theorem.
Theorem
Let f ( z ) be a meromorphic function and let C be a
contour that encloses the zeros of sin z , located at
z = p, p + 1, , n . If we assume that the poles of f ( z ) and
(sin z ) 1 are distinct,

n
(1)m f (m) =
m= p
1
2 i
C sin z f ( z )dz
f
z
Res
(
)
sin z
.
poles of f ( z )
inside C
0- 147
Chapter 0
Therefore, if the poles of f ( z ) inside the contour C, zk

(k = 1, 2,3,) , are all simple poles, then
dz
Re s f ( zk )
f ( z)
2 i
sin z
sin z k
zk
= 2 i
( 1)
f (n).
If, in addition, f ( z ) / sin z 0 at infinity sufficiently

rapidly, the integral around the contour C at infinity also
vanishes. This is not a very restrictive requirement since
( sin z )
0 exponentially except near the real axis.
However, it must be checked for each case.
Assuming that the size the integral approaches infinity and

as the limit
0 , we have:
( 1)
n =
f ( n) =
zk
Re s f ( zk )
sin zk
0- 148
Chapter 0
Example:
Consider the series
( 1)
n =
In this case, f ( z ) =
1
.
n
1
and there is only one simple pole
z
located at zk = . Besides,
Res f ( zk ) = 1 .
From the above theorem, we have:
1
=
, or
( 1)
n
sin
n =
1
1 2
=
sin
( 1)
n
n =1
This is the partial fraction expansion of csc .
0- 149
Chapter 0
Applications of Complex Analysis

Electrostatic field in two dimensions
The electrical force between charged particles is governed
by the Coulombs law. This force is expressible as the
gradient of a function , called the electrostatic potential.
In a region where the charge density is zero, is a
solution of Laplaces equation.
2 = 0.
The surfaces
= constant
are called the equipotential surfaces. At each point P the
gradient of is perpendicular to the surface = constant
through P ; that is, the electrical force has the direction
perpendicular to the equipotential surface.
If the potential is independent of one of the space
coordinates, say z , then the Laplaces equation becomes
2 2
= 2 + 2 = 0.
x
y
2
This is called the 2-dimensional Laplaces equation.
0- 151
Chapter 0
Examples:
Potential between parallel plates
( x)
=
1
1
( 2 1 ) x + ( 2 + 1 ) .
2
2
Here 1 and 2 are the

potentials
on
the
plates,
respectively located at x = 1.
Potential between coaxial cylinders
Assume that the cylinder axis is along the z -axis.

Then the potential is independent of z and as
well. The Laplaces equation in 2-dimensional
cylindrical coordinates reads as:
r 2 rr + r r + = 0
Since = 0 , we then have:
r + = 0.
0- 152
Chapter 0

1
= ,
r
d ln
1
= ,
dr
r
ln = ln r + C ,
a
,
r
= a ln r + b.
Potential in an angular region

We consider the potential
between two conducting
plates, which are kept at
potentials 1 (the lower
plate) and 2 , and make an
angle , where 0 < .
The Laplaces equation:
r 2 rr + r r + = 0
We will make use of the theory of harmonic
functions to find the potential.
0- 153
Chapter 0
Consider the function

( x, y ) = a + b Arg z ,
where a and b are real constants. It is a harmonic

function satisfying the Laplaces equation since it is
the real part of an analytic function:
a - ibLn z = -ibLn z + ( a + bArg z )
In polar coordinates, = Arg z. Therefore, is

constant on rays = constant , which satisfies the
boundary conditions on the plates:
1
a + b = 1 ,
2
1
a + b = 2.
2
From these two equations, the constants a and b can be

found, and the potential is given by:
( x, y ) =
1
1
+
( 2 1 ) ( 2 1 ) ,
where
= arctan
y
.
x
0- 154
Chapter 0
Complex Potential
Let ( x, y ) be harmonic in some domain D and ( x, y )
a conjugate harmonic of in D . The complex function
F ( z ) = ( x , y ) + i ( x, y )
is an analytic function of z = x + iy . This function F is
called the complex potential corresponding to the real
potential .
Theorem:
The curves = constant intersect the equipotential lines
= constant in the xy-plane at right angles except where

