Unit 11 Complex Integration

UNIT 11
COMPLEX
INTEGRATION
Structure
11.1 Introduction 11.4 Cauchy’s Integral Formula
Expected Learning Outcomes Derivatives of an Analytic Function
11.2 Integration of a Function of a 11.5 Series Representation of a
Complex Variable Complex Function
Arcs and Contours Sequence and Series
Contour Integrals in the Complex Plane Taylor Series
Upper Bound for the Absolute Value of Laurent Series
Complex Integral: Darboux Inequality 11.6 Summary
11.3 Cauchy’s Integral Theory 11.7 Terminal Questions
Some Consequences of Cauchy’s 11.8 Solutions and Answers
Integral Theorem Appendix A: Taylor Series
Appendix B: Laurent Series
11.1 INTRODUCTION
In earlier Unit 10 you learned about functions of a complex variable. You now
know how to calculate the derivative of a function of a complex variable and to
identify whether or not a function is analytic. In this unit we have extended the
study of analytic functions to integration in complex plane. By mastering the
theory of complex integration, as a physics student you will be equipped to
handle many difficult real integrals encountered in advanced topics.
In Sec.11.2 we will introduce you to the meaning of a contour in the complex
plane and the idea of contour integration involving complex functions. In
Sec. 11.3 and 11.4 you will learn two fundamental integral theorems in relation to
analytic functions  the Cauchy’s integral theorem and the Cauchy’s integral
formula. You will note that the Cauchy’s integral formula enables us to show that
an analytic function possesses derivatives of all orders, a remarkable result!
Further, it enables us to develop power series representation  Taylor series and
Laurent series for functions of a complex variable. The Laurent series is around a
singular point of the analytic function. Its utility stems from the fact that it is
particularly useful to study the nature of the related singularity. These are
discussed in Sec. 11.5.
The proofs of Taylor series and Laurent series are given in the appendices A
and B, respectively.
Expected Learning Outcomes
After completing this unit, you should be able to:
 evaluate contour integrals involving complex functions;
37
Block 3 Complex Analysis
 apply integral theorems to evaluate integrals and derivatives of an analytic
function;
 write an analytic function near a non-singular point as an infinite Taylor
series; and
 express an analytic function near a singular point as an infinite Laurent
series.

11.2 INTEGRATION OF A FUNCTION OF A

COMPLEX VARIABLE
In the preceding unit you learnt to evaluate the derivative of a complex
function. Now we consider integration of a function of a complex variable in
the complex plane. This may be regarded as a generalization of the method of
Riemann integration for a real function along the real x-axis. For a real
function f (x ) bounded and defined on the interval x = a and x = b, the integral
b
is written as  f ( x ) dx . Geometrically, this is equal to the area between the
a
curve y  f (x ) and the x-axis and bounded by the ordinates x = a and x = b. In
the case of complex integration, we generalize from a part of x-axis to a curve
in the z-plane. Since a complex integral has to be defined in relation to a curve
in the z-plane, it is natural to ask: What types of curves do we consider in
defining a complex integral? To enable you to appreciate this question, it is
important to first learn about arcs and contours in the complex plane.

11.2.1 Arcs and Contours

A continuous arc is the set of points (x, y) that satisfy parametric equations
x = (t), and y = (t), a  t  b, where (t) and (t) are continuous functions of
the real parameter t in [a, b]. If each point (x, y) on the arc corresponds to a
single value of t, the arc is called a Jordan arc (see Fig. 11.1a).

If it happens that (a) = (b) and (a) = (b), but no other two values of t
correspond to the same point (x, y), then the continuous arc is called a simple
closed curve or a Jordan curve.

An arc or curve is smooth if the derivatives '(t) and '(t) are continuous in
[a, b] and do not vanish simultaneously. The length of a smooth arc or curve is
given by
b
L  (t ) 2   (t ) 2 dt (a  b) (11.1)
a

A contour is a continuous chain of a finite number of smooth arcs. For a

contour, ' and  are sectionally continuous. The length of a contour is the
sum of the lengths of smooth arcs. A piece-wise smooth closed curve without
points of self-intersection is called a closed contour.

In (Fig. 11.1a-d) we have shown a few typical arcs and closed contours.
Arrows indicate the sense of describing a curve.
38
Unit 11 Complex Integration

(a) (b)

(c) (d)

Fig. 11.1: a) Jordan arc; b) A closed contour with four pieces of smooth arcs;
c) A smooth closed curve; d) Large semi-circle with a small semi-
circle. It is an example of a closed contour with piece-wise smooth
arcs.

Example 11.1
Refer to Fig. 11.2. It shows a circle of radius R, centred at z0 ( x0 , y 0 ). Let
z ( x, y ) be a point on the circle.
Solution :

Fig. 11.2: Parametric representation of a circle of radius R in the z-plane with

centre at z0 .

Then z  x  iy and z 0  x 0  iy 0 so that z  z 0  x  x 0  i ( y  y 0 ). In terms

of radius R and angle , where  is measured from an axis passing through
the point z  z 0 and parallel to the x-axis, we can write 39
Block 3 Complex Analysis
x  x 0  R cos  and y  y 0  R sin , 0    2

 z  z0  R(cos   i sin )  R e i

This is the parametric equation of a circle.

For a circle centred at the origin of the z-plane, z 0  0, the parametric

equation of a circle takes the form

z  R e i and z R

You should note that in the complex plane, a circle has a simple parametric
representation.

SAQ 1
Obtain parametric representation of an ellipse defined by

x2 y 2
 1
a2 b2

11.2.2 Contour Integrals in the Complex Plane

Consider a piece-wise smooth curve C in the complex plane connecting two

points a and b (Fig. 11.3). Let f (z ) be a continuous function which is defined at
each point of C. We sub-divide C into n parts by selecting n1 intermediate
points: a  z0 , z1, z 2 ,..., z n  b. On each portion of sub-division of C we select
an arbitrary point; say j is a point on the curve between z j and z j 1. We next
form the sum
n
Sn   f ( j )  z j (11.2)
j 1

where z j  z j  z j 1 is a chord connecting the points z j and z j 1. Let us

increase the number of sub-divisions infinitely in such a way that the largest of
the chords z j approaches zero, i.e., n  so that z j  z j 1  0 for all j.

If limn  Sn exists and is independent of the manner in which the curve has
been sub-divided or the intermediate points  j chosen, then this limiting value
is called the contour integral of f (z) along the curve C from point a to point b:

n
 f (z) dz  lim f
n  j 1
( j ) (z j  z j 1)
C

max zj  zj 1  0 (11.3)

The curve C is called the path of integration.

40
Unit 11 Complex Integration

Fig. 11.3: C is a smooth curve. The curve is divided into n parts by the points
z1, z2,...,zn 1.z j  z j  z j 1 is the chord between the points z j and
z j 1.

Conversion of a Contour Integral to Real Integrals

The complex contour integral can be expressed in terms of line integrals

involving real part u(x, y) and imaginary part v ( x, y ) of f(z). To this end, we
write

f ( j )  u(j ,  j )  i v (j ,  j )

where  j   j  i  j and z j  z j 1  x j  x j  1  i ( y j  y j  1 )  x j  i y j .
On substituting this in Eq. (11.3) we obtain

n
 f ( z ) dz  lim
n 
 u( j , j )  iv ( j , j )(x j  iy j )
C j 1
max x j  x j 1 0
max y j  y j 1 0

n
 lim  u ( j ,  j ) x j  v ( j ,  j ) y j  

n  
j 1

n 
i  u( j ,  j )y j  v ( j ,  j ) x j  

j 1

Since f is continuous, so are u and v. Hence as n  , x j  0 and y j  0

and each sum becomes a real line integral:

 f (z) dz   (u dx  v dy )  i  (u dy  v dx ) (11.4)
C C C

This suggests that the value of the line integral is independent of the choice of
sub-divisions and intermediate points. 41
Block 3 Complex Analysis
Let us pause for a while and think as to what have we achieved. You will note
that we have essentially reduced a complex integral to a complex sum of two
real integrals. But this may not be possible always, say if f ( z )  log z. We may
consider such cases separately.

Conversion to a definite integral

An equally useful way of representing a contour integral defined by Eq. (11.3)

is by the parametric representation of the curve C. Let a point z on C satisfy
the parametric equations x = (t) and y = (t), (t a  t  t b ). On substituting
these in Eq. (11.4), we obtain

tb
 
f ( z )dz  u [(t ), (t )] dt  v [(t ), (t )]dt 
C ta

tb

i u[(t )(t )]dt  v [(t ), (t )]dt (11.5)
ta

You will note that the contour integral is expressed as a definite integral
between the limits t a and t b . We illustrate this through a simple example.