F ( z ) = 0 .
Proof:
Assume that the curves = A and = B intersect at a
point z0 = x0 + iy0 . The gradient vectors of these two
functions are respectively given by:
= i
+j
x
y
and
= i
.
+j
x
y
0- 155
Chapter 0

i
i
j
i = i
+j
+
x
y

+
=
x x y y
From Cauchy-Riemann relations
=
=
and
,
x
y
y
x
i =
The
gradient

=0 .
+
y y y y
vectors
and
hence
the
curves
are
perpendicular to each other at the point of intersection.
Note:
The gradient vectors vanish if F ( z ) = 0 and the proof
sketched above becomes invalid.
With the complex potential F = + i we can
simultaneously
handle
equipotential
lines
= constant and the lines of force = constant .
0- 156
Chapter 0
Examples:
Consider the complex potential
F ( z ) = az + b = ax + b + iay.
The corresponding real potential is ( x, y ) = ax + b ,
which gives the potential in between two parallel
plates. The lines of force are given by:
= ay = constant
The complex potential

F ( z ) = a Ln z + b
yields the real potential between coaxial cylinders:
= a ln r + b = a ln z + b .
The family of lines of force are given by:
= aArg z = constant
= constant .
F ( z ) = a ibLn z = ( a + bArg z ) ibLn z

is associated with the real potential
( x, y ) = a + b ,
representing the potential between two metal plates.

The lines of force are given by: Ln r = const .
0- 157
Chapter 0
Application of Conformal Mapping

Boundary value problems for the Laplaces equation:
To find a solution of the Laplaces equation in some
domain with prescribed values on the boundary.
Method of conformal mapping

Conformal mapping is used to map a given complicated
domain onto a simpler one where the solution of the
Laplaces equation is known or can be found more easily.
Theorem Harmonic functions under conformal mapping

Let * (u , v) be harmonic in a domain D* in the w-plane.
Suppose that w = u + iv = f ( z ) is analytic in a domain D in
the z-plane and maps D conformally onto D* . Then the
function
( x, y ) = * (u ( x, y ), v( x, y ))
is harmonic in D.
0- 158
Chapter 0
Proof:
Let * (u , v) be a harmonic conjugate of * (u , v) . The
complex function
F * ( w) * (u , v) + i * (u , v)
is then an analytic function of w. Since the composite of
analytic functions is also analytic, the composite function
F ( z ) F * ( f ( z ))
is an analytic function in D. Consequently, the real part of
F ( z ) , given by
Re F ( z ) ( x , y )
Re F * ( f ( z )) = * (u ( x, y ), v( x, y ))
is a harmonic function.
0- 159
Chapter 0
Linear fractional transformations

Definition:
Linear fractional transformations (bilinear transformations
or Mbius transformations) are one-one mappings
represented by:
w=
az + b
cz + d
where
a b
.
c d
Note:
If
a b
= = k (i.e. ad bc = 0 ), the mapping becomes
c d
w=
kcz + kd
=k ,
cz + d
a constant independent of z. In the following discussion

we will consider nontrivial cases with ad bc 0 .
Special cases:
Identity
w= z
Translation
w = z+b
Rotation
w = az with a = 1
Linear transformation
w = az + b
Inversion in the unit circle w = 1/ z

0- 160
Chapter 0
Inversion in the unit circle
w = 1/ z
Fixed points:
Fixed points of a mapping are points that are mapped onto
themselves. That is,
w = f ( z) = z .
For linear fractional transformation w =
az + b
, the fixed
cz + d
points satisfies the algebraic equation
z=
az + b
.
cz + d
The fixed points are the roots of the quadratic equation:
cz 2 ( a d ) z b = 0 .
This equation has at most two solutions except when
c = 0 = b and a = d .
0- 161
Chapter 0
Theorem
A linear fractional transformation, not the identity, has at
most two fixed points. If a linear fractional transformation
is known to have three or more fixed points, it must be the
identity mapping w = z such that c = 0 = b and a = d .
Theorem Uniqueness of linear fractional mapping