Example 11.2


Evaluate the integral z dz along the straight line from 0 to 1 + i.
C

Solution : For any path C between the two points in the z-plane, we can
write

 f (z)dz   ( x  iy ) (dx  idy )

C C

 
= ( xdx  ydy )  i ( ydx  xdy )
C C

The given path has a parametric representation x = t, y = t with 0 t  1.

Hence

1 1
  
z dz  (tdt  tdt )  i (tdt  tdt )
C 0 0

Since the first integral on the right hand side vanishes, we obtain

1
 z dz  2i  tdt  i
C 0

In defining the contour integral with respect to a closed contour, we say that
the contour should be described in the positive sense, i.e., anticlockwise
42
Unit 11 Complex Integration
direction. For this the region enclosed by the closed contour should always lie
to the left as we move along the contour. A circle described in the anti-
clockwise sense is a familiar example. This is illustrated in the following
example.

Example 11.3

dz dz
Evaluate the integrals (i)  z
and (ii)  zn
(n  2, 3,... ) around a unit circle

with centre at the origin.

Solution : A point on a unit circle has parametric representation z  e i so

that dz  e iid.

2 i  2
dz e id
i)  z
  ei 
 i d  2i
0 0

The symbol  has been used to emphasize that the path of integration is a
simple closed contour traversed positively.

2 i 2
dz e id
ii)     i e  i n 1d

zn e in 
0 0

e i (n 1) 2
i
 i (n  1) 0

1
(n  1)
e i (n 1)2  e0 
1
 1 1  0
(n  1)

Properties of Contour Integrals

Some properties of contour integrals which follow naturally from Eq. (11.3) and
(11.4) are as follows:

 If f (z ) and g (z ) are analytic functions, then integral of their sum along C is

equal to the sum of their integrals:

 f (z)  g(z)dz   f (z)dz   g(z)dz

C C C

 If A is a complex constant, then

 A f (z)dz  A  f (z)dz
C C

 If a and b are two points on C, then

b a
 
f ( z )dz   f ( z )dz
a b 43
Block 3 Complex Analysis
 If m is an intermediate point between a and b, then

b m b
 
f ( z )dz  f ( z )dz   f (z)dz
a a m

11.2.3 Upper Bound for the Absolute Value of a Complex

Integral: Darboux Inequality

Consider a function f (z ) which is continuous on C and satisfies the

inequality f (z)  M. The contour integral on C is denoted by I  f ( z ) dz. From
C
Eq. (11.3) we recall that

n
I  lim
n
 f ( k )(zk  zk 1)
k 1
max zk  zk 1 0

The absolute value of the contour integral is given by

n
I   f ( z ) dz  lim  f (k ) (zk  zk 1)
n  k 1
C

Since the modulus of the sum of complex numbers is less than the sum of
their moduli, we find that

n
 f (z)dz  nlim

 f ( k ) zk  zk 1
C k 1

 n 
 M  lim
 n  
zk  zk 1 
 (11.6)
 k 1 

as f (z)  M.

n
Since lim  zk  zk 1
n  k 1
defines the length, L, of the contour, Eq. (11.6) can

be rewritten as

 f (z)dz  M.L  max f L (11.7)

C

This inequality is called the Darboux's inequality. It is very useful particularly in

the evaluation of contour integrals. Let us demonstrate this inequality with the
help of an example.
44
Unit 11 Complex Integration

Example 11.4

2 i
Show that  z 2dz  10, where the integration path is a straight line joining
i
z = i to z = 2+ i.
Solution : Length of the integration path L  2  i  i  2 . Maximum value of
the modulus of z 2 , M  (2  i ) 2 = 5. By Eq. (11.3)

2 i
 z 2dz  ML  5  2  10
i

Having discussed line integrals in the complex plane, we will now introduce
you to two basic theorems  Cauchy's theorem and Cauchy's integral formula.
To this end, you need to know some basic concepts regarding simply
connected and multiply connected regions in a complex plane.

Simply and Multiply Connected Regions in a Complex Plane

A region R in the complex plane is said to be simply connected if every simple
closed curve in R can be shrunk to a point without going out of the region R
(Fig. 11.4 a). The interior of a circle, an ellipse or a square is simply
connected. In general, the interior of a simple closed curve is simply
connected. Is a circular ring simply connected?

Fig. 11.4: a) The shaded area is a simply connected region. The closed contour
C can be shrunk to a point without going out of the region; b) The
shaded area is a multiply connected region because of the presence
of holes which lie outside the region. A closed contour C enclosing a
hole cannot be shrunk to a point without leaving the region; c)The
shaded area which is the annular region between two concentric
circles is a doubly connected region; d) A general multiply-connected
region. 45
Block 3 Complex Analysis
A region which is not simply connected is said to be multiply connected
(Figs. 11.4b, c and d).
You should note the basic difference between the aforementioned two types of
regions. In Fig. 11.4b, if you try to shrink C to a point, you cannot do it without
leaving the region and entering the hole. The annular region between two
concentric circles is also an example of a multiply-connected region
(Fig. 11.4c). A more general multiply-connected region containing more than
two non-overlapping holes, which lie outside the region, is shown in
Fig. 11.4d.

SAQ 2
State whether the following regions are simply connected or multiply
connected:
a) Re z  0 b) 0 < z  1
c) 0  arg z  /4 and z > 1

11.3 CAUCHY’S INTEGRAL THEOREM

In BPHE-104 course Cauchy's integral theorem forms the basis of the theory of the behaviour of
on Mathematical functions of a complex variable. It may be stated as follows:
Methods in Physics-I,
you have learnt to
If a function f (z) is analytic and its derivative is continuous at each point
transform a line
integral to a double
within a simply connected region R, then for every closed path C in R the line
integral for contours in
integral of f (z) around C is zero. Mathematically, we write
the x-y plane. You will
recall that for a vector 
O f ( z )dz  0 (11.8)
A in the x-y plane C
given by
If we write z = x + iy and f ( z )  u  iv , where x, y, u, and v are real, then from
A  P( x, y ) ˆi  Q( x,y )ˆj, Eq. (11.4) we have
we can write  f (z) dz  (udx  vdy )  i  (vdx  udy ) (11.9)

 (Pdx  Qdy ) Since f (z ) is analytic, it is real and imaginary parts u and v satisfy the Cauchy-
C Riemann conditions (Eq. 10.13). Moreover, continuity of f (z ) assures that
 Q P 
     dxdy
 x y 
partial derivatives of u and v are also continuous in the region bounded by C.
R We can apply Green's theorem in a plane to both the integrals on the right
where C is a simple hand side of Eq. (11.9) to obtain
closed curve oriented
 v u 
in the positive
direction and enclosing
 udx  vdy      x  y  dx dy (11.10a)
C R
the region R. P and Q
are continuous
and
 u v 
functions with
continuous first partial
 vdx  udy     x  y  dx dy (11.10b)
C R
derivatives in R. This
is referred to as where R is the region bounded by the curve C.
Green's Theorem. From Cauchy-Riemann conditions, you will recall that
u v u v
 and 
46 x y y x
Unit 11 Complex Integration
On using these conditions, we observe that both the surface integrals vanish.
This completes the proof of Cauchy's integral theorem. You will realise that it
uses the condition that f (z ) is continuous. Actually, as shown by Goursat, this
condition is not necessary. The proof is rather elaborate and we shall not
discuss it here. In case you are interested in details, see Arfken (the
reference is given at the end of the block). The revised version of Cauchy's
theorem, sometimes referred to as Cauchy-Goursat theorem, is stated in the
following way:

If f (z ) is analytic in a region R and C is a simple closed curve lying in this

region, then

 f (z)dz  0
C

SAQ 3
Determine the domain over which the functions given below are analytic:
z2
a) f ( z )  , b) f (z) = sech z and c) f (z) = tan z
z2  2

Using Cauchy's theorem show that  f (z)dz  0 where C is a closed contour,

C
z  1.

Cauchy’s Theorem for a Multiply-connected Region

So far Cauchy's integral theorem has been proved for a closed contour in a
simply connected region. However, it is easy to extend this theorem to a
multiply connected region. For simplicity, we first consider a doubly connected Fig. 11.5: a) A doubly-
connected region;
region as shown in (Fig. 11.5a). Suppose that the function f (z) is analytic in
b) Conversion of a
this region including the boundary curves C1 and C2 . The simple closed curve doubly connected
C1 forms the outer boundary and the simple closed curve C 2 , which lies region into a simply
entirely within C1 , forms the inner boundary of the region. The multiply connected region;
connected region enclosed by a contour B using a cross-cut connecting C1 c) The region
described in
and C2 is shown in (Fig. 11.5b). A cross-cut in the form of thin slit MNN' M' is Fig. 11.5b) remains
introduced between C1 and C2 . MN and M' N' are line segments. Starting simply connected
from a point P we move continuously along C1 to M reaching N on C2 by the even as the width of
line segment MN, describe C2 in the anti-clockwise sense and return to P from the slit is made zero
so that N' coincides
M' on C2 via the line segment N'M'. When this path is described, the region
with N and M'
always lies on the left, so that with reference to the full contour PMNQN'M'P coincides with M
(call it contour B) the region is simply connected. giving a line segment
MN.
Following (Fig. 11.5c) we can write

 f (z)dz   (anti
f ( z ) dz
-clockwise)
  f ( z )dz   f( z )dz   f ( z )dz  0
(clockwise)
B C1 MN C2 NM
But
 f (z) dz =   f (z) dz
NM MN 47
Block 3 Complex Analysis
and

 (clockw
f ( z ) dz   
ise)
f ( z ) dz
(anticlock w ise)
C2 C2

  f (z )dz   f (z )dz  0
C1 C2
or

 f (z) dz   f (z)dz (11.11)

C1 C2


With regard to the value of the integral f ( z ) dz, the closed contours C1 and C2
are equivalent, because the contour integrals over C1 and C2 are the same.
This is called the principle of deformation of contours. It may be stated as

The integral of an analytic function of a complex variable f (z ) over a closed

curve C1 has the same value over a closed curve C2 into which C1 can be
continuously deformed such that in the process of deformation, the curve
does not pass over any singular point of f (z ).

We therefore observe that Cauchy's integral theorem enables us to replace an

integral about an arbitrary simple closed contour by an integral about a more
conveniently shaped region, say a circle. It may be emphasized here that the
cut line is a matter of mathematical convenience and allows application of
Cauchy's integral theorem.
Using this principle we can considerably simplify the evaluation of integrals
involving analytic functions. Extension of the above results for the case of a
multiply-connected region is a straight forward exercise. Let a function f(z) be
analytic in a multiply-connected region shown in Fig. 11.6. Then, as in the
previous example, the multiply connected region can be converted into a
simply connected region bounded by a single closed contour B in which f(z) is
analytic.

Fig. 11.6: The region R is bounded by the closed contour C externally and
internally by a series of non-overlapping closed contours
C1,C2 ,C3 ,C 4 . This multiply-connected region is converted into a
simply-connected region by introducing cross-cuts like
L1N1, L2N2 , L3N3 and L4N 4 . A single closed contour B can be imagined
consisting of C described in the anti-clockwise sense, C1,...,C 4
described in the clockwise sense and the cuts L1N1, L2N2 , L3N3 , L4N4
each described twice in opposite directions. Such a closed contour B
then encloses the whole region.
48
Unit 11 Complex Integration
From Cauchy's integral theorem we recall that the integral of f (z) over the
closed contour B is zero:

4
 f ( z ) dz   (anticlockwise) 
f ( z ) dz
i 1
 (clockw
+
f ( z ) dz +
ise)
B C Ci

4 4

i 1
 f (z) dz  i1  f (z) dz  0
Li Ni Ni Li

But

 f (z) dz    f (z) dz
Li Ni Ni Li

Therefore, the sum of the third and fourth terms vanishes and we finally obtain

4
 f ( z ) dz   f ( z ) dz (11.12)
C (anticlockw ise) i 1 C1 ( anticlockw ise)

SAQ 4
dz
Show that  z  z0
 2i
C

dz
and  ( z  z0 ) n
 0; n  2, 3,...
C

where C is a closed contour enclosing the point z 0 .

11.3.1 Some Consequences of Cauchy’s Integral

Theorem

Cauchy integral theorem characterizes an analytic function of a complex

Fig. 11.7: Two paths
variable. We now show some important consequences of this theorem. from a to b enclosing
a simply-connected
a) Consider a closed contour C lying in a simply-connected region in which region.
f (z) is an analytic function of z. We divide the closed curve C into two
parts, as shown in Fig. 11.7, by inserting two points a and b on it. From
Cauchy's integral theorem, we can write

b a
 f ( z ) dz =  f ( z ) dz +  f (z) dz = 0 (11.13)
C a b
along C1 along C2

a b
But  f ( z ) dz    f (z) dz
b a
along C2 along C2
49
Block 3 Complex Analysis
Using this result in the above expression, we obtain

b b
 f ( z ) dz   f (z ) dz  0
a a
along C1 along C 2

or

b b
 f ( z ) dz   f (z ) dz (11.14)
a a
along C1 along C 2

This shows that the value of a line integral between two points is
independent of the path if f(z) is analytic throughout a simply-connected
region containing the paths.

SAQ 5

Show that  z2dz is independent of the path C connecting two points (0,0) and
C
(2,2).

i) Take C to be the straight line connecting (0,0) and (2,2), and

ii) Take C to be the path (0,0) to (2,0) along the real axis and then from (2,0) to
(2,2) along the imaginary axis.

b) From the result contained in Eq. (11.14) we may say that the integral
z
 f (z) dz depends only on the initial and final values of the variable of
z0
integration and hence is independent of any path lying entirely in R. That
is, it defines a function of z excepting a constant. So we may write

z
F (z)  z 0
f ( z) dz  A (11.15)

For z  z0 , Eq. (11.15) yields F ( z0 )  A. So we can rewrite the above

relationship as

z
F ( z )  F ( z0 )  z 0
f ( z) dz (11.16)

If we evaluate the derivative of F (z ) using Eq. (11.16), we may reasonably

expect that F ( z )  f ( z ). From this we note that the value of a line integral
of an analytic function f (z) is equal to the difference in the values of an
indefinite integral (primitive) at the end points of the path of integration.
This greatly simplifies the evaluation of line integrals of analytic functions
once the primitive of f (z) is known.
50
Unit 11 Complex Integration

Example 11.5

z 3z2  1dz.
z2
Evaluate the integral
1

Solution : Here f ( z)  3z2  1 which is an analytic function of z. We note that

the primitive of f (z ) is z3  z. Therefore

z2 z2
z1
(3z 2  1) dz  ( z3  z )
z1

 z23  z13  z2  z1.

You should note that Cauchy’s theorem enables us to evaluate the line
integral of an analytic function around a closed contour in a simply connected
or multiply connected region. You may now ask: What if the integrand were
not analytic and the value of the function is required at an interior point?
Obviously, Cauchy’s theorem will not apply. In such a situation, we resort to
Cauchy’s integral formula, which we discuss in the following section.

11.4 CAUCHY’S INTEGRAL FORMULA

Cauchy’s integral formula allows us to determine the value of an analytic
function at an interior point in a simply-connected region by integrating the
function around the curve surrounding the region. This means that any change
in the value of the function on the boundary will correspond to a change in the
value of the function at an interior point. It states that:

If f (z ) is analytic throughout a simply-connected region R and C is a

simple closed contour in R, then the value of the function at a point z0
interior to C given by
1 f ( z ) dz
f ( z0 )  
2i C z  z0
(11.17)

where the integration along C is taken in the positive (counter-clockwise)

sense.

To prove this we note that z is on the contour C while z0 is in the interior,

z  z0  0 and the integral in Eq. (11.17) is well defined. Further, although
f (z)
f (z ) is assumed to be analytic within and on C, the integrand is not
z  z0
analytic at z  z0 . Therefore, Cauchy's integral theorem is not directly
applicable. To circumvent this problem, we exclude the point z0 by enclosing it
by a small circle C of radius r, as shown in Fig. 11.8. Simple closed curve C
forms the outer boundary and the circle C with radius r and centre at z0
forms the inner boundary of the doubly connected region. By means of the
cross-cut MN joining C and C , the doubly connected region is converted into
a simply connected region with regard to contour B which is the sum of C
(anti-clockwise), C (clockwise) and line segments MN and NM. Then for the 51
Block 3 Complex Analysis
f (z)
function , the region bounded externally by C and internally by C is a
z  z0
doubly-connected region and Cauchy’s Integral Theorem applies.
Since C and C are equivalent contours, from Eq. (11.11) we can write

f (z) dz f (z) dz
C z  z0
 
C z  z0
0

z0

Fig. 11.8: Exclusion of a singular point

Because of the circular path around z0 , we can use polar representation and
write z  z0  rei so that dz  ireid where r is radius of the circle.

f ( z ) dz f ( z0  r ei )i r eid
 C  z  z0

C
 r ei
2
i  f ( z0  r ei ) d
0

Since f (z ) is analytic and therefore continuous at z  z0 , in the lim r0 we can

write
lim f ( z0  r ei )  f ( z0 )
r 0

Therefore, as r  0, the above expression reduces to

f ( z )dz 2
 z  z0
 i f ( z0 ) d 
C 0

 2i f ( z0 )

or
1 f ( z ) dz
f ( z0 ) 
2i  z  z0
(11.18)
C
This is known as Cauchy’s integral formula. It allows us to calculate the value
of the function at any interior point z0 from its specified boundary values on
the contour C. This is illustrated in the following example.
52
Unit 11 Complex Integration

Example 11.6

If C is a circle z   described in the positive sense and

(3z 2  6)
g ( z0 )   z  z0
dz
C

calculate g (2).