Let f ( z ) and g ( z ) be two linear fractional mappings. If
there exist three distinct values z1 , z2 , z3 such that
f ( zi ) = g ( zi ) for i = 1, 2,3 ,
then these two mappings are identical.
Theorem
Three given distinct points z1 , z2 , z3 can always be mapped
onto three prescribed distinct points w1 , w2 , w3 by one, and
only one, linear fractional transformation w = f ( z ) . This
mapping is given implicitly by the equation
w w1 w2 w3 z z1 z2 z3
.
w w3 w2 w1 z z3 z2 z1
0- 162
Chapter 0
Proof:
(a) It is obvious that the mapping
w w1 w2 w3 z z1 z2 z3
w w3 w2 w1 z z3 z2 z1
is a linear fractional mapping from z-plane to the wplane.
(b) It can be easily shown that
F ( w1 ) = 0 = G ( z1 ) ,
F ( w2 ) = 1 = G ( z2 ) ,
F ( w3 ) = = G ( z3 ) ,
where
F ( w) =
w w1 w2 w3
, and
w w3 w2 w1
G( z) =
z z1 z2 z3
.
z z3 z2 z1
(c) Therefore, this transformation maps zi onto wi for
i = 1, 2,3 .
(d) By the uniqueness theorem there is only one linear
fractional mapping that can map three distinct zi onto
wi and hence the theorem is proved.
0- 163
Chapter 0
Method constructing linear fractional mappings

Theorem
All bilinear transformations are composites of linear
transformation w = az + b and inversion w = 1/ z .
Theorem
A bilinear transformation always maps circles into
circles.
Note: Here circles include straight lines as a limiting
case.
Construction of bilinear mapping

To obtain a bilinear transformation that maps a domain D
in the z-plane onto a domain D* in the w-plane, one can
choose three distinct boundary points z1 , z2 , z3 of the
domain D and prescribes the three corresponding images
w1 , w2 , w3 on the boundary of the domain D* . The bilinear
transformation is then fixed by the values of z1 , z2 , z3 and
w1 , w2 , w3 .
0- 164
Chapter 0
Standard mappings:
Mapping of a half-plane onto a disk
If we let z1 = 1 , z2 = 0 , z3 = 1 ; w1 = 1 , w2 = i ,
w3 = 1, then from
w w1 w2 w3 z z1 z2 z3
w w3 w2 w1 z z3 z2 z1
we have:
w ( 1) i 1
z ( 1) 0 1
w 1 i ( 1)
z 1 0 ( 1)
z i
w=
.
iz + 1
Mapping
z i
w=
iz + 1
0- 165
Chapter 0
Mapping of the unit disk onto right half-plane

Find the linear fractional transformation that maps
z1 = 1 , z2 = i , z3 = 1 onto w1 = 0 , w2 = i , w3 = ,
respectively.
w w1 w2 w3 z z1 z2 z3
,
w w3 w2 w1 z z3 z2 z1
Since
we have:
w (0) i z ( 1) i 1
w i (0)
z 1 i (1)
w=
z +1
.
z 1
Note that the following conventions are useful:
z
w
= 1,
= 0 = , z = , etc
Mapping of the unit disk onto another unit disk

The mapping
z c
w=
cz 1
maps z 1 onto w 1 with z = c mapped onto

w = 0.
0- 166
Chapter 0
If c =
1
2z 1
, then w =
.
2
z2
Applications of conformal mapping in potential theory

We can first solve the Laplaces equation in a simple
domain and then use conformal mapping (e.g bilinear
mappings) to map the solution to another domain where
explicit solution of the Laplaces equation is not available.
Examples:
Potential between noncoaxial cylinders
Find the potential between the cylinders C1 : z = 1
(potential U1 = 0 ) and C2 : z 2 / 5 = 2 / 5 (potential
U 2 = 110 volts).
0- 167
Chapter 0
Using the bilinear mapping

w=
z b
, (b is a real number)
bz 1
we map the unit disk z = 1 onto the unit disk w = 1 in

such a way that C2 is mapped onto some cylinder
C2* : w = r0 . In particular, we demand that:
z = 0 w = r0 ,
z = 4 / 5 w = r0 .
r0 =
0b
= b,
0 1
r0 =
4 / 5 r0
4/5 b
.
=
4b / 5 1 4 r0 / 5 1
r0 = 1/ 2 , (The other solution r0 = 2 is discarded.)
and b = 1/ 2 .
The mapping is given by
w = f ( z) =
2z 1
.
z2
0- 168
Chapter 0
The complex potential in the w-plane is