Solution : From Cauchy’s integral formula, we know that

1 f (z)
f ( z0 ) 
2i  z  z0
dz
C

where C is a closed contour enclosing z0 . We rewrite the given function in a

form amenable for application of Cauchy’s integral formula. To this end, we
multiply and divide it by 2i. Then

1 2i (3z 2  6)
g ( z0 ) 
2i  z  z0
dz
C

C is the circle z   .

On comparing this expression with that given in Eq. (11.18) you will readily
note that

g ( z)  2i (3z2  6)

Hence g (2)  2i (3  4  6)

= 12i.

We now want you to solve the following SAQ.

SAQ 6
cos z
Calculate the value of the integral  z
dz when C is the circle z  2.
C

11.4.1 Derivatives of an Analytic Function

Cauchy’s integral formula can also be used to obtain the derivative of an

analytic function f (z ) . Let us first consider a simply connected region R within
which and on whose boundary C the given function is analytic. Consider
points z0 and z0  h which lie inside C. Then according to Cauchy’s formula,
we can write
53
Block 3 Complex Analysis
1 f ( z ) dz
f ( z0  h) 
2i  z  ( z0  h)
C

and

1 f ( z ) dz
f ( z0 ) 
2i  z  z0
C

On combining these expressions we can write

f ( z0  h )  f ( z0 ) 1 1 1 1 
h

2i    f ( z ) dz
h  z  ( z0  h) z  z0 
C

On simplifying,

1 f ( z ) dz

2i  ( z  z0  h) ( z  z0 )
C

To simplify the integrand, we rewrite the denominator as

1 z  z0  h  h

( z  z0  h) ( z  z0 ) ( z  z0  h) ( z  z0 )2

1 h
 
( z  z0 )2 ( z  z0  h) ( z  z0 )2

That is, we have written the given expression in partial fractions. Using this
expression in the above integral, we obtain

f ( z0  h)  f ( z0 ) 1 f ( z ) dz h f ( z ) dz
h

2i  ( z  z0 ) 2

2i  ( z  z0  h)(z  z0 )2
C C

We now show that the second term on the right hand side approaches zero as
h0. Let the maximum value of f (z) on C be M. If the length of C is denoted
by L and the shortest distance from z0 to C is d, then

h f ( z ) dz h ML
lim
h 0 2i  ( z  z0  h )(z  z0 )2
 lim
h  0 2 ( d  h ) d 2
C

=0

Proceeding further we note that by definition of a derivative

f ( z0  h )  f ( z0 )
f ( z0 )  lim
h 0 h

1 f ( z ) dz

2 i  ( z  z0 )2
(11.19)
54 C
Unit 11 Complex Integration
That is, the value of the first derivative of an analytic function at an interior
1
point is given by an integral of the product of the given function with z  z0 
taken along a contour enclosing that point.

You can prove that f (z ) is also an analytic function within C by showing that
f (z ) exists at each point within C. To do so, you should repeat the technique
used in arriving at Eq. (11.19). That is, we write f ( z0  h ) and f ( z0 ), subtract
and divide the resultant expression by h :

f ( z0  h)  f ( z0 ) 1 f ( z ) dz  1 1 
h

2 i    
h  ( z  z0  h)2 ( z  z0 )2 
C

1
f ( z ) dz. 2( z  z0  h)
1

2 i  2
( z  z0 )2 ( z  z0  h )2
C

As h  0, this expression reduces to

2 f ( z ) dz
f ( z0 ) 
2i  ( z  z0 )3
(11.20)
C

Since f ( z0 ) has a unique derivative, it is an analytic function.

Continuing in this way, you can show that, in general, the nth derivative of
f (z ) at z0 is given by

n! f ( z ) dz
f (n ) ( z0 ) 
2i  ( z  z0 )n 1
(n  1, 2, 3, ...) (11.21)
C

Since n is not restricted to any finite value, we note that an analytic function of
a complex variable has derivatives of all orders. The existence of the (n+1)th
derivative of f (z ) at a point means the nth derivative of f (z ) is an analytic
function at that point. From the above discussion we may conclude that

If a function f is analytic at a point, its derivatives of all orders f , f , f ,... are
guaranteed. Moreover, the derivatives of f are analytic at that point.

11.5 SERIES REPRESENTATION OF A COMPLEX

FUNCTION
11.5.1 Sequence and Series
Let us first define a sequence and the related concepts.

Definition of a Sequence: A sequence is a set of numbers z1, z2, z3 ,... in a

definite order of arrangement and formed according to a definite rule. Each
number in the sequence is called a term and zn is called the nth term. They are
represented by curly brackets { zn }. The finite sequence possesses finite
number of terms while there are infinite terms in the infinite sequence.
55
Block 3 Complex Analysis
Limit of a Sequence: A number l is termed as the limit of an infinite sequence
{ zn } if for any number  0 , there exist a positive number N which is function of
 and z such that zn  l  for all n > N. In other words, we write

lim zn  l
n 

Such sequences whose limit exists are called convergent. Otherwise, it is called
a divergent sequence. It is important to note that a given sequence possess a
unique limit.

Cauchy’s Convergence Criteria for a Sequence: A necessary and sufficient

condition that a sequence { zn } converges is that for a given  > 0, there exist
a number N such that zp  zq  for all p > N, q > N.

Infinite Series and Convergence Criteria: Let { zn } be a given sequence.

Form a new sequence { Sn } defined as

S1  z1, S2  z1  z2, S3  z1  z2  z3,..., Sn  z1  z2  ...  zn

where Sn is the sum of the first n terms of the sequence { zn } and is called the
nth partial sum. The sequence { Sn } can be represented as


z1  z2  ...  zn  ...   zi
i 1

which is called an infinite series. The infinite series is a convergent series if the
limit of nth partial sum exists i.e.

lim Sn  S
n 

where S is the sum of the infinite series. A necessary but not sufficient condition
for a given infinite series to be convergent is given by

lim zn  0
n 

If a series converges for all values in a region R, it is called the region of

convergence of the series.

Power Series: An infinite series having the form


a0  a1 ( z  a )  a2 ( z  a )2  ...  an ( z  a )n  ...   an (z  a)n
n 1

is called a power series in (z – a). Clearly the power series converges for
z = a. If we draw a circle C of radius R with centre at z = a, such that the series
converges at all points inside C and diverges at all points outside C, while it
may or may not converge on the circle C, the corresponding circle is called the
circle of convergence of the power series. The radius R called its radius of
56 convergence.
Unit 11 Complex Integration

Example 11.7

Show that the geometric series  azn converges for z  1 , where a is a real
n 0
number.
Solution : The nth partial sum of the geometric series can be written as
a (1  z n )
Sn  a  az  az 2  ...  az n 
1 z
This is convergent if the limit of the sequence Sn exists. Thus
lim z n
a (1  z n ) a
lim Sn  lim   
n
n  n  1  z 1 z 1 z
For the limit lim zn to be finite, z  1 lest the limit tends to  . Hence for
n 
a
z  1 lim Sn  . Thus, the geometric series converges for z  1 with the
n  1 z
a
sum as .
1 z

Example 11.8

Show that the series  Z n (1  z) converges for z  1 and find its sum.
n 1
an 1
Solution : Let an  zn (z  1). It can be shown that  z which when
an
applied to ratio test converges for z  1.

To find its sum, the nth partial sum of the series is

n
Sn   zi (1  z)  z(1  z)  z2(1  z)  z3 (1  z)  ...  zn (1  z)
i 1

Sn  z  z 2  z 2  z 3  z 3  ...  z n 1  z  z n 1
It is observed that all the terms in the series except the first and last term are
cancelled out. Hence the required sum is S  lim Sn  z  lim zn 1. For
n  n 
z  1, lim zn 1  0. Hence the sum is
n 

S  lim Sn  z
n 

Absolute and Uniform Convergence: A series  zn is called absolutely

convergent if the series of absolute values, i.e.,  zn converges. An
absolutely convergent series is always convergent which implied that if  zn
converges,  zn too shall converge.

For an infinite sequence zn  , if for any number  > 0, there exist a positive
number N which is function of  but independent of z, such that zn  l  for 57
Block 3 Complex Analysis
all n > N, then it is said to converge uniformly to l. Similarly, for an infinite series
 zn , if the sequence of partial sums Sn (z ) converges uniformly to S(z) in a
region, we say that the infinite series is uniformly convergent to S(z) in the
region.