F * ( w) = aLn w + k ,
which is associated with a real potential:
* (u, v ) = Re F * ( w) = a ln w + k .
We can determine a and k from the boundary conditions.

If w = 1, then * = a ln1 + k = 0 , hence k = 0 .
If w = r0 = 1/ 2 , then * = a ln(1/ 2) = 110 , hence
a = 110 / ln(1/ 2) = 158.7 .
Hence, the corresponding complex potential in the w-plane

is
2z 1
F ( z ) = F ( f ( z )) = aLn
.
z2
*
which has a real potential:

2z 1
( x, y ) = Re F ( z ) = a ln
.
z2
0- 169
Chapter 0
Potential between two semicircular plates

Find the potential between two semicircular plates P1 and
P2 having potentials 3000 and 3000 volts, respectively.
We map the unit disk onto the right half of the w-plane by
using the linear fractional transformation
w = f ( z) =
1+ z
.
1 z
The boundary z = 1 is mapped onto the boundary u = 0

(the v-axis), with z = 1 , i , 1 going onto w = 0 , i , ,
respectively, and z = i onto w = i . Hence the upper
semicircle of z = 1 is mapped onto the upper half, and the
lower semicircle onto the lower half of the v-axis.
0- 170
Chapter 0
From the boundary conditions

3000 for v > 0
(u = 0, v) =
3000 for v < 0
*
we can determine the potential * (u, v ) in the right half

plane of the w-plane:
* ( u, v ) =
6000
, where = Arg w .
F * ( w) =
6000i
Ln w .
The potentials in z-plane are then obtained by expressing

w in terms of z:
F ( z ) = F * ( f ( z )) =
6000i
Ln
1+ z
,
1 z
and
( x, y ) = Re F ( z )
1+ z
=
.
Im Ln
1 z
6000
1+ z
=
Arg
1 z
6000
0- 171
Chapter 0
Applications of complex analysis in fluid mechanics

Theory of fluid mechanics
Consider a fluid flow in a 2-dimensional domain D. The
velocity of the flow is described by a 2-dimensional vector
field (V1 , V2 ) .
The fluid is incompressible if
V1 V2
+
= 0.
x
y
The flow is said to be irrotational if

V2 V1
= 0.
x
y
If the domain of flow is simply connected and the flow is

irrotational, then by Greens theorem
V2 V1
dxdy = 0 ,
y
x
C (V1dx + V2 dy ) =
D
where C is any simple closed curve in D . Hence, the line

integral
( x, y ) =
( x, y )
( a ,b )
(V1dx + V2 dy )
is independent of path in D .
0- 172
Chapter 0
This line integral defines the velocity potential ( x, y ) .

Since
d = V1dx + V2 dy =
dx +
dy ,
x
y
we have:
V1 = / x ,
V2 = / y .
Complex analysis and fluid flow

For an incompressible and irrotational flow, the velocity
potential satisfies the Laplaces equation:
2 2
= 2 + 2 = 0,
x
y
2
and is a harmonic function. In this case, the flow can be

described by a complex potential F ( z ) ,
F ( z ) = ( x , y ) + i ( x, y ) ,
which is an analytic function. is called the stream
function and it also satisfies the Laplaces equation:
2 2
= 2 + 2 = 0.
x
y
2
0- 173
Chapter 0
Since the curves ( x, y ) = const are perpendicular to the

curves ( x, y ) = const at points of intersection and the
velocity field is given by the gradient of ( x, y ) , the
curves ( x, y ) = const in fact shows the direction of the
flow and hence the name stream function.
Besides, a complex velocity function