Theorem 1: A power series converges uniformly and absolutely in any region

that lies entirely inside its circle of convergence.

Special Tests for Convergence:

Theorem 2 (Ratio test)

zn 1
If lim  l then for
n  zn

a) l  1,  zn converges

b) l  1,  zn diverges

c) l = 1, the test fails

Theorem 3 (Alternating series test)

If for n =1, 2, 3, ….

a) zn  0
b) zn 1  zn

Then the series z1  z2  z3  z4  ...   ( 1)n 1zn converges.

Example 11.9

z n 1
Prove that the series  n
converges for z  2 and find its sum.
n 1 2

z n 1
Solution : Let zn  . Applying the ratio test
2n
z n 1 z
lim 
n  z n 2
z
For the series to converge  1. Hence the series converges for z  2 .
2
n
z i 1 1 z z 2 z n 1
The nth partial sum Sn   2i
  
2 4 8
 ... 
2n
is a geometric series
i 1
1   z n    z  
n
1      1  lim   
2   2    n  2   1
with sum as S  lim Sn  lim  
n  n 
1
z 2z 2z
2
n
z
as lim    0 for z  2 .
n  2 

58
Unit 11 Complex Integration

11.5.2 Taylor Series

You are familiar with the power series representation of a real function  the
1
Taylor series. For example f ( x )   1  x  x 2  x 3  ... is a power series
1 x
1
representation of valid for x  1. For an analytic function of a complex
1 x
variable, the Cauchy’s integral formula opens the channel for derivation of
Taylor series for an analytic function f (z ) about a point z0 .

Taylor series

If f (z ) is analytic at all points within a circle C0 with centre at z0 , then for

every z inside C0 , f ( z ) can be represented by an infinite power series of
the form
f ( z0 )
f ( z )  f ( z0 )  ( z  z0 ) f ( z0 )  ( z  z0 )2  ...
2!
f (n ) ( z0 )
 ( z  z0 )n   (11.22)
n!

In other words, the infinite power series converges to f (z ) for every z for
which z  z0  r0, where r0 is the radius of the circle C0 (Fig. 11.9).
Geometrically speaking, z should lie within the circle of radius R which
signifies the distance of the nearest singular point of f (z ) from z0 . Then R is
the radius of convergence of f (z ) about z0 .

The proof of Taylor series representation is given in Appendix A. However you

will note that for z0  0, the Taylor series given by Eq. (11.22) reduces to
Maclaurin series:
Fig.11.9: Region of
convergence for

f ( n ) (0)z n

Taylor series. z0 is the
f ( z )  f (0)  (11.23)
n 1
n! common centre of two
concentric circles C0
and C.
For instance, the Taylor series for cos z about z0  0, is given by

z 2n
cos z  1   (1)n (2n )! , when z  . The radius of convergence of cos z
n 1
about z0  0 is z  .

Example 11.10

Obtain a Taylor series expansion of sin z about z = /4.

Solution : Taylor series of f (z ) about z = a is given by

f (a)
f ( z )  f (a)  ( z  a)f (a)  ( z  a)2  ...
2! 59
Block 3 Complex Analysis
Here
f ( z )  sin z, a   / 4
Therefore,
1
f (  / 4)  sin (  / 4) 
2
1 1
f (  / 4)  cos (  / 4)  , f (  / 4)   sin (  / 4)  
2 2
1
f (  / 4)   cos (  / 4)  
2
1 1   1 1   2 1 1   3
 sin z    z     z     z    ...
2 2 4  2! 2  4 3 2 4
   2   3 
  z    z   
1   
1  z    
4

4
  ...
2  4 2! 3! 
 
 

SAQ 7
Obtain the Taylor series representation of log (1+ z) about z = 0.

Many a time we wish to expand a function f (z ) in a series about a singular

point. For instance, a function may be analytic in an annular region and there
may be singular point(s) inside the smaller circle or outside the larger circle.
Since the function f (z ) must be analytic at all interior points, we can not apply
Taylor series. In such cases, a different series expansion  Laurent series  is
used. We shall discuss it now.

11.5.3 Laurent Series

Consider the annular region between two concentric circles C1 and
C2 centred at z0 . The function f (z ) is analytic on the circles and in the annular
region (Fig. 11.10a) but one or more singularities occur inside the smaller
circle C2 . (There may be singularities outside the larger circle C1 .)

At each point z in the annular region, f (z ) is represented by a convergent

series of positive and negative powers of ( z  z0 ) :

 
  (z  zn0 )n
b
f (z)  an ( z  z0 )n  (11.24)
n 0 n 1
where
1 f ( z ) dz 
an 
2i  ( z  z0 )n 1
, n  0,1, 2,... (11.24a)
C1

and
1 f ( z ) dz 
bn 
2 i  ( z  z0 ) n 1
, n  1, 2, 3,... (11.24b)
C2
60
Unit 11 Complex Integration
Each integral is taken in the anticlockwise sense. Eq. (11.24) is the Laurent
expansion of f (z ). The proof of Laurent series expansion is given in
Appendix B.

Fig. 11.10: a) The annular region between the two concentric circles C1 and C2
centred at z0 ; b) The doubly-connected region is converted into a
simply connected region by means of the closed contour B which is
the sum of C1 (anti-clockwise), C2 (clockwise) and line segments MN
and N M  of the cross-cut.

Since the function f (z ) is analytic throughout the annular region, in place of

the circle C1 and C2 , we can choose any closed contour C around the
annulus. Then Laurent series can be written as

f (z)   A j (z  z0 ) j (11.25)
j  

where
1 f ( z) dz
Aj 
2i  ( z  z0 ) j 1
( j  0,  1,  2,...) (11.25a)
C

In the expanded form, we can rewrite Eq. (11.25) as

f (z)  A0  A1(z  z0 )  A2 (z  z0 )2  ...

A1 A 2
.
   ... (11.26)
( z  z0 ) ( z  z0 )2

You should note that Laurent series has two parts. The first part consists of
the infinite series with positive powers of ( z  z0 ); A0  A1( z  z0 ) 
A2 ( z  z0 )2  ... This is called the analytic part of Laurent series. The second
part consisting of the inverse powers of ( z  z0 ); A1( z  z0 )1 
A2 (z  z0 )2  A3 (z  z0 )3 is called the principal part of f (z ) .

Isolated Singular Points

In Sec. 10.6 you learnt about different types of singularities of a function of a

complex variable. Here we shall confine ourselves to isolated singularities of
an analytic function in the light of its Laurent series expansion. Suppose that
f (z ) has an isolated singularity at the point ( z  z0 ) and is analytic within a 61
Block 3 Complex Analysis
circle centred at this point. In this case, the Laurent series expansion for f (z )
converges for all points on and inside the circle C1 except at the point z  z0 .
We note that the radius of the circle C2 is zero in this case.

We consider two different cases:

i) If the principal part contains a finite number of terms such that all A j  0
for j > m, then the singularity of the function f (z ) at z0 is called a pole of
order m. Expressing f (z ) in a Laurent series about z0 , we have


 An (z  z0 )n  z z10  (z  z20 )2  ...  (z  zm0 )m
A A A
f (z) 
n 0

 
 ( z  z0 )m   An ( z  z0 )n  m  A1( z  z0 )m 1 
n 0

A2(z  z0 )m2  ...  Am   ( z  z0 )m g ( z) 

g(z)
( z  z0 )m
(11.27)


where g ( z )   An (z  z0 )n  m  A1(z  z0 )m 1  ...  Am is an analytic
n 0
function at z  z0 and g(z)  0.

When m = 1, f (z ) has a simple pole. If m = 2, f (z ) has a double pole.

ii) If the principal part of f (z ) contains an infinite number of terms, the function
f (z ) is said to have an isolated essential singularity at z  z0 . To
understand this, consider the function e1/ z . It has a singularity at z = 0.
Writing u = 1/z, we find that e1/ z  eu . Further

u 2 u3
eu  1  u    ... for u  
2! 3!

1 1 1
 1    ... for z  0
z 2! z 2 3! z 3

The principal part contains an infinite number of terms. Therefore, the

function e1/ z has an isolated essential singularity at the origin.

SAQ 8
ez
Write the Laurent series expansion of about z = 1. Determine the type
( z  1)2
of singularity and the region of convergence.

Let us now sum up what you have studied in this unit.

62
Unit 11 Complex Integration

11.6 SUMMARY
 The integration of a function of a complex variable in the z-plane is
the generalization of the Riemann integration for a real variable.

 The line integral of f (z ) along a curve C can be written as

n
 f ( z ) dz  lim
n 
 f ( j ) z j
C j 1

This line integral can also be written as a sum of two real integrals:

 f (z) dz   (u dx   dy )  i  (u dy  v dx )
C C C

where f ( z )  u  iv .

 The upper bound for the absolute value of a complex integral is given
by Darboux inequality:

 f (z ) dz  max f L
C

 Cauchy’s integral theorem states that if a function f (z ) is analytic and

its derivative is continuous at each point within a simply connected
region R, then for every closed path C in R, the line integral of
f (z ) around C is zero:

 f (z) dz  0
 The principle of deformation of contours states that the integral of
an analytic function f (z ) of a complex variable over a closed curve
C1 has the same value over any other closed curve C2 into which C1
can be continuously deformed so that the curve C1 does not pass over
any singular point of f (z ) .

 If f (z ) is analytic throughout a simply-connected region R and C is a

simple closed contour in R, then for a point z0 interior to C, Cauchy's
integral formula states that by integrating the function around the
curve, we can determine the value of the function at an interior point:

1 f ( z ) dz
f ( z0 )  
2 i C z  z0

 The derivatives of an analytic function are also analytic and are

given by
n! f ( z ) dz
f n ( z0 )  
2 i C ( z  z 0 ) n 1
n  1, 2, 3,...

 The infinite Taylor series representation of an analytic function at a

point in the complex plane within its domain of analyticity is given by: 63
Block 3 Complex Analysis
(z  z0) 2
f (z)  f (z0)  (z  z0) f (z0)  f (z0)  ...
2!
(z  z0) n n
 f ( z0)
n!

 The infinite Laurent series representation of an analytic function about

an isolated singular point of the function is given by
 
 an (z  z0) n   (z  zn0) n
b
f (z) 
n 0 n 1

where

1 f ( z ) dz 
an  
2 i C1 ( z   z0 ) n 1
n  0,1, 2,...

and

1 f ( z) dz
bn  
2 i C2 ( z  z0) 1n
n  1, 2, 3,...

11.7 TERMINAL QUESTIONS

1. Evaluate the integral

i)  z* d z where the path C1 is the upper half of the unit circle from z =1
C1
to z = 1.

ii)  z* d z where the path C2 is the lower half of the unit circle from z =1
C2
to z = 1.

2. a) Use the indefinite integral to show that for every contour C extending
from a point z1 to a point z2 .

1
 zndz  n  1(z2n 1  z1n 1) n  0,1, 2,...
C


b) Evaluate the integral e z dz over some path C joining z = 0 and z = i.
C

Also use the indefinite integral as a second method.

3. Evaluate the following integrals over the closed contour C formed by the
lines x = 1, y = 1.
sin z
a)  z2
dz
C

tan z dz
b)    2
C
 z 
64 4 
Unit 11 Complex Integration
e zdz
c)  z3
C

dz
4. Evaluate the integral  1  z2 , where C is the circle z  3.
C

5. Obtain the Taylor series expansion of

i) cos2 z about z = 0
1
ii) f ( z )  about z = 1 and z = i.
z 1
6. Verify the expansions


1
 (n  1) ( z  1)n , for z  1  1
z 2 n 0

1
7. Determine the Laurent series for f ( z )  2 valid in each shaded
z  4z  3
region shown in (Fig. 11.11a, b, c).

1
Fig. 11.11: Region of validity of Laurent series for f ( z )  .
z2  4z  3


8. Prove that the series  nz n converges for z  1 and find its sum.
n 1

9. Prove that the alternating series z  z2  z 4  z6  z8  ... z  1 and find its

sum.

11.8 SOLUTIONS AND ANSWERS

Self-Assessment Questions
x2 y2
1. Parametric representation of an ellipse   1 is x  a cos,
a2 b2
and y  b sin ; 0    2.

2. a) Re z  0 means x  0, y  0 or y < 0. It represents the whole region to

the right starting from the y-axis. It is simply connected.
b) The origin is excluded from the region bounded by a circle of radius
less than one. Therefore, this is a doubly connected region.

c) The region is simply connected. 65

Block 3 Complex Analysis
z2
3. a) The singularities of f ( z )  are determined by the condition
z2  2
z 2  2  0 or z  i 2. But for these two isolated singularities, the
function is analytic throughout the z-plane. f (z ) is analytic throughout
C, the unit circle z  1 and so  f ( z ) dz  0.
C

b) The singularities of f (z ) = sech z are determined by the condition

i
e z  e z  0 or e 2z  1  e i  2ni . Hence, z   ni , n  0, 1, 2...
2


The nearest singular points are at a distance away from the origin.
2
These lie outside the unit circle. So  f ( z ) dz  0.
C

sin z i
c) tan z  . As in (b) above, cos z = 0 for z   ni .
cos z 2
So  f ( z ) dz  0 .
C

4. C is a closed contour enclosing z0 . Draw a circle C with radius r and

1
centre at z0 such that C lies wholly inside C. The function is
z  z0
analytic in the annular region bounded by C and the circle C . Therefore,

dz dz
 z  z0   z  z0
C C

A point on C is represented by z  z0  rei and dz  rieid. On

substituting it in the above integral, we find that

2
dz i r e id
 z  z0
  r e i
 2 i
C 0

dz
 C z  z0
 2i

For the other case

2
dz dz d
C z  z0  n
 C z  z0 n i  r n 1e( n 1)i
= 0; n = 2,3,…
0

5. i) 
I  z 2dz
C

where C connects the points (0,0) and (2,2).

Since z2  ( x  iy )2  x 2  y 2  2ixy , we find that u( x, y )  x 2  y 2 and
66
v ( x, y )  2xy .
Unit 11 Complex Integration
( 2,2) ( 2,2)
  z 2dz   ( x 2  y 2  2ixy ) (dx  idy )   ( x 2  y 2 ) dx  2xy dy 
C (0,0 ) (0,0 )

( 2,2)
i  2xy dx  x 2  y 2 dy 
( 0,0 )

Take C to be the straight line connecting (0,0) and (2,2). Then along C,
x = y and the line integral simplifies to

2 2
 
z 2dz  2 y 2dy  2i x 2dx 
C 0 0

8 8 16
 2   i 2  ( 1  i )
3 3 3

The same result could have been obtained easily by the method of
indefinite integrals.

z2 z 2  2i
z3 2 z3 (2  2i )3
 z 2dz  z 2dz 3 z

3

3
C z1 1 0

16
 ( 1  i )
3

ii) Here the path C consists of two line segments: (0,0) to (2,0) along the
real axis (call it AP) and then from (2,0) to (2,2) along the imaginary
axis (call it PB). Along AP, y = 0 along PB, x = 2. Moreover, dz = idy
along PB.

2 2
 z 2 dz   z 2 dz   
z 2 dz  x 2 dx  i 2  iy  2 dy 
C AP PB 0 0

2 2 2
  x 2dx  2  2 y dy  i  (4  y 2 ) dy
0 0 0

8 44 i  8 16
   i ( 4  2)   ( 1  i )
3 2 3 3

We observe that  z2dz is independent of the path C connecting the

C
given points.

6. From Cauchy’s integral formula we recall that

1 f ( z ) dz
f ( z0 ) 
2i  z  z0 67
C
Block 3 Complex Analysis
To be able to use this result, we rewrite the given integral as
cos z 1 2i cos z
 z
dz 
2i  z
dz
C C

Here C is a circle defined byz= 2, f (z) = 2i cosz and z0  0.

cos z
  z
dz  2 i cos 0  2i
C

7. f ( z )  log (1  z ), f (0)  log 1  0

1
f ( z )  , f (0)  1
1 z

1
f ( z )   , f (0)  1
1  z 2

( 1)(2)
f ( z )  , f (0)  2!
1  z 3

n!
f ( n 1) ( z )  ( 1)n , f ( n 1) (0)  ( 1)n n!
1  z n 1
We know that Taylor series representation of a function f (z) is given by

z2
f (z)  f (0)  f (0)z  f (0)  ...
2!

On inserting the above values, we get

z 2 z3 z 4
f (z)  z     ...
2 3 4

ez
8. We have to obtain Laurent series expansion of f ( z )  about z = 1.
( z  1)2
Here z = 1 is the singular point. Let us put z – 1 = u so that z = u + 1.

ez eu 1  eu  e  u2 u3 
   e   1  u    ...
( z  1)2 u 2 2
u  u 2 2! 3! 

e e e e.u
     ...
u 2 u 2! 3!

e e e e( z  1)
     ...
( z  1)2 ( z  1) 2! 3!