V = V1 + iV2
is defined, giving the magnitude V = V12 + V22 and the
direction Arg V of the velocity field at each point
z = x + iy . The complex velocity field is directly related to

the complex potential through the relation:
V = V1 + iV2 = F ( z )
where the overbar denotes the complex conjugate.
Proof:
,
V1 =
x
V2 =
.
y
F ( z ) = x i x = x + i y = V1 + iV2 = V .
0- 174
Chapter 0
Applications:
Flow around a 90 corner
F ( z ) = z 2 = x 2 y 2 + 2ixy
describes an incompressible irrotational flow around a 90
corner.
Equipotential lines = x 2 y 2 = const (Hyperbolas)
Streamlines
= 2 xy = const
Velocity
V = ( z 2 ) = 2 z = 2( x iy ) ,
(Hyperbolas)
V1 = 2 x , V2 = 2 y .
Note: In fluid flow a boundary across which fluid cannot

flow must be a streamline.
In this case the boundary is

given by the streamline:
= 2 xy = 0 ,
i.e. the x and y axes.

0- 175
Chapter 0
Flow around a cylinder

F ( z ) = ( x, y ) + i ( x, y ) = z +
1
z
describes the flow around a cylinder. In polar coordinates,

z = rei , and
1
1
1
F ( z ) = rei + e i = r + cos + i r sin .

r
r
r
The streamlines are

1
( x, y ) = r sin = const .
r
( x, y ) = 0 gives r 1/ r = 0 or sin = 0 . Hence this

streamline consists of the unit circle ( r = 1/ r gives r = 1)
and the x-axis ( = 0 and = ). Therefore, this flow can
be interpreted as a flow around a unit cylinder.
0- 176

Ch0 Total

Uploaded by

Copyright:

Available Formats

Ch0 Total

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Ch0 Total

Uploaded by

Copyright:

Available Formats

Chapter 0

Theory of complex variables

x and y are respectively the real part and the imaginary

PHY 4602 Methods in Theoretical Physics II

Theory of complex variables

Addition of two complex numbers

Multiplication of two complex numbers

In practice, complex numbers z = ( x, y ) are written as:

( x1 + iy1 )( x2 + iy2 ) = x1 x2 + ix1 y2 + iy1 x2 + i 2 y1 y2

x1 + iy1 ( x1 + iy1 )( x2 iy2 )

PHY 4602 Methods in Theoretical Physics II

Theory of complex variables

The xy-plane in which the complex numbers are

The addition and subtraction of complex numbers acquire

PHY 4602 Methods in Theoretical Physics II

Theory of complex variables

Complex conjugate numbers

PHY 4602 Methods in Theoretical Physics II

Theory of complex variables

In the complex plane, z and z form a pair of mirror

Polar Form of Complex Numbers

Theory of complex variables

r is called the absolute value or modulus of z and is

is called the argument of z and is denoted by arg z .

The value of that lies in the interval < is

PHY 4602 Methods in Theoretical Physics II

Theory of complex variables

lengths of any two

The generalized triangle inequality:

PHY 4602 Methods in Theoretical Physics II

Theory of complex variables

Multiplication and Division in Polar Form

= [cos(1 2 ) + i sin(1 2 )].

arg( z1 z2 ) = arg z1 + arg z2 .

PHY 4602 Methods in Theoretical Physics II

Theory of complex variables

Obviously, there are n possible distinct values of w ,

Hence the symbol

PHY 4602 Methods in Theoretical Physics II

Theory of complex variables

Open circular disk

An open circular disk z a < is called the

PHY 4602 Methods in Theoretical Physics II

Theory of complex variables

An open annulus (circular ring):

(Open) upper half-plane:

PHY 4602 Methods in Theoretical Physics II

Theory of complex variables

Theory of complex variables

A complex function f(z) is equivalent to a pair of real

Theory of complex variables

in which z approaches should not

affect the value of the limit. If a limit exists, it is then

Theory of complex variables

In other words, if 0 as 0 , then

f ( z ) is said to be continuous in a domain if it is

PHY 4602 Methods in Theoretical Physics II

Theory of complex variables