This is the required Laurent series expansion about z = 1. The principal part
e
contains terms upto . So z = 1 is a pole of order two.
( z  1)2

68 The series converges for all values of z  1.

Unit 11 Complex Integration

Terminal Questions

1. i) To evaluate  z * dz where C1 is the upper half of the unit circle from

C1
z = 1 to z = 1, we use the parametric equation of a circle:
z  x  iy  rei where r is radius of the circle. For unit circle
r  1, z  ei, dz  ieid and z *  e i . Further, z = 1 corresponds to
 = 0 whereas z = 1 corresponds to  =  on C1.
 
  z*dz   ei i eid   id   i
C1 0 0

ii) For the lower-half of the unit circle,  lies between  and 2.
 
 z *dz  2 e  iie id  i 2 id  i (   2)
C2

 i
You should note that the integral of z* depends upon the path. This is
because z* is not an analytic function of z.
2. a) From Eq. (11.15), we know that the indefinite integral of an analytic
z
function f ( z ), F ( z )   f (z) dz, satisfies the relation F (z)  f(z) . Here
z0
f( z)  z n (n  0,1, 2,...) is an analytic function and it is possible to
zn 1
identify F (z)  , because F (z)  z n .
n 1
z2
z n 1 z2 z2n 1  z1n 1
  z ndz 
n  1 z1

n 1
z1


b) As e z is an analytic function, e zdz is independent of the path C
C
joining z = 0 and z = i. So we can write

 ezdz   e x iy (dx  i dy )

C C


 e x (cosy  i siny ) (dx  i dy )
C

 
 (e x cosy dx  e x siny dy )  i (e x siny dx  e x cosy dy )
C C

A simple path C from z = 0 to z = i is along the y-axis. Here x = 0 and

y changes from 0 to  so that dx = 0 and e x  e0  1. Hence
 
 
e zdz   siny dy  i cos y dy 
C 0 0

 cosy  i siny 0  cos  cos 0  2
0 69
Block 3 Complex Analysis
Using the indefinite integral method, you can obtain the result more
quickly in this case:
i i
 e zdz   e zdz  e z 0  e i  e0  2
C 0

3. a) We know from Cauchy’s integral formula that

n! f ( z ) dz
f ( n ) ( z0 ) 
2 i  ( z  z0 )n 1
C

where C encloses the point z0 . Since f ( z )  sin z, we find that

sin z d 
 z2
dz  2i  sin z 
 dz  z  z0 0
C

For n = 1 and z0  0, we get

sin z
 z2
dz  2i cos z z  z0 0  2i
C

tan z dz d
b)  z  (  / 4) 2
 2i
dz
tan z z  z  / 4
0
C

 2 i sec 2 z z  z0   / 4

= 4i

ez 2i d 2e z
c)  z3
dz 
2! dz 2 z  z 0
0

=  i e z z  z  0  i
0

1
4. The given integral is  f ( z ) dz with f ( z ) 
1  z2
and C is the circle z  3.
C
We rewrite f (z ) as

1 1
f (z)  
1 z 2 ( z  i )(z  i )

1 1 1 
   
2i  z  i z  i 

From this we note that the given function has two singular points, z = i and
z = i. Both of these lie within the given circle z  3. If we exclude these
points from the circular region bounded by C by drawing sufficiently small
circles C1 and C2 around z = i and z = i respectively, then f (z ) will be
analytic in the triply connected region exterior to C1 and C2 and interior to
C. We can use Cauchy theorem for multiply connected region and using
Eq. (11.12) write

 f ( z ) dz   f ( z ) dz   f ( z ) dz
70 C C1 C2
Unit 11 Complex Integration
Now we substitute the value of f (z ) to obtain

1 dz 1 dz
 f ( z ) dz 
2i  
z  i 2i  zi
C1 C1 C1

1 dz 1
Since C1 encloses only z = i, we note that
2i  
z  i 2i
( 2i )  . But
C1
1
the second integral on the RHS of above expression vanishes as
zi
has no singularity within C1 . You can similarly show that

1 dz 1 dz
 f ( z ) dz 
2i  
z  i 2i  zi
C2 C2 C2

1
0  2i  
2i

dz
  1  z2     0
C

1
5. i) f ( z )  cos2 z  (1  cos 2z )
2

We have

f ( z )   sin 2z, f ( z )  2 cos 2z

f (z)  22 sin 2z, f iv (z)  23 cos 2z

f (0)  0, f (0)  2, f (0)  0, f iv (0)  23

Hence, Taylor series of cos 2z about z = 0 is

f (0) z2 f (0) z3 f iv (0) z 4

f (z)  cos2 z  f (0)  f (0)z     ...
2! 3! 4!

2 2 23 4
 1 z  z  ...
2! 4!

1
ii) f (z) 
z 1

To obtain Taylor series about z = 1, we have to find the proper

binomial series. Let us first rewrite the given function as

1 1 1 1
f (z)    
z 1 1 z 2  (z  1) 2u

1 1  u  1 1  u  u  2  u 3 

  1    1         ...
21  
u 2  2  2  2 2 3 

 2
71
Block 3 Complex Analysis

u
for  1 or z  1  2
2

where u = z + 1.

In terms of z, the required Taylor series about z = 1 is

1 1 1 z  12 z  13 
f (z)    1  ( z  1)    ...
z 1 2  2 4 8 

To obtain Taylor series about z = i, we write

1 1 1 1 1
f ( z)    
z  1 z  i  1  i  (1  i )  z  i 1 i  zi 
1  
 1 i 
zi
A binomial series exists if < 1 or z  i  1  i  2 . The required
1 i
Taylor series about z = i is

1  z  i  z  i 2 
f (z)   1     ...
1  i  1  i  1  i  

with z  i  2.

6. The Taylor series expansion is given about z0  1. We have to verify this
expansion. We write
1 1 1 1
  
z 2 (z  1  1)2 (1 (z  1))2 (1  u )2

where u = z + 1.
A binomial expansion in terms of the powers of u is possible if u  1.
Therefore,
1 1
  (1  u )2 u 1
z 2 (1  u )2

23 2 23 4 3
 1  2u  u  u 
2! 3!
 1  2( z  1)  3(z  1)2  4( z  1)3  

  (n  1)(z  1)n , z 1 1
n 0

The given expansion is verified.

7. a) First we consider the Laurent series expansion in the shaded region of
(Fig. 11.11a). The shaded region corresponds to 1  z  3 and f (z ) is
expressed in terms of appropriate partial fractions. Thus,

 
1 1 1 1 1 
  
z2  4z  3 (z  3) (z  1) 2  z  3 z  1

72
Unit 11 Complex Integration
1 1 1
  .
 z  2 (1  z )
61  
 3
1 1 1
  .
 z 2z  1 
61   1  
 3  z

The first term has a singularity at z = 3. The Taylor series is valid for
z  3 . The second term has a singularity at z = 1. The appropriate
1 1
Taylor series is valid for  1 or  1. So the series valid for
z z
z  3 is

1 z 2 z 3 
  1         
1 1 z
 .
6 1  z  6  3 3 3 

 
 3

1 z z 2 z3
    
6 18 54 162

For the second term, we have

1 1 1  1 1 1 
 .   1   2  3  
2z  1  1  2z  z z z 
 
 z 

1 1 1 1
    
2z 2z 2 2z 3 2z 4

Hence, the required Laurent series is obtained by adding these two

series:

1 1 z z2 z3
    
z 2  4z  3 6 18 54 162

1 1 1 1
    
2z 2z 2 2z 3 2z 4

which is valid in the region 1  z  3.

b) To obtain a series expansion of the given function in the shaded region

0  z  1  2 (see Fig. 11.11b), we expand it about the point z = 1 as
follows:
1 1 1 1
  
z2  4z  3 ( z  3)( z  1) 2( z  1)  z  1
1  
 2 

1  z  1  z  1 2  z  1 3 
 1       
2( z - 1)  2  2   2  

1 1 ( z  1) ( z  1)2
    
2( z  1) 4 8 16
73
Block 3 Complex Analysis
This is the Laurent series valid for 0  z  1  2.

c) In (Fig. 11.11c), the shaded region is exterior to the circle z  3 .

Therefore,

 
1 1 1 1 
 
z2  4z  3 2  z  3 z  1 

To proceed, we note that

1 1 1 3  3 2 
  1      
z3  3 z  z z 
z 1  
 z

1 3 32
   
z z 2 z3

which is valid for z  3. Similarly, the series valid for z  1 is given by

1 1 1 1 1 1 
  1     
z  1 z1  1  z  z z 2 z3 
 
 z

1 1 1 1
    
z z 2 z 3 z4

Therefore, the Laurent series expansion valid for z  3 is given by

1 1  3  1  1  3 2  1  1  33  1 
        ...
z 2  4z  3 2  z 2  2  z 3  2  z 4 

1 4 13
    ...
z 2 z3 z 4

 lim 1   z  z .
zn 1 1
8. Taking zn  nzn and applying the ratio test, lim
n  zn n  n

Hence the series converges in the region z  1. Computing the nth partial
sum
n
Sn   izi  z  2z2  3z2  4z4  ...  (n  1) zn 1  nzn
i 1

Multiplying the above series by z, we get

zSn  z2  2z3  3z4  4z5  ...  (n  1)zn  nzn 1

and subtracting from Sn

z(1  zn )
(1  z) Sn  (z  z2  z3  z4  z5  ...  zn )  nzn 1   nzn 1
1 z

z(1  z n ) nzn 1 z  1 1 
Sn     z n 1  
(1  z )2 1  z (1  z )2  (1  z )2 1  z 
74
Unit 11 Complex Integration
z  1 1 
Therefore, the sum S  lim Sn   lim zn 1  .
n  (1  z )2 n   (1  z )2 1 z

z
For z  1 lim zn 1  0. Hence S  .
n  (1  z )2


9. The series is  (1)n z2n. Taking zn  (1)n z 2n , let us compute
n 0

zn 1
lim  lim z 2  z 2
n z n n 

From ratio test, the above series converges in the region z  1.

The nth partial sum can be written in the form

n
Sn   (1)i z2i  1  z2  z4  z6  ...  (1)n 1z2n 2  (1)n z2n
i 1

Multiplying the series z 2 and adding to Sn , we arrive at the following

result

(1  z 2 )Sn  1  z 2  z 2  z 4  z 4  z 6  z 6  ...

 (1)n 1z 2n  (1)n z 2n  2

Except the first and last terms, rest of them cancel out in pairs.

1  ( 1)n z 2n  2
Hence Sn  . Putting in the expression of the sum
(1  z 2 )
1 z 2n  2
S  lim Sn   lim ( 1)n and applying the result
n  (1  z )2 n  (1  z )2
lim z2n  2  0 for z  1, the sum of the series returns the value
n 

1
S  lim Sn 
n  (1  z )2

75
Block 3 Complex Analysis
APPENDIX 11A TAYLOR SERIES

Let C0 be a closed loop is which f (z ) is analytic and z0 is a point inside it. To

prove the theorem, we draw a circle C centred at z0 with radius r   r0 such
that C encloses z. If z is a point on C , then z  z0  r  . Let z  z0  r .
Then r  r   r0 (see Fig. 11.9). Since f (z ) is assumed to be analytic within
C0 we can use Cauchy's integral formula to write

1 f ( z)
f (z) 
2i  z  z
dz (11A.1)
C

The denominator in the integrand of this integral can be rewritten as

1 1 1
 
z  z z  z0  ( z  z0 )  z  z0 
( z  z0 )1  
 z  z0 

1 1
 . ,

( z  z0 ) 1  

z  z0 z  z0 r
where   and     1.
z  z0 z  z0 r 

You may note that  is a complex number other than 1. Using the algebraic
identity

1 n
 1    2    n 1  (  1)
1  1 

we can write

 2 n 1 
1 1 z  z0  z  z0   z  z0 
 1     ...     Tn 
z  z z  z0   z  z0  z  z0   z  z0  

where

n
 z  z0  1 ( z  z0 )n 1
Tn    .  .
    
 z  z0  1   z z0  ( z  z0 )
n 1 z z
 
 z  z 0

1
Substituting the expansion of in (11A.1), we get
z  z

f ( z)dz 
2
1 z  z0  z  z0 
f (z) 
2i  
1  
z  z0   z  z0  z  z0 
 
C

n 1 
 z  z0 
   Tn 
 z  z0  
76
Unit 11 Complex Integration
1 f ( z) dz z  z0  f ( z) dz

2i  z  z

2i  ( z  z)2

C C

( z  z0 )2 f ( z ) d z 
2i  ( z  z)3
C

( z  z0 )n 1 f ( z) dz ( z  z0 )n f ( z ) dz 

2i  ( z  z )n

2i  ( z  z )(z  z0 )n
C

Now we use the Cauchy's integral formulae (11.21) and obtain

f ( z0 )
f ( z )  f ( z0 )  ( z  z0 )f ( z0 )  ( z  z0 )2 
2!

f ( n 1) ( z0 ) ( z  z0 )n f ( z  ) dz 

(n  1)!
( z  z0 )n 1 
2i  ( z  z )(z  z0 )n

(11A.2)

In arriving at this expression, we have used Eq. (11.20). The integral term
represents the remainder term Rn (z ) after n terms:

( z  z0 )n f ( z  ) dz 
Rn ( z ) 
2i  ( z  z )(z  z0 )n
C

This term tends to zero as n tends to infinity. To show this, we put

z  z0  r ei and note that z  z0  r  . Further, we write
z  z  z  z0  z0  z  z  z0  z  z0 or z  z0  r   r . Then

z  z0 n f ( z) dz
Rn ( z ) 
2  z  z z  z0 n
C

z  z0 n f ( z) dz

2  z  z 0  z  z0 z  z0 n
C

r n 1 M.2r 
   .
 r   2 r   r

where M is the maximum value of f (z) on C .

r n
 1, the right hand side consisting of   multiplied by a finite constant
r
As
r  r
Mr 
term tends to zero as n   implying that lim Rn ( z )  0.
r  r n 

We therefore establish that as n  , the Taylor series converges to f (z ).

Finally we can write
77
Block 3 Complex Analysis
f ( z0 )
f ( z )  f ( z0 )  ( z  z0 )f ( z0 )  ( z  z0 )2 
2!

f (n ) ( z0 )
 ( z  z0 )n  
n!

f n ( z0 )
=  (z  z0 )n n!
(11A.3)
n 0

which is our desired Taylor series expansion.

78
Unit 11 Complex Integration

APPENDIX 11B LAURENT SERIES

Refer to (Fig. 11.10b), which depicts an imaginary cut to convert a doubly

connected region to a simply connected region bounded by the contour B.
With reference to the closed contour B, we can apply Cauchy's integral
formula:

1 f ( z)
f (z) 
2i  z  z
dz (11B.1)
B

f ( z)dz
Since the integral  z  z
vanishes over the cut, the contour integral over B

consists of the integral over C1 in the anti-clockwise sense and on C2 in the

clockwise sense. Therefore
1 f ( z ) 1 f ( z )
f (z) 
2i  z  z
dz 
2i  z  z
dz (11B.2)
C1 C2

The negative sign signifies that the contour C2 (like C1 ) is traversed in the
anticlockwise sense. Now we wish to express the denominator in the
integrand of Eq. (11B.2) in a form that results in positive powers of ( z  z0 )
when integration is carried out along C1 and in negative powers in the C2
integration. When z is a point on C1, z  z0  z  z0 . Therefore, we write

1 1 1
 
 
z  z z  z0  ( z  z0 )  z  z0 
( z  z0 )1  
 z  z0 
n 1 n
1 z  z0 ( z  z0 )  z  z0  1
     .
z  z0 ( z  z0 )2 ( z  z0 )n  z  z0  ( z  z )

Hence the first integral in (11B.2) can be rewritten as

1 f ( z) 1  1 z  z0
2i  z  z
dz 
2i  f ( z) dz


z  z0 ( z  z0 )2

C1 C1

1 
n
( z  z0 )n 1  z  z0 
   
( z  z0 )n  z  z0  ( z  z ) 

n 1
  a j (z  z0 ) j  Rn
j 0

where
1 f ( z )
aj 
2 i  ( z  z0 ) j 1
dz
C1

and

( z  z0 )n f ( z )
Rn 
2i  ( z  z0 )n ( z  z )
dz
C1 79
Block 3 Complex Analysis
As in the proof of Taylor series, we can readily show that here also Rn  
as n  .

f ( z)

1

2i  z  z
dz  a j ( z  z0 ) j (11B.3)
B j 0

Again, when z is on C2, z  z0  z  z0 and we can write

1 1 1
  
z  z z  z z  z0  (z  z0 )

1

 z  z0 
( z  z0 )1  
 z  z0 

1 z  z0 ( z  z0 )n 1 ( z  z0 )n
   
z  z0 ( z  z0 )2 ( z  z0 )n ( z  z0 )n ( z  z)

Hence, the second integral in (11B.2) takes the form

f ( z) n bj

1

2i  z  z
dz  
( z  z0 ) j
 Qn
C2 j 1

where
1 f ( z )
bj 
2i  ( z  z0 ) j 1
d z
C2

and

1 1 ( z  z0 )n
Qn 
2i ( z  z0 )n  ( z  z)
dz.
C2

It is interesting to see that Qn  0 as n  . You should try to obtain this

result yourself. Hence,

f ( z) bj

1

2i  z  z
dz 
( z  z0 ) j
(11B.4)
C2 j 1

On combining Eqs. (11B.2), (11B.3) and (11B.4), we obtain the Laurent series
expansion of f (z ) as given in Eq. (11.24).

80